Lumbar-sacral bone marrow dose modeling for acute hematological toxicity in anal cancer patients treated with concurrent chemo-radiation

The aim of the study was to model acute hematologic toxicity (HT) and dose to pelvic osseous structures in anal cancer patients treated with definitive chemo-radiation (CT-RT). A total of 53 patients receiving CT-RT were analyzed. Pelvic bone marrow and corresponding subsites were contoured: ilium, lower pelvis and lumbosacral spine (LSBM). Dose-volume histograms points and mean doses were collected. Logistic regression was performed to correlate dosimetric parameters and ≥G3 HT as endpoint. Normal tissue complication probability (NTCP) was evaluated with the Lyman-Kutcher-Burman (LKB) model. Logistic regression showed a significant correlation between LSBM-mean dose and ≥G3 leukopenia (β coefficient 0.122; p = 0.030; 95% CI 0.012-0.233). According to NTCP modeling, the predicted HT probability had the following parameters: TD50: 37.5 Gy, γ 50: 1.15, m: 0.347. For node positive patients, TD50: 35.2 Gy, γ 50: 2.27, m: 0.176 were found. Node positive patients had significantly higher PBM-V15 (Mean 81.1 vs. 86.7%; p = 0.04), -V20 (Mean 72.7 vs. 79.9%; p = 0.01) and V30 (Mean 50.2 vs. 57.3%; p = 0.03). Patients with a mean LSBM dose >32 Gy had a 1.81 (95% CI 0.81-4.0) relative risk to develop ≥G3 leukopenia. For node positive patients, those risks were 2.67 (95% CI 0.71-10). LKB modeling seems to suggest that LSBM-mean dose should be kept below 32 Gy to minimize ≥G3 HT in anal cancer patients treated with IMRT and concurrent chemotherapy. The contribution of LSBM dose in the development of HT above 25 Gy seems steeper in node positive patients.