AI in Medicine News — 2026-09-26 (Sat)

Latest news on AI and cancer treatment

Overview

Movements are noticeable in the field of screening and diagnosis (breast cancer, pancreatic cancer, liver cancer blood tests, and actual operational data for breast AI screening). In the area of treatment selection, topics included immunotherapy selection models for head and neck cancer and AI-designed molecules for multiple myeloma. On the regulatory front, there are reports that a cancer blood test has received support from an FDA advisory committee.

Breast cancer

- Report that AI reveals "hidden patterns" within breast cancer tissue (Technology Org). Details of target and method are not listed in the heading. - It is reported that Lord's Mark is advancing AI for breast cancer screening to the clinical trial stage (The Economic Times). The phase and scale of the study are not stated. - A report states that the introduction of breast cancer screening AI in South Korea has shown a reduction in interval cancers (cancers discovered after negative screening) and faster diagnosis (Korea IT Times). The level of evidence (single center or multicenter, prospective or retrospective) is unclear from the headline.

lung cancer

- Report that CT-based AI model predicts radiation pneumonitis in lung cancer patients (AuntMinnie). The target is lung cancer patients, and the target is treatment-related complications. The evidence stage is not stated in the heading. - diagnosticimaging.com ``Advances in AI — September 2026'' is a monthly summary of industry trends, and individual findings specific to lung cancer are not specified. - The article "Medical Dialogues" on ``Vertebral Bone Density Assessment and Cognitive Function Decline Using AI'' was classified in the lung cancer category, but the content is about bone density and cognitive function, and a direct relationship with lung cancer cannot be gleaned from the headline. Since there is a possibility of misclassification, this information is treated as reference information.

Malignant brain tumor

- Report that "AI tools can support malignant brain tumor treatment before recurrence" (Mirage News). The specific cancer type (glioblastoma or not), target, and evidence stage are not listed in the heading.

Leukemia/lymphoma

- Encinas Mayoral comments on OncLive about the AI-designed TRiTE (T cell guidance molecule) that targets BCMA/FcRL5 for multiple myeloma (a type of blood cancer). The development stage (preclinical or clinical trial stage) is not stated in the heading. - The "STLN Stock" article (intellectia.ai) is a stock price/investment analysis and does not include medical information, so it is not covered by this report.

Colon cancer/digestive organ

- Mayo Clinic's AI model that predicts pancreatic cancer risk three years before diagnosis was reported (Medical Xpress and Investing News Network reported on the same study). The data underlying the prediction and the verification method (retrospective cohort, etc.) are unclear from the headline. Headings related to colorectal cancer are not applicable.

Rare cancer

- Bill Ackman (investor) reportedly described a case in which an AI model found a drug combination for a critically ill patient as "incredible" (TradingView). However, the headline only says ``rare-disease'' and does not specify the specific cancer type, such as adenoid cystic carcinoma. Please note that the category notation and the text content may not match.

Diagnosis/Pathology/Image

- Leica Biosystems adds pathology products to Aperio AI Store (Imaging Technology News). This is an expansion of the commercial platform, and no clinical accuracy data is listed in the headline. - Research aimed at detecting liver cancer cells in blood by combining AI image analysis and flow cytometry has been reported (bioengineer.org). Development stage is not stated in the heading. - "Which Australian AI Healthcare Stocks Are Shaping the ASX?" (Kalkine Media) is an article for the stock market and does not include medical information, so it is not included.

Treatment selection/drug discovery

- A transparent AI model reportedly showed 95% accuracy in selecting immunotherapy for head and neck cancer (bioengineer.org). The target cancer type is head and neck cancer, and the accuracy is 95%. The scale of the study (single center/multicenter) and validation method (retrospective/prospective) are not listed in the heading.

Regulations/Approval

- Cancer blood test reportedly received support from FDA advisory committee (KGW). Please note that this is an advisory committee endorsement and not an approval by the FDA itself. The specific item name for inspection is not listed in the heading. - Reports (healthexec.com) that U.S. state attorneys general (State AGs) are seeking AI regulation at the federal level. This is not a request for the approval or designation of a specific product, but for AI policy in general, and is not limited to cancer treatment.

Clinical trial implementation status

No applicable information (individual data from ClinicalTrials.gov is not included in this message).

Japanese movements

- CareNet.com series ``Part 78: ``Smart AI'' alone will not change medical care. The reality of AI agent treatment after 1 year'' has been published. The content is a review of the actual status of medical treatment operations using AI agents over the past year, but it is unclear from the headline any specific outcome indicators or references to cancer types. - Stella Pharma Co., Ltd.

4888

The Yahoo! Finance article about this article is not applicable because it is an explanation of stock price trends using AI and does not contain information on the use of AI in medical and clinical settings.

Remarks from VIPs

- Encinas Mayoral (title and affiliation not listed in heading): Commented on OncLive about the AI design TRiTE (BCMA/FcRL5 target) for multiple myeloma. - Bill Ackman (Investor, Pershing Square. Please note that I am not speaking as a medical expert): It is reported that the case of rare disease treatment using AI has been evaluated as "incredible".

Please note: This report is based on headline information only. No conclusions are made regarding efficacy, safety, or indications unless explicitly stated in the heading. The clinical trial section cannot be described because there was no provided data.

Latest news on AI and rare and incurable diseases

Overview

The three headlines collected this time are all focused on the areas of ``diagnosis of rare diseases'' and ``development of therapeutic drugs,'' and do not include announcements regarding regulatory authorities or systems in Japan. Information that can identify individual disease names or notification numbers is not included in the headings.

