Wui-Jin Koh and Lindsay R. Sales
Radiotherapy plays an integral role in the care of many gynecologic cancers and can be used for definitive management, adjuvant therapy, or palliation. The principal basis of therapeutic radiation lies in its ability to cause ionization, or the creation of free electrons and free radicals, when absorbed by biologic matter. These highly reactive chemical species interact with critical molecules in a cell (in particular deoxyribonucleic acid [DNA]) and, if unrepaired, lead to loss of cellular reproductive capacity and eventual cell death. Ionizing radiation can be emitted from radioactive isotopes, both naturally occurring and man-made, or created using specialized high-voltage but nonradioactive equipment such as linear accelerators.
Optimal radiation for gynecologic malignancies often combines both teletherapy (external beam radiotherapy) and brachytherapy (internal radiation) with careful clinical judgment required to determine the proper weighting of each component. The challenge in radiation delivery is to deliver intended full dose(s) to selected target(s), while minimizing exposure to adjacent normal tissues. Sophisticated developments in imaging, computer-based treatment planning, and linear accelerator technology provide for ever-greater sophistication and accuracy in radiotherapy. However, such precision in dose delivery has to be accompanied by improvements in patient set-up immobilization, reproducibility, and regular tumor tracking to prevent marginal misses of the intended target volume. Advances in radiotherapy for gynecologic malignancies will be based on further integration with systemic agents (for both spatial cooperation and chemosensitization), as well as developments in targeting, tracking, and adaptive processes, featuring radiation plans that may be modified during a course of therapy to conform to changes in patient and tumor geometry.
FUNDAMENTALS OF RADIATION PHYSICS
Structure of Matter
All matter is composed of individual units called elements. Each element is defined by the physical and chemical properties of its basic component—the atom. The atom consists of a central core, the nucleus, made up of positively charged particles, called protons, and neutrons, which have no charge. The nucleus is surrounded by a “cloud” of negatively charged particles, or electrons, which move in orbits around the nucleus. In the basic “resting” state of an atom, the number of protons in the nucleus is equal to the number of orbiting electrons, making the atom electrically neutral.
The formula is used to identify each atom. X is the chemical symbol for the element, A is the mass number or number of nucleons (the number of neutrons and protons in the nucleus), and Z is the atomic number (the number of protons in the nucleus). The number of protons (Z) in an atom determines its chemical properties and its elemental name. Within the periodic table of elements, as Z increases, the number of accompanying neutrons increases proportionately more (ie, A:Z ratio > 2) to maintain nuclear stability. Atoms with the same Z, but with different numbers of neutrons, share the same element name and chemical properties but are called isotopes. When the A:Z ratio varies from the baseline (or lowest energy) form, these isotopes are often unstable and seek to achieve nuclear stability by giving off excess energy in the form of radiation, and are thus called radioisotopes (or radionuclides).
Hydrogen, carbon, oxygen, and nitrogen are the main elements that make up the human body. Each has a relatively small atomic number and mass number. The neutron-to-proton ratio for each of these elements is unity, and each exists at baseline in an electrically neutral and stable configuration.
The mass of subatomic particles is measured in terms of the atomic mass unit (AMU). An AMU is defined as one-twelfth of the mass of the carbon atom. In metric units of mass, . The mass of a proton is 1.00727 AMU, and that of a neutron is very similar at 1.00866 AMU. Electrons are significantly smaller, with a mass of 0.000548 AMU.
According to the atomic model proposed by Niels Bohr, the negatively charged electrons revolve around the positively charged nucleus, held in place by Coulombic force of attraction, in fixed orbits at specific distances from the nucleus. Electron orbits are referred to as shells, with the K shell being the inner most shell, followed radially by the L, M, N, and O shells. The maximum number of electrons in an orbit is defined as 2n2, where n is the integer specific to each shell and called the principle quantum number. In reality, the actual configuration and location of orbital electrons are rather complex and dynamic; however, this simplified model provides for an understanding of the basic concepts of atomic structure.
Electron orbits can also be considered as energy levels within an atom. When an electron moves to an orbit closer to the nucleus, energy is released. For an electron to move to an orbit farther from the nucleus, energy is required. The energy required to remove an electron completely from an atom (or ionization) is termed the binding energy of the electron. The binding energy of the electron depends on the magnitude of the Coulomb force of attraction between the nucleus and the electron. The binding energy for the higher Z atoms is greater because of the greater nuclear charge. Additionally, the binding energy is greater for electrons closer to the nucleus. However, on average, removal of an electron from an orbit requires 33 to 35 eV, where 1 eV, or electron volt, is defined as the kinetic energy acquired by an electron in passing through a potential difference of 1 V. This 33 to 35 eV range reflects the minimum amount of energy that an incident beam of radiation, or photon, must have to cause ionization.
Types of Radiation
The different forms of radiation are usually categorized into 2 groups: electromagnetic and particulate.
Electromagnetic radiation exists on a spectrum and is defined by its energy, or corresponding wavelength. In general order of increasing energy, the electromagnetic spectrum includes radio waves, microwaves, infrared, visible light, ultraviolet, and x-rays and γ-rays. All radiation within the electromagnetic spectrum has the same velocity (the speed of light, or 3 × 108 m/s). Although electromagnetic radiation has no mass or charge, it exists in a duality that can be considered either as a waveform (expressed as wavelength or frequency) or as packets of energy called photons (expressed as eV). These 2 properties of photons can be readily converted from one form to another using the following equation:
where E is the photon energy (eV), v is the frequency of the radiation (s–1), and h is Planck’s constant (4.1357 × 10–15 eV • s).
The frequency and wavelength of a photon are inversely related, with the correlation given by:
where v is the frequency of the radiation (s–1), c is the speed of light (3 × 108 m/s), and λ is the wavelength (m).
As the wavelength of the photon becomes shorter, the frequency increases in inverse proportion. Hence, electromagnetic radiation of shorter wavelengths has higher energies. For the purposes of radiotherapy, it is only the photons that have sufficient energy to overcome the binding energy of electrons in biologic matter that are of specific interest; these are the ionizing x-rays and the γ-rays that exist in the higher energy portion of the electromagnetic spectrum. The lower energy forms of electromagnetic radiation do not cause ionizations but can result in heat and/or light, which are usually considered less injurious to biologic matter. Typically, electromagnetic radiation is considered to be ionizing when the photon energy exceeds 124 eV (or has a wavelength < 10–8m).
There are no intrinsic differences in characteristics between x-rays and γ-rays—their names refer only to the specific photon source. Gamma rays (γ-rays) arise from within the nuclei of radioactive atoms, whereas x-rays come from extranuclear sources. In general, γ-rays are emitted from radioactive isotopes as they decay, whereas x-rays are produced “artificially” in high-voltage equipment, via bombardment of a target with high-speed electrons. Otherwise, they share the same physical properties and, if of similar energy, result in identical biologic effects.
Particulate radiation can be charged (electrons, protons, helium ions, carbon ions) or uncharged (neutrons). The charged particles have a defined and finite depth of penetration in matter, determined by their incident energy. This characteristic is exploited clinically to limit dose to the target range, sparing tissues beyond a specified depth. Of these particulate types, only electrons are commonly used in clinical radiotherapy for gynecologic cancers, although proton radiotherapy is an emerging technology that holds promise for ultraprecise delivery of radiation.
Radioactive decay is a phenomenon in which radiation is given off by the nuclei of elements. As previously mentioned, certain combinations of protons and neutrons are less stable than others for a given element, with heavier elements requiring a specific neutron-to-proton ratio greater than 1 (or ) for maximal stability. When the A:Z ratio varies from the explicit baseline (or lowest energy) form, these isotopes (whether naturally occurring or artificially created by bombardment with neutrons in a reactor) are often unstable and seek to return to stability by nuclear disintegrations that result in the emission of ionizing radiation. This process is known as radioactive decay, in which a “parent” radioisotope achieves a more stable form by transforming to a lower energy “daughter” isotope of the same atom (same Z), or even to a completely different element (different Z). In general, the elements with high atomic numbers tend to be “naturally” radioactive, and all elements beyond lead (with a Z of 82) are radioactive, even if the decay rate is so slow as to be undetectable except with the most sophisticated equipment. However, any element, if bombarded with neutrons to perturb its baseline A:Z ratio, can be rendered radioactive, such as tritium (hydrogen 3) or carbon 14.
A common concept used in describing radioactive decay is the half-life (T½), which is the time required for half the atoms of a specific parent radioisotope to have transformed into its daughter nuclide. The daughter nuclide, in turn, could be stable or remain unstable with further nuclear disintegrations, resulting in its own subsequent T½ of decay. A related concept is the “activity,” or the number of disintegrations per second of a given amount of a specific radioisotope. The greater the activity, the shorter the T½, as the parent radionuclide is transformed more rapidly. The half-lives of radioisotopes vary tremendously and can range from femtoseconds (10–15 seconds) to billions of years.
There are 3 distinct types of radiation emitted by radioactive decay: α-particles, β-particles, and γ-rays. Each nuclear disintegration of a radioisotope can result in the emission of 1 or more of these types of radiation. Alpha decay occurs when the ratio of neutrons to protons is lower than the stable baseline, particularly in radio-nuclides with atomic numbers above 82. The emitted particle in α decay is a helium nucleus (2 protons and 2 neutrons) with a positive charge. These α-particles are relatively “heavy” and of low kinetic energy, such that their effective range is at most only a few centimeters of air or a few millimeters of tissue. In β decay, a β-particle is emitted; it typically has a negative charge and is also known as an electron. The distance range of electrons from nuclear decay is variable, depending on the kinetic energy with which they are ejected. Sometimes, the β-particle has a positive charge, and is then called a positron. Because positrons are readily attracted to surrounding negatively charged electrons in matter, they travel very short distances (typically < 1 mm) before reacting (or annihilating) with an electron, producing a pair of opposing 511-eV photons that can be detected by specialized scintillators. This forms the basis for positron emission tomography (PET), where fluorine 18, a positron emitter, is used to label glucose uptake in tissues. Finally, γ decay occurs when a nucleus undergoes a transition from a higher to lower energy state and a high-energy photon (or γ-ray) is emitted. The penetration capability of the γ-ray in tissue is dependent on the specific energy of the photon; these γ-rays can be used for diagnostic or therapeutic purposes.
The “hotness” of a radioisotope is a complex function of nuclear activity (disintegrations per second), the type(s) and energy of radiation emitted by that specific isotope, and the distance at which the radioactivity is measured.
Inverse Square Law
The inverse square law is represented as:
where I is the intensity of radiation (or a force such as gravity) and d is the distance from the source of the radiation (or force).
When applied to ionizing radiation, the inverse square law strictly applies only to electromagnetic radiation (not to particulate radiation) and where the origin of radiation is a “relative” point source in relation to the distance at which the radiation is measured. When these conditions are met, it means that an increase in the distance from a radiation source results in a proportionately greater decrease in the radiation exposure. For example, doubling the distance from a radiation point source would result in only one-quarter of the dose at the original distance. When the radiation source is not at a “point” relative to the distance at which dose is measured (eg, when calculating vaginal mucosal and paravaginal doses with brachytherapy using a “line source” in the cylinder), the inverse square law does not hold, and the exposure may then be more proportional to 1/d. Nonetheless, the impact of distance on dose remains, and this concept is critical in helping design shielding for rooms housing linear accelerators and/or brachytherapy radioisotope sources. It also explains why, when handling radioisotopes, that in addition to shielding, long-handled equipment is used where possible to maximize source distance and minimize dose exposure.
Interaction of Radiation With Matter
Electrons carry a negative charge and have mass. As a result, they rapidly interact with other electrons found in matter, resulting in rapid transfer of energy, ionization of atoms, and a relatively short, finite range (typically up to a few centimeters), depending on the energy of the incident electron.
When photon radiation enters matter, it is possible that it will pass through without interaction, or it may interact in 1 of several ways, including photoelectric effect, Compton effect, and pair production. In clinical radiotherapy, with the contemporary clinical use of high-energy, megavoltage photons, Compton effect is the dominant interaction of ionizing radiation with biologic matter.
In Compton interaction, the photon interacts with an orbital electron, where it provides enough energy to overcome the binding energy of the electron to the nucleus and further transfers additional kinetic energy to the “free” electron, which is then emitted from the atom. The photon is likewise deflected but continues with reduced energy and may react with additional electrons in a similar fashion along its new path, as long as it has sufficient residual energy to overcome the binding energies of other electrons. Figure 19-1 depicts the Compton effect, which thus results in ionization of biologic molecules and the production of fast electrons. The electrons produced by the Compton effect can also go on to cause further ionization of additional atoms by interacting with other orbital electrons.
