AI in Medicine — Cancer & Rare Disease — 2026-09-28 (Mon) Evening News

Latest news on AI and cancer treatment

Overview

In the diagnostic field, progress has been made in the implementation of image diagnostic AI, with Lunit's AI breast cancer screening being rolled out across the entire Stockholm prefecture in Sweden. In the drug discovery and clinical trial field, Nutshell Therapeutics' AI-generated compound received FDA IND (investigational new drug) clearance, and the cost-reducing effects of AI-based clinical trial monitoring tools were also reported. There are limited reports regarding formal approval and designation by regulatory authorities (FDA/EMA/PMDA).

Breast cancer

It has been reported that Lunit (Korea)'s AI interpretation solution for breast cancer screening will be supplied to the entire breast cancer screening business in Stockholm County, Sweden. The company is said to have expanded into weekly medical check-up business after accumulating experience in clinical research and actual medical treatment at a hospital in the prefecture for five years. In some news reports, there is a headline that says ``It replaces the interpretation of one radiologist,'' but this is just a media outlet's expression, and specific figures such as accuracy and detection rate are not shown in the headline. Separately, there is a research report in Japanese reporting that ``AI evaluates the risk of late recurrence of breast cancer from pathological images and genetic information,'' but the number of patients targeted, verification method, and stage of evidence (retrospective/prospective, etc.) are unclear from the headline.

lung cancer

We were unable to find any content directly targeting lung cancer among the headings classified in this category. There are reports that Alibaba DAMO has announced an "AI model for cancer and abdominal CT," but details of the targeted organs and cancer types are not listed in the headline. Reports on Xsight Group's AI agent collaboration and DeepTek's cost reduction of tuberculosis (TB) screening AI do not mention lung cancer itself.

Malignant brain tumor

No applicable information

Leukemia/lymphoma

None of the headlines classified in this category (Raspberry AI's application to the apparel industry, Bill Gates' AI safety remarks, Peter Diamandis' moonshot theory, Meta "Muse" AI agent) are directly related to blood cancer diagnosis, stratification, or treatment selection.

Colon cancer/digestive organ

There are reports that an AI that predicts the risk of pancreatic cancer three years in advance has been developed (medicaldialogues.in). However, the size of the target cohort, verification method (preclinical/retrospective/prospective, etc.), and indicators of predictive accuracy are unclear from the headline.

Rare cancer

We were unable to confirm content specific to rare cancers such as adenoid cystic carcinoma. The headline of the related category is that OpenEvidence, which claims to be ``ChatGPT for Doctors,'' has raised $250 million to expand its business into the drug discovery field, but the headline does not mention any specific application to the rare cancer field.

Diagnosis/Pathology/Image

- It has been pointed out (news-medical.net) that biomarkers identified by AI "may appear accurate at first glance, but may fail in the biggest tests". The specific cancer type and verification method are unclear from the headline. - It is reported that the Radiological Society of North America (RSNA)'s AI certification program has been "significantly revamped." - DeepTek introduces global access pricing for 138 countries for AI for tuberculosis screening. - Indian newspaper The Hindu publishes a special feature on the potential of medical AI and patient safety risks.

Treatment selection/drug discovery

- It is reported that a clinical trial monitoring tool using AI has reduced clinical trial costs (stage unknown, the-scientist.com). - Nutshell Therapeutics announced that it has received FDA IND (investigational new drug) clearance for NTS231, a compound that induces NRF2 degradation discovered using AI. This corresponds to the transition stage from preclinical to clinical (assumed Phase I). - Deep Origin announced that it will deploy an ADMET (absorption, distribution, metabolism, excretion, toxicity) prediction AI model to the entire drug discovery program (the stage is presumed to be the entire drug discovery process and the preclinical stage, but it is not specified in the headline). - South Korea's Hits plans to release an AI drug discovery support "AI co-scientist" tool in November, raising 18.3 billion won. - It is reported that a cancer drug maker in which Nvidia invests is planning an initial public offering (IPO). There is no specific information in the heading regarding the efficacy or approval status of the drug. Evaluations and recommendations regarding stocks are outside the scope of this report.

Regulations/Approval

It is reported that the FDA has granted IND (Investigational New Drug, notification before the start of clinical trials) clearance for Nutshell Therapeutics' NRF2 degradation-inducing compound "NTS231." This is a regulatory procedure that allows the start of clinical trials (Phase I) and does not imply approval of efficacy or safety. Others (such as AI infrastructure infrastructure development by a Korean investment company, AI drug discovery formulation design by Okuma Pharmaceutical, and government AI infrastructure business order received by Samsung SDS) are reports on investment and business projects that are different from the approval and designation of drugs and medical devices by the FDA, EMA, and PMDA, and do not fall under this item.

Clinical trial implementation status

The information provided this time does not contain specific information on the individual trial name, target cancer type, phase, or registration status derived from ClinicalTrials.gov, and there is no relevant information.

Japanese movements

Japanese media (BigGo Finance, The Medical AI Times, Japanese reprints of 매일경제, etc.) have introduced overseas developments such as Lunit (South Korea) supplying AI for breast cancer screening in Stockholm Prefecture and Hits (South Korea) raising funds to develop AI drug discovery support tools. We could not confirm any reports regarding the introduction or approval of unique AI by domestic medical institutions or regulatory authorities. Even in a report that states, ``AI assesses the risk of late recurrence of breast cancer from pathological images and genetic information,'' it is not possible to identify the entity conducting the research (domestic or overseas) from the headline.

Remarks from VIPs

- Bill Gates (no title listed, reported by Forbes): He reportedly said that concerns about the safety of AI are "not a hoax," indicating a position contrary to Trump's. The detailed context and target area of the statement (whether it is limited to cancer treatment or not) is unclear from the headline. - There are also reports about Peter Diamandis (moonshot theory) and the new CEO of South Korea's Sangmidang (former Kakao Pay COO, Han Seo-woo), but the headlines do not mention statements directly related to cancer medical AI.

Latest news on AI and rare and incurable diseases

Overview

Only one headline was provided, which was a university announcement regarding emergency endoscopic evaluation support using AI, and no direct movement toward diagnosis, therapeutic drugs, or designated intractable diseases itself could be confirmed. In terms of systems, there are no concrete developments in the headlines regarding medical expense subsidies for designated intractable diseases or research teams. Overall, during this collection period, there is a paucity of AI-related news specializing in rare and incurable diseases.

Diagnosis of rare diseases

No applicable information

Gene analysis/genome

No applicable information

Development of therapeutic drugs

No applicable information

Nerve/muscular diseases

No applicable information

Immunology/Collagen disease

No applicable information

Rare metabolic/endocrine diseases

No applicable information

System/Patient Support

- Meijo University announced an initiative to use AI to quickly determine the "need for emergency endoscopy" in emergency patients (Infoseek). Expectations were expressed for reducing the burden on doctors on duty and maintaining the local medical system. However, the target disease, number of cases, and verification stage (preclinical/case report/observational study, etc.) are not listed in the heading and are unknown. A direct relationship with designated intractable diseases cannot be confirmed from the headline.

Japanese movements

- The above announcement by Meijo University is the only confirmed domestic movement. This is a report on the use of AI for emergency endoscopy diagnosis, and is introduced as an initiative aimed at improving local medical systems (particularly to address the shortage of doctors on duty).

Remarks from VIPs

No applicable information