AI in Medicine — Cancer & Rare Disease — 2026-10-02 (Fri)

Latest news on AI and cancer treatment

Overview

- The main focus of the movement is business development in diagnosis and risk assessment (breast cancer) and partnership in AI biomarkers and companion diagnostics (CellCarta/Imagene AI). In drug discovery, we have released platforms and partnered with Talus Bio and Rakovina, and in Japan, Shionogi has published articles on AI drug discovery. - The only regulatory aspect is that Clairity's breast cancer risk assessment AI is labeled as "FDA-Authorized." The type of approval is unclear from the heading. - All headlines are at the announcement stage of products, partnerships, and conference presentations, and there are no numbers indicating clinical efficacy in the headlines.

Breast cancer

- Risk Assessment (Clairity/Everlywell): Clairity's "FDA-Authorized" AI breast cancer risk assessment will be rolled out nationwide to women in the United States through Everlywell. Numerical values such as accuracy and scope/type of approval are not listed in the heading. - Direct-to-Consumer Screening (Eve Wellness): Announcement of expanding AI breast cancer screening to Walnut Creek. The evaluation method and performance are unclear from the headline. - Mammography Quality Control (KMUH / MacroInsight): Announcement that they have developed a real-time AI edge system. At the development stage, performance data is not in the headlines. - General coverage (CBS News): Covers the potential for AI to help diagnose breast cancer. The specific findings are unclear from the headline.

lung cancer

- This content uses an AI model to extract smoking history and utilize it for lung cancer screening (European Medical Journal). The sampling accuracy and research design are unclear from the headline. - The collaboration between CellCarta and Imagene AI will be discussed in the Diagnostics and Drug Discovery section.

Malignant brain tumor

No applicable information

Leukemia/lymphoma

- Although AI can detect leukemia cells on blood smears, it reportedly continues to miss certain types of blast cells (Bioengineer.org). This shows the limits. It is unclear from the headline which blast type, detection rate, and study design.

Colon cancer/digestive organ

- Mayo Clinic's initial AI model reportedly showed pancreatic cancer risk from medical records up to three years before diagnosis (Medical Daily). This is an "Early" model, and the sensitivity, specificity, and stage of validation (retrospective or prospective, etc.) are unclear from the headline. - There are no headings regarding endoscopic detection.

Rare cancer

No applicable information

Diagnosis/Pathology/Image

- CellCarta and Imagene AI: Announcement of expanding collaboration to validate, implement, and expand AI biomarkers and companion diagnostic programs across drug discovery (prnewswire, Unite.AI). Performance levels are not listed in the heading. - Lunit: We are planning to present 5 research studies on AI biomarkers at ESMO 2026. The content and target cancer types are unclear from the headline. - Workflow: There is an article that says “AI extends from note-taking to clinical workflow” (Healthcare IT News). It is unclear from the headline whether it is cancer-specific or not. - There are two market size reports for radiology AI and image diagnosis AI, but they are not included in the evaluation because they do not provide clinical knowledge.

Treatment selection/drug discovery

- Talus Bio: "Ptarmigan-1" has been released. It is said to be a drug discovery AI model that does not use structural information and targets the entire proteome. At the drug discovery platform stage, there is no mention of individual drugs or clinical trial stages. - Rakovina Therapeutics x Celvion Therapeutics (Boston): Collaborated on precision cancer drug discovery using AI and real-world clinical data. The target drug and development stage are unclear from the headline. - Shionogi & Co.: It is reported that AI drug discovery will halve the time it takes to enter clinical trials to three years (Nihon Keizai Shimbun). It is said that the company will utilize the human resources acquired through the acquisition of JT's pharmaceutical division. The target cancer type is unclear from the headline.

Regulations/Approval

- FDA: Clairity's breast cancer risk assessment AI is labeled as "FDA-Authorized". The type of approval/approval and scope of application are unclear from the headline. - There are no headlines regarding EMA and PMDA.

Clinical trial implementation status

This input does not include the data obtained from ClinicalTrials.gov, so it is assumed that there is no relevant information. We do not make assumptions about the trial name, phase, or registration status.

