Cancer in Children: Clinical Management, 5th Edition

Chapter 23. Germ cell tumours in children and adolescents

Ulrich Göbel

Gabriele Calaminus

Dominik T. Schneider

Introduction

Germ cell tumours (GCTs) constitute a highly heterogeneous group of tumours that vary significantly with respect to clinical presentation, histology, and biology. In adolescents and adults, they most commonly develop within the gonads. In young men, testicular GCTs represent the most frequent solid tumour. In contrast, childhood GCTs predominantly arise at extragonadal midline sites such as the sacrococcygeal region, the central nervous system (CNS), and the anterior mediastinum. This heterogeneous clinical presentation requires a multimodal treatment that includes the paediatric oncologist in cooperation with appropriate surgical disciplines (paediatric surgeon, urologist, gynaecologist, thoracic surgeon, and neurosurgeon) and the radiotherapist.

During the past two decades, a dramatic improvement in the prognosis of malignant GCTs has been achieved in both the adult and paediatric populations. This progress can be attributed to national and international cooperative studies that utilize cisplatin-based chemotherapy as part of a multimodal therapeutic approach. The first paediatric trials revealed the particular clinical and biologic features of childhood GCT. Moreover, the early observations have allowed therapy to be tailored more specifically to the paediatric setting and stratification of chemotherapy according to defined risk groups to be introduced.

In this chapter, we shall review the rapid developments of recent years, and describe what should be considered as state-of-the-art therapy for paediatric GCT.

Epidemiology

GCTs can become clinically apparent in all age groups, ranging from the fetal period to adulthood. Among children aged <15 years, GCTs account for 3–4 per cent of cancers. Thus the annual incidence can be estimated as approximately 0.5 per 100 000 children aged <15 years. However, it should be taken into account that teratomas may be under-reported to cancer registries, resulting in an underestimation of the incidence rate.

The clinical data reported to the German trials for testicular and non-testicular GCTs show a distinct distribution pattern with regard to site and tumour histology1. During childhood, most GCTs present at non-gonadal sites close to the body axis, such as the sacrococcygeal region, the mediastinum, or the pineal gland. Figures 23.1 and 23.2 show that, in general, two incidence peaks can be distinguished. The first peak includes teratomas (in neonates) and yolk sac tumours (during infancy and early childhood). These predominantly arise in the sacrococcygeal region and testis, and less frequently in the mediastinum or retroperitoneum. During adolescence there is another peak of incidence which is due to gonadal, mediastinal, and CNS tumours. Histologically, these tumours resemble germinoma (also known as seminoma or dysgerminoma) or non-seminomas including embryonal carcinoma, choriocarcinoma, and yolk sac tumour. Spermatocytic seminomas are not observed in adolescents and young adults. In conclusion, the age–characteristic distribution of GCTs has a significant impact on differential diagnostic considerations and should guide therapeutic decisions.

Fig. 23.1 Age distribution with respect to primary tumour site in 1307 children and adolescents with germ cell tumours registered in the MAKEI protocols. Reproduced from D.T. Schneider et al.

Fig. 23.2 Age distribution with respect to histology in 1307 children and adolescents with germ cell tumours registered in the MAKEI protocols. YST, yolk sac tumour. Reproduced from D.T. Schneider et al.1

Table 23.1. Histologic classification of germ cell tumours (WHO)

 

Synonyms

1.

Seminoma (SE)

Dysgerminoma (DYS) (ovary), germinoma (GE) (brain)

2.

Yolk sac tumour (YST)

Endodermal sinus tumour

3.

Embryonal carcinoma (EC)

4.

Choriocarcinoma (CHC)

5.

Teratoma (TER)

5.1. Mature

   5.1.1. Solid

   5.1.2. Cystic

Dermoid cyst

5.2. Immature (IT)

5.3. With malignant transformation

5.4. Monodermal

6.

Tumours with mixed histology (MGCT)

7.

Spermatocytic seminoma (SS)

8.

Polyembryoma (POLY)

Histologic classification of germ cell tumours

GCTs are characterized by a profound heterogeneity in histologic differentiation. They are classified according to the World Health Organization (WHO) classification of testicular, ovarian, and intracranial tumours (Table 23.1). In our experience, these classifications constitute a more precise morphologic description than the traditional British classification as they permit recognition of all histologic components of mixed malignant GCT. As intratumour heterogeneity may be subtle, the initial diagnostic workup should include evaluation by a pathologist experienced in GCT histology in order to achieve a standardized and reliable histopathologic diagnosis and grading (Tables 23.1 and 23.2).

According to the holistic concept of Teilum, GCTs arise from totipotent primordial germ cells which are capable of embryonic and extraembryonic differentiation (Fig. 23.3). Yolk sac tumours and choriocarcinoma follow an extra-embryonic differentiation pattern and are characterized by significant secretion of a1-fetoprotein (AFP) or human choriogonadotropin (HCG or β-HCG), respectively (Table 23.3). Embryonal carcinomas represent tumours of immature totipotent cells. Teratomas display embryonic differentiation and may mimic organ structures of all germ layers. In teratoma, the histologic grade of immaturity is defined by the extent of immature (predominantly neuroepithelial) elements (Table 23.2).2 The histologic grading of immaturity correlates with the risk of locoregional relapse, particularly after incomplete resection, and therefore is essential for risk assessment of pure teratomas.3 Finally, germinomatous tumours [seminoma (testis), dysgerminoma (ovary), germinoma (brain)] display morphologic features of undifferentiated germinal epithelium. In contrast with testicular GCTs of adult patients, paediatric GCTs do not develop from carcinoma in situ.

Table 23.2. Biological characteristics of histologic germ cell tumour subentities

 

Histological grading

Tumour marker

Sensitivity to

AFP

β–HCG

CT

RT

Seminoma/germinoma

Malignant

(+)

+++

≥24 Gy

Embryonal carcinoma

Malignant

+++

≥45 Gy

Yolk sac tumour

Malignant

+++

+++

≥45 Gy

Choriocarcinoma

Malignant

+++

+++

≥45 Gy

Teratoma, mature/immature

Benign/potentially malignant

–/(+)

?

