Steve P. Lee
I. INTRODUCTION
A. Radiation oncology is a medical discipline that specializes in the utilization of radiation for therapeutic purposes. It is synonymous with therapeutic radiology, an older name created to distinguish the specialty from diagnostic radiology. Radiation therapy (RT) is a treatment modality in which ionizing radiation is used for patients with cancers and other diseases.
The first therapeutic use of radiation dates back to early 1896, almost immediately after the discovery of x-rays. For over a century, RT continues to play a significant role in the fight against cancer, with its progress intimately supported by the advances in modern science and technology. Among the main cancer treatment modalities, RT and surgery aim to provide local-regional tumor control, while chemotherapy addresses systemic metastases in addition to serving frequently as a radiation sensitizing agent.
B. A radiation oncologist is a physician trained in cancer medicine who uses radiation for the treatment of cancer and occasionally nonmalignant conditions. The radiation oncologist analyzes and discusses the benefits and risks of the proposed treatment, designs and controls the treatment process, cares for the patient’s treatment-induced side effects, and continuously monitors the patient’s disease status. The delivery of RT necessitates the efforts of other health care professionals.
1. Medical physicists ensure the proper functioning of radiation-producing machines and maintain treatment planning hardware and software.
2. Dosimetrists and physicists perform treatment planning for individual patients to the specifications of radiation oncologists.
3. Radiation therapists operate treatment machines to irradiate patients according to specific treatment plans.
4. Collaborators. To deal with a patient’s cancer problem, a radiation oncologist must also collaborate closely with diagnostic radiologists, pathologists, surgeons, medical oncologists, and other physician colleagues. Other frequent supporting specialists include nurses, dieticians, dentists, physical therapists, geneticists, psychiatrists or clinical psychologists, social workers, administrative assistants, and so on.
II. PHYSICAL, CHEMICAL, AND BIOLOGIC BASIS OF RADIATION ACTION
A. Ionizing radiation. Radiation in the energy range used for RT can cause the ejection of orbital electrons and result in the ionization of atoms or molecules. The amount of energy deposited within a certain amount of tissue is defined as the absorbed dose with a unit of gray (Gy; 1 Gy = 1 joule/kg). The older unit of rad is equivalent to 1 centigray (cGy; 1 rad = 1 cGy). The types of radiation commonly used clinically are:
1. Photons are identical to electromagnetic waves. The energy range used for RT pertains to either x-rays, which are commonly produced by a linear accelerator (LINAC), or γ-rays, which are emitted from radioactive isotopes. Photons of different energies interact with matter differently: from low to high energy, the absorption mechanism ranges from photoelectric effect, Compton effect, to pair production. Modern therapeutic machines produce photon beams with energy of megavoltage (or million electron volt, MeV) range rather than kilovoltage (kilo-electron volt, KeV) as used in diagnostic radiology. In general, the higher the photon energy, the greater the depth of penetration into the body and the more “skin-sparing” effect with less radiation-induced dermatitis.
2. Electrons dissipate their energy rapidly as they enter tissue. Thus, they have a relatively short depth of penetration and generally are used to treat superficial lesions. Their effective range in tissue also depends on their energy. Each modern LINAC usually provides one or two energy levels of photon beams and several levels of electron beam energy.
3. Other radiation particles used for RT include protons, neutrons, and heavy ions such as carbon anions. These particles are characterized by the so-called linear energy transfer (LET), a quantity measuring the rate of energy loss per length of path traversed. Heavy ions have high LET and thus are “densely ionizing,” in comparison with the low-LET photons and electrons that are “sparsely ionizing.”
a. Relative biologic effectiveness (RBE) is defined as the ratio of doses needed to produce the same biologic endpoint between a standard low-LET photon beam (by convention 250 KeV x-ray) and another radiation of different LET. The term relates LET to an actual biologic effect, such as cell death. In general, higher LET particles have higher RBE up to a certain level (approximately 100 KeV/μ), beyond which RBE actually declines due to “wastage” of further energy transfer.
b. Another factor that depends on LET is the efficiency of oxygen molecules to enhance radiation cell killing (mediated by oxygen radical formation). Lower LET radiation tends to depend more on the presence of oxygen to effect cellular damage and thus have a higher oxygen enhancement ratio (OER, defined as the ratio of doses needed for a particular radiation to produce the same cell survival between anoxic and oxic environments) in comparison with higher LET particles.
c. Protons have a level of LET or RBE similar to photons and thus have no significant biologic advantage over high-energy photons or electrons. A proton beam, however, has a special physical property of releasing very little energy as it traverses into tissue until a fixed depth is reached where almost all the dose is deposited (called a “Bragg peak”). The depth of this peak can be manipulated electronically to coincide with the target by varying the incident energy of the protons. Thus, proton radiation has a dosimetric advantage when treating a deep seated tumor next to a critical normal structure.
d. Neutrons do not possess the dosimetric advantage of protons since they do not have the depth-dose characteristic of a Bragg peak. Nevertheless, neutrons have high LET and low OER; that is, their cell-killing function does not depend significantly on the presence of oxygen. Thus, they have the biologic advantage of treating tumors that are relatively resistant to photons due to the presence of significant hypoxia (see Section III.C.1).
e. Heavy ions have high LET and low OER, as well as the presence of a Bragg peak. Thus, if used properly, they possess both the biologic and physical advantages seen in neutrons and protons, respectively.