Japanese movements

- It is reported that Shionogi & Co., Ltd. will use AI drug discovery to halve the time it takes to enter clinical trials to three years (Nihon Keizai Shimbun). It is said that the company will utilize the human resources acquired through the acquisition of JT's pharmaceutical division. There are no headlines regarding the use of AI in cancer treatment in clinical settings in Japan.

Remarks from VIPs

No relevant information (list of important people being tracked is empty).

Note: This report is based on headlines only. Research design, accuracy, etc. have not been confirmed in the main text, and points not listed in the heading are marked as "unknown."

Latest news on AI and rare and intractable diseases

Overview

This quarter, developments in diagnostics and systems cannot be confirmed from the headlines, and the main focus is on therapeutic drugs and drug discovery platforms (AI-designed enzymes, early clinical data on AI-designed small molecules, and investment in gene transfer technology). The only headline directly related to the designated intractable disease was one on immunology/collagen diseases (AI analysis of ANCA-related vasculitis), and the headline on nerve/muscle diseases was unrelated to the disease. All information is only in the headlines, and efficacy, safety, and number of cases cannot be read.

Diagnosis of rare diseases

No applicable information (There are no headings indicating a reduction in time to diagnosis or diagnosis rate.)

Gene analysis/genome

- Regarding Tempus AI (TEM), there is a page with stock prices and news, and a headline that says ``Pricing and data growth support long-term expansion.'' It is not possible to confirm from the headline what was determined by mutation interpretation or whole genome analysis. - It is reported that Anthropic has added A$10 million to its investment in Australian science, bringing the total to A$13 million. The relationship with genome analysis cannot be confirmed from the headline. - All information is business/investment information and does not fall under the clinical evidence stage.

Development of therapeutic drugs

- There are reports that BioMarin will work on creating a new enzyme therapy using Profluent's AI model (it appears that BioMarin has adopted the AI model as a partner, but the target disease and development stage cannot be confirmed from the headline). It is not possible to confirm whether the disease falls under a designated intractable disease. - Insilico Medicine plans to present first-in-human data for its AI-designed pan-TEAD inhibitor ISM6331 at ESMO 2026. The stage is an early clinical announcement (equivalent to phase 1), and efficacy has not been shown. It is not possible to confirm from the headline whether the subject is a rare disease or not, nor can it be determined whether the disease is a ``rare disease that is not a designated intractable disease.'' - Regarding gene transfer technology, there are reports that AI is changing the technology and that VC investment will reach $258.7 billion, marking a structural shift. This is an investment trend and does not indicate the effectiveness or safety of individual treatments. - Here's an article discussing what Anthropic's "wet labs" mean for pharmaceuticals. This is a commentary, and the results of therapeutic drug development are not shown.

Nerve/muscular diseases

No applicable information (The headline Huntington Ingalls Industries appears to be a stock article about a shipbuilding/defense-related company, and was judged to be unrelated to Huntington's disease (a neuromuscular disease listed) and was excluded from the analysis. There is a possibility of misclassification at the time of collection.)

Immunology/Collagen disease

- There is a report that "AI-based analysis shows four different patterns of organ damage in AAV (ANCA-associated vasculitis)" (ANCA Vasculitis News). - The list includes microscopic polyangiitis, granulomatosis with polyangiitis, and eosinophilic granulomatosis with polyangiitis, all of which are included in ANCA-related vasculitis, but it is not possible to confirm from the heading which disease range the heading "AAV" refers to. No notification number will be attached. - The stage of evidence (retrospective analysis or cohort size), number of cases, and validation status of the analysis cannot be determined from the headline. It has not been shown whether pattern classification leads to prognosis or treatment selection.

Rare metabolic/endocrine diseases

No applicable information (It is not possible to confirm from the headline whether the diseases targeted by BioMarin's enzyme therapy are metabolic diseases.)

System/Patient Support

No applicable information

Japanese movements

No applicable information

Remarks from VIPs

No applicable information (This is because the list of important people being tracked is empty.)