FIGURE 19-1. The Compton effect. (Reproduced, with permission, from Khan FM. The Physics of Radiation Therapy. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2010:60, Figure 5.7.)
Ultimately, the primary basis of ionizing radiation is the production of these free electrons and ions, which are highly reactive and energetic. These ultimately react with important biologic molecules, especially DNA, and result in disruption of vital chemical bonds that may lead to cellular damage and death.
An important consequence of the Compton effect is that its interaction with matter is nearly independent of the atomic number (Z) of the absorbing matter. This is in contrast to the photoelectric effect, which prevails in lower energy photon–matter interactions, where absorption is proportional to Z3. Diagnostic radiology uses lower energy photons (in the kilovoltage range), where photoelectric effect is significant, such that heavier elements (eg, calcium in bones) absorb much more radiation than soft tissue, giving rise to the characteristic tissue contrasts seen in radiographs. However, for therapeutic radiation, preferential absorption by bone (eg, the pelvic girdle in gynecologic cancers) would be detrimental. The relative independence of radiation absorption with respect to Z for the Compton effect (which predominates in megavoltage radiotherapy) means that although a radiograph taken with higher energy photons has reduced contrast, the absorbed dose is very similar in soft tissue, muscle, fat, and bone and, as such, allows better dose delivery to central pelvic structures.
Units Used in Radiation Oncology
Historically, the unit for radiation exposure has been the roentgen (R), defined as the amount of x-rays or γ-rays required to liberate positive and negative charges of 1 electrostatic unit of charge in 1 cubic centimeter of air. With the advent of Système International d’Unités (SI units), the roentgen is no longer used, and exposure is now expressed as coulomb per kilogram (C/kg), which is equivalent to approximately 3876 R. The SI-based definition for absorbed dose in tissue is measured in joule per kilogram (J/kg), which is also known as gray (Gy). As a unit of energy, 1 J is equal to 6.24 ×1018 eV. Rad had previously been used for absorbed dose and was equivalent to 0.01 J/kg. Hence 1 Gy equals 100 rad, and thus, the units centigray (cGy) and rad are often used interchangeably.
For radioisotopes, activity describes the number of disintegrations per unit of time interval. Curie (Ci) had been the historical unit of activity and was based on the number of nuclear events in 1 g of radium 226 (3.7 ×1010disintegrations per second). Becquerel (Bq) is the SI unit now used for activity and is equal to 1 nuclear disintegration per second of any given radionuclide .
Radiation Damage in Biologic Matter
Radiation biology is the study of the effects of ionizing radiation on biologic systems. The most important biologic effect of radiation appears to result from DNA damage, which can result in genetic mutations, chromosome aberrations, disturbed cell proliferation patterns, cell death, neoplastic transformation, or teratogenesis.
As previously mentioned, when radiation (in particular photons or electrons) enters a biologic system, it results in creation of kinetically energized free electrons (ionization). These electrons ultimately impact the cell by disrupting vital chemical bonds, either directly or indirectly via a cascade of free radical formation.
In direct action, the free electron (eg, produced via Compton effect) itself results in ionization of the nearby DNA strand, thus leading to DNA molecular damage. For x-rays or γ-rays, this accounts for only about one-third of the DNA damage produced. In indirect action, the electron interacts with water molecules (which comprise the vast majority of a cell volume), which become ionized as follows: . The H2O+ ion radical has a relatively short lifetime, on the order of 10–10 seconds, but can react with surrounding water molecules to form additional free radicals . These added free radicals, particularly the hydroxyl radical (OH·), are highly reactive and have a longer lifetime of about 10–5 seconds, so they can diffuse some distance to the DNA and cause damage. Indirect action is estimated to account for two-thirds of the DNA damage produced by x-ray or γ-rays. Because of its “lengthier” course, indirect action is also the component of biologic effect that can be modified by chemical sensitizers and protectors. Figure 19-2 depicts both direct and indirect action of free electrons produced by ionizing radiation. This DNA damage occurs through physical and chemical processes that occur in fractions of seconds. If unrepaired, it then leads to a cascade of biologic events that may take hours, days, months, years, or even generations to be expressed.
FIGURE 19-2. Direct and indirect action of free electrons created by ionizing radiation. (Reproduced, with permission, from Hall EJ, Giaccia AJ. Radiobiology for the Radiologist. 7th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2012:9, Figure 1.8.)
Biologic Effects of Radiation: Repair and Cell Death
There is strong laboratory evidence that DNA is the principal target for the biologic effects of ionizing radiation. DNA is a very large and long molecule, which consists of 2 complementary strands of sequential bases held together by an alternating sugar and phosphate “spine,” entwined in a double helical structure. Primarily, damage from radiation results in strand breaks in the DNA spine or backbone, although the induction of abnormal cross-links between DNA strands or between DNA and nuclear proteins can also play a role in the loss of normal replication and transcription. DNA strand breaks can be either single- or double-stranded. Single-stranded breaks are considered of little biologic consequence because they can be repaired readily using the opposite strand as a template, assuming no disruption of normal cellular repair mechanisms. Additionally, breaks in both strands that are well separated in the nucleotide base sequence are also readily repaired as independent breaks. However, DNA strand breaks that are directly opposite one another or separated by only a few base pairs lead to double-strand breaks, which are more difficult or impossible to repair and can result in the separation of the chromatin into 2 or more pieces.
It may be noted that cells are generally very efficient at repair of DNA damage, even with double-strand breaks, using a variety of repair pathways and enzymes. However, a great number of double-strand breaks can overwhelm a cell’s repair processes. Individuals with defective DNA repair mechanisms, such as in ataxia telangiectasia, are more vulnerable to the effects of ionizing radiation. Cancer cells themselves often have mutations that can impact their DNA repair pathways; this increased susceptibility to unrepaired DNA double-strand breaks, as compared to normal cells, is one of the factors exploited with radiotherapy.
Unrepaired fragmented chromatin ends may reassort and rejoin other broken ends to give rise to grossly distorted structures or may fail to rejoin, ultimately giving rise to chromosomal aberrations. Several of these aberrations are lethal to cells. Others are not lethal but can lead to carcinogenesis and other mutations.
When exposed to a dose of ionizing radiation, a cell may survive or die. For nonsurviving cells, death while attempting to divide, known as mitotic or reproductive death, is the dominant process. In this case, the unrepaired DNA/chromosome damage in the parent cells does not allow creation of viable progeny cells. Other mechanisms of inactivation include apoptosis or programmed cell death (typically seen in lymphoid cells) and cellular senescence. Nevertheless, the most important end point for radiation-induced lethality is mitotic death, or the cell’s loss of reproductive capability. The hallmark of a cancer cell is its immortalization, or the ability to continuously proliferate. Using this definition of cell kill, a tumor cell may remain physically and morphologically intact for some time, but if it ultimately dies at the time of attempted mitosis, it has lost its reproductive integrity and is no longer clonogenic.
Cell death, or conversely, survival, following exposure to ionizing radiation is often described by means of a cell survival curve. For a single cell type population, typically determined in vitro, the cell survival curve is represented on a semi-logarithmic scale, where the x-axis represents radiation dose and the y-axis denotes the log of the proportion of surviving cells (Figure 19-3). This is described as a linear-quadratic model of cell survival. As noted earlier, cell death is linked to the frequency of unrepaired DNA double-strand breaks and chromosomal aberrations. At very low doses, both strand breaks may be caused by the same electron, in which case the probability of chromosomal disruption and cell death is linearly proportional to the dose. However, at higher doses, the 2 double-strand breaks are more likely to result from 2 separate electrons acting independently and stochastically, in which case the probability of an interaction is proportional to the square of the dose. The linear-quadratic cell survival curve is a simplified but useful model for evaluating radiation effects on proliferating cell systems and is described mathematically as:
FIGURE 19-3. Linear-quadratic cell survival curves for tumor and early-responding tissues compared to late-responding tissues.
where S is the surviving fraction of cells, D is the dose of radiation delivered, α is the linear component of ionizing radiation cell kill, and β is the quadratic component of ionizing radiation cell kill.
The α and β values used in the linear-quadratic cell survival curve model vary by cell type and are measures of the DNA repair capacity of that cell. A high α means that for a given dose of radiation, there is less repair compared to cells with a high β. Because these 2 components of repair exist to a variable degree in most cells, this repair capability is often expressed as the ratio α/β. Cellular systems with high α/β ratios have a “steeper” cell survival curve, indicating less repair, and thus greater cell kill, for a given radiation dose than cells with lower α/β ratios. In general, tumors and acute-responding, rapidly cycling normal tissues (eg, hematopoietic cells, skin, hair follicles, and gastrointestinal mucosa) are considered to have high α/β ratios, approximating 10, whereas late-responding normal tissues (eg, visceral parenchyma and stroma) have lower α/β ratios of approximately 3. This difference in α/β ratios, representing the differential repair of tumor and various tissues to the effects of ionizing radiation, is critical and is exploited when using a course of fractionated radiotherapy of multiple lower-dose applications, rather than a single large dose of radiation (Figure 19-4). During fractionated radiotherapy, assuming sufficient time between doses to allow for repair, the initial slope of the cell survival curve is reprised with each subsequent dose of radiation. This allows the relatively small differences in the single-fraction cell survival curves to be magnified over a protracted course of radiation, allowing greater sparing of late-responding, dose-limiting tissues relative to tumor. Although fractionation may provide relatively little sparing of early-responding tissues, acute toxicity in clinical practice is often supported by compensatory proliferation and migration of neighboring cells from outside the radiation field. The linear-quadratic model and α/β ratios can also be used to “convert” one dose fractionation scheme to its biologic equivalent using a different fraction size.
FIGURE 19-4. Effect of dose fractionation on radiation cell survival. As compared with single large doses, fractionation magnifies the difference in survival, or “sparing,” of late-reacting normal tissues relative to tumor.
A similar corollary to radiation dose size on cell survival, and the relative sparing impact of fractionation, can be seen for the effect of radiation dose rate (Figure 19-5). This is particularly germane in the use of brachytherapy sources.
FIGURE 19-5. Effect of dose rate on radiation cell survival. For a given dose, a higher dose rate results in greater cell kill than lower dose rates.
Beyond ionization, DNA double-strand breaks, and cellular repair capacity, several other factors impact on the biologic effects of radiation. All proliferating cells, including tumors, travel through the cell cycle, with phases that have been defined as mitosis (M), gap1 (G1), DNA synthesis (S), and gap2 (G2). The nature of the DNA molecule itself, the presence of cell cycle checkpoints to accommodate repair time, and the relative levels of repair, replicative, and transcriptive enzymes vary during the course of the cell cycle. In general, cells are considered most radiosensitive in M and late G2 phases, but are most resistant during the S phase. The presence of oxygen during radiation has profound impact on the subsequent biologic effects. It has been postulated that oxygen prolongs or perpetuates the free radical process initiated by ionization, leading to an approximately 3-fold increase in radiosensitivity in the presence of oxygen (known as the oxygen enhancement ratio) compared to hypoxic or anoxic conditions. Because many tumors have aberrant vasculature and regions of hypoxia, the study of oxygen effect on ionizing radiation, as well the potential to enhance tumor oxygenation during radiotherapy, has commanded a significant proportion of basic and clinical research efforts over the past few decades.
Much of the biologic effects of ionizing radiation described earlier relate to low linear energy transfer (LET) radiation, which include photons, electrons, and also protons. High-LET radiation, such as neutron and carbon ions, causes very dense ionization trails in biologic matter, which result in decreased DNA repair and reduction of both the cell cycle and oxygen effects. However, the expense, lack of broad clinical benefit, and often unacceptable normal tissue injury (because reduced repair affects both tumor and normal tissues) of high-LET radiation have led to very limited clinical use and applicability.
Genetic Effects of Radiation and Radiation’s Impact on the Human Reproductive Process
In distinction to the reproductive or mitotic cell death in any tissue that can be induced by ionizing radiation, it also has direct impact on the human reproductive system. These effects can take the form of reduction or ablation of fertility or germline mutations that can be transmitted to future, yet unconceived offspring, or they can affect a developing embryo or fetus in a pregnant woman.
In women, the ovarian dose associated with permanent sterility is age-dependent, reflecting the maximum number of oocytes present at birth (with no further production) and accelerated decrease following the onset of menarche. In the prepubertal female, a radiation dose of 12 Gy or more is generally required to cause sterility, whereas a dose of only 2 Gy may cause the same consequence in a premenopausal woman. Unlike in men, ovarian gonadal function is tightly linked to hormonal production, such that the same doses that cause permanent sterility also result in menopause.
