DISCUSSION
Ever since Einhorn4 in his seminal paper elucidated the role of multi-agent chemotherapy in germ cell tumors in adults, it has formed the basis of treatment across all age groups. The above regimen has been employed to paediatric GCTs albeit with modifications13, owing to concerns about cumulative toxicities if given at the same schedule as in adults. Precise comparison of outcomes in paediatric GCTs treated on multimodality regimens across various trials has been fraught with challenges because of its inherent heterogeneity, relative rarity and lack of uniform risk stratification across major study groups10. In resource-limited countries, these tribulations are further compounded by factors like flawed surgical approach in inexperienced centres leading to imprecise staging, late presentation, significant attrition rate, and higher incidence of sepsis-related deaths.
This retrospective study is an attempt to highlight the highly favourable outcome of germ cell tumors especially in low and intermediate risk groups treated at our centre. Our series showed an overall bimodal distribution of tumors with a female preponderance which is consistent with previous experiences1,2. This trend is secondary to higher incidence of ovarian tumors as compared to testicular tumors in the adolescent age group. Extragonadal tumors, especially those arising from mediastinum, however showed a male predominance similar to that in earlier studies3. In literature1,13,14, extragonadal tumors have been shown to be of slightly higher prevalence as compared to gonadal counterparts because of inclusion of intracranial GCTs as an extragonadal site. The occurrence of extracranial-extragonadal GCTs was otherwise similar to that in our cohort.
Notwithstanding the necessarily adaptive nature of our risk stratification (Figure2), the outcomes of our cohort parallels the results of similar studies11,13,15,16. Gonadal tumors had a better outcome as compared to extragonadal tumors. Amongst ovarian GCTs, inability to ensure strict compliance to the surgery guidelines can result in incomplete staging in up to 57% patients17. Nearly 75% patients in patients who were operated prior to presentation were upstaged to stage III as a suitable compromise because of incomplete surgical details.
Testicular tumors had an excellent outcome irrespective of the stage. The adherence to high inguinal orchiectomy and diligent monitoring of tumor marker trend post-operatively cannot be overemphasized in this group as stage I tumors require only surgery. Several limitations exist in extrapolating this approach to LMICs like lack of accurate tumor marker measurement at presentation, inappropriate primary surgery often at a non-oncological centre and non-compliant follow-up during observation period. Nearly 3/4th of orchiectomies in our cohort were performed prior to presentation of which only half were HIO. While there was no statistically significant survival difference of because of this flawed surgical approach, an Indian study in adult GCTs has reported scrotal orchiectomy to be an adverse prognostic factor in testicular GCT18. Several investigators in the United States, Germany, France and United Kingdom report a similar outcome in testicular GCTs irrespective of the stage13,19-21.
Extragonadal tumors across all sites were found to be associated with relatively dismal EFS, which was statistically significant on univariate analysis (p<0.001) and a trend towards statistical significance was noted on multivariate analysis (p–0.07). Baranzelli et al in a study classifying all extragonadal tumors as high risk demonstrated a 43% 3-year failure-free survival and 77% overall survival rate, treated on a protocol comprising of alternating cycles of Cyclophosphamide-Dactinomycin and Vinblastine-Bleomycin-Etoposide22. Recently, the outcomes of 165 extragonadal tumors treated on a more contemporary protocol showed a 5-year EFS and OS as 79.0% and 83.4%,respectively with age ≥ 12 years and mediastinal site tumors being poor prognostic variables16. The superior EFS of extragonadal tumors in the said cohort can be imputed to the usage of high dose PEB in nearly half of the patients. Outcome of sacrococcygeal tumors treated on a German Trial MAKEI 83/8623with total cumulative doses of cisplatin and etoposide comparable to the former study15, cite a 5-year EFS of 76% with a better outcome in tumors with delayed but complete tumor resection. Considering that the above chemotherapy regimens15,23come with their share of significant toxicities, extending these regimens to developing countries to achieve similar outcomes would be a challenge. A high background prevalence of malnutrition and higher incidence of both rectal carriage and bloodstream infection with multi-drug resistant organisms are hurdles specific to LMIC countries, that prevent this extrapolation24,25. Furthermore, dose intensification of chemotherapeutic agents that has been described in phase I/II studies to improve outcomes in high risk GCTs may not have the same applicability in LMIC populations26.
Prognostic variables used for adult GCTs do not hold the same pertinence in children on account of differences in histology, primary site and age distribution27,28. With an attempt to risk stratify malignant extracranial paediatric GCTs, the Malignant Germ Cell Tumor International Collaborative (MaGIC) did a recent meta-analysis on seven GCT trials, which identified age ≥ 12 years, ovarian stage IV disease and extragonadal stage III to IV disease as conferring a significantly worse prognosis29. Elevated levels of pre-treatment AFP level, initially reckoned to be a poor prognostic indicator was not linked to inferior outcome in this meta-analysis22,30. On implementing the MaGIC risk stratification to our cohort of malignant GCTs (n=192)29, we found that the SR-1 (Standard Risk – 1) subset had an unexpectedly inferior survival compared to SR-2 (Standard Risk – 2). While all the 31 patients in SR-2 had ovarian GCT and none had extragonadal GCT, 78/136 (57.4%) SR-1 patients had tumors in extragonadal sites. Though this association was not statistically significant, we presume that a higher frequency of unfavourable site tumors adversely influenced the survival in SR-1. The lack of association of age in our cohort could also be due to fewer numbers of patients aged over 11 years in our study.
Both PEb and JEb regimens in our cohort was reasonably tolerated with a toxicity profile akin to published studies15,21,23. There was no survival difference between the two regimens in our cohort as is reported in literature31. A few patients, in whom hybrid regimen (PEb and JEb) was employed, had worse outcomes. This could be attributed to the use of this hybrid regimen in a relatively morbid cohort (extragonadal GCTs or large ovarian GCTs with obstructive uropathy at presentation). Nonetheless, as outcomes were similar on both JEb and PEb, we believe switching between either regimen midway during treatment may not be a worthwhile strategy. While the incidence of anaemia was higher in PEb regimen, the incidence of neutropenia and thrombocytopenia was expectedly higher in JEb. The incidence of myelosuppression and septic episodes noted in an Indian study was substantially less, probably because of a longer cycle interval (28 days) and a higher neutrophil count cut-off to administer chemotherapy32. Protocols employing carboplatin in lieu of cisplatin had a higher incidence of non-fatal haematological toxicity but ototoxicity, pulmonary toxicity and nephrotoxicity were rare21. One child had fatal bleomycin toxicity suggesting that although rare, even 1 day of bleomycin in the regimen can be detrimental. Overt oto-toxicity or nephrotoxicity was not found in any child, which could partly be attributable to inconsistent gathering of data on audiograms of all children and lack of a methodical long-term follow-up. Despite the overall promising outcomes of GCTs in a resource-limited centre like ours, we do understand the limitations that come with a retrospective study and relatively less number of adolescent patients. A systematic study encompassing long-term effects would be more capable of unmasking the ostensibly absent nephrotoxicity, pulmonary toxicity and the frequency of second malignant neoplasms.