Dosimetric impact of bone marrow sparing for robustly optimized IMPT for locally advanced cervical cancer.

Abstract

To investigate the trade-off between bone marrow sparing (BMS) and dose to organs at risk (OARs) for intensity modulated proton therapy (IMPT) for women with locally advanced cervical cancer (LACC). Twenty LACC patients were retrospectively included. IMPT plans were created for each patient using automated treatment planning. These plans progressively reduced bone marrow mean doses by steps of 1 GyRBE, while constraining target coverage and conformality. The relation between bone marrow dose and bladder, small bowel, rectum, and sigmoid doses was evaluated. A total of 140 IMPT plans were created. Plans without BMS had an average [range] bone marrow mean dose of 17.3 [14.7-21.6] Gy_RBE , which reduced to 12.0 [10.0-14.0] Gy_RBE with maximum BMS. The mean OAR dose [range] increased modestly for 1 Gy_RBE BMS: 0.2 [0.0 - 0.6] Gy_RBE for bladder, 0.3 [-0.2 - 0.7] Gy_RBE for rectum, 0.4 [0.1 - 0.8] Gy_RBE for small bowel, and 0.2 [-0.2 - 0.4] Gy_RBE for sigmoid. Moreover, for maximum BMS, mean OAR doses [range] escalated by 3.3 [0.1 - 6.7] Gy_RBE for bladder, 5.8 [1.8 - 12.4] Gy_RBE for rectum, 3.9 [1.6 - 5.9] Gy_RBE for small bowel, and 2.7 [0.6 - 5.9] Gy_RBE for sigmoid. Achieving 1 Gy_RBE BMS for IMPT is feasible for LACC patients with limited dosimetric impact on other OARs. While further bone marrow dose reduction is possible for some patients, it may increase OAR doses substantially for others. Hence, we recommend a personalized approach when introducing BMS into clinical IMPT treatment planning to carefully assess individual patient benefits and risks.