If the female is not rendered infertile, radiation can cause nonlethal damage and mutations in oocyte DNA that may be inherited by subsequent generations, resulting in heritable, or genetic, diseases in offspring. It is important to remember that genetic mutations occur with some baseline frequency in the general population, independent of any radiation, and that not all mutations result in clinically apparent disease. Ionizing radiation does not result in unique or bizarre heritable diseases, but rather increases the frequencies of the same genetic mutations that already occur spontaneously. Hereditary diseases can be classified as single gene–based (eg, Huntington chorea and sickle cell anemia), chromosomal (eg, Down syndrome), or multifactorial (eg, neural tube defects, cleft lip). Information on the added hereditary effects of radiation exposure comes almost entirely from animal experiments; based on mouse data, the dose required to double the spontaneous germ cell mutation rate is approximately 1 Gy (100 cGy).1
The effect on ionizing radiation on an existing embryo or fetus depends on its gestational stage. In the first week to 10 days following fertilization and zygote formation (preimplantation phase), doses as low as 10 cGy can result in loss of the embryo. However, if the embryo survives, it may then develop normally, with few consequences (hence termed an “all-or-nothing” effect). Radiation exposure during the period of organogenesis, from approximately weeks 2 to 6, results in the highest likelihood of congenital malformations (teratogenesis), neonatal death, and growth retardation. After approximately week 6 (fetal stage of gestation), radiation can lead to permanent growth retardation and also mental retardation, as the central nervous system matures later in utero. Although the effect of radiation on the embryo and fetus is dose dependent, there is no threshold dose below which ionizing radiation can be stated to have no impact. Thus, the use of medical radiation in a pregnant patient should be avoided whenever possible or, if absolutely required, be given only after a full discussion of risk and informed consent. Although controversial, it has been suggested that 10-cGy in utero exposure, at least during the early first trimester, be used at the cutoff point beyond which a therapeutic termination of pregnancy should be considered.
Beyond 20 to 25 weeks (third trimester), low doses of radiation to the fetus may be “relatively” safe, although some have reported an increase in childhood malignancies. Obviously, pelvic or abdominal radiotherapy, as is used in gynecologic cancers, results in an unacceptably high dose to the fetus and would be associated with spontaneous abortion or require an evacuation. However, there are many case reports and small series of successful and healthy infant deliveries following radiotherapy to other sites during late pregnancy (eg, for breast cancer or supradiaphragmatic lymphomas). Careful blocking and patient set-up to limit scatter dose to the uterus, medical physics consult to optimize shielding and dose monitoring, and patient informed consent are clearly required.
Ionizing radiation has been shown to cause an increased risk of both solid tumors and hematologic malignancies. The induction of secondary cancer is considered to be a stochastic effect, that is, the probably of occurrence increases with dose, but there is no threshold dose, and the severity of the malignancy, once induced, is independent of the dose. The risk is also age dependent; younger patients may have developing tissues that are more susceptible to radiation carcinogenesis and may have a longer life span in which to manifest the secondary malignancy. There is often a long interval between exposure to radiation and the appearance of the induced malignancy. The shortest latencies are seen in leukemias, which may occur 5 to 7 years after radiation, whereas secondary solid cancers may take 10 to 40 years or more to develop. Radiation-induced malignancies also tend to appear at the same age as spontaneous malignancies of the same type.
Risk estimates of secondary malignancy after therapeutic radiation are somewhat controversial because patients undergoing radiation are often already at a higher risk of developing a second cancer based on lifestyle and/or genetic predisposition. However, studies do indicate that there is an increased risk of radiation-induced malignancies in cancer patients, regardless of underlying biases. Fortunately, the absolute risk of secondary malignancy induction by ionizing radiation is very low. In a recent large-population analysis of more than 485,000 irradiated patient survivors from the US Surveillance, Epidemiology, and End Results (SEER) database, it was estimated that therapeutic radiation resulted in a risk of 3 and 5 excess cancers per 1000 individuals at 10 and 15 years, respectively, after diagnosis and treatment. Germane to the field of gynecologic oncology, radiotherapy did result in a small risk of induced secondary malignancies in patients treated for cervical cancer (hazard ratio [HR], 1.34) and endometrial cancer (HR, 1.14), but this risk disappeared in patients radiated when they were 60 years of age or older.2
Up to about 1950, most external beam radiation was primarily carried out with x-rays generated from kilovoltage machines. These low-energy x-rays suffered from poor penetration for deeply seated tumors and resulted in excess dose deposit on the skin surface, as well as in bones (high photoelectric component of photon interaction). Indeed, skin toxicity was often the limiting factor in delivering adequate doses of kilovoltage therapeutic radiation.
The introduction of cobalt 60 (60Co) therapy units in the 1950s represented a big step forward in radiation oncology. 60Co, a radioactive isotope that is artificially created by neutron bombardment of the stable cobalt 59 in a reactor, undergoes nuclear decay with the emission of 2 clinically useful γ-rays of 1.17 MeV and 1.33 MeV and a half-life of 5.3 years. 60Co radiotherapy units allowed the first routine use of megavolt-age photons, with improved depth dose penetration, skin sparing, and abrogation of excessive bone absorption (via high-energy Compton effect). Other advantages of the 60Co units were the relatively constant beam output, lack of day-to-day output fluctuations, well-defined half-life allowing for predictable decay, and introduction of an isocentric gantry system, in which the radiation source rotated about a stationary patient. However, the quest for even higher energy photons, the poor dose homogeneity for large fields, the need to replace the radioactive 60Co isotope every 4 to 5 years, strict Nuclear Regulatory Commission requirements for shielding and securing of a source that cannot be “turned off,” and costly licensing fees have led to its widespread replacement by linear accelerators (at least in developed countries) over the past 2 to 3 decades.
The linear accelerator (linac) has now become the dominant radiotherapy treatment unit (Figure 19-6). It is a megavoltage machine and uses high-frequency electromagnetic waves to accelerate electrons to high energies through a linear tube, with the electron energy determined by the strength of the accelerating electromagnetic field. The units used to determine the energy of the produced radiation are MeV for electrons and MV for photons—both units refer to the millions of volts in electrical potential difference that is created for electron acceleration in a linac. The accelerated high-energy electron beam can be used by itself for treating superficial tumors, or it can be made to strike a metal target (typically tungsten), within the accelerator head, to produce x-rays for treating deeper tumors. This type of machine is often referred to as a multimodal linac because it can provide multiple electron beam energies (6-25 MeV), as well as 2 or 3 x-ray energies (6, 10, and 18 MV). The effective source of radiation in a linac is mounted on a gantry and can rotate 360° around a stationary patient. The point around which the gantry rotates is called the isocenter. It is typically sited in a patient within a tumor volume so that the target can be readily radiated from multiple different directions.
FIGURE 19-6. A contemporary linear accelerator (linac). Note that the gantry is mounted on a rotational mechanism that allows it to completely rotate about an isocentric point. Opposing the gantry is the electronic portal imaging device (EPID), which allows capture of digital radiographic portal images, predominantly emphasizing bony anatomical landmarks. Perpendicular to the gantry/EPID axis is a computed tomography (CT) imager and detector set, which allows for 3-dimensional cone beam CT (CBCT) of specified soft tissue targets for enhanced image-guided radiotherapy.
Within the accelerator tube, electrons are accelerated to extremely high speeds (and energies) via an electromagnetic microwave field. As the high-energy electrons emerge from the accelerator structure, they are monoenergetic and in the form of a pencil beam of about 3 mm in diameter. To produce x-rays, which are the most commonly used form of ionizing radiation, the electrons are directed toward a water-cooled metal target, typically consisting of tungsten. As the electrons “crash into” the tungsten target, they lose energy, resulting in the production of a spectrum of x-rays with a range of different energies, but with the maximum photon energy equal to the incident electron energy. These x-rays are typically referred to as Bremsstralung radiation, which results from the “braking,” or rapid deceleration, of the high-energy electrons on encountering the metal target. The average photon energy of the beam thus created is approximately one-third of the monoenergetic incident electron energy. Photon energy is designated as MV because of its heterogeneous energy; the MV refers to the maximum photon energy. If electrons, rather than x-rays, are selected for clinical use, the tungsten target of the linac is withdrawn. Instead, as the thin electron beam exits the accelerator structure, it is made to strike a scattering foil to widen the beam to a clinically useful dimension and to get a uniform electron fluence across the treatment field. Electron beam energy is designated by MeV because it is monoenergetic.3
The treatment head of a linac also contains the collimation (or aperture) system for defining, or “shaping,” the radiation beam. The primary collimator is created by pairs of heavy metal jaws (or blocks) that move in perpendicular directions, thus determining the length and width of the radiation field. Previously, secondary collimation to further shape the radiation beam into asymmetric shapes was achieved by poured blocks that were then mounted on the outside of the gantry head. New linacs have built-in secondary collimation that consists of multiple thin leaves, which can be designed to achieve the desired geometry of almost any radiation field shape. This system of multiple-leaf collimators (MLCs) now allows near-infinite computer-assigned positions that can remain static during a radiation exposure or can change shape in real time and, when coordinated with gantry and patient couch rotations, produce highly dynamic, conformal dose delivery such as with intensity-modulated radiation therapy techniques.
A treatment simulator is an apparatus that uses diagnostic x-rays to display the treatment fields so that the target volume may be appropriately encompassed without delivering excessive radiation to the surrounding normal tissues. It duplicates the physical set-up characteristics of the linac itself, in terms of isocentricity and the ability to recreate patient and treatment machine alignments. Historically, simulators were based on fluoroscopic units, with plain radiographs obtained of the region to be treated. Areas to be avoided were then defined on the 2-dimensional films, which were then used as a template for cut blocks to shape the radiation beam. Today, computed tomography (CT) scans have almost completely replaced fluoroscopic simulators in developed countries. CT simulation allows the capture of more anatomic information and creation of 3-dimensional volumes that can be used to refine tumor coverage and normal tissue shielding.
Treatment of gynecologic cancer with radioactive material inserted against, or into, the tumor, has been used effectively for many decades. With rare exceptions, the radiation sources are not placed directly into the patient, but rather are contained indirectly within hollow, specialized applicators that are positioned within the target volume. Early applicators were preloaded with the radiation sources and were thus inserted “hot” into the patient, raising concerns for operator safety. Today, all gynecologic applicators are first inserted without radioactivity, allowing the clinician to take time to achieve optimal geometrical positioning. These hollow applicators are then “after-loaded” with the appropriate radioisotope, either manually (using long-handled equipment) or, increasingly, remotely by specialized machines.
The original radiation source used for gynecologic implants was radium 226 (226Ra). 226Ra itself does not produce any γ-rays appropriate for treatment. However, 226Ra undergoes nuclear decay and transforms into radon gas (radon 222 [222Rn]). The daughter products of the 222Rn actually produce the higher energy γ-rays necessary for effective treatment. Due to these properties, the radium, along with the radon, had to be encapsulated and sealed to produce a suitable brachy-therapy source. 226Ra is a rare, naturally occurring isotope with an extremely long half-life. Several tons of ore material have to be refined to obtain enough 226Ra for a single clinically useful source.
The typical activity for a 226Ra source is 10 to 20 mCi. This corresponds to a mass of 10 to 20 mg of radium by definition (because 1 Ci is defined as the activity, or number of disintegrations per second, of 1 g of 226Ra). Most treatments used between 2 and 5 sources, lasted about 48 hours, and were defined as low dose rate (LDR) brachytherapy. These 226Ra sources had to be hand loaded into the applicators, and the patient was confined to bed during the entire treatment.
Much of our understanding regarding radiation implants, from applicator geometry, dose and dose rate, tumor control probably, and normal tissue tolerance, is based on the historical use of 226Ra. However, the hazards of securing 226Ra and 222Rn and the disastrous consequences if the integrity of a source capsule is compromised became large-scale problems for institutions and regulators. Other radioisotopes were investigated as replacements, but with the intention of mimicking the geometry and dose distribution of 226Ra. With the advent of nuclear reactors, a good clinical substitute was found in cesium 137 (137Cs). This was a readily available reactor by-product refined from spent fuel rods. The sealed 137Cs sources, without a gaseous daughter molecule, were much safer than the 226Ra capsules. To recapitulate the size, shape, and activity of the 226Ra sources, 137Cs sources were calibrated in milligram-Ra-equivalent units (mgRaEq). For example, a 20-mgRaEq 137Cs source did not specifically detail the amount of 137Cs itself in the source; it was simply the amount of 137Cs that gave the same amount of dose and activity, at a specified distance, as the previously used 226Ra source. Because the size, shape, and activity of the 137Cs sources reprised those of its corresponding 226Ra capsule, the applicators generally were unchanged, and source loading and duration of treatment remained the same (ie, continued LDR brachytherapy).
