RESULTS
Clinical characteristics: A total of 336 children were treated, of which 39 patients were excluded in view of prior chemotherapy before presentation. The median age of the analysable cohort (n-297) was 4 years (1.4 months – 16.6 years), with a bimodal age distribution and female: male ratio of 1.72:1 (Figure 3). Forty-three patients (14.5%) had pure teratomas (27 mature; 16 immature) of which 25 were in extragonadal sites. All pure teratomas were with surgery alone followed by observation.
Table 1 summarises the patient demographic and clinical characteristics of patients with malignant GCTs (n-254). Gonadal GCTs were more common than extragonadal tumors with ovary being the primary site in 128 children (43%) and testis in 52 children (17.5%). While testicular and extragonadal tumors were more common in children<3 years (82.6% and 77.7% respectively), ovarian tumors were seen in a higher proportion in children>11 years (61%). Metastatic disease was noted in 75 patients (25.3%) with lungs being the commonest site (Isolated-41; in combination-10). Non-pulmonary visceral metastases (NPVM) were seen in the remaining patients, of which liver was the most common site (Isolated-14; in combination-12). Pre-chemotherapy histopathological subtype was available in 253 (85%) cases, of which yolk sac tumor was the commonest (109/253; 43%). In the remaining patients, the tumor was presumptively treated as germ cell tumor based on imaging and elevated TM, and histology was established post-chemotherapy as residual teratomas. Amongst all yolk sac tumors (n-151), AFP was uniformly elevated with a median value of 16839 IU/mL (99-856496 IU/mL). AFP upwards of 1000 IU/mL was seen in 129 patients (85%).
  1. Risk grouping and treatment received: As per the institutional risk stratification, the study cohort consisted of 60 (20.2%) low-risk patients, 125 (42.5%) intermediate-risk patients and 112 (37.7%) high-risk patients. The cohort of malignant GCT (Table 1b) also displayed a similar risk distribution. On applying the recently drafted MaGIC risk stratification (excluding Stage I tumors, dysgerminomas and chemo-naïve teratomas were excluded)12 to the same cohort, 136 (71%), 31 (16%) and 25 (13%) had Standard risk 1, Standard risk 2 and High risk tumors respectively. Amongst the malignant GCTs, 231/254 patients received chemotherapy; PEb regimen in 181 patients (71.2%) and JEb regimen in 27 patients (10.6%). Twenty-three patients received carboplatin in the first two cycles and completed their remainder of cycles on PEb.
  2. Outcomes and Prognostic Variables: The 5-year EFS and OS of the entire cohort at a median follow-up of 52.1 ± 37.3 months was 79.30% and 84.40% respectively. The 5-year EFS and OS of malignant GCTs at a median follow-up of 51.9 ± 38 months was 72.50% and 82.70% respectively. Figure 4 shows the group-wise Kaplan-Meier curves of malignant GCTs in a single panel. These outcomes were compared with known prognostic variables like gender, age brackets, pre-treatment AFP level, site, stage & risk grouping of the primary tumor. Chemotherapy regimen and the MaGIC risk stratification were two additional variables compared in the subset of malignant GCTs.