Diagnosis of rare diseases

There is a report that prominent investor Bill Ackman described a case in which an AI model suggested a combination of drugs for a terminally ill patient as "incredible" (TradingView). The headline does not include the disease name, diagnosis rate, survival period, or verification stage (case report or observational study, etc.), so details cannot be confirmed. Additionally, there is a report (New York Post) that Anthropic has set up a biology lab as part of its AI drug discovery program, but this is about building a drug discovery system rather than diagnosis, and there are no specific disease or outcome indicators in the headline.

Gene analysis/genome

No applicable information

Development of therapeutic drugs

There is an article (BioXconomy) that claims that AI has paved the way for customizing mRNA cancer treatments. The headline only states the word "cancer," and it is not possible to confirm whether the disease falls under the category of rare disease or designated intractable disease. The stage (preclinical/clinical trials, etc.) and specific figures are not listed in the heading and cannot be confirmed.

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

Japanese movements

No applicable information

Remarks from VIPs

Bill Ackman (investor and CEO of Pershing Square Capital Management) reportedly called the case of an AI model finding a drug combination for a terminally ill patient "incredible" (TradingView). The name of the disease that was the subject of the comment and the specific clinical results are not listed in the headline, so it is not possible to confirm which of the 348 designated intractable diseases it falls under.

Latest news on AI and cancer treatment

Overview

Several developments were seen in the diagnostic and imaging fields (breast cancer screening equipment, skin cancer detection, glioblastoma recurrence prediction). While there were reports on the preclinical stage of drug discovery and treatment selection, there were no headlines regarding regulatory approvals and designations. The provided data does not include specific ClinicalTrials.gov information supporting the status of the trial.

Breast cancer

- India's PGIMER (medical research institute) plans to soon introduce AI-equipped equipment to support breast cancer detection (detailed accuracy and target data are not listed in the headline). - It is reported that Lord's Mark is developing an AI-based breast cancer screening device that does not use radiation and is moving toward clinical trials. At this point, it can be said that the company is at a pre-clinical trial stage.

lung cancer

No applicable information

Malignant brain tumor

- Report that an AI-based model predicts recurrence of glioblastoma and supports decision-making for targeted treatment. The stage of evidence (retrospective/prospective, single institution, etc.) is not stated in the heading. - Report that a new phthalazine drug candidate for malignant brain tumors has been identified using machine learning. It is positioned as drug discovery research at the preclinical stage.

Leukemia/lymphoma

- A report on rare cancer diagnosis introduces a case in which a doctor used an AI chatbot to diagnose and make decisions about his wife's rare cancer (this is an individual anecdote, not a report on clinical evidence). There are no descriptions of specific research results regarding diagnosis, stratification, and treatment selection specific to blood cancers.

Colon cancer/digestive organ

- Report that an AI model can predict pancreatic cancer three years in advance. Details of the target population and level of evidence (prospective/retrospective, cohort size, etc.) are not listed in the heading. - There are no specific reports regarding detection using an endoscope.

Rare cancer

- Reports related to adenoid cystic carcinoma are investment analysis articles (discussing the relationship between Royalty Pharma's investment strategy and the expansion of AI), and are not about clinical diagnosis or treatment.

Diagnosis/Pathology/Image

- Kinki University and others have developed an AI that uses images from a common camera to identify eight types of skin lesions, including age spots, moles, and skin cancer, with 95.9% accuracy. In addition, it also has a function to visualize "where the AI focuses." The evidence stage (details of the development/verification stage) is unclear from the headline alone.

Treatment selection/drug discovery

- An editorial article stating that immunotherapy, AI, and the aging of the population are changing the way surgery is performed in the field of esophageal cancer surgery. Specific numbers and evidence levels are not listed in the heading. - Phthalazine drug candidate for malignant brain tumors (preclinical stage, see [4]).

Regulations/Approval

No applicable information

Clinical trial implementation status

There is no information that can be listed because the provided data does not include the specific trial name, target cancer type, phase, or registration status from ClinicalTrials.gov.

Japanese movements

- Kinki University and others have developed an AI that uses images from ordinary cameras to distinguish eight types of lesions, including skin cancer (see 8 for details). Reports on the use of AI in areas such as breast cancer and brain tumors in Japan are not included in this list.

Remarks from VIPs

The list of important people to be tracked is blank, so no relevant statements are recorded.

Latest news on AI and rare and incurable diseases

Overview

Only two headlines were collected this time, and both of them were about international discussions about diagnostic support AI and genome data sharing, and there was no information about therapeutic drug development or domestic system trends in Japan.

Diagnosis of rare diseases

According to a report from Bioengineer.org, a method has been demonstrated that uses AI to analyze newborn facial images and detect Prader-Willi syndrome earlier than conventional methods. Prader-Willi syndrome is included in the ``endocrine diseases'' list of designated intractable diseases (the notification number is unknown as it is not listed in the heading). The headline only says ``early'' and does not indicate specific detection rates, shortened time to diagnosis, or number of targeted cases. The research stage (preclinical/case report/clinical trial) is also not specified and cannot be confirmed.

Gene analysis/genome

It has been reported that FAO (Food and Agriculture Organization of the United Nations) has participated in an international dialogue on genomic data sharing and AI utilization in the food safety field. This is a framework for the food safety field, and does not cover mutation interpretation or whole-genome diagnosis of individual patients with rare diseases.

Development of therapeutic drugs

No applicable information

Nerve/muscular diseases

No applicable information

Immunology/Collagen disease

No applicable information

Rare metabolic/endocrine diseases

Reports on Prader-Willi syndrome are in the context of diagnostic support AI, as described in 2. above, and information regarding therapeutic drugs and pathological mechanisms is not included in the headlines. The disease is listed in the endocrine disease group in the list of designated intractable diseases.

System/Patient Support

No applicable information

Japanese movements

No applicable information

Remarks from VIPs

No applicable information