?

AFP, α-fetoprotein; β–HCG, human chorionic gonadotrophin; CT, chemotherapy; RT, radiotherapy.

Fig. 23.3 Characteristic histologic morphology of paediatric germ cell tumours. (a) Mature sacrococcygeal teratoma displaying bone, cartilage, bone marrow, fibrous tissue, glial tissue and squamous epithelium (H&E, 10). (b) Immature sacrococcygeal teratoma with immature neurotubuli (H&E, 10). (c) Immature sacrococcygeal teratoma with microfoci of yolk sac tumour (AFP immunohistochemistry, 10). (d) Mediastinal yolk sac tumour with reticular and papillary growth pattern (H&E, 10). (e) Pineal germinoma with clear cytoplasm and rounded nuclei with prominent nucleoli; scattered nests of lymphocytes (H&E, 40). (f) Mediastinal choriocarcinoma with syncytiotrophoblastic giant cells (H&E, 40).

In most patients, the response to the different therapeutic modalities can be predicted from the histologic appearance and the tumour marker profile (Table 23.3). About 25 per cent of all paediatric GCTs present as tumours with more than one histologic type. In this situation therapy and prognosis depend on the component with the highest malignancy.4

Table 23.3. Histologic grading of teratomas (Gonzalez-Crussi et al.2)

Mature teratoma (grade 0)

All tissue components appear well differentiated

Immature teratoma
   Grade 1
   Grade 2
   Grade 3

Some components consist of incompletely differentiated tissue
Occasional foci ≤10% of the sampled surface
10%–50% of the sampled surface
>50% of the sampled surface consist of undifferentiated tissue

Biology

Molecular studies of the imprinting status substantiate the hypothesis that non-gonadal GCTs develop from germ cells that have mislocated during their embryonic development.5 While no consistent correlation between cytogenetic aberration and primary tumour site has been observed, it is apparent that histology (teratoma versus malignant GCT) and age (prepubertal versus postpubertal) both significantly correlate with distinct genetic profiles.6 More than 80 per cent of malignant testicular GCTs of young males display a distinct and specific chromosomal aberration, the isochromosome 12p.7 The remaining isochromosome-12p-negative tumours frequently show amplification of 12p (homogeneously staining regions or tandem repeats), and candidate genes have recently been identified in this region. These aberrations have been observed in both testicular and ovarian tumours, and in mediastinal GCTs. In mediastinal GCTs, there is a considerable association with constitutional Klinefelter syndrome, which constitutes a major risk factor for the early development of malignant mediastinal GCTs.

Ovarian GCTs show a significant association with constitutional aberrations of the sex chromosomes, i.e. lack of the second X chromosome (Ullrich–Turner syndrome) or presence of the Y chromosome at the microscopic or submicroscopic level. In such patients, ovarian dysgerminomas may develop within gonadoblastomas, which may therefore be considered a premalignant stage. Furthermore, these tumours may sometimes be associated with abnormal development of the external and internal genitalia (e.g. streak gonads in Swyer syndrome) and testicular feminization. In characteristic cases, these ‘ovaries’ histologically resemble testes. A prophylactic removal of the second ovary is mandatory in these patients because of the significant risk of contralateral tumours and the obvious lack of physiologic reproductive function.

In contrast with adult patients, isochromosome 12p has rarely been found in malignant GCT in children aged <10 years. On the other hand, aberrations at chromosomes 1, 6, and 20 and the sex chromosomes have frequently been found. Studies of childhood yolk sac tumours have revealed frequent loss of heterozygosity at the distal part of the short arm of chromosome 1 and the long arm of chromosome 6 in a region that includes the IGF-2 receptor gene.

Finally, virtually all pure teratomas are cytogenetically normal. However, cystic teratoma of the ovary may present with isodisomic karyotype, consistent with its origin from postmeiotic germ cells.

Diagnosis

In general, GCTs tend to occur as indolent masses and clinical symptoms are mostly related to local tumour growth. For instance, mediastinal GCTs may cause airway obstruction or superior vena cava syndrome, while pelvic GCTs may lead to obstipation. A defined programme of clinical, radiographic, and laboratory investigations has to be followed in a timely fashion in order to address the central strategic questions as to whether a biopsy must be obtained for diagnostic purposes, and whether a primary tumour resection or preoperative chemotherapy followed by delayed resection is preferable.

Tumour markers

Depending on their histologic differentiation, GCT tend to secrete the tumour markers AFP and/or HCG/β-HCG (Table 23.3). These facilitate clinical diagnosis in tumours that present at typical sites.8However, it must be taken into account that serum AFP levels may be greatly elevated in neonates and infants.9 It is important to note that, in contrast with other reports, this series of normal neonates and infants has demonstrated that in a significant percentage of healthy children, AFP does not decline to the normal range of adult patients before the end of the second year of life.9 Therefore, in the first 2 years of life, only AFP levels significantly above the age-related normal value can be regarded as diagnostic for a secreting GCT.

In general, the tumour marker profile is highly specific for the histologic differentiation of the tumour (Table 23.3).8 However, there may be secretion of β chains of HCG in seminoma/ germinoma (<50 IU/l), which is often related to syncytiotrophoblast-like giant cells. Some patients with immature teratoma show a moderate elevation of the AFP level (<100µg/l), sometimes associated with histologically detectable small yolk sac tumour foci within the teratoma. On the other hand, microscopic yolk sac tumour foci (Fig. 23.3) may also be present and not result in clinically apparent elevation of AFP. In CNS GCT, tumour markers may frequently be elevated at different levels in CSF and serum. Their values can be within normal range in CSF and elevated in serum or vice versa. Therefore measurement in both compartments is mandatory at diagnosis. About 20 per cent of germinoma may secrete placenta-like alkaline phosphatase (PLAP), which can be used as an additional diagnostic tool if elevated.

Diagnostic imaging

Appropriate radiographic procedures must be performed according to tumour site and potential routes of tumour dissemination. In most patients with extracranial tumours, the initial radiographic assessment of the tumour will be made by ultrasound. During ultrasound, the tumour should be measured in three dimensions, and the abdomen and the lymph nodes should also be screened for metastases. The next step is to perform CTor preferably MRI scans of the tumour.

