PURPOSE

To develop a prognostic model in patients with germ cell tumors (GCT) who experience treatment failure with cisplatin-based first-line chemotherapy.

PATIENTS AND METHODS

Data from 1,984 patients with GCT who progressed after at least three cisplatin-based cycles and were treated with cisplatin-based conventional-dose or carboplatin-based high-dose salvage chemotherapy was retrospectively collected from 38 centers/groups worldwide. One thousand five hundred ninety-four (80%) of 1,984 eligible patients were randomly divided into a training set of 1,067 patients (67%) and a validation set of 527 patients (33%). Seminomas were set aside for posthoc analyses. Primary end point was the 2-year progression-free survival after salvage treatment.

RESULTS

Overall, 990 patients (62%) relapsed and 604 patients (38%) remained relapse free. Histology, primary tumor location, response, and progression-free interval after first-line treatment, as well as levels of alpha fetoprotein, human chorionic gonadotrophin, and the presence of liver, bone, or brain metastases at salvage were identified as independent prognostic variables and used to build a prognostic model in the training set. Survival rates in the training and validation set were very similar. The estimated 2-year progression-free survival rates in patients not included in the training set was 75% in very low risk, 51% in low risk, 40% in intermediate risk, 26% in high risk, and only 6% in very high-risk patients. Due to missing values in individual variables, 69 patients could not reliably be classified into one of these categories.

CONCLUSION

Prognostic variables are important in patients with GCT who experienced treatment failure with cisplatin-based first-line chemotherapy and can be used to construct a prognostic model to guide salvage strategies.