The 137Cs supply has now become increasingly restricted due to decreased reactor production. Additionally, techniques have been developed, using remote afterloading equipment, to effectively treat patients in a much shorter time, removing the necessity of a prolonged hospital stay. This required a very high activity source and a shielded procedure suite. This method of implant radiation delivery has become known as high dose rate (HDR) brachytherapy. The source developed for HDR brachytherapy was iridium 192 (192Ir). 192Ir is activated inside a nuclear reactor, without having to disturb the fuel core or spent fuel rods. To accommodate outpatient-based HDR brachytherapy, a source would have to be highly radioactive, as well as suitably small, such that it could fit in small-diameter tubes. This combination of source attributes is defined by an isotope’s specific activity, which is the maximum achievable activity of an isotope per gram of material. The specific activity of 192Ir is 100 and 10,000 times that of 226Ra and 137Cs, respectively, allowing it to have a very compact size while maintaining high activity. For HDR brachytherapy delivery, the 192Ir source activity typically ranges between 4 and 10 Ci. For this activity range, 226Ra or 137Cs sources would have to be the size of grapes or larger, but 192Ir, with its high specific activity, can be constructed into sources that measure 4 to 5 mm in length and < 1 mm in width, making it eminently suitable for HDR systems. A comparison of the physical properties of the 3 radioisotopes historically and currently used for gynecologic brachy-therapy is provided in Table 19-1.
Table 19-1 Properties of Brachytherapy Isotopes Used in Gynecologic Cancer
In addition to the sealed radiation material discussed earlier, “unsealed” radioisotopes may also be injected parenterally into a patient, with resulting large-volume or whole-body distribution. Of these, phosphorus 32 (typical dose of 15 mCi) was previously used intraperitoneally in the treatment of ovarian and endometrial cancer but has now lost favor secondary to high toxicity and questionable efficacy. Other current systemically administered unsealed radioisotopes include iodine 131 for thyroid cancer and strontium 89 for diffuse bone metastases.
CLINICAL RADIOTHERAPY CONSIDERATIONS: DOSE RESPONSE AND MODIFIERS OF DOSE EFFECT
Radiation Dose Response and the Therapeutic Ratio
The goal of therapeutic radiation is to maximize locoregional tumor control, while minimizing injury to surrounding normal structures. It has been previously discussed that increased doses of ionizing radiation cause increased DNA damage, and corresponding cell death, in both tumors and normal tissues. However, this relatively straightforward laboratory description of radiation effects on isolated cell populations does not take into account that in clinical radiotherapy, the ultimate impact on tumor and normal tissues is influenced by complex interactions among multiple cell types in both cancer and surrounding stroma, as well as volumetric and patient-specific considerations (see later section, Radiation Toxicity).
A useful clinical concept that guides the application of radiotherapy is the dose-response curve. This is an overall function that implicates the probabilistic impact of radiation dose on tumor control probability, as well as on normal tissue injury (Figure 19-7). In general, this function follows a sigmoid-shaped curve. There are 3 components to this curve: (1) a minimal dose below which radiation is “wasted,” with no chance of tumor effect; (2) a steep portion of the curve where an increase in dose is associated with a significant increase in tumor response; and (3) a plateau region where further escalation in dose would result in little incremental tumor control probability. Note that the x-axis does not specify actual doses; it is the relative dose that matters in this simplified concept. Actual doses will depend on the tumor type (intrinsic radiosensitivity), tumor volume, and the outcome being measure (long-term control and cure vs. palliative effect). Likewise, a similar sigmoid curve describes the risk of normal tissue injury. Empirically, the normal tissue injury dose-response curve often “sits to the right” of the tumor control curve, indicating that in many situations, a therapeutic benefit can be achieved at a correspondingly “lower risk” of complications. This difference, which forms the basis for clinical radiotherapy, is expressed as the therapeutic ratio (or index), defined as the probability of tumor control achieved for a specified dose, for a given incidence of normal tissue damage.
FIGURE 19-7. Dose-response curves for control of tumor and late-reacting normal tissues. The therapeutic ratio (TR) is defined as the probability of tumor control achieved, for a specified dose, for a given incidence of normal tissue damage. Sensitization of tumor to radiotherapy results in a “leftward” shift of the tumor control curve and increases the TR, as long as the normal tissue dose-response curve is not likewise shifted or is shifted to a lesser degree.
Much clinical effort has focused on improving the therapeutic ratio, in essence by separating the curves further apart. As noted previously, one approach is to optimize radiation fractionation to exploit the relatively small differences in radiation repair between cancer cells and late-responding normal tissues. Decades of clinical experience have established that, for most tumor systems, a fractionated dose of 1.8 to 2 Gy/d is feasible and effective for tumor control and normal tissue recovery, without the hazard of further prolonging treatment duration if smaller fraction sizes are used. However, other means of modifying the therapeutic ratio, both physically and biochemically, have been tried.
Physical Modifiers of Dose Response: The Impact of Surgery
Radiation is often combined with surgery in the treatment of gynecology cancers. Examples include radiotherapy following hysterectomy for cervical or endometrial cancers and wide local resection and inguinofemoral node dissection for vulvar cancer. By removing the bulk of tumor, the dose of radiation to achieve tumor control may be reduced in the adjuvant setting, in effect shifting the tumor dose-response curve to the left. Furthermore, susceptible normal tissues may be displaced (such as with transposition of the ovaries outside the radiation field). There are 2 caveats to this role of surgery. First, cell survival following radiation exposure is based on a logarithmic basis, and it has been estimated that a dose of approximate 6 to 7 Gy is required to achieve 1 log of clonogenic cell kill in a typical epithelial cancer. Hence, subtotal resection of a tumor (eg, of 50%-80%), or the presence of grossly positive margins, does not significantly impact on the dose of radiation required to maximize tumor control. Using this logic, it may be argued that surgery should remove at least 1.5 to 2 logs of existent tumor cells, or at least 95% to 99% of all disease, to meaningfully reduce the subsequent radiation dose prescribed. Second, surgery may itself cause injury to adjacent normal tissues or displace more normal tissue into the radiation volume (eg, adhesions and/or increased small bowel in the pelvis following hysterectomy), so that the normal tissue damage curve is also shifted leftward, indicating increased risk of late complications. Therefore, careful assessment of patient and tumor characteristics should be undertaken when integrating surgery and radiotherapy in clinical management.
Chemical Modifiers of Radiation Response and the Role of Chemotherapy
The historical definition of a radiosensitizer is that of a compound that in and of itself has little or no direct effect on tumor cells but, when combined with ionizing radiation, significantly enhances the lethality of the radiation. In the laboratory, oxygen is the archetypal radiosensitizer; it is a simple, widely available substance, and the presence of oxygen reduces the dose of radiation required to achieve a specified level of cell kill by approximately 3-fold compared to a hypoxic cell populations (sensitization ratio, or oxygen enhancement ratio, of 3). It was historically thought that the oxygen effect was mediated via promulgation of free radicals initiated by ionizing radiation, based on the affinity of oxygen to accept free electrons and thus contribute to the oxidative process. It was also shown that multiple experimental and human tumors had hypoxic elements, either due to aberrant vasculature or other factors, and this was felt to contribute to clinical radioresistance. In this scenario, molecular oxygen, or a chemical substitute, had to be present exactly at the time of radiation to interact with the free radicals formed. Multiple approaches at radiosensitization by trying to overcome hypoxia were developed in the laboratory and extended to clinical trials over the past 4 decades, including the use of hyperbaric oxygen, red cell enhancement by transfusion and erythropoietin administration, and the use of hypoxic cell radiosensitizers (chemicals that mimic the chemistry of oxygen by having high electron affinity), but unfortunately, these have overall not been found effective. More recent information indicates that the impact of hypoxia in tumor biology and radiation resistance is highly complex. In addition to its acute effect on free radical reactions during radiation, hypoxia induces other biochemical changes (eg, increased vascular endothelial growth factor production and induction of hypoxiainducible factor) that create a cascade of downstream effects that lead to increased tumor aggressiveness and metastatic potential, regardless of whether the inciting hypoxic environment is subsequently corrected.
More recent attempts at radiosensitization have focused on the use of cytotoxic chemotherapy agents concurrent with radiotherapy. The initial rationale for combining radiation and chemotherapy was based on the concept of “spatial cooperation,” where the drug and the ionizing radiation acted independently. This notion recognized that radiation could provide local control, but chemotherapy might address micrometastases not targeted by radiation. In this construct, the impact of chemotherapy and radiation on the local tumor was additive (ie, overall tumor kill was the sum of the effect of the individual modalities of chemotherapy and radiation).
More recent clinical efforts have focused on the potential synergistic interactions between chemotherapy and radiation. Although both modalities retain their independent influence, certain chemotherapeutic agents demonstrate a synergistic, or supra-additive, effect on local tumor control when combined with radiotherapy. This synergism results in tumor control rates that are higher than would be expected from the sum of the effect of each modality if used alone. The specific mechanisms for this interaction are complex and myriad, but they broadly relate to chemotherapy inhibition of repair following DNA damage induced by ionizing radiation. This has now led to the adoption of the term “chemosensitization” when discussing the integrated effects of concurrent chemotherapy and radiation.
For gynecologic cancers, the prototype drug used for chemosensitization is cisplatin. Cisplatin-based chemoradiation has become the standard of care in the management of advanced cervical cancer and has also been used in vulvar and endometrial cancers. Other chemotherapy agents that have shown benefit in combined-modality approaches include carboplatin, 5 fluorouracil, paclitaxel, and gemcitabine. Although the synergistic aspects of these drugs with radiation are clear, it is noteworthy to remember that spatial cooperation may remain an important component (especially in management of micrometastatic disease), because these are also generally the most active cytotoxic agents in gynecologic cancers.
Accelerated Repopulation and Its Detrimental Impact on Prolonged Duration of Radiation
Within a tumor, a variable proportion of the clonogenic cells exist in a rest, or G0, stage. Studies have indicated that during a course of radiation, many of these cells are recruited into the active cell cycle, causing accelerated repopulation of the cancer. This leads to an increased growth fraction of cycling clonogenic cells, which may occur even as the tumor is grossly shrinking. If the treatment ultimately fails to eradicate the clonogenic cells, then tumor persistence, or recurrence, occurs.
The phenomenon of accelerated repopulation implies that radiation, once started, should be completed as expeditiously as possible, with no elective treatment breaks. For example, in locally advanced cervical cancer, which arguably represents the most aggressive/complex use of radiotherapy for a gynecologic malignancy, the recommendation is for completion of all radiation in 8 weeks. The use of concurrent cisplatin-based chemotherapy is indicated during radiation, but care should be taken to monitor and mitigate acute combined-modality effects (including aggressive supportive care) such that treatment breaks for toxicities are avoided or minimized.
TECHNICAL ASPECTS OF RADIOTHERAPY
Teletherapy (tele meaning from afar, Greek origin) is radiation delivered from a distance beyond the body (whether from a linac or a radioisotope such as 60Co) and is now more commonly termed external beam radiotherapy (EBRT). The most common forms of EBRT used are photons, for deep structures within the body, and electrons, for superficial regions that do not extend more than a few centimeters below the skin surface (eg, superficial groin nodes).
For practical purposes, a linac radiation source may be considered a “point source,” as the tungsten target used for producing x-rays is only a few millimeters in size, while the patient is treated, typically, at an isocentric distance of 100 cm. The most direct component of the radiation beam, which exits the source with 0° of deviation and usually reaches the patient at the shortest, perpendicular distance, is known as the central axis, or central ray. Within a radiation field, other portions of the beam diverge from this central axis, with the angle of divergence increasing as the field dimensions are increased.
Any given single beam of external radiation has an identifiable distribution pattern, known as the percent depth dose (PDD) characteristic, or beam profile. This is dependent on the type of radiation (clinically speaking, either photons or electrons), energy of the beam, size of the radiation field, and specific features of the linac that produces the beam. Therefore, these beam characteristics have to be individually measured for each energy, and varying field sizes, for each linac. The PDD of a beam describes the sequential decrease in measured doses at greater depths in tissue, as a result of both the inverse square law (increased distance from the radiation source) and, more importantly, absorption of some of the radiation itself by more superficial layers of tissue (known as attenuation). This PDD characteristic is determined along the central axis of the beam and is usually characterized by absorption in water, which accounts for the overwhelming proportion of human tissue composition. Examples of PDD curves for photons and electrons are shown in Figures 19-8 and 19-9. There are 3 notable features of the PDD curves: (1) the depth of penetration of a beam increases with energy; (2) electrons have a finite depth of dose deposition, whereas photons technically have infinite, but constantly decreasing, depth dose distribution; and (3) higher energy photons show increasing skin and superficial dose sparing. Rudimentary calculations of dose may be accomplished “by hand” from these depth dose distribution curves and measured patient parameters, as was done historically, but this was confined to dose along the central axis, with limited accounting for “off-axis” structures or correction for tissues of nonwater density. In treatment of gynecologic cancers, single photon beams are rarely used. Although high-energy photons have deep tissue penetration, multiple beams are required to appropriately concentrate dose to deep pelvic tumors without overdoing more superficial structures.
