Résumé
5034 Background: Since 2014, adaptation of chemotherapy based on tumor marker decline after the cycle of BEP (Bleomycin, Etoposide, Cisplatin) is a standard for patients with IGCCCG poor-risk NSGCT based on data from the GETUG-13 phase 3 trial (Fizazi et al, Lancet Oncol 2014; J Clin Oncol 2024). The aim of this multicenter study was to evaluate the GETUG-13 algorithm when used in routine practice. Methods: We collected data from all poor-risk NSGCT patients treated consecutively in 13 expert centers from 2013 to 2019. After one cycle of BEP, tumor marker levels were assessed at day 18-21. Pts with a favorable decline continued with BEP for 3 additional cycles (Fav group), whereas those with an unfavorable decline received up to 4 cycles of dose-dense chemotherapy (Unfav group). Data was analysed descriptively, and the Kaplan-Meier method was used to estimate progression-free (PFS) and overall survival (OS). Results: Data from 146 patients were collected (46 with PS ≥ 2, 35 with mediastinal NSGCT): 111 (76%) had an Unfav decline and 35 (24%) had a Fav decline. More pts with hCG > 50 000 UI/L and AFP > 10 000 were in the Unfav group (44.9% vs 17.6%, p=0.0045, and 26.9% vs 8.8%, p=0.0282). Surgery of residual masses was performed in 85.7% and 74.3% in the Fav and Unfav groups, and the procedure was complete in 83.3% and 59%, respectively. With a median follow-up of 5.8 years (95% CI,63.2-77.2), 5-year PFS rates were 68.6% (95% CI, 50.5-81.2) and 61.1% (95% CI, 51.2-69.6) in the Fav and Unfav groups, respectively. Five-year OS rates were 73.8% (95% CI, 55.6-85.4) and 64.6% (95% CI, 54.6; 73.0), respectively. In the short term, neuropathy, anemia and thrombopenia were more frequent in the Unfav group. Treatment-related deaths were reported in 2 (5.7%) and 5 (4.5%) (including 2 post-surgery deaths) pts in the Fav and Unfav groups, respectively. Peripheral neuropathy evolved favorably, with 4 (5.9%), 2 (3.8%) and no pts in the Unfav group reporting grade 3 toxicity at 6 months, 1 year and at last follow-up, respectively. Long-term side effects were infrequent with only one pt with grade 3 cardiovascular toxicity in the Unfav group. Late grade 2 toxicities included cardiovascular toxicity (1.4%), hypoacousia (1.4%), peripheral neuropathy (4.2%), chronic renal failure (CRF) (9.9%) in the Unfav group, and grade 2 CRF (8.3%) in the Fav group. In both groups, almost 80% pts had returned to work. Among pts with progression or relapse, salvage high-dose chemotherapy with stem-cell transplant was used in 4/11 (36.4%) and 13/33 (43.3%) in the Fav and Unfav groups, respectively. Conclusions: The GETUG-13 algorithm can be safely used in routine practice by expert centers, with a high cure rate similar to that reported in the original phase 3 trial and rare long-term toxicity. This data confirms that this algorithm is standard for poor-risk NSGCT.