B. Mechanism of damage to cellular targets. Where water is abundant, short-lived (10−10 to 10−12 seconds) hydroxyl radicals can be formed by ionizing radiation (via a process called radiolysis) and impact a nearby (approximately 100 Å) macromolecule such as DNA to damage its chemical bonds (indirect action). Alternatively, such chemical bond damage can result readily from the direct deposition of radiation energy (direct action). Evidence suggests DNA to be the main target of radiation action in cells. The “elementary lesions” include base damages, cross-links, single-strand breaks, and double-strand breaks. Studies have shown that “complex clustered damages” or “multiply damaged sites” are created upon irradiation, each of which consists of multiple elementary lesions spanning a few nanometers, or about 20 base pairs of DNA. These damages can remain unrepaired and lead to eventual cell “lethality,” which is defined operationally as the loss of reproductive integrity (i.e., inability to maintain clonogenicity). Such a mode of mitotic death is considered a predominant mechanism of radiation killing, but other processes such as interphase death and apoptosis (programmed cell death) also play important roles.
C. Cellular and tissue response to radiation damage. Many molecular mechanisms have been identified to govern a cell’s response to radiation damage, with an intricate network of signal transduction pathways leading to either cell death or survival. The ultimate outcome is a result of a complex chain of events dictated not only by the biophysical interaction between radiation particles and DNA, but also molecular and genetic determinants such as oncogenes, tumor suppressor genes, and cell-cycle regulation. In addition, extracellular and tissue conditions such as hypoxia, cell–cell interaction, and extracellular matrix can also modify the final expression of radiation effects on cells and tissues. Despite these complexities, certain clinical-pathologic outcomes can still be predicted due to the fundamental action of radiation directly upon critical cellular targets.
III. BIOLOGIC BASIS FOR RADIATION THERAPY
A. Target cell hypothesis. A biophysical interpretation of the radiation mechanism can be made between a given dose, D, and the observed fraction of surviving clonogenic cells (surviving fraction, SF). The basic assumption is that critical targets exist within each cell; when ionizing radiation particles hit these targets, loss of cellular clonogenicity may result. This is the essence of the so-called “target-cell hypothesis” or “hit theory.”
B. Cell survival curves. When plotted as a semi-log graph of logeSF versus D, almost all mammalian radiation survival curves reveal a very similar shape with a “curvy shoulder” at the low-dose region, in contiguity with a relatively linear tail toward high-dose region. This suggests that at least two biophysical mechanisms seem to be operating simultaneously to produce such a result.
1. The linear component signifies that radiation action on the critical target within a cell is a random (specifically, Poisson) process, resulting in logarithmic decrease in survival such that equal dose increments cause a constant logarithmic proportion of cell deaths. It is commonly described as a “single-hit” killing, which results in nonrepairable damage leading directly to cell death.
2. The curvy shoulder, however, reflects a much more complex situation whereby interactions of more than one target lesion are at work to result in eventual cell lethality (thus described as “multitarget” killing). Before subsequent interacting events can take place, the initial lesion may be repaired. Thus, the ultimate expression of this mode of cell killing depends on the kinetics and efficiency of the repair process.
3. Quantitative models. Mathematical models are very useful in clinical radiation oncology to help quantify the biologic effects of radiation. The abovementioned dual processes that characterize the observed survival curve have been formulated as the single-hit, multitarget (SHMT) model. However, its mathematic expression is rather too cumbersome to be used for routine clinical applications. Another mechanistically oriented theory, the linear-quadratic (LQ) model, has become widely popular due to its relatively simple mathematics. Based on this model, the surviving fraction (SF) after a single treatment of radiation dose (D) can be characterized by the following equation:
SF = exp(−αD − βD2)
where α and β are tissue-specific parameters governing intrinsic radiation sensitivity. The LQ model can be used to explain the differential sensitivities of malignant tumors versus normal tissues to fractionation radiotherapy (i.e., dividing the overall therapy into numerous fractions of small-dose irradiation). The α/β ratio is used clinically to characterize how various tissues respond to fractionation treatment.
a. Acute-responding tissues, such as cancers and fast-dividing normal cells, typically have high α/β (approximately 8 to 10 Gy), and upon irradiation will express acute effects (e.g., tumor shrinkage, dermatitis, mucositis, etc.).
b. Late-responding tissues (normal cells which rarely proliferate but can express late effects such as fibrosis, xerostomia, and nerve damage) have low α/β (approximately 2 to 5 Gy).
c. Upon fractionation using a small dose per fraction repeatedly, an “effective” survival curve results from the repetition of the initial curvy shoulder primarily due to repair of cellular damages (see Section III.C.3). Ultimately, it is seen that acute-responding cells are killed predominantly, while the late-responding tissues are relatively spared due to their higher capacity to repair.
d. It should be noted that, on a typical survival curve, the LQ model seems to predict well the observed survival result at the low-dose range, up to about 6 to 8 Gy. With a much higher dosage, the LQ model seems to overestimate cell lethality, whereas the surviving fraction is seen to be better fit by the SHMT model (see Section IV.H.7).
C. Fractionation radiobiology. The biologic processes occurring in-between the treatment fractions have been summarized as the “4 R’s of fractionation radiobiology”:
1. Reoxygenation. The damage of tissues by radiation depends largely on the formation of hydroxyl radicals, which in turn depend on the availability of oxygen molecules in close proximity. Fractionation allows oxygen to diffuse into the usually hypoxic center of an expanding tumor during the interval between fractions and thus enables more tumor cell killing during the subsequent treatment.
