Prof. Dr. Kazunori Akimoto

Prof. Dr. Kazunori Akimoto is a Professor of Medicinal and Life Sciences at the Tokyo University of Science. A medical scientist specialising in molecular oncology and medical data science, he leads research on cellular signalling mechanisms, cancer progression, and data-driven therapeutic innovations.

Data-Driven Integrative Cancer Research: Bridging Computational Discovery, Molecular Mechanisms, and Clinical Outcomes

Cancer is a highly heterogeneous disease in which clinical outcomes arise from complex interactions between tumor-intrinsic programs and the tumor microenvironment. To address this complexity, we present a data-driven integrative research framework that connects computational analysis, molecular mechanism studies, and clinical interpretation, while emphasizing that these approaches can operate both independently and complementarily. In the first part, we introduce MI-POG (Mutual Information–based Prognostic Omics Gene framework), an information-theoretic approach for unbiased genome-wide discovery of prognostic biomarkers. MI-POG captures nonlinear dependencies between gene expression and clinical outcomes, enabling the identification of clinically relevant gene sets beyond conventional Cox-based or predefined gene-panel approaches. Applications to large-scale patient cohorts demonstrate its utility in extracting prognostic signals that are difficult to detect using traditional methods. In the second part, we present an independent line of investigation focusing on cell–cell junction (CCJ) biology in breast cancer. Through multi-cohort transcriptomic analyses, we show that subtype-specific CCJ gene expression signatures stratify patient prognosis, particularly in Luminal B and Basal-like tumors. These findings reveal that coordinated regulation of adhesion-related genes reflects subtype-dependent tumor behavior and microenvironmental interactions. Importantly, this work is conceptually distinct from the MI-POG framework and is based on hypothesis-driven biological investigation of junctional systems. In the third part, we describe our ongoing research on tight junction–associated mechanisms in cancer metastasis. While tight junctions have traditionally been viewed as static structural components, our findings suggest that their dynamic remodeling contributes to metastatic dissemination. We will present emerging evidence supporting a novel molecular framework linking epithelial junction architecture to invasive and metastatic potential. Together, these studies highlight two complementary but distinct research strategies: data-driven computational discovery and mechanistic molecular investigation. By integrating insights from both approaches, we aim to advance a more comprehensive understanding of cancer biology and to contribute to the development of precision oncology strategies that bridge discovery and clinical application.