Adaptation of chemotherapy to the decline tumor markers in patients with poor prognosis nonseminomatous germ cell tumors:Real-world French experience.

Abstract

385 Background: Personalized chemotherapy based on tumor marker decline is the new standard in poor prognosis germ-cell tumor in Europe since 2014 (GETUG 13, Lancet, Fizazi et al). The purpose of this study was to analyze the reproducibility of the princeps study in patients not selected in clinical routine between 2014 and 2018. Methods: Patients (pts) were eligible if they had at least one criteria of IGCCCG classification for poor prognosis group. They had to be treated according the study terms of GETUG 13 study and did not received prior treatment. They had to received 1 BEP (Bleomycin, Etoposide, Cisplatin). Tumor markers (HCG and AFP) were dosed between day 18 and 21. Then, they received 3 additional BEP if they had favorable tumor marker decline or intensive chemotherapy if they had unfavorable decline. Results: This retrospective study included 104 patients in 14 french centers treated between 2013 and 2018: 22,1 % (n = 23) in the favorable group (Fav), 77,9 % (n = 81) in the unfavorable group (Unfav). Thirty-two pts had PS ≥ 2. In Unfav, there were more pts with HCG > 50 000 UI/L (44,2 % vs 13 %, p = 0,0067), neutrophil-to-lymphocyt ratio was also higher (median 6,4 vs 4,5, p = 0,0199). At cycle 1, all pts received BEP in Fav and 87,5 % (n = 70) in Unfav. After chemotherapy and surgery, 65,2 % in Fav and 41,3 % in Unfav obtained complete response. At 30 months (median follow-up), Fav-OS was 80,5 % (IC95% 55,8 – 92,2) and Unfav-OS was 64,4 % (IC95% 52 – 74,4). At 30 months, rates were 69,6 % (IC95% 46,6 -84,2) and 63.5 % (IC95% 51,9 – 73) respectively. In GETUG 13 study, 3-years OS was 84 % in Fav and 73 % on Unfav; 3-years PFS was 70 % and 59 % respectively. Seven pts died because of toxicity in Unfav (No one in Fav). Neuropathy, anemia and thrombopenia were more frequent in Unfav. Salvage high-dose chemotherapy with stem-cell transplant was required in 4 (66,7 %) pts in Fav and 8 (36,4 %) pts in Unfav. Conclusions: This study showed a reproducibility of the princeps study in terms of PFS and OS. Toxicity seemed more important in real world. For the congress, results will be reported with 50 additional pts.