2. Repopulation. All living cells have the potential to repopulate by mitotic (clonogenic) growth. If normal tissue progenitor cells repopulate more adequately than malignant cells during the treatment course, then a therapeutic gain may be obtained by fractionation. Moreover, a phenomenon of accelerated repopulation, which is stimulated by cytotoxic intervention such as radiation, has been described for fast growing malignant and normal cells. Thus, the overall treatment time is an important clinical variable affecting the chance of tumor control. As cancer cells quickly repopulate, further protraction of overall treatment time is disadvantageous. Hence, once radiation treatment commences, unnecessary interruption of treatment course is discouraged.
3. Repair. Repair machinery within cells can reverse initial partial damages caused by a relatively small fraction of radiation dose. The cells would die if such damages fail to be repaired sufficiently while being accumulated by further radiation insults. One such contributing repair mechanism is called sublethal damage (SLD) repair. As the dose per fraction decreases and interfractional time interval increases enough to allow for complete SLD repair, the total dose required to achieve a certain level of cell death would be increased. Thus, fractionation can help spare cells from radiation killing in comparison with single-dose irradiation. Furthermore, late-responding tissues, which have higher SLD repair capacity, would be spared preferentially over acute-responding malignant cells which might lack adequate repair mechanisms.
4. Redistribution. Cells exhibit differential sensitivities toward radiation at different phases of the cell cycle. Most mammalian cells are more sensitive at the junction between G2 and M phases. After an initial fraction of dose, the cells at a more resistant phase (e.g., late S) may survive but then progress eventually to the sensitive phases in time, allowing more efficient killing during the next fraction. Thus, fast cycling cells (like skin or mucosal cells and most cancers) are more prone to radiation killing than slow or dormant ones (such as muscle or skeletal cells).
D. Dose rate effect. The biologic effect of a given radiation dose also depends on the rate at which it is delivered. With decreasing dose rate, the cell survival increases due to SLD repair. Cell survival is also enhanced due to repopulation as the treatment course gets further protracted until a limit (characterized by nonrepairable single-hit killing) is reached. Beyond this limit, further decrease of dose rate will in fact give rise to more cell death (inverse dose rate effect) due to cell cycle arrest at G2 phase where cells are most vulnerable to radiation killing.
1. Once daily fractionated RT typically utilizes a dose rate of about 1 Gy/min.
2. For continuous low dose rate (LDR) brachytherapy (or implant; see Section IV.C.4), radioactive seeds are typically inserted into a patient’s body for a long period of time, with radiation dose emitting at a rate of about 1 cGy/min.
3. High dose rate (HDR) brachytherapy has gained wide popularity, using a dose rate of about 1 Gy/min, as is used with external beam teletherapy.
E. Altered fractionation. The fractionation scheme used in conventional RT usually utilizes 1.8 to 2 Gy per fraction to a total dose that is required for a particular malignancy (e.g., about 70 Gy for gross epithelial cancers, or substantially less for more radiosensitive tumors such as lymphomas). However, by exploiting the radiobiologic principles as listed above, therapeutic benefit can be increased via altered fractionation regimens.
1. Since fractionation preferentially spares late-responding normal tissues, a strategy of hyperfractionation can be used to enhance tumor cell killing while maintaining the same degree of late normal tissue damage. More fractions are delivered (usually twice daily) using a smaller dose per fraction but culminating in a higher total dose, while keeping the time of overall treatment course about the same as conventional fractionation.
2. In order to overcome the potential bottleneck for tumor control due to accelerated repopulation (see Section III.C.2) of cancer cells, a strategy of accelerated fractionation can be used to deliver a conventional level of total dose, while shortening the time of overall treatment course with more intensely fractionated patterns. A lower dose per fraction is delivered two or three times daily.
3. Based on the LQ model, a quantity termed biologically effective dose (BED) has proven convenient in quantifying radiobiologic effects and has enabled comparisons among various clinical trials using different fractionation schemes. For late responding tissues,
BED = D · {1 + [d/(α/β)]}
where D is the total dose and d is the dose per fraction. BED is versatile since it is linearly additive; that is, one can sum up directly all BED values for partial treatments with various fractionation regimens or special techniques such as brachytherapy, in order to predict the net biologic effect to a tissue characterized by a specific α/β.
4. Despite the biologic advantage of increasing fractionation supported by decades of clinical observations, single-dose treatments as well as hypofractionation have gained popularity in recent years due to advancing physics and technology in facilitating ultraprecision-oriented irradiation techniques (see Sections IV.C to IV.H).
F. Dose response curves. The terms tumor control probability (TCP) and normal tissue complication probability (NTCP) can be used to assess clinical consequences of RT quantitatively as functions of dose.
1. Defined as the probability of killing all tumor cells so that none can survive, TCP is expressed via Poisson statistics as
TCP = exp (−M · SF)
where M denotes the number of clonogenic cells, and SF is an explicit function of dose as discussed in Section III.B.3.
2. Similarly, NTCP can be formulated if a complication is assumed to arise due to the radiation-induced depletion of the so-called functional subunits (FSUs) within the specific normal tissue. Thus,
NTCP = exp (−N · SF)
where N denotes the number of FSUs. In reality, the structural organization of these FSUs may introduce substantial complexity in determining the ultimate clinical outcome (see Sections III.Gand III.Hbelow).
3. As probability curves, both TCP and NTCP exhibit rising sigmoid shapes from 0 to 100% when plotted linearly against increasing dose. Only when the NTCP curve is sufficiently situated to the right (higher dose region) of the TCP is it warranted clinically to irradiate. Almost all innovations in clinical radiation oncology have been based on attempts to separate these two curves.