In pelvic tumours, imaging includes the upper abdomen in order to detect lymphatic spread to the nodes at the renal veins or liver metastases. Sacrococcygeal tumours must be evaluated with MRI including sagittal images to exclude spread into the sacral and spinal canal.

In CNS GCTs MRI examinations should include axial, sagittal, and coronal images of the whole brain and the complete spine, and these must be supplemented by CSF cytology in order to detect CSF metastases. There is a consensus that if a bifocal CNS tumour is diagnosed radiologically (pineal and suprasellar region), the radiologic features together with negative markers in CSF/serum and negative CSF cytology are sufficient for diagnosis of a germinoma.

Skeletal metastases have been observed in 10 per cent of patients; however, their occurrence has not adversely affected prognosis. Metastases to the CNS, which may be found in 4 per cent of young males with malignant testicular GCTs are extremely rare (<1 per cent) during childhood. Therefore MRI scans of the CNS are not considered mandatory in extracranial GCTs.

Laboratory studies (pretreatment)

In addition to the tumour markers AFP and β-HCG, serum lactate dehydrogenase (LDH) has proved to be a prognostic marker in adult patients with GCTs. In germinomas, PLAP can be measured in the serum and may then serve as a marker of treatment response during follow-up. In addition to the routine blood tests before chemotherapy, special attention should be paid to renal function (creatinine clearance, urine electrolytes), as several cytotoxic agents such as platinum-compounds and ifosfamide may interfere with this. Finally, a cytogenetic analysis of blood lymphocytes is recommended in female patients with malignant ovarian GCTs and male patients with mediastinal GCTs in order to exclude constitutional aberrations such as Ullrich-Turner or Klinefelter syndromes.

Therapy

Surgery

If the initial radiographic assessment demonstrates infiltration into adjacent organs and/ or metastases, initial (up-front) chemotherapy followed by delayed tumour resection is recommended, as preoperative chemotherapy will facilitate complete resection at delayed surgery.10 Tumour marker measurement in combination with imaging permits a clinical diagnosis at most sites, except the liver and the retroperitoneum.8 In equivocal cases (i.e. non-diagnostic markers, hepatic or upper retroperitoneal tumours), a diagnostic biopsy is recommended, which should include obtaining fresh tumour tissue for genetic and biologic studies.

If the radiographic assessment indicates a localized tumour without metastatic spread, the treatment of choice is primary tumour resection, except in the CNS as this region carries specific risks of surgical morbidity. In general, there is no role for debulking surgery in paediatric GCT. Surgery must always aim for a microscopically complete resection, as incomplete resection carries the risk of recurrence, even with adjuvant chemotherapy.10 The criteria of complete resection are the resection of the tumour with capsule and adjacent organs en bloc. If this is not achieved, tumour resection has to be regarded as incomplete and an increased risk of recurrence must be considered.

In patients with tumour residues after initial tumour resection, second-look surgery is essential to achieve secondary complete resection. This is also the case in malignant nongerminomatous CNS GCTs. Second-look surgery may at least partially overcome the otherwise unfavourable prognostic impact of incomplete resection.10 Finally, surgery of metastases is not indicated unless they show insufficient response to chemotherapy.10,11

Surgical techniques:

Rescorla12 has recently reviewed surgical techniques in paediatric GCTs.

Testis

The resection of testicular tumours is performed by unilateral orchidectomy after high inguinal incision. Trans-scrotal biopsy is obsolete and such a procedure will be considered an indication for adjuvant chemotherapy, even in otherwise stage I tumours. Retroperitoneal lymph node dissection is not advocated because of the increased risk of retrograde ejaculation and the overall favourable response to adjuvant chemotherapy in childhood GCTs.

Ovary

The resection of ovarian tumours should be performed after midline laparatomy. Because of the risk of bilateral tumours, a biopsy of the contralateral ovary should be performed in cases of macroscopically suspicious appearance. Intraperitoneal fluid or ascites must be examined cytologically to exclude malignant ascites.

Coccyx

Coccygeal tumours frequently present as large tumours with intrapelvic and sometimes extrapelvic components. Therefore a dorsal approach is used in most patients. Infiltrated skin areas should be removed en bloc with the tumour. Postoperatively, bowel and bladder function recover in the vast majority of patients. The coccyx should be resected en bloc to avoid tumour rupture. The so-called hourglass formation of coccygeal tumours refers to a huge tumour partially sited in the bony pelvis incorporating the coccygeal region. In these tumours, an additional ventral approach is useful to achieve a complete en bloc resection.

Central nervous system

As most CNS GCTs are located centrally, surgery is often difficult and prone to complications. In addition, it is almost impossible to achieve a complete resection of an intracranial tumour.13 Therefore radiotherapy has become an integral part of the treatment of intracranial GCTs. As a consequence, surgical interventions should now be reserved for diagnostic biopsies in non-secreting tumours or for tumours with significant residues after chemotherapy and/or radiotherapy. Only neonatal teratomas of the CNS should be considered an exception from this recommendation, as tumour resection constitutes the only meaningful therapeutic approach for these patients who often present with large tumours and hydrocephalus.

The surgical approach to the other extragonadal sites such as the mediastinum or retroperitoneum must be planned according to the presenting situation.

Cisplatin-based chemotherapy

Until 1980, the prognosis of children with malignant GCTs was poor, and outcome was determined by the parameters of age, site, histology, and stage. The modern era of GCT chemotherapy began in the mid-1970s with the identification of the efficacy of cisplatin in testicular GCTs. In 1977, Einhorn and Donohue14 reported a complete response rate of 85 per cent in patients with metastatic testicular GCTs treated with a combination of cisplatin, vinblastine, and bleomycin (PVB) and tumour resection. Most importantly, in contrast with a previously reported regimen using only vinblastine and bleomycin, the overall good response was also translated into durable remissions.

