AI in Medicine — Cancer & Rare Disease — 2026-10-01 (Thu) Evening News

Latest news on AI and cancer treatment

Overview

- The areas where there was notable movement were the areas of diagnosis, imaging, and pathology, with reports including detection of pancreatic and gastric cancer, biomarker prediction from pathology slides, and breast cancer risk assessment. - In drug discovery, multiple media reported that Dong-A ST and Galux have begun joint research on AI-designed ADCs. - No regulatory/approval headings and no clinical trial data provided.

Breast cancer

- NYT reported that women can now use AI to assess their breast cancer risk. The target, accuracy, and verification stage are unclear from the headline. - WCYB reported that Simbiosys demonstrated advances in AI in breast cancer care. The specific details are unknown. - diagnosticimaging.com has published a podcast (Part 1) covering the current status, challenges, and possibilities of AI in breast imaging diagnosis. - Lunit plans to present five AI biomarker studies across lung, breast, and pancreatic cancers at ESMO 2026 (Korea Biomedical Review). The content of breast cancer is unknown.

lung cancer

- Lunit's AI biomarker presentations for ESMO 2026 (a total of 5 topics across lung, breast, and pancreatic cancers) were reported in Korea Biomedical Review and Businesskorea. At the time of publication, the results are not in the headlines. - No other headings specific to lung cancer.

Malignant brain tumor

- Bioengineer.org reported that AI has mapped mechanisms of treatment resistance in pediatric brain tumors. The target is pediatric brain cancer, and it is unclear from the headline whether it is a glioblastoma, the research stage, and the results.

Leukemia/lymphoma

No applicable information

Colon cancer/digestive organ

- Earth.com reported that AI predicts pancreatic cancer years before diagnosis. Accuracy, target population, and type of validation (retrospective or prospective) are unclear from the headline. - Bioengineer.org reported that an AI model detected stomach cancer in patients with psychological symptoms. Similarly, the numbers and verification stage are unknown. - There is no heading regarding endoscopic detection of the colon.

Rare cancer

No relevant information (no heading specific to adenoid cystic carcinoma/sarcoma). As a reference, there is an article (finance.biggo.com) by Daniel McKinnon, who deals with diagnosis of rare diseases, but it is unclear whether the content is limited to cancer.

Diagnosis/Pathology/Image

- Bioengineer.org reported that AI predicted cancer biomarkers in 12 tumor types from routine pathology slides. The accuracy and verification stage are unclear from the headline. - The same media also reported on AI that calculates senescence scores from blood proteins and predicts death and disease. It is unclear whether the content is limited to cancer. - Radiology Business reported that the introduction of AI in private practice has brought efficiency gains. It says "measurable" improvement, but there are no numbers in the headline. - XAID selects top 7 AI radiology platforms for remote reading providers (marketscreener.com). This is a vendor rating, and the basis for clinical performance is unknown. - AuntMinnie (AI in clinical practice), Cardiovascular Business (Mass General's cardiovascular imaging workflow), and ET HealthWorld (AI native hospital) are not cancer-specific.

Treatment selection/drug discovery

- Dong-A ST and Galux have started joint research on ADC using AI antibody design (Seoul Economic Daily, Asian Economy, Chosunbiz). There is no mention in the heading that this is the beginning stage of joint research and that it is in the clinical trial stage. - Bioengineer.org reported an AI framework that removes batch noise from DNA methylation data and refines cancer subtype classification. This is a method report, and its clinical usefulness is unclear from the headline. - EurekAlert! reported on AI-driven analysis and data mining strategies in drug discovery (which appears to be a general overview). - Collaborations Pharmaceuticals published case studies of AI applications from the past 10 years (EIN News). - Microsoft announces Quine, an AI for scientific discovery (HPCwire). It is unclear from the headline whether it is cancer-specific.

Regulations/Approval

No relevant information (no headings regarding FDA/EMA/PMDA approval/designation/guidelines).

Clinical trial implementation status

No applicable information. ClinicalTrials.gov data is not included in the input and trial name, phase, and registration status are not inferred.