On univariate analysis, gonadal sites, low risk tumors and non-metastatic tumors were associated with better EFS and OS, which was statistically significant (p<0.05). Amongst the malignant GCTs (n-254), low risk and standard risk 2 groups had a better OS compared to other groups. In addition, the type of chemotherapy regimen (PEb or JEb) did not have any impact on EFS or OS. However, a small proportion of patients who received the hybrid regimen had unfavourable outcome. Table 2 shows the survival outcomes against the known prognostic variables. On multivariate analysis, Stage IV GCT had an inferior EFS (RR: 2.96; 95% CI: 1.24 - 7.08; p=0.015) and OS (RR: 19.93; 95% CI: 2.46 - 161.25; p=0.005). Additionally, extragonadal GCTs had a poor EFS (RR: 1.98; 95% CI: 1.20 - 3.29; p=0.007) and OS (RR: 3.09; 85% CI: 1.65 - 5.79; p<0.001). The aforementioned prognostic variables failed to retain their prognostic significance when computed for, in the subset of malignant GCT. Children who received hybrid regimen had an inferior OS (RR: 2.86; 95% CI: 1.30 - 6.29; p=0.009). Age, gender and baseline AFP did not have any bearing on EFS or OS in either subsets. The demographic and survival correlates of individual group of malignant GCTs is described below:3a. Ovarian GCTs: One hundred and fourteen patients were eligible, of which 55 (48.2%) underwent surgery prior to presentation. Lack of thorough surgical steps, including failure to collect peritoneal washings/ascitic fluid for cytology led to upstaging in 41 patients (75%) in this subset. Twelve patients relapsed. At a median follow-up of 54 ± 38.27 months, the 5 year EFS and OS at a median follow up of 52.5 ± 32.1 months was found to be 85.9% and 90.3% respectively.3b. Testicular GCTs:Forty-eight patients were eligible for analysis. Thirty-five (73%) patients underwent surgery outside, of which 23 (47.9%) underwent high inguinal orchiectomy. Adjuvant chemotherapy was needed in 30 patients. Seven patients relapsed (2 in Stage I; 1 each in Stage II & III; 3 in Stage IV). At a median follow-up of 52.82 ± 39.3 months, the 5-year EFS/OS was 82.3%/91.7%. Patients who underwent surgery/biopsy prior to presentation had an inferior EFS, but this association was not statistically significant (79.5% v/s 90.9%; p=0.388).3c. Extragonadal GCTs:Amongst the extragonadal GCTs (n-92), sacrococcygeal tumors were the commonest (41; 44.6%) followed by retroperitoneum (14; 15.2%), mediastinum (15; 16.3%) and pelvis (5; 5.4%). A small minority of patients had GCTs in other locations (n-17; 18.5%) like vagina, maxilla, orbit and the parapharynx. While 53 (57.6%) patients were treatment-naive, 23 (25%) underwent surgical resection and 16 (17.4%) underwent biopsy prior to presentation. Eighteen patients with intra-abdominal GCTs (sixsacrococcygeal, eightpelvic and four retroperitoneal) presented with features of obstructive uropathy of varying severity, necessitating urgent procedural intervention in four of them (ureteral stenting in 2 and nephrostomy in 2 patient). Eighteen patients (19.5%) received JEb instead of PEb because of underlying renal dysfunction. Additionally, 14 patients (15.2%) received JEb for the first two cycles before continuing on PEb. Two patients had involvement of the sacral nerve plexus with neurogenic bladder and one of them had long-standing neurological complications including clubfoot. The 5-yr-EFS and OS of extragonadal tumors at a median follow-up of 45.9 ± 36.1 months was 69.0% & 70.7% respectively.
  1. Events: During the study period amongst all GCTs (n-297), there were a total of 41 relapses and 6 progressions. The median time to relapse was 8.9 months (3.32 - 71.56 months), with 22 (53.7%) patients, 8 (19.5%) patients and 11 (26.8%) patients having local, metastatic and combined relapses respectively. A total of 43 children died during the entire study period (disease related – 31; sepsis related – 7; suspected bleomycin toxicity – 1; unknown – 1). There were no cases of second malignant neoplasm in our cohort. Twenty-two patients (7.4%) were lost to follow-up at various time-points: 6 patients (2%) during primary therapy; 13 patients (4.4%) after completion of therapy but unavailable for contact; 2 patients (0.7%) at first relapse and 1 patient (0.3%) at second relapse. Outcomes and prognostic variables concerning relapsed/refractory GCTs will be discussed elsewhere.
  2. Toxicity of chemotherapy regimen: As per the chemotherapy regimen, the frequency of various toxicities is tabulated in Table 3. While Grade 1/2 anaemia was more common with PEb regimen (p=0.002), grade 3/4 neutropenia and thrombocytopenia of all grades were common in JEb regimen (p=0.006; p<0.001 respectively). Audiogram reports were available in 102 patients (PEb-86; JEb-16). In the PEb cohort, Chang 1a and 1b ototoxicity was present in six (7%) and seven patients (8%) patients respectively. None of the patients on JEb had deafness. While the proportion of fatal septic episodes was more in the JEb arm, the difference was not statistically significant (10% v/s 3.96%; p=0.24). Bleomycin related pulmonary toxicity led to one death in the entire cohort.