4. The therapeutic ratio is the relative degree of tumor control (measured as TCP) over normal tissue damage (NTCP). There exists an optimal dose for which this ratio, or the uncomplicated TCP (UCTP), is maximized:
UTCP = TCP · (1 − NTCP)
5. With a sharply rising sigmoid shape, TCP dictates that once a certain dose is deemed necessary to achieve adequate tumor control, the treatment should not be terminated prematurely since almost no therapeutic gain would be expected until the whole course is near completion (i.e., “all or nothing” response).
6. Note that both TCP and NTCP depend not only on the dose but also on the size or volume: that is, the number of clonogenic cells in a tumor for TCP or the number of FSUs within a normal tissue for NTCP, respectively. The higher the number of cells or FSUs to irradiate, the more either dose response curve shifts to the right.
7. For a specific malignancy or normal tissue, factors that may “flatten” (i.e., decrease slope of) the respective sigmoid TCP or NTCP curves include the wide variation of radiation sensitivity in a patient population.
8. For the prophylactic control of subclinical micrometastases by radiation, the dose response curve is flattened due to the heterogeneous distribution of metastatic tumor burdens in the targeted volume. Thus, a small amount of dose can still be beneficial, in contrast to the substantially larger dose required for the control of bulky tumors.
G. Tissue organization. The structural organization of the FSUs in a normal tissue may be critical in determining the kinetics of its damage expression, as well as the effect of heterogeneous dose distribution across its volume.
1. Based on physiologic and cellular kinetics reasoning, some normal tissues can be separated structurally into type-H (hierarchical) or type-F (flexible) tissues.
a. Type-H tissues (e.g., bone marrow, skin, and gastrointestinal tract) contain stem cells that are destined to mature into functional cells. As they lose clonogenicity in the process, these cells become radioresistant because only the rapidly proliferating stem cells are likely to be sensitive to radiation killing.
b. Type-F tissues (e.g., lung, liver, and kidney) contain cells that can simultaneously maintain their proliferation capacity (thus are radiosensitive) and serve their normal physiologic function.
c. Upon irradiation, type-F tissues can exhibit a dose-dependent kinetics of damage expression—the higher the dose, the earlier the time of expression. In contrast, the kinetics of damage for type-H tissues is relatively independent of the dose.
2. The spatial orientation of normal tissue can be separated into parallel and series structures. Parallel structures are typified by kidney, liver, lung, and tumors, while series structures include the GI tract, spinal cord, and peritoneal sheath. Most normal tissues have mixed characteristics of both parallel and series structures. The concept of relative seriality has been proposed, based on the perceived organization of FSUs. This concept is useful when dealing with heterogeneous dose distribution across a structure of interest (see Sections III.H.3 and III.H.4).
H. Heterogeneous dose distribution. The biologic effect of radiation on a particular normal “organ at risk” (OAR) can vary significantly with the amount of volume as well as which portion or region of the organ is treated. Clinicians have been trained to be familiar with the overall radiation effect (e.g., NTCP) of an OAR being irradiated with uniform dose distribution. However, modern treatment techniques using inverse planning and IMRT (see Section IV.C.2) often involve heterogeneous dose distribution.
1. The cumulative biologic effects of partial volume irradiations that might not amount to the effect predicted based on the same total physical dose assumed to be uniformly deposited for the whole organ is called the volume effect.
2. For a grossly heterogeneous dose distribution across a significant portion of an OAR, the degree of such heterogeneity can be quantified using the dose volume histogram. It takes a shape of a monotonically decreasing sigmoid curve on a fractional volume versus dose plot.
3. Tissues in parallel have been modeled along the so-called “critical volume” argument. The total volume irradiated has a direct impact on NTCP. Irradiating a significant volume of such tissue, even if with moderate doses, would be more detrimental than giving an extremely high dose but to only a small volume of the organ. Thus, the bulk of the volume irradiated does matter significantly.
4. Tissues in series have “critical elements” arranged in chains upon which irradiating even a small volume of the structure to a sufficiently high dose might incur a complication. The prime example would be spinal cord, which needs only a hot spot at a given segment to manifest transverse myelitis. The incidence of a complication increases in proportion to the volume irradiated for series tissues.
5. Equivalent uniform dose (EUD) has been defined to address heterogeneous dose distribution and is the dose which, when distributed uniformly across the target volume, gives rise to the same biologic effect.
IV. CLINICAL PRACTICE OF RT
A. Consultation. Patients are usually referred by surgeons, medical oncologists, primary care physicians, or other specialists for consultation, during which the history and physical examination are done by radiation oncologists. The indications for RT, as well as its associated short- and long-term side effects, are explained to the patient in detail.
1. Indications. Like surgery and chemotherapy, RT has definite indications and contraindications for clinical application. RT can be used alone or in combination with other methods, as the major component of treatment or as an adjuvant modality. Approximately 50% to 60% of all patients with cancer receive RT during the course of their illness. Properly used, the intent of RT for about 60% of these patients will be curative. For others incurable by any current method of treatment, palliation of symptoms and signs by RT can improve their quality of life.
a. Treatment with curative intent is often complicated and requires professional skills and facilities that may be distant from the patient’s home. Usually the doses are higher and, consequently, the risks of sequelae are greater than for palliative treatment.
b. Palliative treatment should have a specific objective. Inconvenience, cost, discomfort, risk, and overall treatment time should be minimized. Objectives of palliative RT include relief of pain, usually from metastases to bone; relief of neurologic dysfunction from intracranial metastases; relief of obstruction of the ureter, esophagus, or bronchus; promotion of healing of surface wounds caused by tumor, such as breast cancer; preservation of the weight-bearing skeleton compromised by metastases; or preservation of vision by controlling metastases to or tumor invasion of the eye or its orbit.
