Evaluation of radiation-induced bone marrow toxicity using artificial intelligence-based image analysis in mice.

Abstract

Radiation exposure profoundly affects bone marrow, with even low-dose exposure inducing notable biological changes. Traditional pathology methods often face challenges in detecting subtle alterations. However, artificial intelligence (AI) offers enhanced sensitivity and precision, enabling a more detailed analysis. This study aimed to evaluate the effect of whole-body irradiation (WBI) on blood parameters, bone marrow density, and apoptosis in mice, employing AI-based image analysis for accurate and efficient quantification of cell density and apoptosis. Female C57BL/6 mice (n = 120) were divided into four groups: a control group (Group 1) and three irradiated groups (Groups 2, 3, and 4), exposed to 0.5 Gy, 1 Gy, and 2 Gy of WBI, respectively. Mice were sacrificed at 1, 3, and 7 days post-irradiation for analysis. Blood samples were assessed for hematological changes, and the sternum was histopathologically evaluated. Despite stable body weights, WBI significantly altered blood parameters, reducing white blood cell and red blood cell counts while increasing platelet counts. Histopathological examination revealed a marked reduction in bone marrow cellularity in Groups 3 and 4 on day 1 post-irradiation; however, cellularity appeared to recover in these groups by days 3 and 7. AI-based image analysis of the sternum provided precise quantification, confirming a significant decrease in bone marrow cellularity in Groups 3 and 4 compared to Group 1 on day 1 (p < 0.01). Apoptosis analysis also demonstrated a significant increase in the apoptotic index in Group 4 sternum samples on days 1, 3 (p < 0.01), and day 7 (p < 0.05) relative to Group 1. In summary, WBI induced hematological and histopathological alterations in mice, characterized by changes in blood cell parameters, bone marrow cellularity, and apoptosis. The integration of AI-based image analysis provides a robust and efficient tool for quantifying these changes, offering considerable potential for advancing radiation biology and pathology research.