Nevertheless, relapses or refractory cancers, although rare, established the need for secondline therapies. Etoposide soon emerged as an active drug with a single-agent efficacy superior to vinblastine. Moreover, etoposide shows a more favourable acute toxicity profile with less neuromuscular toxicity. On the other hand, the use of etoposide may be associated with therapy-related leukaemias in 1–2 per cent of patients.

In addition, the efficacy of ifosfamide in cisplatin-refractory GCTs has been documented. The combination of cisplatin with etoposide and ifosfamide for recurrent testicular GCTs results in a 30 per cent durable remission rate and can now be considered standard treatment for relapse. These observations initiated studies that included etoposide and/or ifosfamide in the first-line treatment of GCTs.

The combination of cisplatin, etoposide and bleomycin (BEP) is now considered standard chemotherapy for GCTs in adult patients. In most patients, a total of three cycles is considered appropriate. The inclusion of ifosfamide in first-line treatment (PEI) did not result in significantly improved outcome but was associated with higher bone marrow and renal toxicity compared with BEP.

In relapsing and refractory GCTs, the therapeutic value of high-dose chemotherapy with autologous stem cell transplantation has been investigated. These analyses have shown only limited efficacy in prognostically unfavourable tumours such as cisplatin-resistant mediastinal GCTs with high β-HCG or multiple relapses. Nevertheless, introduction of high-dose chemotherapy into first-line treatment of high-risk tumours may be beneficial in some patients. Finally, ‘modern’ drugs, such as paclitaxel and gemcitabine, are currently under investigation.

Development of cooperative protocols for paediatric GCTs

Encouraged by the data discussed above, prospective protocols for gonadal and non-gonadal GCTs in children and adolescents were initiated. The first published trial was conducted by the US Children's Cancer Group (CCG) and included 54 children with malignant nonseminomatous GCT. Patients underwent initial tumour resection followed by VAC + PVB chemotherapy over a 2-year period, second-look resection 4 months after diagnosis, and irradiation if there was residual tumour15. Fifteen of 20 evaluable patients with ovarian nonseminomatous GCTs, including a substantial proportion of high-stage tumours, achieved continuous clinical remission. The prognosis of children with non-gonadal GCTs was worse but still encouraging compared with all other previous studies.

The analysis of the following CCG protocol included 93 children and showed that patients with ovarian GCTs had a better prognosis [4-year event-free survival (EFS) of 63 per cent] than children with non-gonadal GCTs (4-year EFS of 42 per cent).16 This difference was mainly attributed to a higher rate of incomplete tumour resections in non-gonadal tumours.

In the US Intergroup protocol, a watch-and-wait policy was followed in stage I testicular GCTs. Intermediate-risk patients (testicular stage II, ovarian, and non-gonadal stage I–II) received four cycles of cisplatin, etoposide, and bleomycin (reduced to one infusion per cycle compared with three infusions in corresponding adult regimens). Furthermore, the therapeutic impact of cisplatin dose intensification at 200 mg/m2/cycle was evaluated in high-risk patients (stage III–IV). The analysis of both gonadal and non-gonadal GCTs revealed that higher doses of cisplatin may result in higher response and complete remission rates (9 per cent benefit), but with significantly higher renal and auditory toxicity. Notably, two-thirds of patients in the high-dose cisplatin arm required hearing aids after completion of treatment. More recent investigations of this study group aimed to evaluate amifostine protection during cisplatin therapy at escalated doses; however, no significant benefit with regard to ototoxicity has been demonstrated, and amifostine therapy was associated with significant electrolyte imbalances, particularly hypocalcaemia (data presented at the ASCO Conference, 2003).

The analysis of different chemotherapy regimens administered in the British UKCCSG GC I and GC II protocols also demonstrated the high therapeutic efficacy of platinum-based regimens such as BEP or JEB (carboplatin 600 mg/m2/cycle, etoposide, and bleomycin), which resulted in a 5-year EFS of 57 per cent and 87 per cent, respectively, in non-gonadal GCTs.17 The recent analysis of the UKCCSG GC II study emphasizes the high efficacy of the JEB regimen, with a 5-year EFS of 88 per cent and a favourable toxicity profile.18

The French study group reported 35 children with ovarian and non-gonadal advanced stage GCTs who were treated with a VAC + PB regimen.

19 The French cooperative protocol TGM 85 used a similar chemotherapeutic approach, and in the next TGM 90 protocol cisplatin was replaced by carboplatin (400 mg/m2/cycle).20 The results were less favourable with this regimen than with the British JEB regimen. This difference was mainly attributed to the lower single and cumulative doses of carboplatin. In the recent French protocol, alternating combinations of cisplatin with etoposide or ifosfamide are administered, resulting in a superior response rate compared to the previous carboplatin based strategy.

In both the French TGM 90 and the British GC II studies, the analysis of prognostic factors revealed the prognostic impact of high AFP serum levels at diagnosis, a finding that could not be confirmedin other studies using cisplatin-based regimens10,21 or in theongoing Frenchprotocol.

From 1982 onwards, the German protocols for testicular (MAHO) and non-testicular (MAKEI) GCTs included cisplatin- and etoposide-based chemotherapy regimens integrated into a multimodal approach that also recommended delayed tumour resection after preoperative or neoadjuvant chemotherapy. As early reports indicated an increased risk of bleomycin-associated pulmonary toxicity in children (particularly infants) compared with adults, the German MAKEI protocol recommended the sequential administration of bleomycin and cisplatin. Furthermore, from 1989 onwards, the bleomycin dose was reduced in children aged <2 years and withheld in infants. As a result of the excellent EFS of >80 per cent achieved with the first MAKEI and MAHO protocols, the cumulative chemotherapy was reduced stepwise from a maximum of eight to six cycles, and currently to four or five cycles. This reduction of cumulative chemotherapy did not affect outcome.10 In addition, a risk stratification of chemotherapy according to age, site, histology, stage, and completeness of resection has been introduced: The favourable outcome for infants with malignant GCTs from whom bleomycin was withheld because of their age, indicated the higher chemosensitivity of malignant GCTs in this age group and encouraged the reduction of chemotherapy to a two-agent regimen with cisplatin and etoposide in some low-risk patients. Moreover, according to the current MAKEI 96 protocol, an expectant watch-and-wait strategy is recommended for patients with completely resected low-stage tumours. This approach makes it possible to avoid chemotherapy in 75 per cent of patients with stage Ia malignant GCTs. However, patients who relapse during the surveillance require a more intensive regimen with four cycles of three-agent chemotherapy, and therefore therapy is intensified in these patients. In locally advanced and/or metastatic tumours a neoadjuvant approach appears beneficial as it facilitates complete tumour resection and thereby reduces the need for second-look surgery.