2. Negative side effects. Most normal tissue effects of RT are related to cell killing, and expected to occur only within the irradiated volume. Some effects, such as nausea, vomiting, fatigue, and somnolence, remain unexplained, although there may be a relationship to radiation-induced cytokines. Some late effects may be related to radiation-induced proliferative responses such as gliosis or fibrosis.
Side effects of RT are influenced by other treatments. Acute skin or mucous reactions may be accentuated by the concurrent or even later administration of chemotherapy agents (e.g., doxorubicin). Radiation-induced bowel damage may be accentuated by prior surgery. The appearance of clinical injury after cell killing will depend on factors such as cell turnover time and differentiation kinetics. For convenience, these time-related responses are arbitrarily divided into
a. Acute responses that appear usually within 2 to 3 weeks after treatment commences, such as mucositis and diarrhea, are secondary to the depletion of stem cells (esp. for type-H tissues, see Section III.G.1) and are expected to subside gradually once the treatment course is over.
b. Subacute responses, such as Lhermitte syndrome (electric shock-like sensation down the periphery upon sudden flexion of neck—presumably due to demyelination) or somnolent syndrome, occur after several months and are nearly always transient.
c. Late responses are secondary to the depletion of slowly proliferating cells and are nearly always permanent. These are usually the critical structures which limit the dose prescribed by the radiation oncologists.
B. Treatment preparation. Once the patient agrees and provides the informed consent to proceed with RT, the next step is usually a simulation process and subsequent treatment planning.
1. Simulation is a procedure when the radiation oncologist tries to determine how to aim the beams of RT according to the patient’s anatomy and the locations of the target lesions, as well as the OARs. A conventional simulator may be used, which has the geometric construct of the beam source and patient couch movement identical to the actual treatment machine. Such machines are being replaced by CT simulators, where tomographic scans form the basis of subsequent three-dimensional (3-D) oriented treatment planning. During simulation, the patient is placed on a treatment couch and certain immobilization measures are often implemented, since the positioning must be reproducible for subsequent daily treatments with acceptable precision. Frequently, permanent tattoos are marked on the patient’s body surface.
2. Treatment planning. Other than fairly simple cases, for which old techniques continue to serve well, modern RT requires complex treatment planning. Computerized data are essential to this process in order to produce a finalized plan which can be transferred seamlessly to computer-controlled therapy equipment. It requires the integrated efforts of radiation oncologists, medical physicists, dosimetrists, and radiation therapists.
a. The first step of treatment planning is the identification of essential anatomic structures relevant to the goal of the treatment. The 3-D extent of each structure of interest can be traced in contoured forms, section by section, on the tomographic images.
b. Conceptually the treatment volume incorporates the gross tumor volume (GTV), which represents the detectable extent of the tumor target, the clinical target volume (CTV), which includes microscopic tumor extensions, and the planning target volume (PTV), which includes margins around the CTV to allow for positional uncertainty.
c. Normal structures of interest are also identified, contoured, and designated as OARs.
C. Precision oriented RT. Computerized treatment planning now allows for delivering RT with ultraprecision.
1. Conformal RT. Since the structures of interest are often irregular in shape, devices such as metal blocks have been constructed manually and customized for individual patients, and the treatment planning has been rather rudimentary with a 2-D dosimetry plan used as a basis for 3-D extrapolation. When CT-based imaging became available and adapted for simulation, truly 3-D oriented conformal radiotherapy became common practice.
a. Computer technology allows for a very slick 3-D treatment planning approach using a machine-driven beam-shaping device called multileaf collimators (MLC). The target structures are basically sliced one beam-path at a time with a width measuring from a few millimeters to a centimeter, and the radiation dose within each slice is calculated to precision and spatially conformed to the edge of the desired target. One useful tool is the “beam’s eye view,” which simulates looking along the axes of multiple radiation beams to plan the best arrangement.
b. During the treatment, the automated MLCs restrict the delivered dose to tightly conform to the target. Thus, precise dosimetric determination for any 3-D irregularly shaped tumor target or OAR is theoretically feasible. Tools are available to maximize the therapeutic ratio by a process of optimizing various treatment parameters.
2. Intensity modulated radiation therapy (IMRT) is a technique that allows photon beam intensity to be modulated by MLCs, in order to deliver specific doses to irregular-shaped target volumes while sparing nearby OARs.
a. The essence of IMRT is inverse planning, such that the treatment outcome could be optimized. The physicist feeds in the anatomic information of tumors and OARs, specifies the desired outcome with dose constraint for each structure of interest, and then lets the computer search for the best solution to achieve such goal. The answer will dictate how the treatment machine might adjust (or “modulate”) the radiation beam intensity in an automated fashion by moving the MLCs rapidly across the irradiated target, while constantly avoiding the OARs.
b. The computer-generated IMRT plan allows for a much higher dose within and tighter dose distribution around the irregular target border, while the adjacent normal tissues receive relatively little dose.
c. Higher doses can be given via IMRT as a “boost” to the primary tumor bed sequentially after a course of RT aimed at a broader coverage of the head and neck region. This follows the traditional practice of the shrinking-field technique, with the dosages of various structures (including the tumor) prescribed to commonly accepted values.
d. IMRT is now often used from the very beginning of the treatment course with the so-called simultaneous integrated boost (SIB) technique. For each fraction, the subclinical spread of cancer cells in the broad area is treated to a relatively lower dose, while the primary tumor is irradiated simultaneously with a higher dose. Therefore, the total dose received at any structure of interest and its subsequent clinical effect can vary widely depending on the fractionation schemes. It is less meaningful to use total physical doses for intercomparison of treatment results using different SIB techniques. Instead, some sort of quantitative biologiccorrection is usually needed.
e. IMRT has the potential of introducing dose heterogeneity (see Section III.H) within a specific structure because of intensity modulation. The biologic consequence due to such effect is still not very well understood, since clinicians have traditionally been trained to be familiar with the consequences of only homogeneous dose distribution across an anatomic object.