FIGURE 19-8. Percent depth dose curves for photons of different energies. Note the greater depth penetration and associated increased skin and superficial sparing with higher energy photons.
FIGURE 19-9. Percent depth dose curves for electrons of different energies. As opposed to photons, electrons have a relatively finite, and short, range in tissue.
To more fully describe beam characteristics within a 3-dimensional (3D) volume, isodose curves are now used, which describe dose distribution at multiple points and depths within the total path of the beam (as opposed to just the central axis), and include the effect of beam divergence and corresponding scatter. The isodose characteristics for each beam are programmed into the dosimetry computer. Contemporary dosimetric planning allows the calculation of 3D dose distribution, integrating the contribution of multiple radiation beams (which may be of differing energies, fields sizes, and angles), when matched to anatomic structures and densities from the treatment planning CT, to create composite isodose distribution within a specified volume (see later section, Dosimetry).
EBRT allows for homogenous dose coverage of large target volumes. In practice, it is used when the volume of disease is diffuse and cannot be covered by brachytherapy alone. This would include nodal groups deemed at risk, as well as bulky primary tumors. For nodal coverage, EBRT can be extended beyond a standard pelvic volume (extended-field radiation) in patients with known extrapelvic retroperitoneal metastases or in patients with multiple pelvic nodes who are felt to be at high risk of harboring occult para-aortic disease. Although providing the widest coverage, a potential downside of EBRT is that all structures within the high-dose radiation volume, including normal tissues, receive essentially the same prescribed dose. The challenge with EBRT is to precisely determine the volume at risk, so that as much normal tissue as possible can be excluded, or shielded. In the management of intact, locally advanced cervical cancer, a key role of EBRT is to shrink the primary lesion to such as size where it can be adequately encompassed by brachytherapy dosing. Other gynecologic sties may not be as amenable to brachytherapy, such as locally advanced vulvar primaries. In this situation, escalation of EBRT may be required to achieve optimal local control, but the volume of irradiated tissue needs to be carefully assessed. To avoid unacceptable normal tissue injury, the volume taken to increasingly high doses must be carefully minimized; this can be accomplished through a technique known as “sequential shrinking field boosts,” whereby the external beam fields are decreased in size as the primary tumor regresses.
Brachytherapy (brachy meaning from a short distance, Greek root) is a term that has persisted in common clinical use and refers to radiation delivered from sealed radioactive sources placed on or within a patient’s body. In the treatment of gynecologic cancers, this is synonymous with “internal radiation.” Brachytherapy allows a high dose of radiation to be delivered locally (to the implanted or juxtaposed target volume), with rapid dose fall-off with small increases in distance. Unlike EBRT, brachytherapy dose distribution is inherently nonhomogenous within a given volume or target, but this rapid decrement in dose, or inhomogeneity, is exploited to escalate central dose while sparing nearby normal tissues.
As noted earlier, brachytherapy now involves the use of specialized applicators placed within a patient, which are then afterloaded, either manually or remotely, with the radiation source(s). The majority of brachytherapy procedures for gynecologic cancers are intracavitary, which means that the applicator is placed within an existing body space, such as the uterine cavity, endocervical canal, or vaginal tract (Figures 19-10 to 19-12). In rare instances where the tumor geometry is bulky, distorted, or asymmetrically oriented with respect to the central pelvis, especially in a posthysterectomy patient, an interstitial approach, where applicator needles are pierced into tumor stroma or connective tissue, may be required (Figure 19-13).
FIGURE 19-10. Components of a low dose rate, manually afterloaded Fletcher-Suit tandem and ovoids set, used for brachytherapy in intact cervical cancer. To accommodate variable tumor and patient geometry, tandems of variable angles and ovoid caps of different sizes are provided.
FIGURE 19-11. Anterior radiograph of a tandem and ovoids implant within a patient. The approximate position of the uterus is illustrated. Dose points corresponding to points A, points B, and bladder and rectum are defined.
FIGURE 19-12. Vaginal cylinder applicator. A. Assembled high dose rate vaginal cylinder, with cylinder segments of varying diameters selected clinically. B. Radiograph of vaginal cylinder placement within a patient for adjuvant therapy of the vaginal cuff following hysterectomy for early-stage endometrial cancer.
FIGURE 19-13. Interstitial brachytherapy apparatus. A. Perineal template-guided interstitial needles, combined with a central vaginal obturator/tandem. B. Radiograph of a manually afterloaded, low dose rate interstitial implant in a posthysterectomy patient with bulky vaginal canal tumor recurrence. The volume of intended tumor coverage is emphasized on the image.
Historically, brachytherapy was performed using LDR exposure, with typical dose rates of 0.4 to 2.0 Gy/h to a selected point or volume. Treatment times for LDR brachytherapy range from 1 to 4 days and require the patient to be hospitalized and bed-bound for the entire duration. In the past 2 decades, HDR brachy-therapy has gained widespread use in gynecologic and other cancers. With dose rates exceeding 12 Gy/h and radiation exposure times of only a few minutes, HDR brachytherapy can be performed in an outpatient setting and holds the promise of greater patient comfort, avoidance of prolonged bed rest, and possible better implant stability and reproducibility compared to protracted LDR exposures. Initial concerns were raised over the potential radiobiologic disadvantages of HDR brachytherapy (specifically in relation to normal tissue injury, because higher dose rates of radiation allow less cellular repair; see Figure 19-5). However, broad clinical experience and 4 randomized trials in intact cervical cancer have not substantiated this fear, and it is now generally accepted that HDR brachytherapy, when performed with care, proper fractionation, and technical attention to details, is a safe and effective alternative to traditional LDR approaches. An example of HDR intrauterine brachytherapy, using a tandem and ring applicator, with 3D dosimetry planning is illustrated in Figure 19-14.
FIGURE 19-14. Anterior (A) and lateral (B) radiographs of high dose rate (HDR) intracavitary brachytherapy in intact cervical cancer, using a tandem and ring applicator. C. Computerized dosimetry planning allows viewing of the dose distribution in multiple planes. Note that this HDR application reprises the “pear-shaped” dose distribution of traditional low dose rate brachytherapy by using only the lateral source dwell positions in the ring colpostat.
Recognizing that historical understanding of tumor dose response and normal tissue injury risk are based on LDR experience, recommendations in textbooks and published guidelines continue to use LDR-based doses. Because HDR dose rates are intrinsically more biologically effective, it would be a mistake to apply these LDR-based dose guidelines to HDR brachy-therapy without an appropriate dose modification. This conversion, from an HDR dose to an “LDR-equivalent” dose, or vice versa, can be achieved using a modification of the linear-quadratic equation previously discussed:
BED = Total dose [1 + dose per fraction/α/β]
where BED is the biologically effective dose and α/β is the measure of cellular radiation repair (~10 for tumor and acute-responding normal tissue and ~3 for late-responding normal tissues).
Brachytherapy may be used by itself, independent of any EBRT, for management of gynecologic cancer. The most common example is when vaginal brachy-therapy alone is given to reduce the risk of isolated cuff failures following hysterectomy for endometrial cancer. Additionally, tandem-based brachytherapy is sometimes used alone to treat nonoperable patients with small, very early stage cervical cancer.
Brachytherapy, combined with EBRT, is considered a critical component of radiation for locally advanced, intact cervical cancer. Historically, and persisting to today, brachytherapy dose is specified using the point A system. Point A identifies a paracervical location 2 cm superior and 2 cm lateral to the external cervical os, along the axis of the intrauterine tandem (Figure 19-15). In a highly simplified way, the dose at point A serves as a surrogate for minimal lateral dose to a cervix of no more than 4 cm in diameter. Point B identifies a lateral parametrial location, defined as 2 cm superior to the external os and 5 cm lateral to the midline of the patient. Specific designations for representative rectal and bladder dose points have also been developed by international consensus.
FIGURE 19-15. Schematic definitions of points A and B for intact cervix brachytherapy.
Although much has been made about the specific technique and system of brachytherapy for intact cervical cancer, it can be argued that whether intracavitary brachytherapy is used may be just as important, or even more so, than how it is placed. What is clear is that the introduction of radioactive sources into the endocervical and endometrial canal permits a central tumor dose intensification that is unachievable by any EBRT technique, regardless of the nuances of source placement (Figure 19-16). It is this very high radiation dose, placed in direct juxtaposition to tumor, within a radiation-tolerant organ, that allows for control of even large cervical cancers.
FIGURE 19-16. The central dose escalation achieved when combining intracavitary brachytherapy with external beam radiotherapy in intact cervical cancer. The “0” on the x-axis marks the position of the tandem, with dose fall-off as one moves laterally away from the tandem. This illustration assumes “standard” dose prescriptions of 45 Gy to the whole pelvis, 40-Gy low dose rate equivalent to point A using brachytherapy, and bilateral parametrial boosts of up to 9 Gy (with shielding of midline structures).
The dose distribution from a typical intracavitary implant for cervical cancer is centrally symmetrical and classically described as “pear-shaped.” Recent efforts in image-guided brachytherapy, in particular using magnetic resonance imaging (MRI), have suggested that improved tumor control and decreased normal tissue injury may be achieved in some cases where the dose distribution is deviated from this traditional pattern and biased toward an asymmetrical region of residual tumor. However, these early experiences remain to be validated in routine clinical use, and concerns remain that such image-guided approaches would result in underestimation of the residual volume at risk or result in underdosage of tumor due to strict but indiscriminate adherence to normal tissue constraints.
In its most elemental form, radiation treatment planning is the technical calculation of dose (in Gy) to be delivered to a region in the body. However, this definition of treatment planning excludes the myriad variables that have to be considered in radiotherapy planning. The contemporary treatment planning process encompasses a series of steps, performed both sequentially and iteratively, to arrive at a comprehensive plan that will deliver specified doses(s) to carefully defined target(s), while minimizing exposure to adjacent normal tissues, often based on different correlative imaging tests and taking into account variable patient presentation, anatomy, and the impact of multimodality interventions.
The initial question posed in this expanded concept of treatment planning is whether the patient should be considered for radiotherapy based on clinical evaluation of tumor type, stage, location, other therapeutic interventions, and patient comorbidities. Localization of the tumor and/or the region(s) at risk follows and includes integration of physical examination findings, assessment of surgical and pathologic reports, and often, tomographic imaging studies.
A major recognizable part of treatment planning is simulation, where patient anatomy and geometry are “captured” for further computerized evaluation and manipulation. Although fluoroscopic simulation is still used at times, CT simulators have become standard equipment in most radiation oncology centers. In addition to obtaining detailed 3D anatomic information, CT simulation also provides the tissue densities (in Hounsfield units) that determine radiation absorption characteristics used in subsequent dose calculations. Although a patient may perceive CT simulation as similar to a diagnostic scan, implicit in the simulation process is specific patient alignment and immobilization, such that the same position can be recapitulated each day during external beam therapy. The patient is usually simulated while supine; however, when clinically indicated, a prone or different position can be used, as long as that position is deemed reproducible on a daily basis. For pelvic malignancies, the lower extremities and/or torso are typically fitted to a custom-formed cradle to assist in daily positioning and setup. Often, oral and/or intravenous contrast, as well as skin or internal markers, are used to help ascertain structures or volumes at risk (eg, a thin wire around a suspicious groin node or a cervical or vaginal clip to define the most inferior extent of a tumor). Simulation is usually set up to accommodate isocentric radiation treatment, centered on an axis or point around which the gantry, collimators, and couch of the linac will rotate. This isocenter is projected onto the patient surface by light lasers built into the CT simulator or within the imaging room. At this point, several skin marks are typically made on the patient, corresponding to the isocenter projections, and are used for daily treatment alignment. To prevent the marks from being erased, small permanent tattoos are usually etched to define the skin alignment points.
The information and images collected from simulation are next transferred to a work station for computerized dosimetric planning. Within the dosimetry computer, all relevant information (ie, isodose distributions) for each radiation beam that may be used on a specific linac has been programmed (see earlier section, Teletherapy, for further details). As noted, ultimate radiotherapy dose distribution also depends on the CT images and the associated Hounsfield units that determine radiation absorption by tissues of different densities.