In summary, the PEI, BEP, CarboPEI, and JEB regimens (Table 23.4) have synergistic cytotoxic activity and can be considered as standard regimens of comparable efficacy. These regimens are used in the paediatric GCT protocols currently open.

Side effects of chemotherapy

Pulmonary toxicity of bleomycin, which had also been reported as a problem in combination with impaired kidney function or enhanced by anaesthesia, led to the introduction of a regimen without this drug. However, the current US Intergroup Study has demonstrated that a reduction of bleomycin from three to one infusions per cycle does not affect outcome. In this study, no deaths due to pulmonary toxicity have been observed. On the other hand, the highly efficient combination of cisplatin, etoposide, and ifosfamide (PEI) is associated with a higher degree of myelosuppression and carries the risk of tubular nephropathy, particularly in small children. In our experience with this regimen, we observed clinically apparent hearing impairment in 5 per cent of patients. In contrast, two-thirds of patients treated with high-dose cisplatin (200 mg/m2 per cycle) require hearing aids. Although the auditory toxicity and the renal toxicity of carboplatin regimens are smaller, carboplatin at effective doses (600 mg/m2/ cycle) causes substantial myelotoxicity.18 The risk of therapy-related secondary leukaemia is dependent on the therapeutic modalities used, with an estimated cumulative risk of 1.0 per cent (3/442 patients) (Kaplan–Meier method at 10-year follow-up) for children treated with surgery and chemotherapy only and 4.2 per cent (3/174 patients) for children treated with combined radio- and chemotherapy.

Therapeutic strategies of current protocols for malignant germ cell tumours

In general, current paediatric GCT protocols apply different strategies. Some protocols define cumulative chemotherapy according to the response to treatment (e.g. one standard chemotherapy regimen to a total of two cycles after achieving complete remission17,18). In other protocols, therapy is stratified according to initial diagnostic parameters, and is only intensified if there is insufficient response to treatment. In the following sections, we summarize the international SIOP-CNS-GCT 96 protocol and the common strategies for the risk-stratified treatment of extracranial GCTs.

Table 23.4. Standard chemotherapy regimens in paediatric GCT

PEI (MAKEI 96, SIOP-CNS-GCT 96, MAHO 98, SFOP)

Cisplatina

20 mg/m2

Over 1 h

Day 1, 2, 3, 4, 5

Etoposide

100 mg/m2

Over 3 h

Day 1, 2, 3

Ifosfamideb

1500 mg/m2

Over 20 h

Day 1, 2, 3, 4, 5

Two to four cycles

PVB (MAHO 98)

Cisplatina

20 mg/m2

Over 1 h

Day 4, 5, 6, 7, 8

Vinblastine

3 mg/m2 or 0.15 mg/kg

IV bolus

Day 1, 2

Bleomycinc

15 mg/m2

Over 24 h

Day 1, 2, 3

Three cycles

BEP (MAHO 98)

Bleomycinc

15 mg/m2

Over 24 h

Day 1, 2, 3

Etoposide

80 mg/m2

Over 3 h

Day 1, 2, 3

Cisplatina

20 mg/m2

Over 1 h

Day 4, 5, 6, 7, 8

Three cycles

BEP (US Childrens Oncology Group)

Bleomycin

15 mg/m2

Over 24 h

Day 1

Etoposide

100 mg/m2

Over 3 h

Day 1, 2, 3, 4, 5

Cisplatina

20 mg/m2

Over 1 h

Day 1, 2, 3, 4, 5

Four cycles

High-dose BEP (US Children's Oncology Group)

Bleomycin

15 mg/m2

Over 24 h

Day 1

Etoposide

100 mg/m2

Over 3 h

Day 1, 2, 3, 4, 5

Cisplatina

40 mg/m2

Over 1 h

Day 1, 2, 3, 4, 5

Four cycles

JEB (UKCCSG GCII)

Carboplatin

600 mg/m2

Over 1 h

Day 2

Etoposide

120 mg/m2

Over 1 h

Day 1, 2, 3

Bleomycinc

15 mg/m2

Over 15 min

Day 3

Five cycles, or two cycles after complete remission

CarboPEI (SIOP-CNS-GCT 96)

Carboplatin

600 mg/m2

Over 1 h

Day 1

Etoposide

100 mg/m2

Over 3 h

Day 1, 2, 3, 22, 23, 24

Ifosfamideb

1800 mg/m2

Over 3 h

Day 22, 23, 24, 25, 26

Two cycles

aPlus mannitol-forced diuresis.
bPlus mesna uroprotection.
cOmitted in children aged <1 year; 7.5 mg/m2 in children aged <2 years.