3. Particle therapy. Treatment planning for protons and heavy ions (see Section II.A.3) may be done in an inverse manner, with intensity modulation amounting to “dose painting.” This may represent the most sophisticated form of precision-oriented RT, with minimal “integral dose” (i.e., the sum of total body dose—taking account the undesirable traces of stray radiation dose scattered elsewhere beyond the intended treatment target within the patient’s body).
a. Clinical cases especially advantageous if treated by particle therapy include ocular tumors and pediatric malignancies.
b. The main disadvantage of particle therapy has been its extremely high cost of production and operation.
c. Due to continuing advances in engineering technology and commercialization, proton therapy has become more accessible in recent years as new treatment facilities are established worldwide.
4. Brachytherapy (implant) is another form of precision irradiation.
a. In this technique, radioactive sources manufactured as tiny metal seeds are inserted within the tumor-bearing area of the patient’s body, either via interstitial (e.g., soft tissue embedded with tumor bulk), intraluminalroutes. (e.g., esophagus, trachea, or rectum) or intracavitary (e.g., vaginal vault) routes. For superficial tumors, seeds can be arranged to be in contact with the lesions or via a specially designed surface applicator.
b. Typical radioactive isotopes used are iridium (192Ir), iodine (125I), palladium (103Pd), cesium (137Cs), strontium (90Sr), cobalt (60Co), and occasionally gold (198Au). Radium sources are nowadays of historic interest only (see Chapter 2).
c. Depending on the rate of radioactive dose delivery (see Section III.D), two general types of brachytherapy are usually available: low dose rate (LDR) and high dose rate (HDR).
d. The LDR radioactive sources are available as individual seeds, which can be further strung together into a ribbon with specified distances in-between the seeds. Source seeds can be inserted manually via a needle instrument for permanent implant, with their radioactivity allowed to decay spontaneously in time. The precise dosimetry is done based on the actual radiographic position of each seed.
e. For temporary implant using the so-called “afterloading” technique, single or multiple hollow catheters or special apparatus are positioned within the body site first, which allows for preimplant dosimetry planning and optimization. The source seeds are then loaded according to the intended plan for short-term irradiation, usually manually in the case of LDR treatment. Personnel radiation safety precautions need to be exercised at all times, especially when inserting and removing the sources.
f. Aiming exclusively for temporary implant, the HDR treatment is done via afterloading technique using an electronically controlled unit that houses a single highly radioactive source seed. In contrast to LDR, personnel exposure to radioactivity is expected to be negligible during the highly computerized, automated procedure.
g. Because the contributing dosage is inversely proportional to the square of the distance from each radioactive seed, normal tissues near the target can benefit from the rapid “fall off” of the dose. Thus, in comparison with focal external beam irradiation, the main advantage of brachytherapy is the relatively low integral dose. Its disadvantages mainly involve the operative risks (anesthesia, bleeding, infection, etc.) associated with invasive (especially interstitial) catheter insertions, and the need to observe radiation precautions for health care personnel and patients.
h. Following comparable radiobiologic arguments favoring increasing fractionation, LDR treatment could be considered more beneficial than HDR theoretically (see Section III.D). However, the prolonged irradiation time is a relative deterrence for LDR temporary implants. HDR is usually done through a hypofractionated scheme, a practice still considered safe due to the limited volume irradiated (see similar discussion in Section IV.D.3.d).
D. Stereotactic irradiation. Precision-oriented treatment planning loses its meaning if the positions of the target and OARs are deviated during actual treatment because of set-up uncertainty or motion. Patient immobilization is thus crucial, especially for tumors in the brain and the head and neck.
1. Stereotactic radiosurgery (SRS). For relatively few and small tumors, ablating each lesion precisely with an exceptionally high level of radiation dose using the so-called SRS technique may be indicated.
a. Originally developed by neurosurgeons to locate brain lesions with pinpoint accuracy using a 3-D coordinate system in reference to a rigid frame attached to the patient’s skull, stereotactic localization technique is utilized when high-dose radiation is used to substitute invasive surgical resection.
b. Two different ways of producing the radiation for SRS are commercially available. For systems such as the Gamma Knife®, about 200 60Cobalt radioisotope sources emitting γ-rays are oriented in a hemispherical fashion or other similar geometrical construct, to focus all the beams on a central point. Alternatively, such focused radiation can be produced via a LINAC which generates an x-ray beam as a single source and can be rotated or moved around a central focus.
c. Feasible SRS targets must be small (usually 3 cm or less in diameter), and the number of lesions to be treated must be few (usually four or less).
d. SRS has gained wide popularity to treat central nervous system tumors (both benign and malignant) and, at times, neurophysiologic disorders such as trigeminal neuralgia.
e. The use of SRS for head and neck tumors is usually limited to administering an additional “boost” beyond conventional radiation treatment or as salvage for local recurrence.
f. Extracranial tumors (such as in the spine, lung or liver) are also being explored for possible SRS treatment as long as a single or few fractions of large-dose irradiation is deemed appropriate and the problem of motion uncertainty can be resolved (e.g., with technical innovation to compensate respiratory motion for lesions in the trunk; see Sections IV.F.1 and IV.H).
