The modernization of traditional Chinese medicine (TCM) confronts a persistent bottleneck: interpreting its dynamic, holistic “formula–human body” system in the language of modern science. The multi-component, multi-target pharmacology of TCM compound formulas resists reductionist analysis, and conventional experimental models fall short of capturing the systemic regulatory effects that underlie clinical efficacy. Recent advances in artificial intelligence (AI) and organoid technology now provide complementary tools to address this long-standing challenge. AI supplies the computational power to model complex, high-dimensional biological networks; organoids furnish biologically faithful platforms to validate predictions in human-relevant contexts. Together, these technologies enable a closed-loop research paradigm—computational prediction, experimental verification, model iteration—that is high-throughput, quantifiable, and predictive. This Comment argues that the convergence of AI and organoids can drive TCM research from an experience-driven tradition toward a data-driven, mechanism-verified framework, bridging millennia of accumulated clinical wisdom with contemporary biomedical science.

Glioblastoma (GBM) has been regarded as one of the most deadly and challenging cancers to treat with extremely poor prognosis. The limited efficacy of current chemotherapies might be attributed to the presence of glioma stem cells (GSCs) as well as the difficulties in passing through the blood–brain barrier (BBB) and targeting tumor cells. Tumor-derived exosomes are emerging as novel and promising drug delivery systems. However, great concerns regarding the biosafety and BBB penetrability remain to be addressed. Herein, we have developed a simple and feasible strategy to engineer GBM cell-derived exosomes with improved biosafety termed “Exo@TDPs” to deliver the cargos of chemotherapeutic agents temozolomide (TMZ) and doxorubicin (DOX) into GBM tissues. Exo@TDPs decorated with angiopep-2 (Ang-2) and CD133-targeted peptides improve the capacity to penetrate the BBB and target tumor cells. Both in vitro and in vivo studies demonstrate that Exo@TDPs can cross the BBB, target GBM cells, penetrate into deep tumor parenchyma, and release the therapeutic cargos effectively. Synergistic delivery of TMZ and DOX by Exo@TDPs exerts therapeutic effects to suppress the tumor growth and prolong the survival time of orthotopic syngeneic mouse GBM models. These findings suggest that our developed Exo@TDPs loaded with chemotherapeutic drugs may bring new possibilities for the application of tumor cell-derived exosomes for brain tumor treatment.

Programmed death 1 (PD-1) and its ligand PD-L1 are two typical immune checkpoints. Antibody-based immune checkpoint blockade (ICB) strategy targeting PD-1/PD-L1 achieved a significant therapeutic effect on cancer. However, due to the impenetrability of antibody drugs and the occurrence of immune-related adverse events, only a minority of patients benefit from this treatment. Peptides multimerization has been widely proved to be an effective method to improve receptor binding affinity through a multivalent synergistic effect. In this study, we report a novel peptide-aggregation-induced emission (AIE) hybrid supramolecular TAP, which can self-assemble into nanofibers through non-covalent interactions such as hydrogen bonds, with a specific nanomolar affinity to PD-L1 in vivo and in vitro. Combined with near-infrared agents, it can be used for tumor imaging and photothermal therapy, which enables photothermal ablation of cancer cells for generating tumor-associated antigen (TAA) and triggering a series of immunological events. Collectively, our work suggests that synthetic self-assembled peptide nanofibers can be developed as attractive platforms for active photothermal immunotherapies against cancer.