SIOP-CNS-GCT 96 protocol for malignant intracranial GCTs

Therapy for malignant intracranial GCTs is stratified according to histologic differentiation (i.e. germinoma versus secreting GCT) and initial dissemination.22,23 The ongoing SIOP-CNS-GCT protocol aims to evaluate and standardize diagnostic procedures, which include measurement of markers in serum/CSF, CSF cytology, and MRI of head and spine in all patients. Two different therapeutic options in intracranial germinoma are being evaluated with regard to both their therapeutic impact and their specific acute and long-term toxicity. For secreting intracranial tumours and embryonal carcinoma, the effect of a combined treatment with PEI (Table 23.4) and radiotherapy adapted to dissemination is examined.22

In pure intracranial germinoma, which accounts for 50 per cent of all intracranial GCTs which do not secrete significant amounts of HCG/β-HCG, histologic verification of the tumour is mandatory. According to the current SIOP-CNS-GCT 96 protocol, patients with germinoma and localized disease can be treated either with craniospinal irradiation with 24 Gy and a tumour boost of 16 Gy or with a multimodal treatment including two cycles of chemotherapy [CarboPEI (Table 23.4)] followed by focal irradiation (40 Gy). In metastatic disease craniospinal irradiation and boost to tumour and metastatic sites is still the treatment of choice. The higher local dose to the primary tumour also aims for the control of potential small foci of non-germinomatous histology such as syncytiotrophoblastic cells that may have been missed by biopsy. To date, data concerning the effect of chemotherapy and localized RT reveal that this approach carries a higher risk of ventricular relapses (Cefalo, data presented at the SIOP Conference, 1995; Alapetite, data presented at the SIOP Conference, 2002). Therefore additional ventricular treatment is implemented in new protocols for localized CNS germinoma. As GCT may arise adjacent to sensitive structures such as the optic chiasma, it is recommended that a paediatric radiotherapist should be consulted about optimal treatment techniques. It has been demonstrated that a 5-year EFS of 91 per cent and 5-year survival of 94 per cent can be achieved with radiotherapy only,24 but because of the higher risk of ventricular failures, the 5-year EFS of patients treated with combined therapy is 85 per cent and survival is 92 per cent. Another important risk factor is incomplete staging in germinoma, particularly with regard to marker evaluation in serum/CSF. More than 50 per cent of relapsing patients show secretion of markers which had not been measured at the initial diagnosis of a germinoma (Calaminus, data presented at the SIOP Conference, 2003). The secreting intracranial GCTs show a poorer prognosis than germinomas. In these patients, four cycles of cisplatin-based chemotherapy [PEI (Table 23.4)] are given, followed by delayed tumour resection and radiotherapy. The radiotherapy is stratified according to initial stage. Non-metastatic tumours receive focal irradiation (54 Gy), whereas patients with intracranial or spinal metastases or tumour cells in the CSF receive craniospinal irradiation (30 Gy plus a tumour boost of 24 Gy). Data from Balmaceda reveal24 that chemotherapy alone is able to cure only 30% of the patients with malignant CNS GCT.

The summary of several cooperative protocols and the preliminary data of the SIOP-CNS-GCT 96 protocol suggest that long-term remission can be obtained in about two-thirds of patients. In the SIOP-CNS-GCT 96 protocol, AFP > 1000 µg/l at diagnosis and residual disease at the end of treatment have been defined as clinical risk factors and will be used for definition of risk groups in the forthcoming SIOP-CNS-GCT II protocol.

Risk-stratified treatment of extracranial germ cell tumours

Both gonadal and extragonadal GCTs are treated according to a similar therapeutic concept. Only in small non-metastatic tumours, in which radiographic evaluation shows no evidence of invasive growth beyond the organ of origin, is a primary resection recommended. In patients with bulky, invasive, or metastatic tumours, preoperative chemotherapy followed by a delayed tumour resection is preferred to avoid the risk of incomplete resection. The fall of the tumour markers according to their serum half-lives indicates a favourable response to chemotherapy.7,8

Adjuvant treatment

In current protocols, patients with completely resected testicular stage Ia teratoma or yolk sac tumour are treated according to a watch-and-wait strategy which includes frequent (weekly) controls of the relevant tumour marker AFP. Disease progression can be detected either as failure to normalize AFP according to its half-life or as lymph node metastases at the renal veins. With this treatment approach, chemotherapy can be omitted in >70 per cent of patients. If progression occurs, all patients can be salvaged by conventional chemotherapy. Therefore chemotherapy is reserved for those patients who show disease progression and for patients with malignant testicular tumours that are stage Ib or higher.

Ovarian tumours are treated according to the same strategy. Stage I tumours are followed expectantly and higher-stage tumours receive adjuvant chemotherapy. In the current German MAKEI protocol, completely resected stage II tumours receive two to three cycles of a two-agent regimen [PE (Table 23.4)]. After incomplete resection a three-agent combination is applied [PEI (Table 23.4)]. In other protocols, all stage II–IV patients receive a minimum of four cycles of three-agent chemotherapy (e.g. JEB or BEP). Ovarian dysgerminomas are treated according to the same strategy. Irradiation is omitted to preserve fertility.

Teratoma

Teratoma represent a distinct histologic entity which shows a significant diversity in the clinical course depending on the histologic grade of immaturity.2,3 Mature teratoma are considered benign, whereas immature teratomas may show clinical features of malignancy. The risk of recurrence can be estimated from the parameters of primary site of the tumour, histologic grade of immaturity, and completeness of tumour resection.3 The role of adjuvant chemotherapy has not yet been established. However, recent reports have shown that chemotherapy may not be indicated after complete tumour resection, even in the presence of small foci of yolk sac tumour.3,26 Incompletely resected tumours have a 10 per cent risk of relapse in mature teratomas and a 20 per cent risk in immature teratomas, with increasing risk with respect to the grade of immaturity irrespective of adjuvant chemotherapy.3 Histologically, half the recurrent tumours may display yolk sac tumour. However, after adjuvant chemotherapy no malignant histology has been observed at relapse, although chemotherapy did not reduce the overall risk of recurrence.

Follow-up

A complete clinical remission is defined as normalization of the tumour markers to within the age-related normal range and the absence of suspicious residual structures, even in patients with normalized tumour markers, as these structures may represent remaining mature teratoma. If any of these criteria are not fulfilled, a diagnostic re-evaluation and change or intensification of treatment, if necessary, is urgently indicated. Most relapses occur within the first 2 years after diagnosis. However, in some patients, late recurrences have been observed up to 5 years after diagnosis of malignant ovarian GCT or intracranial germinoma. Therefore the initial follow-up examinations after completion of chemotherapy must be performed at short intervals, with frequent (i.e. weekly) measurements of the tumour markers AFP and β-HCG early during follow-up. In watch-and-wait patients, the fall in AFP values is evaluated with regard to its serum half-life of approximately 6–7 days. The interpretation of AFP may be difficult, especially in infants aged <2 years, because of the physiologically elevated serum levels. In this context, it has proved helpful to compare the fall in AFP in neonates and infants.9

A delayed decline or a secondary rise in AFP levels strongly indicates incomplete tumour resection or a recurrence of yolk sac tumour.