2. Stereotactic radiotherapy (SRT). For many malignancies, the size of the primary tumor is typically larger than what SRS can accommodate, and more importantly, its edges are often mingled with normal tissues. In such cases, stereotactic technique can be combined with the biologic advantages of conventional fractionation to provide SRT as a treatment option. SRT is done often with removable body-fixation frames for daily treatments.
3. Selection of SRS versus SRT (guidelines):
a. SRT will in general have a theoretical biologic advantage over SRS for most malignancies. SRS is often favored for logistic reasons rather than biologic considerations per se.
b. Whenever an aggressive tumor is found located in close proximity to a critical normal tissue, SRT would probably be more beneficial than SRS since the biologic advantage of fractionation can be exploited.
c. If there is not much biologic difference between tumor (e.g., a benign or low-grade lesion) and the surrounding normal tissue, it may be legitimate to offer patients SRS treatment, serving just like a surgical tool.
d. Perhaps due to the wide acceptance of SRS, or because SRT is simply a more tedious procedure, clinicians might develop a lingering desire to minimize the number of fractions (i.e., hypofractionation) for patient treatment. Only with precision treatment is it safe to do so. By spatially segregating tumors from normal tissues, one can treat the former without too much concern of deleterious biologic effect over the latter.
E. Functional image-guided RT. The development of functional imaging studies like positron emission tomography (PET) or magnetic resonance spectroscopy (MRS) has allowed physicians to contemplate whether dose escalation to metabolically active or radiation-resistant spots within a tumor might help raise the local tumor control rate. These sophisticated imaging techniques may unite modern molecular biology to clinical radiation oncology using IMRT or particle beams for dose painting purpose. As it stands today, much remains to be researched before their clinical application becomes routine.
F. Image-guided radiation therapy (IGRT). IGRT is preoccupied with precise tracing of the radiation target in order to compensate for motion uncertainty. An example is to implement respiratory gating for tumors in the trunk during each fraction of irradiation by synchronizing the treatment field coverage precisely over a target which moves with respiration. Another frequent application of IGRT is for prostate cancer, since the prostate gland can move (mostly depending on the content of the rectum behind it) from day to day through the long course of radiation therapy.
1. The internal soft tissue structures, which ordinarily will escape radiographic detection, can be illuminated. For example, metal seeds could be inserted as “fiducial markers.” For relatively fixed tumors, internal bony landmarks can be used for x-ray positioning.
2. Special imaging devices such as a perpendicular pair of diagnostic x-ray systems or “cone-beam CT” can be added to currently existing LINACs for the purpose of IGRT. Commercially available ultrasound systems (suitable for imaging soft tissue structures) or traceable radiofrequency/infrared signal emitting devices can also be used for daily image guidance.
3. Another option is to acquire a system such as Tomotherapy which can serve the functional duality of both a rotational (or “axial”) RT machine as well as providing frequent tomographic images for IGRT.
G. Adaptive radiation therapy. Initially bulky tumors can often shrink readily during the long course of radiation and chemotherapy treatment. The anatomic uncertainty is thus introduced not because of patient motion or set-up error, but the significant anatomic deviations of relevant internal structures due to the progressive change of the tumor bulk (or the patient’s significant weight loss). Adaptive radiation therapy aims to keep track of this dynamic situation and issue appropriate countermeasures as frequently as possible. The goal is to modify sequentially the treatment plan based on the initial simulation scan and the subsequent daily image verifications, using a sophisticated mathematical algorithm for mitigation of the geometrical incongruities and variations, without actually repeating the laborious simulation and treatment planning.
H. Stereotactic body radiation therapy (SBRT). In recent years, a special technique called SBRT has gained wide popularity.
1. While SRS and SRT remain valid terminologies for radiation treatment of CNS tumors, SBRT pertains to treatment for cancers outside the brain and spine.
2. The hallmarks for SBRT are image-guidance via stereotactic localization and hypofractionation (in the United States commonly defined as five or fewer number of fractions; see Sections IV.D.1.f and IV.D.3.d).
3. With highly accurate target volume irradiation, dose escalation becomes feasible. Like SRS, the biologically equivalent dose delivered with the hypofractionation schemes of SBRT typically exceeds that of conventional fractionated regimens.
4. Since the therapeutic aim of SBRT seems to be tumor ablation in nature, another term has been proposed as a substitute: stereotactic ablative body radiotherapy (SABR).
5. So far, SBRT has been applicable mainly for tumors of the trunk (lungs and liver), as well as for prostate cancer. Other anatomic sites have been treated likewise or are currently under clinical investigation.
6. With the favorable clinical outcomes of SBRT but the apparent contradiction of hypofractionation to the traditional radiobiologic preference toward more fractionation (see Sections III.C and III.E), a new biologic rationale has been proposed. For example, vascular endothelial damage has been postulated as a significant mechanism in addition to mitotic or other modes of cell death.
