PURPOSE: To evaluate the roles of radiation dose, chemotherapy, and other factors in the etiology of stomach cancer in long-term survivors of testicular cancer or Hodgkin lymphoma.

METHODS AND MATERIALS: We conducted a cohort study in 5,142 survivors of testicular cancer or Hodgkin lymphoma treated in the Netherlands between 1965 and 1995. In a nested case-control study, detailed information on treatment, smoking, gastrointestinal diseases, and family history was collected for 42 patients with stomach cancer and 126 matched controls. For each subject, the mean radiation dose to the stomach was estimated. Relative risks (RRs) of stomach cancer and the radiation-related excess relative risk (ERR) per gray were calculated by conditional logistic regression analysis.

RESULTS: The risk of stomach cancer was 3.4-fold increased compared with the general population. The risk increased with increasing mean stomach dose (p for trend, <0.001), at an ERR of 0.84 per Gy (95% confidence interval [CI], 0.12-15.6). Mean stomach doses of more than 20 Gy were associated with a RR of 9.9 (95% CI, 3.2-31.2) compared with doses below 11 Gy. The risk was 1.8-fold (95% CI, 0.8-4.4) increased after chemotherapy and 5.4-fold (95% CI, 1.2-23.9) increased after high doses of procarbazine (>or=13,000 mg) vs. <10,000 mg. The RR of smoking more than 10 cigarettes per day vs. no smoking was 1.6 (95% CI, 0.6-4.2).

CONCLUSIONS: Stomach cancer risk is strongly radiation dose dependent. The role of chemotherapy, particularly of procarbazine and related agents, needs further study, because of the relatively small numbers of chemotherapy-treated subjects.