Japanese movements

- The Nihon Keizai Shimbun published an interview with Yusuke Nakamura titled "Genomic medicine 'Avoiding side effects and diagnostic errors through integration with AI'". The purpose of the headline is to avoid side effects and diagnostic errors through the fusion of genomic medicine and AI, but the underlying data is unclear from the headline. - Global Startup EXPO 2026 will be held in Osaka, with 15 deep tech companies in the fields of AI, advanced robots, and bio/healthcare taking the stage. The content is not specific to cancer medical care. - East Asia ST and Galux's ADC joint research (with article in Japanese) and KT's medical AI standard model (Mainichi Keizai) are moves by Korean companies and are not examples in Japan.

Remarks from VIPs

- The list of VIPs being tracked is empty and not applicable. - For reference, list the people whose statements are included in the headline. In both cases, the titles are not listed in the headings, and the text is unconfirmed. - Yusuke Nakamura: Avoiding side effects and diagnostic errors through the fusion of genomic medicine and AI (Nihon Keizai Shimbun). - Daniel McKinnon: Diagnosis of rare diseases is AI's first "superhuman" area and a $10-per-diagnosis market (finance.biggo.com).

Note: Headlines related to market prices and market conditions (such as Asian shares) were excluded as they were out of scope.

Latest news on AI and rare and intractable diseases

Overview

Movements can be seen mainly in the rare disease diagnosis field (remarks discussing the commercialization of AI-based diagnosis). Genetic analysis and therapeutic drug development are one or two headlines each, and the content remains at a summary level. There are no headings related to systems/patient support and trends in Japan.

Diagnosis of rare diseases

- Daniel McKinnon is quoted as saying, ``Rare disease diagnosis is AI's first 'superhuman' market'' and ``$10 per case'' are the headlines distributed in the English and Japanese versions (BigGo Finance). - A conversation between him and Joel Borgen titled "AI Utopia and Rare Disease Diagnosis" (The Cognitive Revolution) is also in the headlines. - From the headline, it is not possible to confirm how much the time to diagnosis has been reduced, what percentage the diagnosis rate is, the target disease, or the verification method (prospective/retrospective, number of cases). "Superhuman" and "$10" are headline claims, and the level of proof is unclear.

Gene analysis/genome

- Earth.com: "Millions of years of evolution have fostered an AI model that identifies dangerous DNA." The content is that an AI model that has learned evolutionary information can identify dangerous DNA. What was determined (pathological significance of the mutation, sensitivity/specificity), target disease, and validation data cannot be confirmed from the headline, and the level of evidence is also unclear. - Tempus AI (TEM): This is an article for the stock market asking, ``What is the background behind the latest topics?'' The results of diagnosis and analysis cannot be gleaned from the headline.

Development of therapeutic drugs

- LMU Munich (University of Munich) visited CUHK Medicine (Chinese University of Hong Kong). The headlines list CRISPRa gene therapy and healthcare AI, but they appear to be reports of visits and interactions, and do not indicate the results of therapeutic drugs. Target disease and development stage are unknown. - Collaborations Pharmaceuticals shares safe AI practices from the past decade. It cannot be confirmed from the headline whether the drug falls under the category of orphan drug or drug repurposing, and the development stage is also unknown.

Nerve/muscular diseases

No applicable information

Immunology/Collagen disease

No applicable information

Rare metabolic/endocrine diseases

No applicable information

System/Patient Support

No applicable information. The two articles classified as ``systems and support for designated intractable diseases'' are both headings of rare disease diagnosis related to Mr. McKinnon, and the content regarding medical expense subsidies, the Intractable Disease Act, research groups, and patient associations cannot be confirmed.

Japanese movements

No applicable information. Japanese headlines are only available in Japanese for overseas articles provided by BigGo Finance.

Remarks from VIPs

- Daniel McKinnon (title not visible in headline): argues that rare disease diagnosis is the first "superhuman" market for AI, at $10 per case. The basis for the numbers and verification data are unclear from the headline. - Joel Borgen (title not visible in the headline): With McKinnon, he speaks on The Cognitive Revolution on the topic of AI and rare disease diagnosis. The main point of the statement is not clear from the headline.

Please note that the headline for this period does not include the specific name of the disease, so it is not possible to confirm whether it is a designated intractable disease or not. The number of cases and number of facilities are not listed in the heading, so all information needs to be confirmed with primary information.