Colorectal-Specific Radiation Dose and Chemotherapy Risk for Subsequent Colorectal Malignancies in Childhood Cancer Survivors: A Childhood Cancer Survivor Study (CCSS) Report.

PURPOSE: Among childhood cancer survivors, we evaluated not previously explored relationships between colorectal subsequent malignant neoplasm (SMN) incidence and colorectum-specific radiation dose metrics currently used in radiation therapy (RT) planning and expanded upon previously reported chemotherapy associations. METHODS: The Childhood Cancer Survivor Study (CCSS) includes 5-year survivors of childhood cancer diagnosed between 1970 and 1999. RT was assessed as mean colorectal dose (MCD) and the percent volume (VX Gy) receiving ≥5, 10, 20, 30, and 40 Gy. Chemotherapy was assessed as cumulative doses for procarbazine and platinum agents, cyclophosphamide-equivalent doses for alkylating agents, and doxorubicin-equivalent doses for anthracyclines. Piecewise-exponential models and excess rate ratio (ERR) models evaluated dose-response relationships for the incidence of colorectal SMNs. Reference groups were those not receiving the assessed treatment(s). RESULTS: Among 25,723 survivors (median follow-up = 28.5 years; range = 5.0-48.9), 104 colorectal SMNs were identified. A dose-response relationship was observed between MCD and colorectal SMN rates; incidence rate ratios (IRRs) for 10 to <20 Gy and ≥20 Gy were 3.6 (95% CI, 1.9 to 6.9) and 8.3 (95% CI, 3.9 to 17.8), respectively. When ≥20% of the colorectum volume was irradiated, IRRs increased with increasing volume. The V20 Gy IRRs were 3.8 (95% CI, 1.9 to 7.6), 4.9 (95% CI, 2.0 to 12.0), and 8.7 (95% CI, 3.5 to 21.6) for irradiated volumes of 20% to <40%, 40% to <80%, and ≥80%, respectively. The IRR was 1.8 (95% CI, 1.0 to 3.0) for doxorubicin-equivalent dose ≥250 mg/m2, 3.7 (95% CI, 2.2 to 6.4) for cyclophosphamide-equivalent dose ≥6,000 mg/m2, and 4.5 (95% CI, 2.0 to 10.1) for platinum dose ≥450 mg/m2. For procarbazine dose, the IRR was 6.3 (95% CI, 3.0 to 13.2) for 4,200 to <7,036 mg/m2 and 9.0 (95% CI, 4.3 to 18.9) for ≥7,036 mg/m2. In the absence of RT, colorectal SMN rates increased with exposure to any platinum-based agent (IRR, 3.8 [95% CI, 1.1 to 12.7]), alkylator (IRR, 4.8 [95% CI, 1.6 to 14.4]), or procarbazine (IRR, 16.9 [95% CI, 5.9 to 48.8]). Colorectal SMN rates increased linearly with procarbazine dose (ERR per 1,000 mg/m2 = 73.0 [95% CI, 26.4% to 119.6%]) and MCD (ERR per 1 Gy = 20.8 [95% CI, 9.0% to 32.5%]). Quadratic ERR models did not improve data fit compared with linear ERR models. CONCLUSION: These RT and chemotherapy dose-response relationships can better inform contemporary RT planning for pediatric patients and surveillance guidelines for high-risk survivors.