2903 - Artificial Intelligence-Aided Investigation of the Association of Bone Marrow Dosimetric Parameters with Hematological Toxicity in Cervical Cancer Patients Undergoing Radiochemotherapy
Presenter(s)
S. Zhao1, X. Yang2, K. Men3, J. An1, and M. Huang4; 1Cancer Hospital of Chinese Academy of Medical Sciences Hospital, BEIJING, China, 2National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 3Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 4Department of Gynecological Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
Purpose/Objective(s): This study aimed explored the association of bone marrow dosimetric parameters with hematological toxicity in cervical cancer patients undergoing radiotherapy utilizing artificial intelligence (AI) technology to determine organs at risk (OARs). By accurately delineating bone marrow regions and analyzing radiation doses, this study will provide a theoretical foundation for the application of AI in predicting hematological toxicity.
Materials/Methods: A total of 141 cervical cancer patients who received pelvic Volumetric Modulated Arc Therapy (VMAT) and/or chemotherapy (sequential or concurrent) at the Department of Gynecology, Cancer Hospital of the Chinese Academy of Medical Sciences from March 2019 to December 2019 were enrolled in the study. AI-based automatic segmentation of CT images was performed to delineate pelvic bone marrow and its subregions (ilium, lower pelvis, lumbarsacral spine, and head of femur). Specific volumes that received 10-40 Gy (V10, V20, V30, V40) were calculated, and baseline clinical characteristics of the patients were recorded. Hematological toxicity endpoints included grade =2 (HT2+) and = 3 (HT3+) leukopenia, neutropenia, anemia, or thrombocytopenia. The association of dosimetric parameters with hematological toxicity was explored through Logistic regression.
Results: Among the 141 patients,107 (75.8%) experienced HT2+ and 33(23.4%) experienced HT3+. Univariate analysis revealed that chemotherapy and age were correlated with the occurrence of HT2+. Multivariate analysis demonstrated that femur-V30, femur-V40, and chemotherapy were the independent predictors of HT3+.
Conclusion: This study highlights the potential of AI-based delineation of OARs to precisely assess bone marrow dosimetric parameters in patients with cervical cancer. Our results underscore the need to optimize radiotherapy to minimize bone marrow dose and volume, thereby mitigate hematological toxicity and enhance treatment tolerance. However, chemotherapy appeared to have a more significant effect on hematological toxicity than dosimetric parameters, and patients receiving combined neoadjuvant and concurrent chemotherapy are at a higher risk of hematological toxicity. Software Copyright Registration Number 2023SR0150365.