In addition, follow-up examinations include repeated imaging of the primary site of tumour. If there is residual abnormality after chemotherapy of an extracranial GCT, resection is indicated since mature teratoma may have remained and carry the risk of tumour progression. Positron emission tomography has not proved useful in this situation, as often it cannot distinguish between mature teratoma and residual necrosis or scars. Nevertheless, in cases where complete resection after chemotherapy remains difficult or impossible, this new imaging procedure may add important information concerning the remaining metabolic/proliferative activity of the residue.

In intracranial tumours, endocrinologic tests at diagnosis and during follow-up are mandatory, since tumours of the suprasellar region in particular may be associated with endocrinologic symptoms such as diabetes insipidus or panhypopituitrism. A detailed neurologic and if possible neuropsychologic evaluation, with investigation of visuomotor function and cognitive abilities, should be carried out. This information may lead to more structured and individually adapted rehabilitation and reintegration.27 In children treated with cisplatin-containing polychemotherapy (especially with ifosfamide), renal function must be monitored carefully for tubular nephropathy, and audiometry should be performed at diagnosis and before every course of platinum-based treatment.

In children, prolonged phosphaturia may lead to renal rickets with consequent growth retardation, while adolescents are at risk of renal osteomalacia. These long-term sequelae can be avoided by phosphate supplementation. Further attention should be drawn to the risk of therapy-related secondary leukaemia that depends on treatment intensity and type.

Relapse treatment

Cisplatin-based regimens (preferably PEI) have been successfully used in patients with recurrent or refractory tumours who have previously been treated with a non-platinum or noncarboplatin therapy.20Therefore we prefer a cisplatin-containing regimen in patients with relapsed tumours, if the organ toxicities related to the previous treatment allow further cisplatin therapy. On the other hand, patients suffering from severe cisplatin-related toxicity can be treated with a combination of carboplatin and high-dose etoposide (400–600 mg/m2 on 3 days). Otherwise, there is no international consensus on strategies for treatment of recurrent GCTs. However, in Germany, the following standardized strategy is used for patients with recurrent extracranial malignant GCTs.

In our experience, >90 per cent of relapses occur at the primary site of the tumour. For example, in our series of 104 sacrococcygeal yolk sac tumours only two patients had distant recurrence, whereas 17 patients had local relapse and three patients had combined local and distant relapse.11 Therefore relapse chemotherapy should be accompanied by intensive local therapy, preferably complete resection of the recurrent tumour after reduction by preoperative chemotherapy. We have been able to demonstrate that patients with local recurrences and poor response to conventional chemotherapy may profit from locoregional hyperthermia combined with platinum-based chemotherapy. This approach significantly enhanced local tumour control.11,28 However, analysis of treatment of relapsed sacrococcygeal GCTs emphasized the need to use hyperthermia early, as it showed no beneficial effect in late relapses. This probably resulted from cisplatin resistance or delayed chemotherapy due to the myelotoxicity of previous treatment.11 Finally, high-dose local irradiation at doses >45 Gy has shown some beneficial effect after incomplete resection of the recurrent tumour, whereas irradiation at lower doses was ineffective.11 In our experience, high-dose chemotherapy with stem cell support, as applied in adult patients, has resulted in long-term remissions only in those patients in whom a clinically complete remission could be achieved prior to high dose chemotherapy.11 Therefore we regard high-dose chemotherapy as indicated for consolidation treatment only.

In conclusion, failure of local tumour control at the primary site of tumour represents the main problem in most patients. Further significant advances in relapsing GCTwill probably be based on additional improvement in local therapy.

The chance of achieving a second remission for malignant CNS GCTs, especially of nongerminomatous histology, which fail after first line treatment is small. Reports from the French SFOP series and observation within the SIOP-CNS-GCT 96 protocol reveal that, although tumours do respond to chemotherapy again, a continuous second remission could only be achieved in cases with complete biologic remission, removal of any residual tumour, and successful application of high-dose treatment with additional irradiation.

Future perspectives

A multimodal approach that utilizes cisplatin/etoposide chemotherapy as well as tumour resection is highly effective for the treatment of paediatric GCTs. In the light of the high cure rates achieved by current protocols, research must now focus on new aims. First, treatment must be further intensified in those patients with cisplatin-refractory or poorly responding tumours. Locoregional hyperthermia constitutes a promising therapeutic concept in extracranial tumours as cisplatin is a very good thermosensitizer. Thus hyperthermia may overcome cisplatin resistance.28 However, the clinical data show that hyperthermia should be used early. Ideally it should be integrated into first-line treatment of poorly responding tumours.11

On the other hand, patients should be identified who are only at a low risk of relapse, and in whom adjuvant chemotherapy can either be withheld or significantly reduced, thus minimizing the impact on short- and long-term quality of life and treatment toxicity. In this context, molecular genetic studies may also reveal important information that may be utilized for risk stratification. In conclusion, in the future clinical and molecular biologic information may help to distinguish low-risk from high-risk patients accurately, thereby facilitating a multimodal approach to the individual patient. International cooperation is vital not only to achieve a standardized diagnosis and treatment for this rare disease, especially for unfavourable sites such as the CNS, but also to detect risk factors, to define risk groups, and to obtain results in a shorter period of time. Another important requirement is to focus on rehabilitation, reintegration and quality of life of the cured patients to determine their quality of survival and, if impaired, to consider these results in future treatment planning.

References

1. Schneider DT, Calaminus G, Koch S. et al. (2004). Epidemiological analysis of 1442 children and adolescents registered in the German germ cell tumor protocols. Pediatr Blood Cancer 42, 169–75.

2. Gonzalez-Crussi F, Winkler RF, Mirkin DL (1978). Sacrococcygeal teratomas in infants and children: relationship of histology and prognosis in 40 cases. Arch Pathol Lab Med 102, 420–5.

