This was a national study conducted in South Africa across multiple POUs.

All children under the age of 16 years with benign mature teratoma (MT) and immature teratoma (IT) or MEGCT and sex cord stromal tumours (SCSTs) were prospectively enrolled into the MEGCT-2016 study following informed consent by parents or legal guardians, (and, where applicable participants’ assent). Patients were required to be chemotherapy naïve, whether or not they had undergone previous surgery. Children with intracranial GCTs were excluded, as well as those female patients who were pregnant, breastfeeding or who were sexually active but refused contraception.

Staging was conducted pre-operatively with magnetic resonance imaging (MRI) or computer tomography (CT) in conjunction with serum tumour markers and tumour histology. Germ cell tumours were staged according to the COG Staging System,⁷ while those with SCST were staged according to the International Federation of Gynaecology and Obstetrics (FIGO) classification.⁸

Patients with stage 1 disease with completely resected tumours were watched expectantly and followed up. Patient with SCST Stage 1C and above received four cycles of P₁₀₀E₅₀₀b (cisplatin IV 100 mg/m² in five fractionated doses, etoposide IV 100 mg/m² daily for 5 days, bleomycin IV 15 IU/m² on day 3). Standard risk (SR) patients were defined as those under 11 years of age with any stage disease at any site (ovarian, testicular, extragonadal), while HR patients were described as those 11 years or older with Stage 3 or Stage 4 disease at any site (ovarian, testicular or extragonadal). Standard risk patients with MEGCTs received four cycles of J₆₀₀E₃₆₀b (carboplatin IV 600 mg/m² on day 2, etoposide IV 120 mg/m² daily for 3 days, bleomycin IV 15I U/m² on day 3), while HR patients received a hybrid regimen (J₆₀₀E₃₆₀b / P₁₀₀E₅₀₀b) to a total of six cycles.

Study data were entered into a REDCap^™ database. The first treatment visit data included patient data (age, sex, anthropometry, HIV status), socio-economic data (household income and maternal education), disease data (presenting symptoms), tumour data (site, histology, stage and risk assignment), imaging assessment/s, details of surgery and chemotherapy), tumour markers, haematology and renal blood panels and glomerular filtration rate (GFR) and audiology assessments. The World Health Organization (WHO) Nutritional Landscape Information System (NLIS) Interpretation Guide, 2010,⁹ was used to define anthropometric parameters. Socioeconomic status by household status was assigned according to the Provincial Government of the Western Cape’s Federal Acquisition Regulation No. 9 of 2017 (PGWC FAR9/2017) or provincial equivalent. Information for all subsequent treatment visits included further chemotherapy data (treatment visits), all (severe) adverse events (SAEs), interim disease evaluations, end-of-treatment evaluations and follow-up visit data.

Guideline implementation was assessed by evaluating adherence to recommended treatment approaches as well as completion of key baseline investigations specified in the guideline, including GFR and audiology assessments.

Toxicity was included as a predefined component of the interim analysis in order to evaluate the safety and feasibility of guideline implementation across participating centres and to identify any early signals that might warrant protocol review or modification.

A guideline violation was defined as a treatment decision or action that deviated from the MEGCT-2016 recommendation without documented justification and which was not supported by an accepted clinical indication. For example, this included (but was not limited to) incorrect risk assignment resulting in the selection of a chemotherapy regimen not recommended by the guideline.

Patients were grouped into three disease cohorts: SCST, extracranial germ cell tumour (EGCT) which included MT and IT and MEGCTs for rational analysis. IBM SSPS^™ Statistical Software 28.0.1.1 (Version 14), Armonk, New York, US was used for consolidation, uniform recording and descriptive statistics, survival analyses and output graphics. The Kaplan–Meier procedure was used to estimate OS curves and survival at 3 years with associated 95% confidence intervals. A p-value < 5% was used as a cut-off to define statistical significance.

Ethical approval was obtained from the University of Cape Town’s Human Research Ethics Committee (HREC 002/2018) with reciprocal approval from all participating universities.

Eleven of the 13 POUs (Figure 1) in South Africa were ethically approved to participate in the MEGCT-2016 national treatment guideline study. Nine actively recruited patients.

Between 01 January 2020 and 30 June 2023, there were 72 database entries into the REDCap^™ database, of which 68 were complete at the time of this analysis. Three patients were excluded from the analysis leaving 65, which were analysed. The patient attrition rate was 8.3% over 42 months. To determine if there was homology between patients enrolled on the MEGCT-2016 database and those registered with the SAAPHO Tumour Registry (SAAPHO-TR), the number and type of MEGCT diagnoses registered with the SAAPHO-TR from January 2020 to June 2023 were compared. Seventy-two patients were reported to the registry, of which 65 were enrolled on the MEGCT database.

The median age for the entire group was 23 months, and the mean was 42.8 months (1 month–154 months). There were three Stage 1 patients (15.8%), eight Stage 2 patients (42.1%), five Stage 3 patients (26.3%) and three Stage 4 patients (15.8%). Metastatic disease included two patients with lung metastases (one with a malignant effusion) and one patient with multiple sites (lung, kidney, bone and brain). All male patients had testicular yolk sac tumours (YSTs). Eight (53.3%) female patients had YSTs, five (33.3%) had mixed MEGCT and two (13.3%) had dysgerminomas. In the female patients, nine primary tumour sites were ovarian (60%), and six were extragonadal (40%).

