Graduate School of Medicine, Division of Health Sciences
My research centers on developing individualized treatment-prediction models for boron neutron capture therapy (BNCT) by integrating multimodal medical image analysis with radiomics and dosiomics. In conventional BNCT planning, patient eligibility has typically been determined pre-treatment using [18F]FBPA-PET, based on the tumor-to-normal maximum uptake ratio (Tmax/N). Recent reports, however, indicate that the minimum intratumoral uptake ratio (Tmin/N) correlates more closely with therapeutic effect, suggesting that heterogeneity in intratumoral drug distribution may critically influence outcomes. A key challenge is the lack of established methods to quantitatively characterize this spatial distribution.
To address this gap, my study extracts and analyzes morphological and textural descriptors from imaging via radiomics, together with spatial features derived from high-precision dose distributions via dosiomics. By exploring new suitability metrics that improve the prediction of both treatment response and adverse events, I aim to elucidate the relationship between microscopic heterogeneity of boron-drug distribution and clinical results. The goal is to construct models that enhance local control and response-prediction accuracy in BNCT, thereby advancing toward more precise, patient-specific BNCT.