Characterization of hepatocellular carcinoma patient-derived organoids from patients receiving transarterial chemoembolization as a model for preclinical drug response assessment.

Hepatocellular carcinoma (HCC) is the most frequent type of primary liver cancer and is often diagnosed at an intermediate stage in patients with chronic liver disease. Transarterial chemoembolization (TACE) is the first-line treatment in this setting. However, drug response remains variable and optimal chemotherapeutic selection unresolved. Idarubicin (IDA) has previously demonstrated antitumor activity comparable to doxorubicin (DOX), with improved emulsion stability and membrane permeability. 3D patient-derived organoids (PDOs) have also recently emerged as physiologically relevant systems that recapitulate tumor characteristics and inter-patient heterogeneity more accurately than conventional 2D cell cultures. In this study, we characterized HCC PDOs and investigated their utility as an ex vivo model to assess therapeutic responses. We generated organoids from tumor and non-tumor liver biopsies collected from HCC patients prior to TACE-treatment. PDOs partially preserved tumor architecture and phenotypic features. IDA exposure resulted in a concentration-dependent reduction in organoid growth, with marked inter-patient variability in drug response. Comparative analyses showed that IDA was more potent than DOX, with lower IC₅₀ values across all patient-derived samples. IDA also induced transcriptional changes associated with cellular stress, inflammation, and proliferation pathways. However, ex vivo drug‑sensitivity testing in PDOs did not correlate with clinical response to TACE. These findings confirm the feasibility of generating HCC PDOs and using them to evaluate differential sensitivity to TACE‑associated chemotherapeutic agents ex vivo, although their ability to predict clinical efficacy remains unproven. While IDA exhibited greater cytotoxic activity than DOX in organoid cultures, this observation was based on in vitro sensitivity assays and does not establish clinical superiority. Furthermore, the absence of correlation between PDO drug sensitivity and clinical TACE response indicates that PDO-based prediction of treatment outcome remains challenging in this setting.