The median age for the group was 153 months, and the mean was 151.8 months (age range, 123 months–171 months). All patients were female. There were five Stage 3 patients (41.7%) and seven Stage 4 patients (58.3%). Metastatic sites included four patients with lung metastases, two with lung and liver lesions, and one with multiple sites (lungs, kidney and liver). There were five dysgerminomas (41.7%), three YSTs (25%), three mixed malignant germ cell tumours (MMGCTs) (25%) and one choriocarcinoma (8.3%). Most primary sites (n = 9; 75%) were ovarian, two were retroperitoneal (16.7%), and one metastatic patient had an undetermined primary site (8.3%).

Neither being underweight-for-age (p = 0.637) nor stunted (p = 0.388) impacted the OS. Similarly, socio-economic status (SES) did not significantly impact OS (p = 0.247).

Overall, 62 patients of the 65 patients (95.3%) were treated according to the guidelines. Two patients were lost to follow-up, and one patient suffered an accidental death from drowning and asphyxiation. Table 1 shows improvement in OS for limited-Stage patients (100%), especially Stage 3 patients, but less favourable outcomes for Stage 4 patients (37.5%). Only Stage 4 patients with single-site (lung) metastases and normal anthropometry were salvaged. The 3-year OS for the group was 84.4% (p = 0.005) (Figure 2).

A total of 62 patients with mature (n = 19) and immature extracranial (n = 7) GCT were enrolled. The group comprised five male (19.2%) (0.5 month–121 months; median 5 months) and 21 female (80.8%) (0 month–158 months; median 38 months) patients. All male patients had extragonadal tumours (two mediastinal, one head and neck, one retroperitoneal and one sacrococcygeal). Among the female patients, there were 11 (52.4%) ovarian tumours and 10 (47.6%) extragonadal tumours (four retroperitoneal, two head and neck, two sacrococcygeal and two mediastinal). All patients with MTs bar one (n = 18) were Stage 1, and one patient had Stage 3 disease. Five of the patients with ITs had Stage 1, and two had Stage 3 disease, meeting the criteria for a watch-and-wait approach (diagnostic alpha-fetoprotein [AFP] < 1000 ng/mL and no YST elements).

All South African children with mature and immature EGCTs are alive and disease-free except one child who demised after treatment from aspiration many months after definitive surgery for an extensive left-sided hemi-facial tumour. Two children, one from Zimbabwe (MT) and another from Swaziland (IT), returned home following definitive surgery and were lost to follow-up.

Six patients with SCSTs were enrolled. All were female with primary ovarian juvenile granulosa cell tumours (age range, 8 months–154 months; median age 95 months). Three patients had complete resections and were observed. The remaining three with FIGO stage 1c disease and above (one Stage 3 and one Stage 4 with lung metastases) received four courses of PEb. All children are alive and disease-free.

This interim analysis reveals several positive improvements in practice and outcomes for children treated on MEGCT-2016. This was particularly evident for patients with limited-stage disease compared to historical cohorts, with Stage 3 patients making the most noticeable gains.¹⁰ This was encouraging given that HR Stage 3 patients treated with the JEb/PEb (carboplatin, etoposide, bleomycin alternating with cisplatin, etoposide, bleomycin) hybrid appeared to be doing as well as SR Stage 3 patients treated with JEb. The OS for the whole group was marginally better (84.4% vs. 80.3%). However, we note this with caution, considering that the prospective patient cohort is still small and we acknowledge the inherent problems of selection bias when using retrospective cohorts as comparators and the short follow-up period in the prospective cohort.

Similarly, patients with EGCTs had excellent outcomes (MT 89%, IT 100%) compared to historical datasets (5-year OS All EGCT 85.4%, MT 80.8% IT 92.8%).¹¹ The number of Stage 3 patients was too small to draw any inferences, but it was encouraging that all were alive by the end of the reporting period. Part of the improvement in both the MEGCTs and ECGTs could be attributed to the fact that the guideline offered specific diagnostic criteria to differentiate these two groups more clearly. This facilitated more accurate diagnosis and limited inappropriate treatment with undue treatment-related morbidity.

Patients with SCSTs with a previously reported 82.1% 5-year OS¹² were all alive and disease-free at interim assessment. Again, the guideline provided specific advice on chemotherapy, staging and risk assignment, which was not standardised prior to 2020.

There were no recorded relapses at the end of the interim assessment period but this is likely to change with time.

Audiology and GFR testing, irrespective of the methods employed, had improved substantially compared to historical practice, where the monitoring and assessment of hearing in children receiving cisplatin-based chemotherapy was previously inadequate.¹⁰ Again, the guidelines and database offered specific guidance on these evaluations. Although audiology testing was still imperfect because of varying access to audiology services across public sector hospitals, the practice was undoubtedly improved.

Follow-up had also been more rigorous, although the follow-up period was short. It remains to be seen if this will be sustained beyond the constructs of a national study. Only four patients in the combined prospective cohort were lost to follow-up compared to 20% SCSTs, 25% MTs and 15% ITs in the retrospective cohort.^11,12 In Saudi Arabia, a 30-year retrospective report comprising 50 patients with benign and malignant EGCTs reported no loss to follow-up.¹³ Similarly, a 15-year retrospective report from Thailand reported no patients lost among 44 patients with EGCTs.¹⁴ Two studies from Iran reported rates of 5%¹⁵ and 9%,¹⁶ both in mixed cohorts of children with benign and malignant tumours. Differences in settings, the composition of patient cohorts and follow-up periods make direct comparisons difficult, but there is an impression that the rates of attrition seem to be much higher in our cohort than in other non-African mixed-income settings.

Malignant extracranial germ cell tumour-2016 will continue to enrol patients until the end of 2026 to determine 5-year OS and EFS and assess whether any significant prognostic factors impact patient outcomes in a larger cohort.