In Japan, comprehensive genomic profiling (CGP), which uses next-generation sequencing to analyze multiple genomic alterations simultaneously, has been introduced into the national health insurance system, marking the full-scale implementation of cancer genomic medicine. CGP is primarily offered to patients with solid tumors who have completed, or are expected to complete, standard-of-care treatment. A nationwide clinical network comprising more than 300 cancer genomic medicine core hospitals, designated hospitals, and cooperative hospitals has been established.
The Center for Cancer Genomics and Advanced Therapeutics (C-CAT), established at the National Cancer Center, collects not only genomic test results but also detailed clinical information, including treatment histories, treatment responses, adverse events, and patient outcomes. As of the end of October 2026, data from approximately 150,000 patients will be registered in C-CAT. For each patient, C-CAT generates a report known as the "C-CAT Findings Report," which provides the treating hospital with information on approved drugs and clinical trials that may be matched to the patient's genomic and clinical profile. At the hospital that ordered the test, a multidisciplinary expert panel reviews the C-CAT Findings reports and considers appropriate treatment options for the individual patient, including approved therapies and clinical trials. This framework promotes equitable access to high-quality cancer genomic medicine throughout Japan.
The C-CAT database, which integrates genomic data with detailed clinical information, represents a globally unique research resource and is being extensively used for academic research and drug development. Currently, more than 100 organizations, including 17 pharmaceutical companies, are analyzing C-CAT data for research and development purposes. These analyses are beginning to identify genomic alterations characteristic of the Japanese population, as well as areas of substantial unmet medical need in which treatment options remain limited. These real-world data are also being used by pharmaceutical companies for drug discovery and clinical trial design. Several studies using C-CAT data have already supported regulatory decision-making, including the expansion of indications for approved therapies.
Cancer genomic medicine under Japan's national health insurance system has evolved into an integrated framework in which routine clinical care, research, and development reinforce one another. In this presentation, I will describe the current status of cancer genomic medicine in Japan, its contributions to cancer research, drug development, regulatory science, and health policy, and its future prospects.
While common genetic variants associated with lung adenocarcinoma (LUAD) risk have been identified through genome-wide association studies (GWAS), the role of rare germline pathogenic variants (GPVs), particularly in early-onset and inherited disease, remains poorly characterized in Asian populations. We conducted GWAS focusing on lung adenocarcinoma, identifying 19-29 susceptibility genes. Our previous research indicated that risk variants in telomere-related genes are correlated with telomere length. We also demonstrated that polygenic risk scores contribute to the stratification of lung cancer risk.
Furthermore, we examined GPVs in 454 hereditary cancer and DNA repair genes using whole-exome and whole-genome sequencing of 350 early-onset (aged under 40 years) and 1,441 late-onset (aged 41 years or older) LUAD cases from the Japanese population. A case-control study comprising 10,672 LUAD cases and 7,898 healthy controls was performed to identify moderate-risk genetic factors for this disease. An analysis of somatic mutations in 1,280 LUAD patients, including 31 patients with GPVs, was also performed. The frequencies of GPVs in TP53 and BRCA2 were significantly higher in patients with early-onset LUAD than in those with later-onset LUAD. TP53 and BRCA2 GPVs were dramatically enriched in early-onset LUAD (2.9% and 1.7%, respectively) compared with later-onset disease (0.14% and 0.21%; more than a 10-fold increase, P < 0.001). Patients with BRCA1 GPVs have a high incidence of concurrent TP53 somatic mutations. BRCA2 GPV carriers exhibited biallelic inactivation with homologous recombination deficiency, suggesting potential responsiveness to PARP inhibitor therapy. A novel germline ALKBH2 frameshift variant (p.Glu35Alafs54) was associated with early-onset LUAD risk and correlated with smoking-related mutational signatures (SBS4) in proportion to smoking exposure. TP53 and BRCA2 GPVs and the novel ALKBH2 variant are associated with early-onset LUAD in Asian populations. These findings identify actionable genetic markers for early LUAD detection and suggest potential therapeutic vulnerabilities in BRCA2-deficient tumors, with implications for precision medicine in Asian populations.
Recent studies on the association between overdiagnosis and the recent introduction of low-dose computed tomography (LDCT) in East Asia prompt investigations on whether never-smoking females identified by a model as high-risk are more likely to develop late-stage lung cancer or to die of lung cancer. Because overdiagnosis has been a concern in breast cancer screening programs, one would also like to know whether women identified by a model as high-risk are more likely to develop late-stage breast cancer. Although similar concerns exist for other cancers, I will report some of the relevant observations for lung cancer and breast cancer in Taiwan.
Using the linkage of the Taiwan Biobank, Taiwan Cancer Registry (TCR), TCR Long Form, and Cause of Death Database, we observed that Taiwanese never-smoking females with high predicted risk, according to an update of the lung cancer risk prediction model for Taiwanese never-smoking women (TNSF), have a high observed chance of being diagnosed with late-stage lung cancer and of dying from lung cancer, and might benefit the most from LDCT screening. Randomized controlled trials are warranted to confirm its mortality benefit, beyond personalized counseling and screening design. Using the linkage of the Taiwan Breast Cancer Screening Database, TCR, TCR Long Form, Cause of Death Database, and National Health Insurance Research Database, we built breast cancer prediction models for Taiwanese women and observed that Taiwanese women with high predicted risk have a high observed chance of being diagnosed with late-stage breast cancer, interval cancer, or symptom-prompted cancer. Clinical applications of these results deserve further investigation.
Lung cancer remains the primary cause of cancer-related mortality across Asia. While landmark Western trials such as NLST and NELSON established low-dose computed tomography (LDCT) screening for heavy smokers, strict tobacco-based criteria miss a substantial proportion of the Asian disease burden. Asia faces a distinct dual-epidemiological challenge: persistently high male smoking rates in several regions coexist with a high incidence of lung cancer in never-smokers, particularly among women presenting with driver-mutated adenocarcinomas. To address this challenge, the landmark Taiwan Lung Cancer Screening in Never-Smoker Trial (TALENT) evaluated high-risk non-smokers using multi-factorial risk criteria, including family history of lung cancer, passive smoking, cooking behaviors, and chronic lung disease. The TALENT study demonstrated a remarkably high lung cancer detection rate, with more than 95% of detected cases diagnosed at an early, highly curable stage. These decisive findings provided the scientific foundation for the world's first national, government-funded LDCT screening program targeting high-risk non-smokers alongside heavy smokers in Taiwan.
The success of the TALENT framework offers a crucial model for Asian nations seeking to implement risk-stratified screening for non-smoking populations exposed to genetic susceptibility, environmental hazards, and ambient air pollution. However, expanding LDCT screening across broader populations inevitably increases reader workload, false-positive rates, and potential overdiagnosis. Artificial intelligence (AI) serves as a critical enabler in overcoming these scalability challenges by optimizing triage, reducing inter-observer variability, and refining risk stratification. Integrating AI-assisted diagnostics into multi-factorial screening paradigms offers a scalable and sustainable roadmap for lung cancer control worldwide.