摘要：Abstract Background and purpose This study evaluates the potential efficacy and robustness of functional bone marrow sparing (BMS) using intensity-modulated proton therapy (IMPT) for cervical cancer, with the goal of reducing hematologic toxicity. Material and methods IMPT plans with prescription dose of 45 Gy were generated for ten patients who have received BMS intensity-modulated X-ray therapy (IMRT). Functional bone marrow was identified by 18F-flourothymidine positron emission tomography. IMPT plans were designed to minimize the volume of functional bone marrow receiving 5–40 Gy while maintaining similar target coverage and healthy organ sparing as IMRT. IMPT robustness was analyzed with ±3% range uncertainty errors and/or ±3 mm translational setup errors in all three principal dimensions. Results In the static scenario, the median dose volume reductions for functional bone marrow by IMPT were: 32% for V5Gy, 47% for V10Gy, 54% for V20Gy, and 57% for V40Gy, all with p < 0.01 compared to IMRT. With assumed errors, even the worst-case reductions by IMPT were: 23% for V5Gy, 37% for V10Gy, 41% for V20Gy, and 39% for V40Gy, all with p < 0.01. Conclusions The potential sparing of functional bone marrow by IMPT for cervical cancer is significant and robust under realistic systematic range uncertainties and clinically relevant setup errors.