Prognostic reclassification of patients with intermediate-risk metastatic germ cell tumors: Implications for clinical practice, trial design, and molecular interrogation

Objectives: Approximately one-third of the metastatic germ cell tumors (GCT) in patients are classified as intermediate-risk metastatic GCT, and available guidelines recommend the same treatment of poor-risk cases. Yet the prognosis of these patients is heterogeneous, and consequently refining the intensity of treatment is warranted. We aimed to address the heterogeneity of this category by providing a proof of principle for reclassification attempt.

Patients and methods: Data on consecutive patients with intermediate-risk metastatic GCT and who received treatment at Fondazione INT Milano in the time frame between February 1980 and March 2014 were collected. Cox regression analyses were done, evaluating potential prognostic factors for overall survival (OS, primary end point) to first-line therapy. Each factor was evaluated in a multivariable model. Recursive partitioning was performed to define prognostic risk groups.

Results: A total of 224 patients were suitable for the present analysis. Median age was 26 years (interquartile range: 22-31), 11 patients (4.9%) had a retroperitoneal primary tumor, 6 yielded seminomatous histology, 85 (37.9%) had lung metastases, and 58 (25.9%) had bulky (i.e.,≥ 10 cm) retroperitoneal lymph nodes. Patients received cisplatin, bleomycin, and etoposide (PEB, n = 199) or vinblastine (PVB, n = 23); however, 2 patients received other treatments. Median follow-up was 135 months (interquartile range: 81-223). Globally, 5-year progression-free survival and OS rates were 72.8% (95% CI: 67.1-79.0) and 86.2% (81.7-91.0), respectively. In the multivariable model for OS, elevated alfa fetoprotein (AFP) level was the only significant prognostic factor (hazard ratio = 1.48, 95% CI: 1.12-1.96). The 2 separate prognostic groups with differential OS outcomes were identified based on the cutoff level of 6,200 IU/ml. The 10-year OS rate was 55.6% (95% CI: 36.6-84.3), and it was 86.7% (95% CI: 82.0-91.7) for those with AFP levels more than (n = 19, 8.5%) and less than (n = 205, 91.5%) the cutoff, respectively.

Conclusions: A small fraction of patients with highly elevated AFP levels have an OS approximating the poor prognostic category, whereas most of them are close to good-risk cases. This might have implications to select outlier patients for clinical trials and molecular characterization.