7. New quantitative models have also been proposed to explain the apparent success of hypofractionated RT not predicted by the popular LQ model.
a. Most investigators acknowledge the inadequacy of the LQ model to predict single-dose cell survival at the high-dose region (see Section III.B.3.d).
b. Some new models have been proposed as a “hybrid” between the LQ and the SHMT models,—that is, constructed to preserve the applicability of the LQ model (with its mathematic simplicity) at the low-dose region, while adding a modified version for the high-dose region as predicted by the SHMT model. The BED formulation based on the LQ theory (see Section III.E.3) likewise has been revised.
c. Some authors have modified the hypothesized mechanism behind the “curvy shoulder” of the cell-survival curve (see Section III.B.2) in their proposed models. Higher mathematic complexity often accompanies these new models, which makes their routine applications in clinical settings difficult.
V. THE ROLE OF RT IN OVERALL CANCER MANAGEMENT. Among the traditional therapeutic modalities against cancer, surgery and radiation therapy aim to achieve local control of the tumor, while systemic treatment such as chemotherapy aims mainly to eliminate body-wide presence of malignant cells.
A. Some basic oncologic principles warrant reiteration here:
1. Cancer is characterized by a common pathogenic process: starting with neoplastic transformation of single cells to tumor aggregates, which can invade and extend locally, to metastatic shedding and eventual clonogenic establishments at distant sites.
2. The hallmark of most common malignancies is thus the spatial expansion of tumor cell population throughout the host body, in a process which progresses over time.
3. A frequent critical issue is, even when a tumor is diagnosed at a clinically “localized” stage, subclinical or micro-metastasis might have occurred—since the limiting resolution of detecting a tumor remains on the order of 1 cc, which represents a billion (109) cells.
4. Cure of cancer at a late, metastatic stage is often much more difficult to achieve than when it is at the early, localized stage.
5. Despite the importance of systemic control of all cancer cells, local control of the primary tumor remains a sine qua non (i.e., indispensable) condition for cure.
B. Role of surgical resection
1. Malignant tumors are often infiltrative or invasive, with indistinct borders.
2. Any tumor resection should aim for eradication of all tumor cells, that is, no residual tumor grossly and negative resection margins microscopically. Thus radical or en bloc resection is often the procedure of choice for cancer surgery.
3. Note that theoretically a “90% resection” still leaves behind 100 millions (108) cells for a 1 cc tumor, and a “99% resection” still leaves behind 10 millions (107) cells. Thus, for curative purpose, “partial resection” or “debulking” of tumors should be followed by additional “adjuvant” therapy.
4. Radical resection can often result in undesirable sequelae, for example, sacrifice of organs, compromise of physiologic functions, or unacceptable cosmetic defects.
C. Role of RT
1. For primary tumor embedded within or surrounded by critical normal tissues or organs, RT has the biologic advantage (especially via fractionation, see Section III.C) of preserving these structures while aiming for eradication of allcancer cells within the targeted volume.
2. A typical 1 cc epithelial tumor requires about 70 Gy in 7 weeks to achieve a TCP (i.e., the probability of killing all cancer cells so none survive; see Section III.F.1) of 90% or above. This is achievable through routine conventional RT.
3. RT is often used to supplement surgery by mitigating the limiting factors for resection of malignant tumors.
a. Primary RT aims to substitute surgical resection all together.
b. Adjuvant RT aims to supplement the surgical resection (whether radical or conservative) to further eradicate residual disease.
c. Neoadjuvant RT aims to “shrink” the extent of tumor before a planned surgery and enhance the chance of a successful resection (e.g., achieving clear margin). It may attempt to either convert an otherwise unresectable tumor to a resectable one or enable organ-preserving surgery rather than a radical procedure.
4. Examples of radical surgery versus “organ-conservation” treatment in oncology include
a. Breast cancer: modified radical mastectomy versus lumpectomy + RT
b. Advanced laryngeal cancer: total laryngectomy versus primary RT + chemo
c. Anal canal cancer: abdominal-perineal resection with end colostomy versus primary RT + chemo
d. Muscle invading bladder cancer: radical cystectomy with urinary diversion versus primary RT + chemo
e. Esophageal cancer: esophagectomy versus primary RT + chemo
f. Ocular melanoma: enucleation versus primary RT
g. Soft tissue sarcoma of the extremity: amputation versus limb-salvage resection + RT
5. For “regional” spread of cancer along the draining lymphatic chains, RT can likewise play a substantial role to either substitute or supplement surgical nodal dissection.
6. As precision-oriented RT (see Section IV.C) emulates the ablative role of cancer surgery, it would face the same limitation of the latter, that is, uncertainty of the “ablation” margin. However, such primary RT would be depleted of the critical information otherwise available through postoperative pathologic assessment. Thus, its routine use needs to be done judiciously, especially if utilized alone without supplemental wide-field RT.
D. Role of chemotherapy
1. Chemotherapy aims primarily for the control of systemic spread of cancer cells (i.e., metastasis). The timing of such systemic treatment is crucial, depending on the metastatic potency of the specific tumor type.
2. Some primary malignancies, especially those which originate from hematopoietic cell lines such as leukemia and lymphomas or certain germ cell tumors, are highly sensitive to selected chemotherapeutic agents. Thus, they are treated predominantly by such drugs even if the disease presents at a localized stage.
3. Some chemotherapy agents such as cisplatin or 5-fluorouracil may generate synergistic effects if combined with RT. These “radiosensitizing” drugs are invaluable in helping RT further achieve its intended goal of locoregional tumor control as well as facilitating organ conservation approaches (see Section V.C.4).
4. Other systemic agents such as anthracylines are known to be “radiomimmetic”: when given after the completion of RT, these drugs may bring back the once acute RT side effects such as dermatitis or mucositis, described as the “recall” phenomenon. Thus, the full course of such drugs is usually completed before initiating the planned RT.