3. Göbel U, Calaminus G, Engert J, Kaatsch P, et al. (1998) Teratomas in infancy and childhood. Med Pediatr Oncol 31, 8–15.

4. Göbel U, Schneider DT, Calaminus G, et al (2000). Germ-cell tumors in childhood and adolescence. Ann Oncol 11, 263–71.

5. Schneider DT, Schuster AE, Fritsch MK, et al. (2001). Multipoint imprinting analysis indicates a common precursor cell for gonadal and nongonadal pediatric germ cell tumors. Cancer Res 61, 7268–7276.

6. Perlman EJ, Hu J, Ho D, et al. (2000). Genetic analysis of childhood endodermal sinus tumors by comparative genomic hybridization. J Pediatr Hematol Oncol 22, 100–5.

7. Looijenga LH, Oosterhuis JW (2002). Pathobiology of testicular germ cell tumors: views and news. Anal Quant Cytol Histol 24, 263–79.

8. Schneider DT, Calaminus G, Göbel U (2001). Diagnostic value of alpha1-fetoprotein and human chorionic gonadotropic hormone in infancy and childhood. Pediatr Hematol Oncol 18, 11–26.

9. Blohm ME, Vesterling-Hörner D, Calaminus G, et al. (1998). Alpha 1-fetoprotein (AFP) reference values in infants up to 2 years of age. Pediatr Hematol Oncol 15, 135–42.

10. Göbel U, Schneider DT, Calaminus G, et al. (2001). Multimodal treatment of malignant sacrococcygeal germ cell tumors: a prospective analysis of 66 patients of the German cooperative protocols MAKEI 83/86 and 89. J Clin Oncol 19, 1943–50.

11. Schneider DT, Wessalowski R, Calaminus G, et al. (2001). Treatment of recurrent malignant sacrococcygeal germ cell tumors: analysis of 22 patients registered in the German protocols MAKEI 83/86, 89, and 96. J Clin Oncol 19, 1951–60.

12. Rescorla F (1999). Pediatric germ cell tumors. Semin Surg Oncol 16, 144–58.

13. Nicholson JC, Punt J, Hale J, et al. (2002). Neurosurgical management of paediatric germ cell tumours of the central nervous system–a multi-disciplinary team approach for the new millennium. Br J Neurosurg 16, 93–5.

14. Einhorn LH, Donohue JP (1977). Chemotherapy for disseminated testicular cancer. Urol Clin North Am 4, 407–26.

15. Ablin AR (1981). Malignant germ cell tumors in children. Front Radiat Ther Oncol 16, 141–9.

16. Ablin AR, Krailo MD, Ramsay NK, et al. (1991). Results of treatment of malignant germ cell tumors in 93 children: a report from the Childrens Cancer Study Group. J Clin Oncol 9, 1782–92.

17. Mann JR, Raafat F, Robinson K, et al. (1998). UKCCSG's germ cell tumour (GCT) studies: improving outcome for children with malignant extracranial non-gonadal tumours—carboplatin, etoposide, and bleomycin are effective and less toxic than previous regimens. United Kingdom Children's Cancer Study Group. Med Pediatr Oncol 30, 217–27.

18. Mann JR, Raafat F, Robinson K, et al. (2000). The United Kingdom Children's Cancer Study Group's Second Germ Cell Tumor Study: Carboplatin, etoposide, and bleomycin are effective treatment for children with malignant extracranial germ cell tumors, with acceptable toxicity. J Clin Oncol 18, 3809–18.

19. Flamant F, Schwartz L, Delons E, et al. (1984). Nonseminomatous malignant germ cell tumors in children. Multidrug therapy in Stages III and IV. Cancer 54, 1687–91.

20. Baranzelli MC, Kramar A, Bouffet E, et al. (1999). Prognostic factors in children with localized malignant nonseminomatous germ cell tumors. J Clin Oncol 17, 1212–19.

21. Calaminus G, Schneider DT, Bökkerink JP, et al. (2003). Prognostic value of tumor size, metastases, extension into bone, and increased tumor marker in children with malignant sacrococcygeal germ cell tumors: a prospective evaluation of 71 patients treated in the German cooperative protocols Maligne Keimzelltumoren (MAKEI) 83/86 and MAKEI 89. J Clin Oncol 21, 781–6.

22. Calaminus G, Garre ML, Kortmann RD, et al. (1998). CNS germ cell tumors in children. Results of the German MAKEI 89 protocols and SIOP-CNS-GCT 93P/96 study. In: Jones DP, Appleyard I, Joffe JK, Harnden P (eds) Germ Cell Tumours IV. London: Springer, 247–54.

23. Calaminus G, Bamberg M, Baranzelli M, et al. (1994). Intracranial germ cell tumors: a comprehensive update of the European data. Neuropediatrics 25, 26–32.

24. Bamberg M, Kortmann RD, Calaminus G, et al. (1999). Radiation therapy for intracranial germinoma: results of the German cooperative prospective trials MAKEI 83/86/89. J Clin Oncol 17, 2585–92.

25. Balmaceda C, Heller G, Rosenblum M, et al. (1996). Chemotherapy without irradiation—a novel approach for newly diagnosed CNS germ cell tumors: results of an international cooperative trial. The First International Central Nervous System Germ Cell Tumor Study. J Clin Oncol 14, 2908–15.

26. Marina NM, Cushing B, Giller R, et al. (1999). Complete surgical excision is effective treatment for children with immature teratomas with or without malignant elements: a Pediatric Oncology Group/ Children's Cancer Group Intergroup Study. J Clin Oncol 17, 2137.

27. Sutton LN, Radcliffe J, Goldwein JW, et al. (1999). Quality of life of adult survivors of germinomas treated with craniospinal irradiation. Neurosurgery 45, 1292–7.

28. Wessalowski R, Kruck H, Pape H, et al. (1998). Hyperthermia for the treatment of patients with malignant germ cell tumors: a phase I/II study in ten children and adolescents with recurrent or refractory tumors. Cancer 82, 793–800.



If you find an error or have any questions, please email us at admin@doctorlib.info. Thank you!