Volumes, or regions of interest (ROIs), are next defined within the dosimetry program, corresponding to tumor, pathway(s) or volume(s) of potential spread, and selected adjacent normal tissues. Several nomenclature terms have been coined that help in standardization and evaluation of dosimetric variables. The gross target volume (GTV) represents tumor that can be seen and/or palpated (eg, the grossly apparently tumor and bulky nodes in a locally advanced cervical cancer). It is important that physical examination findings, such as vaginal mucosal extension, which may not be well defined on imaging, be carefully integrated into GTV determination. At times, other imaging modalities, including PET or MRI, provide additional useful information, and these other imaging studies can be coregistered spatially and fused to the CT simulation scan to improve delineation of the GTV. A clinical target volume (CTV) is next defined and includes the GTV as well as possible pathways of spread that may harbor microscopic disease (eg, nodal echelons, parametria, vagina). The CTV is not simply a concentric expansion of the GTV, as routes of potential spread in gynecologic malignancies often follow asymmetrical paths away from the primary tumor. Even where there is no GTV (such as in a patient who has undergone radical hysterectomy for a node-positive cervical cancer), a CTV should be defined that encompasses vaginal, parametrial, and nodal tissues at risk. Finally, a planning target volume (PTV) is created, which includes the CTV plus a selected margin, to account for daily set-up variability as well as patient and internal organ mobility. Separate ROIs are also drawn on the computer that correspond with adjacent normal structures, such as rectum, bladder, small bowel, and femoral heads. These normal tissues are sometimes collectively referred to, in dosimetric parlance, as organs at risk (OARs).
Once the various volumes of interest are defined, radiation beams, with appropriate blocking to exclude as much normal tissue as possible while still maintaining tumor coverage, are created. These beams, or fields, are modeled within the dosimetry program and are later recreated on the actual patient on the linac. With the exception of very superficial structures that may be treated with single-field electrons (eg, skin metastases, groin nodes in a slender patient), 2 or more fields are typically used daily in the treatment of gynecologic cancers. When evaluating a plan, further adjustments in beam angle, energy, and blocking can be performed iteratively to refine the radiation dose distribution. In the past, blocking was accomplished on the linac by customized cut blocks; this has now been mostly replaced by MLCs built into the gantry head that can be programmed and shaped to the desired field geometry. Using computerized planning, the patient can be viewed at any selected beam angle, referred to as a beam’s eye view, which may facilitate the clinician’s ability to appreciate tumor and normal tissue juxtapositions distinct from a typical anterior or lateral projection.
With contemporary computer-based dosimetry, dose distribution can be evaluated by viewing isodose lines in multiple projections within a patient. An isodose line indicates a specific dose level within a volume of tissue, much like an isotherm line shows a particular temperature in a geographic region. A series of isodose lines thus show regions of varying doses, as well as their anatomic location and extent. This information can be further supplemented by dose-volume histograms (DVHs), which provide quantitative information regarding the doses received, as a function of fractional volume, by each of the defined targets and neighboring critical structures. The goal is to deliver uniformly high doses to the entire tumor target, while minimizing the dose and/or the volume radiated of the adjacent normal OARs. Figure 19-17 illustrates isodose distributions and a corresponding DVH in a patient treated for advanced pelvic cancer. Figure 19-18 shows isodose curves for a patient with metastatic cervical cancer to the anterior chest wall, demonstrating the use of a single superficial electron beam to avoid deeper underlying structures.
FIGURE 19-17. A. Transaxial isodose distribution in a patient treated with anterior-to-posterior and posterior-to-anterior (APPA) technique for a T4 rectal cancer with invasion into the vagina and bladder. The red-shaded region represents gross tumor volume (GTV) (primary tumor and right iliac nodes). Other delineated regions of interest include bladder (yellow), small bowel (pink), and internal and external iliac vessels as surrogates for nodal basins. Given the extent of the tumor in a thin patient, it was noted that a 4-field approach would not result in significant added sparing of normal tissue (and in fact would expose more pelvic marrow). B. The corresponding dose-volume histogram showing full prescribed dose to the GTV (red) and the vagina (green), with “sparing” of dose to the uninvolved rectum (brown) and small bowel (pink).
FIGURE 19-18. Single anterior electron field used to treat a chest wall metastasis from cervical cancer, illustrating the superficial dose distribution. Isodose curves are shown in the transverse (A) and sagittal (B) planes.
A historic technical approach for radiotherapy used opposing beams of radiation delivered anterior-to-posterior and posterior-to-anterior relative to the patient (often referred to as an APPA approach). To provide better dose distribution within the pelvis, current conventional radiation field arrangements often uses 4 incident photon beams, which typically include an anterior, a posterior, and 2 opposed lateral beams. This is often referred to as a 4-field box technique, although it should be emphasized that patient transaxial geometry is not routinely “box-shaped,” and the beams do not have to be limited to specific right angles. An example of 4-field pelvic radiotherapy is illustrated later in Figure 19-20. The 4-field approach may allow for some increased sparing of normal tissues compared to an APPA approach, including the distal rectum or small bowel in the anterior portion of the pelvis. However, an APPA approach may still be found preferable in some circumstances, such as in a slender patient with limited bowel anterior to the uterus (see Figure 19-17) or in a very obese patient with severe lateral redundant tissues. The use of multiple radiation beams, when designed with specific blocking for individual fields, together with CT-based dosimetry calculations and evaluation of a DVH, is often referred to as 3D conformal treatment planning.
FIGURE 19-20. Transaxial isodose distributions for a patient undergoing adjuvant external beam radiotherapy following radical hysterectomy for stage I cervical adenocarcinoma, but with high-risk pathologic features. The intensity-modulated radiation therapy (IMRT) plan on the left is compared to its corresponding traditional 4-field plan on the right, for selected levels in the upper pelvis (A), mid pelvis (B), and low pelvis (C). The clinical target volume is shaded in pink. Note that higher intensity isodose curves provide greater “shaping” or conformality surrounding the target volume, but the lower isodose lines are more spread out, secondary to the increased number of incident beams used with IMRT (B).
A critical component of treatment planning is the prescription of radiation dose. Obviously, the tumor targets, as defined by the GTV, CTV, and PTV, should be taken to “full” dose, where possible. It is important to note that several doses can actually be stipulated; for example, the GTV often is taken to a higher dose level, whereas the remainder of the CTV (the portion with presumably only microscopic disease) may be prescribed a more moderate dose. The impact of other interventions, including surgery and/or chemotherapy, and patient medical history (eg, inflammatory bowel disease, severe vascular compromise) influence decisions regarding dose. This recognizes that tumor and late normal tissue effects from radiation are not “all-or-nothing,” but rather exhibit dose-response effects that may overlap to some degree (see Figure 19-7).
Linac Radiation Delivery and Treatment Set-Up Verification
When dosimetry planning is completed and approved, the patient is brought to the linac and positioned/immobilized as per the simulation, and the computer-designed radiation fields are then recreated on the patient, typically focused to the skin tattoos that define the treatment isocenter. Even when aligned to skin marks, further verification of accuracy in final targeting is obtained intermittently (at least weekly) by portal imaging, which compares internal landmarks (usually bony structures) on a megavoltage image taken by the linac, correlated with the x-ray or digitally reconstructed radiograph obtained during simulation. The use of radiographic film for port verification has now mostly been replaced by electronic portal imaging devices (EPID). EPID uses a silicon flat panel detector, built onto the linac itself, to capture portal images, provides software for digital enhancement and manipulation of image contrast, and allows the radiation oncologist to easily compare intended versus actual planned radiation fields simultaneously on a computer monitor (Figure 19-19). As radiation delivery has become ever more focused and anatomically constrained, with reduced margin for localization errors, some have proposed that patient field alignment based on bony landmarks is insufficient and does not account for soft tissue mobility and structural deformation within the target volume. To accommodate this increased level of precision in radiation targeting and delivery, newer linacs have built-in kilovoltage x-ray units that can be used to obtain a tomographic image of the volume of interest, including relevant soft tissue, which can then be fully compared to the 3D information from the original CT simulation. This tomo-graphic imaging obtained on the linac is created using cone beam CT (CBCT) technology. The location of the EPID and CBCT apparatus on a contemporary linac is shown in Figure 19-6. Although much of radiotherapy is clearly defined by imaging, even when using “plain” radiographs of bony anatomy, the use of tomographic imaging on a regular basis for reconciliation with the treatment planning CT is now known specifically as image-guided radiotherapy.
FIGURE 19-19. Standard patient alignment check using pelvic bony anatomic landmarks, comparing the computed tomography simulation–generated digitally reconstructed radiograph (DRR) (left) with the digitally captured megavoltage portal image from electronic portal imaging device (right). Note that secondary to the Compton effect of megavoltage photons, the portal image has less contrast compared to the kilovoltage-based DRR (photoelectric effect), but new technology allows digital enhancement of the portal image to improve viewing resolution.
It is of interest that radiation dose is rarely measured in a patient herself. The beam characteristics used in dosimetry planning are calibrated regularly on the linac by medical physicists, using water or other anthropomorphic tissue-equivalent phantoms. However, for highly complicated set-ups, such as when patient anatomy is highly irregular and heterogeneous or in cases of reirradiation where extra precision is required, special dosimeters (eg, ion chambers, thermoluminescent detectors) can be placed on, or in, a patient while receiving treatment on the linac, and the actual delivered dose can be measured to confirm that it corresponds to the dosimetrically planned dose.
Much of the preceding discussion has focused on the delivery of EBRT. However, given the importance of brachytherapy in the management of many gynecologic cancers, comprehensive treatment planning also needs to account for, and integrate, the internal radiation contribution. In particular, for patients with locally advanced cervical cancer, optimal treatment requires a balance between external beam and brachytherapy components, with the relative weight of each determined by consideration of multiple tumor and patient factors. The dose components delivered by EBRT and brachytherapy are sometimes added together to give a cumulative overall dose, but this can be done only for well-defined “surrogate” points or locations, such as at points A or the bladder and rectal points as designated by the International Commission on Radiologic Units (ICRU).
Adaptive (Dynamic) Treatment Planning
Currently, treatment planning mostly follows a sequential algorithm as described earlier, with the plan developed first, and then radiotherapy delivered in adherence to it throughout the entire course of treatment. However, there are nascent efforts in the development of adaptive treatment planning, in recognition of the fact that a plan created on a “static” set of images at a single time point prior to initiation of radiation may not accurately reflect subsequent alterations in patient anatomy, cancer regression leading to changes in volume and geometry, and variations in the relationships of tumor and normal tissues. In this situation, simulation and treatment planning may be repeated at various intervals, and radiation dose and volume may be adapted to the changes observed. Adaptive treatment planning is being investigated for both EBRT and brachytherapy.4,5
Intensity-Modulated Radiation Therapy
One of the most important recent technical advances in radiation delivery has been the development, and subsequent widespread adoption, of intensity-modulated radiation therapy (IMRT). Although most widely applied in prostate, head and neck, and central nervous system tumors, IMRT has found increasing use in gynecologic cancer radiotherapy.
An underlying premise for the use of IMRT is that gynecologic cancer targets are juxtaposed against, or surrounded by, multiple critical normal structures, including rectum, bladder, small bowel, femoral head and necks, and the pelvic bones. Given their anatomic relationships, a 4-field pelvic “box” technique often cannot maximally exclude these normal tissues, and everything within the “box” is radiated to the same dose as the tumor target. It would thus be attractive to have a technique that would permit “bending and shaping” of the radiation isodose lines to better conform to irregular contours. IMRT provides such a dosimetric capability.6
IMRT has been made possible through advancements in high-definition tomographic imaging (including fusion of different imaging modalities), dose algorithm computing, and linac technology (featuring computer-controlled MLCs that can continually change aperture shape in “real time” during an ongoing treatment). Precise definition of all target volumes and OARs is required at the outset of treatment planning. Multiple fields, typically 7 to 9, are used. Each beam is divided into multiple beamlets using the shifting MLCs, such that the fluence profile of each beam is not homogenous (unlike that in standard radiotherapy), and no 1 beam includes the entire target volume in its entirety. Traditional beam blocks are not defined by the clinician. The dose is not prescribed to a single point or even 1 fixed volume, but rather is calculated based on a “best fit” of multiple dose-volume objectives. Doses are specified for each target and may be at different levels for the GTV compared to the CTV. Doses to the OARs are designated with reference to the fractional volume of the normal structure that is permitted to receive that dose level. Once the dose-volume objectives and constraints are specified, the dosimetry program performs inverse planning, determining the linac output and field shapes that will result in a best-fit dose distribution to the defined parameters. In this setting, intuitive control of the radiation field geometry, as well as dose per field (or dose per beamlet within any field), is surrendered to the computer. Although each beam is heterogenous in dose distribution relative to the tumor target, the integrated doses from all beamlets and all beams allow for a highly conformal, shaped distribution that “bends” or “curves” around a nonuniform target contour.
