Longitudinal drug response assessment of tumor organoids based on optical attenuation coefficient and multi-dimensional morphological characterization

Patient-derived tumor organoids (PDTOs) are vital for precision oncology, but standard drug screening methods, including adenosine triphosphate (ATP) assays, are destructive and prevent longitudinal monitoring. Optical coherence tomography (OCT) offers non-destructive 3D imaging, capturing morphological features and tissue attenuation characteristics via the optical attenuation coefficient (OAC). Here, we propose a non-destructive evaluation framework for tumor organoids that fuses OAC and multi-dimensional morphological features. Using intrahepatic cholangiocarcinoma (iCCA) PDTOs treated with icaritin, we found that OAC exhibited a significant dose-dependent increase (up to 32.8% at 80 μM compared to control), accompanied by a morphological transition of cystic organoids into solid phenotypes. By integrating these features via K-means++ clustering and principal component analysis, we constructed a relative growth score. This fusion score correlated strongly with the ATP gold standard (Pearson correlation coefficient r = 0.938), outperforming a morphology-only model ( r = 0.906). Furthermore, independent experiments with first-line chemotherapeutics (e.g., 5-Fluorouracil, Gemcitabine) and combinatorial regimens indicated the model's potential generalizability ( r = 0.887). This method overcomes the limitations of destructive, single-metric evaluations, providing a quantitative and non-destructive platform for high-throughput drug screening and personalized treatment decision-making.