There are challenges to IMRT, including the limitations of various imaging modalities in defining target volumes. Whereas tight isodose lines surrounding a target allow for exclusion of more adjacent normal tissues, they also permit less latitude in tumor coverage margins and run a higher risk of marginal misses. Patient positioning reproducibility and internal organ motion need to be accounted for. IMRT involves more physician and dosimetrist planning time and also requires stringent physics quality assurance of the plan and individualized phantom measures of the dose for each patient. Finally, although IMRT allows excellent shaping of the higher level isodoses, the lower dose levels are more diffusely spread out (given the multiple entry beams), resulting in larger volumes of tissues exposed to lower doses of radiation.
Despite the added technical demands imposed by IMRT, there is emerging consensus regarding its role in the posthysterectomy setting (where the target volume presents a highly concave geometry within the pelvis) for dose escalation to gross para-aortic nodes or bulky sidewall disease or for selected cases of reirradiation. Figure 19-20 illustrate 4-field versus IMRT isodose distributions for a posthysterectomy patient, with the corresponding DVHs shown in Figure 19-21.
FIGURE 19-21. Comparative dose-volume histograms corresponding to the intensity-modulated radiation therapy (IMRT) (solid lines) and 4-field (dotted lines) plans for the patient illustrated in Figure 19-19. A. For the tumor targets, including the clinical target volume, pelvic nodes, and vagina, there is no discernible difference in radiation coverage. B. However, for the organs at risk, there is noticeable sparing (in dose and in volume) of rectum (brown), bladder (yellow), and small bowel (pink).
The effects of ionizing radiation on the embryo and fetus are covered earlier, in the section entitled Genetic Effects of Radiation and Radiation’s Impact on the Human Reproductive Process.
Effects of Radiation on Normal Tissue
Radiation treatment side effects can be divided into acute (during treatment until 3 months afterward) and late or long-term toxicities (beyond 3 months). Radiation reactions are related to the specific normal structure radiated, as well as to the dose, dose per fraction, and volume of normal tissue exposed. Patient comorbid factors, such as hypertension, diabetes, and inflammatory bowel conditions, affect radiation toxicity. Concurrent chemotherapy and smoking also impact on radiation effects.
It has generally been felt that the acute and late complications of radiation are “decoupled.” Early effects relate to acute-reacting normal tissues with rapidly dividing cell systems, such as skin, hair, gastrointestinal mucosa, and hematologic elements. These acute “irritative” symptoms are expected in most patients undergoing radiation and are exacerbated by the use of concurrent chemotherapy. However, when appropriately managed and supported, acute toxicity is expected to resolve within several weeks after completion of radiotherapy. It has also been discussed previously that although fractionation may provide relatively little sparing of early-responding normal tissues, acute toxicity in clinical practice is often “self-limited” by compensatory proliferation and migration of neighboring normal cells from outside the radiation field.
Late complications, on the other hand, are considered secondary to the depletion of normal tissue stem cells, as well as chronic microvasculature injury. These can occur months to year following radiation and are marked by decreased blood supply, stromal changes, and fibrosis. Late toxicity is generally felt to be independent of acute effects, except in cases where the acute complication is so severe that it does not sufficiently recover and then persists into a late effect (termed “consequential late effect”). The incidence of severe late complications in clinical radiotherapy is fortunately relatively infrequent. However, recent research has focused on lower-grade late changes that, although not catastrophic, may be much more frequent and impact on patients’ long-term quality of life. Although concurrent chemotherapy does increase acute radiation toxicities, most studies have indicated that it does not increase late effects compared to radiotherapy alone.
Potential acute side effects that can occur during EBRT to the abdomen and/or pelvis include fatigue; decreased appetite; radiation dermatitis or hair loss in the treatment fields; cystitis causing polyuria, dysuria, or hematuria; bladder spasms potentially resulting in urinary incontinence; enteritis manifesting as diarrhea, nausea, or vomiting; proctitis resulting in rectal pain, bleeding, or incontinence; vaginal mucositis causing pruritus, vaginal discharge, or discomfort; and hematologic or renal dysfunction. In many cases, these acute reactions respond well to supportive measures, including topicals, medications, and at times intravenous fluid and electrolyte replacement. As noted previously, all attempts should be made to complete the course of radiation or chemoradiation without prolonged treatment breaks, to avoid the hazard of accelerated tumor repopulation.
Potential long-term side effects of EBRT to the abdomen/pelvis include infertility, permanent skin changes or hair loss in the treatment fields, subcutaneous induration, vaginal and cervical stenosis, small bowel malabsorption or obstruction, stricture/fistula of bowel or bladder, femoral head necrosis, osteoporosis and pelvic insufficiency fracture, permanent kidney dysfunction, transverse myelitis, and the risk of a secondary malignancy. The generally low risk of high-grade late toxicity is increased in patients with pre-existing vascular disease (such as hypertension or diabetes) and in patients who have undergone previous surgery within the irradiated volume (presumably secondary to vascular compromise and/or the formation of adhesions). The risk of a permanent, severe complication varies according to the dose and volume of radiation delivered; for patients who receive moderate doses of adjuvant radiation, the incidence of a significant late injury is of the order of 5%, but this risk can increase to 10% to 15% following very high–dose radiotherapy, such as that required in definitive treatment of locally advanced cervical cancer.
The brachytherapy component of radiotherapy has its unique set of possible side effects, including the risks associated with sedation/anesthesia and the hazards of the procedure itself, including infection, bleeding, and uterine perforation.
Late radiation reaction is affected by the structural and functional organization of the tissue/organ in question. Conceptually, normal structures are organized into functional subunits (FSUs), which impact on perceived clinical complication by the patient, and influence the volume of an organ that can be “safely” irradiated without global harm. The FSUs of the spinal cord and peripheral nerves are arranged serially, such that the integrity of each link is critical to maintaining function throughout the chain. Radiation effect on such tissues shows a binary response, with a threshold dose below which there is normal function, but above which there is clearly appreciated and clinically relevant loss of function. The liver, lung, and kidney, on the other hand, have FSUs arranged in parallel. Inactivation of a proportion of parallel FSUs may not lead to whole-organ dysfunction, and thus, these organs have an increased tolerance to high-dose partial-volume radiation (similar to pulmonary lobectomy, partial hepatectomy, or partial nephrectomy). Clinically relevant functional damage to parallel systems does not occur until a critical number of FSUs are inactivated, implying a threshold volume below which no global functional injury is measured, but above which damage occurs as a graded response. The bowel and skin have no well-defined FSUs but are considered to share characteristics of both serial and parallel FSUs.
Tolerance of Organs to Radiation
Table 19-2 shows the widely accepted normal tissue tolerances of organs typically radiated in gynecologic malignancies. The data are based on the general consensus among expert radiation oncologists in the early 1990s.7 Toxicity dose (TD) 5/5 is the dose that is estimated to result in a 5% risk of severe adverse outcome at 5 years. TD 50/5 refers to the dose that would cause an estimated 50% risk of high-grade injury at 5 years. The tolerance doses are presented as a function of the fractional volume of the normal organ irradiated, either 1/3, 2/3, or 3/3. The tolerance doses for skin are quantified in relation to the area of skin exposed, and for the spinal cord, to the length irradiated.
Table 19-2 Tolerance Doses of Organs Typically Radiated in Gynecologic Malignancies
These normal tissue tolerance estimates have guided clinical radiotherapy for several decades but may be considered conservative, given the relative lack of shielding, dose conformality, and sophisticated dosimetry that was available at that time. More recently, the QUANTEC study (Quantitative Analysis of Normal Tissue Effects in the Clinic) was released, which suggests that potentially higher doses may be tolerable in several normal tissues.8
With particular relevance to brachytherapy, experience suggests that the upper one-third of the vaginal mucosa can tolerate radiation doses as high as 120 to 140 Gy, whereas the lower third of the vaginal mucosa should usually receive no more than 80 to 90 Gy. The combined dose of external beam and brachytherapy to the ICRU bladder reference point should not exceed 75 to 80 Gy, and the ICRU rectal point should ideally be limited to 70 to 75 Gy to minimize the risk of late complications.
CLINICAL USES OF RADIATION IN GYNECOLOGIC MALIGNANCIES
The role of radiotherapy in cervical cancer is fairly well defined. Surgery is typically the standard of care for patients with early-stage disease (stage IA, IB1 tumors). However, for patients who have major illness precluding an operation or who decline surgery, radiation is an effective alternative. For patients with stage IA or small IB tumors (typically < 2 cm), brachytherapy alone is an option, which might allow a medically compromised individual to be spared the potential toxicity of EBRT. When brachytherapy alone is used, doses of 70 to 80 Gy (LDR-equivalent specification) to point A are usually prescribed, depending on the volume of presenting disease.
For patients undergoing primary surgery, surgico-pathologic risk features may dictate the need for further adjuvant radiotherapy to reduce the likelihood of relapse. Conventionally, such patients are grouped into 2 risk categories. Those deemed at high risk (positive lymph nodes, parametrial invasion, or positive surgical margins), where collectively, the cancer may be considered to have extended beyond the anatomic limits of the cervix, have demonstrated survival benefit with adjuvant pelvic radiation (generally to a dose of 50 Gy) and concurrent cisplatin-based chemotherapy. For those at intermediate risk, where the pathologic findings remain “confined” to the cervix itself (large primary tumor size, deep cervical stromal invasion, or lymphovascular space invasion), adjuvant pelvic radiation alone to 50 Gy is appropriate and associated with decreased relapses and improved progression-free survival. Brachytherapy is typically not used in the posthysterectomy adjuvant radiation setting unless there are positive vaginal mucosal surgical margins, in which case a cylinder boost to the vaginal cuff, in addition to EBRT, may be considered.
For patients with more locally advanced disease (stage IB2 to IVA tumors), the standard of care is considered to be definitive chemoradiation. This entails the careful combination of EBRT and brachytherapy boost to deliver high doses to the cervix (point A dose of 80 to 85 Gy or higher), while limiting exposure (both in terms of dose and volume) to adjacent critical normal structures such as the bladder and rectum.
Hysterectomy remains the primary initial therapy for almost all patients with uterine cancer. Over the past 2 decades, the use of radiation in endometrial cancer has appropriately diminished. This decrease has been predicated on the impact of surgical staging, a better understanding of patterns of failure, and an appreciation of the contribution of systemic chemotherapy in advanced disease. Nonetheless, radiation continues to play a considerable role in the management of endometrial cancer.
For the infrequent medically inoperable patient, primary radiotherapy, using brachytherapy alone or in combination with EBRT, allows good relief of symptoms (such as bleeding or pain) and provides for a reasonable likelihood of local tumor control. The intrauterine brachytherapy component of therapy may need modification from that used in cervical cancer (where dose is concentrated in the cervix and lower uterus by tandem and colpostat applicators), because endome-trial cancer typically involves more of the superior fundus and is often associated with an enlarged uterine cavity. In this setting, the use of a double or split tandem set-up allows for broader dose distribution in the superior fundus.
In posthysterectomy patients with uterine-confined disease, the presence of certain pathologic features (high grade, deep myometrial invasion, cervical extension, and lymphovascular space infiltration) predicts for an increased risk of local failure, of which the majority are limited to the vaginal cuff. Although historically managed with EBRT, contemporary evidence suggests that most can be effectively and equally well treated with vaginal brachytherapy alone, with excellent prevention of cuff relapses and little associated acute and late toxicity. Brachytherapy, typically using a cylinder applicator (see Figure 19-12), delivers a high dose to the vaginal mucosa, but with rapid fall-off, exposure to bladder, rectum, and other pelvic tissues is minimized. Historically, LDR brachytherapy delivered a dose of approximately 60 to 70 Gy to the vaginal mucosa over 48 to 72 hours as an inpatient. Today, most vaginal brachytherapy is done with outpatient HDR; a common dose scheme prescribes 7 Gy for each of 3 cylinder insertions, with dose specified to a depth of 0.5 cm deep to the mucosal surface. A typical HDR vaginal brachytherapy application is often completed in an hour or less.
For patients with more advanced uterine cancers (stage III and IV), the use of chemotherapy has, to a great degree, supplanted the historical role of radiotherapy, especially when one considers the use of a single adjuvant treatment modality. Despite the benefit of chemotherapy, outcomes for these patients remain relatively poor. There remain unanswered questions regarding whether tumor control and survival for these patients (especially those with stage IIIC or nodal improvement) might be further improved by integrating appropriate tumor-directed EBRT, in addition to systemic chemotherapy. This question has formed the basis for a current phase III clinical trial (Gynecologic Oncology Group trial 258).
In the preplatinum era, whole abdominal-pelvic radiotherapy (WAPRT) was often used in patients with ovarian cancer. Long-term survivals were seen in selected patients, in particular those with lower-grade disease and with minimal residual tumor burden after initial surgery. WAPRT requires significant technical expertise and great appreciation for the radiation tolerance of multiple normal tissues, especially in the upper abdomen (small bowel, liver, kidneys) and even the low thorax (lung base, heart). Ovarian cancers are generally radioresponsive, but this is balanced against the low doses achievable in the upper abdomen. The use of WAPRT has largely been abandoned, given its known increased risk of long-term bowel toxicity, the great difficulty of performing secondary surgery and instituting second-line therapy, and the very high response and progression-free survival rates now seen for primary platinum and, more recently, platinum-and taxane-based chemotherapy.
Radiotherapy may be considered in 2 scenarios in ovarian cancer, both using less extensive radiation coverage than that of WARPT. For patients with localized suboptimally debulked primary disease, the integration of locoregional radiation in addition to chemotherapy may increase the likelihood of tumor control. In this particular setting, radiation may be thought of as an “extension” to surgery, to complete an optimal tumor debulking that was surgically unachievable (eg, in a patient with residual but localized gross retroperitoneal adenopathy). Radiotherapy is also an effective, and perhaps underused, palliative treatment for patients with recurrent or metastatic disease. The response rate of locally treated tumors is high, and radiation can successfully provide durable palliation of symptoms such as bleeding, pain, and localized obstruction.
In the radiotherapeutic management of vulvar cancer, the approaches to the primary lesion and to the nodes are often “uncoupled,” taking into account the different anatomic and clinical constraints. This paradigm also recognizes the general acceptance that localized vulvar relapses following surgery are readily salvageable, whereas metachronous groin failures following initial therapy carry a grave prognosis and are often ultimately fatal.
Primary therapy for early vulvar cancer is typically surgical resection, with associated inguinal-femoral node dissection. Patients found to have groin node metastases, especially if there is gross nodal disease or ≥ 2 positive nodes, clearly benefit from adjuvant radiation to the groin and pelvis. Various pathologic risk factors also predict for primary site relapse following surgery (including deep invasion, close surgical margins, lymphovascular space involvement, and “spray pattern” histology), but the use of adjuvant radiotherapy to the vulvar resection bed has not been systematically evaluated.
To avoid the morbidity (especially lymphedema) of combined groin node dissection and postoperative radiotherapy, there have been attempts to evaluate “prophylactic” groin radiation (without surgery) in patients deemed to be at risk for occult nodal disease. Although criticized for technical shortcomings in EBRT delivery in previous clinical studies, there remains doubt about the effectiveness of prophylactic groin node radiotherapy. Ongoing attempts to improve on this approach include the use of contemporary high-resolution tomographic imaging as well as sentinel node evaluation.
Patients with advanced vulvar cancer were historically treated with ultraradical surgery, including exenteration. To reduce the morbidity of such debilitating surgery, investigators evaluated the role of neoadjuvant radiation and showed that radiotherapy could lead to significant tumor debulking and the successful incorporation of more limited, viscera-sparing resection. More recent studies have established the tolerability, benefit, and role of concurrent chemoradiation (using cisplatin with or without 5-fluorouracil), followed by biopsies or conservative excision, with avoidance of exenterative surgery in more than 90% of cases.
Vulvar cancer presents with highly diverse spatial geometry, where the volume at risk changes significantly in shape and distribution from the vulva to the groins to the pelvic nodes. This variability in geometry results in extra radiation dose to surrounding normal tissues when treated with conventional APPA or 4-field techniques and may lend itself well to the adaptable and conformal coverage provided by IMRT (Figure 19-22).
FIGURE 19-22. Comparison of “standard” 4-field dosimetry (left) with intensity-modulated radiation therapy (IMRT) plan (right) for a patient with locally advanced vulvar cancer and vaginal extension. The different spatial presentations of the volume at risk at varying transaxial levels within the patient lend to greater conformal sparing of normal tissues by IMRT, as seen in selected computed tomography slices: (A) mid pelvis, for coverage of pelvic nodes; (B) low pelvis, for coverage of the vagina and groin nodes; and (C) for coverage of the vulva.
Given its rarity and varied clinical presentations, few prospective data and no phase III trials address therapeutic interventions in vaginal cancer. Due to the desire to spare closely adjacent bladder and/or rectum, with the exception of very small, localized tumors, radiotherapy (rather than surgery) is used for the primary treatment of most vaginal cancers. The treatment algorithm is extrapolated from that of cervical cancer and generally uses a combination of EBRT and brachytherapy. Vaginal cancers are frequently asymmetric in location, and the vagina does not have the geometric and biologic advantage of an intact uterus to accommodate extremely high intracavitary doses, while “displacing” adjacent normal structures. Hence, brachytherapy for vaginal cancers often emphasizes the role of interstitial implants to optimize radiation dose distribution, especially for tumors with residual thickness exceeding 0.5 to 1 cm. The role of chemoradiation has not been widely evaluated in vaginal cancer, but based on the cervical cancer data, it would be reasonable to consider this in patients fit enough to tolerate cisplatin-based chemotherapy.
NOVEL RADIATION MODALITIES
It has long been recognized that heat can denature proteins and lead to cellular disruption and depth. As a complement to radiotherapy, hyperthermia has been attractive because it is not impacted by the S-phase fraction or tumor hypoxia, which are 2 conditions that render tumor cells relatively resistant to ionizing radiation. However, there is no consistent difference in inherent heat sensitivity between normal and malignant cells, unlike the disparities in cellular repair that can be exploited with ionizing radiation (see Figure 19-4). Furthermore, the localized and accurate delivery of heat, and monitoring/maintenance of a suitable stable temperature, in tissue has been hampered by technical limitations until recently, thus restricting the broad use of hyperthermia in oncologic applications.
Recent advances in heat delivery systems and in vivo thermometry, especially using noninvasive techniques, have created renewed interest in hyperthermia. Although heat to a temperature exceeding 50°C to 60°C will ablate tissues on its own, work has also focused on lower levels of hyperthermia (approximately 40°C-43°C) that may not be intrinsically cytotoxic but can enhance the effects of ionizing radiation and certain chemotherapeutic agents. Hyperthermia in combination with cytotoxic agents has been evaluated in isolated limb perfusions and intraperitoneal chemotherapy studies. In a prospective Dutch randomized trial of patients with advanced cervical cancer, the combination of hyperthermia and radiotherapy resulted in improved local control and overall survival compared to radiotherapy alone.9 This approach has not been widely adopted, because chemoradiation has now been established as the standard of care for locally advanced cervical cancer. A new, ongoing international trial seeks to clarify the contribution of hyperthermia by randomizing patients with advanced cervical cancer to chemoradiation versus hyperthermia plus chemoradiation.
Stereotactic Radiosurgery and Stereotactic Body Radiotherapy
The use of stereotactic radiosurgery (SRS) is well established in the treatment of patients with intracranial tumors. This consists of a single ablative dose of radiation, delivered via a specialized gamma knife (which has approximately 200 individual 60Co sources arranged in a semi-spheroidal helmet, with a focal point at an isocenter) or a modified linac. The spheroidal geometry of the skull lends itself well to these approaches, because it permits a centralized lesion to be targeted by multiple, surrounding, noncoplanar beams of radiation.
Stereotactic body radiotherapy (SBRT) is an adaptation of the principles and techniques developed in SRS, but focused on delivery of radiation to extracranial sites. The radiation is given in 1 to 5 fractions and is accomplished either with a specialized CyberKnife unit (which has the radiation source mounted on a free robotic arm that can move about the patient) or a modified linac. SBRT allows for more focal and precise delivery of radiation than conventional EBRT or even IMRT and is typically used for smaller, isolated lesions (< 5 cm). It is important to remember that although SBRT delivers very high doses of radiation to an identified tumor, it, by definition, does not provide prophylactic or adjuvant coverage to adjacent tissues at risk of harboring microscopic disease. A primary challenge to SBRT planning is that because extracranial sites are not spheroidal in geometry, the degrees of freedom, or “angles” of radiation beam entrance, are limited compared to SRS for brain tumors. Furthermore, without the rigid head frame used in SRS, patient positioning reproducibility and internal organ motion (especially from respiration) have to be taken into account. Nonetheless, improvements in immobilization techniques, the use of 4-dimensional CT imaging (3D CT with integration of target volume over respiratory cycle time), sophisticated dosimetric and treatment planning tools, and near real-time image guidance such as CBCT on the linac have allowed for increasing application of SBRT in management of certain disease presentations. Germane to gynecologic malignancies, SBRT may be useful in treating oligometastatic lung or liver lesions or as a localized boost (in combination with more general EBRT) in addressing unresectable nodes or pelvic sidewall recurrences.
Protons are charged particles that can be very tightly focused to a target volume. The technique makes use of the Bragg peak effect, in which the proton beam has an extremely sharp, well-defined range in tissue, with little side scatter, to deliver radiation precisely to the intended target, with markedly reduced doses to the surrounding normal tissues. Proton beams have lower entry, and no exit, doses compared to standard megavoltage photons, giving rise to their dosimetric advantage. Furthermore, protons improve on the dose conformality of current technologies, by providing highly concentrated dose contours around tumor, without the “scatter” of lower doses to large volumes of surrounding tissues inherent with IMRT. The precision of protons requires extreme care in target definition and tracking of tumor volumes, similar to that for other highly conformal techniques such as IMRT. When absorbed by tissue, protons have the same ultimate biologic effect as that of photons for any given dose; the advantage of protons rests entirely on their superior physical dose distribution.
There is significant current interest and momentum in the installation of proton centers and the use of proton radiotherapy. Protons have shown clinical benefit in pediatric cancers, as well as in certain brain and spine tumors, but these represent relatively uncommon clinical presentations. It has been argued by some that protons represent the ultimate in physical radiation dose delivery and should be the radiotherapy technique of choice for many, if not most, cancers. However, there remain concerns regarding the widespread use of protons, mostly surrounding its high costs, indications, and the lack of randomized studies that confirm its clinical superiority.10 Nonetheless, in an age of increasing combined-modality management of cancer, protons may allow for enhanced systemic therapy by reducing the locoregional toxicities of radiotherapy.
1. Hall EJ, Giaccia AJ. Radiobiology for the Radiologist. 7th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2012.
2. de Gonzalez AB, Curtis RE, Fry SF, et al. Proportion of second cancers attributable to radiotherapy treatment in adults: a cohort study in the US SEER cancer registries. Lancet Oncol. 2011;12:353-360.
3. Khan FM. The Physics of Radiation Therapy. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2010.
4. van de Bunt L, Van der Heide UA, Ketelaars M, de Kort GA, Jurgenliemk-Schulz IM. Conventional, conformal, and intensity modulated radiation therapy treatment planning of external beam radiotherapy for cervical cancer: the impact of tumor regression. Int J Radiat Oncol Biol Phys. 2006;64(1):189-196.
5. Potter R, Kirisits C, Fidarova EF, et al. Present status and future of high-precision image guided adaptive brachytherapy for cervix carcinoma. Acta Oncologica. 2008;47(7):1325-1336.
6. Loiselle C, Koh WJ. The emerging use of IMRT for treatment of cervical cancer. J Natl Compr Canc Netw. 2010;8(12):1425-1434.
7. Emami B, Lyman J, Brown A, et al. Tolerance of normal tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys. 1991;21(1): 109-122.
8. Bentzen SM, Constine LS, Deasy JO, et al. Quantitative Analyses of Normal Tissue Effects in the Clinic (QUANTEC): an introduction to the scientific issues. Int J Radiat Oncol Biol Phys. 2010;76(3 suppl):S3-S9.
9. Franckena M, Stalpers LJ, Koper PC, et al. Long-term improvement in treatment outcome after radiotherapy and hyperthermia in locoregionally advanced cervix cancer: an update of the Dutch Deep Hyperthermia Trial. Int J Radiat Oncol Biol Phys. 2008;70(4):1176-1182.
10. Glatstein E, Glick J, Kaiser L, Hahn SM. Should randomized clinical trials be required for proton radiotherapy? An alternative view. J Clin Oncol. 2008;26(15):2438-2439.