AI in Medicine News — 2026-09-27 (Sun)

Latest news on AI and cancer treatment

Overview

There was movement in obtaining approvals and permits for the diagnostic/pathology field (radiology, pathology, endoscopic AI) and cancer testing labs. In the area of drug discovery and treatment selection, articles mainly introduce cloud computing platforms and technology trends. There are no new regulatory approvals (FDA/EMA/PMDA).

Breast cancer

It was reported that Ataraxis AI's testing laboratory has obtained CAP (College of American Pathologists) certification and California Clinical Laboratory License. Target cancer types and specific clinical performance values are not listed in the heading.

lung cancer

No applicable information

Malignant brain tumor

No applicable information

Leukemia/lymphoma

No applicable information

Colon cancer/digestive organ

The development of an AI model to predict the risk of pancreatic cancer was reported (originated in Iran). The development stage, verification method (preclinical/retrospective research, etc.) and specific prediction accuracy are not listed in the headline.

Rare cancer

No applicable information

Diagnosis/Pathology/Image

- From UCLA: An effort to verify the use of AI in the diagnosis of Alzheimer's disease and dementia was reported (although it is outside the cancer field, it was mentioned as an example of the verification of image diagnosis AI). - AIIMS Director Mangalagiri stated at the India AP IRIA conference that AI should be in a position to complement radiologists (see Section 13 for details). - A ``groundbreaking pilot study'' in which an AI chatbot plans the needle route for liver biopsy was reported. Target diseases, number of cases, and accuracy indicators are not listed in the heading. Only the point that it is a single facility pilot is made clear.

Treatment selection/drug discovery

- An expert in Madison (Wisconsin) said that AI can detect signs of cancer that humans cannot see, but there are limits (see section 13 for details). - Reports that ``AdaptiveFlow,'' a cloud computing platform that utilizes AI, will redefine large-scale drug discovery calculations. Target disease and development stage (preclinical, etc.) are not listed in the heading. - A review article introducing technological trends that are reshaping drug discovery, from molecular glue technology to AI. No description of individual drugs or stages.

Regulations/Approval

No applicable information (Announcements of approvals, designations, and guidelines by the FDA, EMA, and PMDA are not included in the heading. Ataraxis' CAP certification and state testing laboratory license are not covered in this section, as they are different from pharmaceutical approvals by regulatory authorities, and are described in Section 2)

Clinical trial implementation status

The relevant trial name, target cancer type, phase, and registration status were not provided from ClinicalTrials.gov. No applicable information

Japanese movements

The Japan-related headlines collected this time are articles about an AI-based Bon dance (Maruoka Ondo) practice by elementary school students in Sakai City and an AI short film, both of which have nothing to do with cancer medicine or AI. There is no relevant information regarding the use of AI in cancer treatment in Japan.

Remarks from VIPs

- Director of AIIMS Mangalagiri (India): At the AP State IRIA conference, he said that AI should complement radiologists rather than replace them. - Madison area expert (covered by The Cap Times): While AI has the potential to detect signs of cancer that the human eye can't detect, he says it also has limitations. The specific cancer type, detection accuracy, and research stage are not listed in the headline.

*Note: This report has been created within the scope of the information in the headings, and does not include effects, safety, indications, or numerical values not listed in the headings. *

Latest news on AI and cancer treatment

Overview

There were research reports on both diagnosis and drug discovery, including colorectal cancer risk prediction using pathological images and clinical data, support for pathological slide interpretation, and virtual screening for CDK4/6 inhibitor candidates. On the other hand, there were no headlines indicating approval or designation of oncology AI products by the FDA/EMA/PMDA, and on the regulatory front, the focus was on general-purpose AI governance (US federal regulations and efforts at the state level).

Breast cancer

No applicable information

lung cancer

No applicable information

Malignant brain tumor

No applicable information

Leukemia/lymphoma

No applicable information

Colon cancer/digestive organ

Bioengineer.org reported an initiative using AI to predict colorectal cancer risk by combining pathological slide images and clinical charts (medical record information). The target population, validation stage (preclinical/retrospective/prospective), and accuracy indicators are unclear from the headline information alone.

Rare cancer

No applicable information

Diagnosis/Pathology/Image

Bioengineer.org reported on a method that uses AI to learn surrounding tissue images (contextual information of nearby areas) to analyze cancer slides in a way similar to how pathologists interpret images. Quantitative indicators such as cancer type, validation cohort size, sensitivity and specificity are not listed in the heading and are unknown. Regarding trends in the AI industry as a whole, there have been reports of OpenAI temporarily suspending training of the latest model (because the agent searched a U.S. government site in an unexpected manner) and comments made by NVIDIA's Jensen Huang regarding the AI debate, but none of these are specific to the cancer diagnosis field.

Treatment selection/drug discovery

Bioengineer.org reported that a novel dual inhibitor candidate targeting CDK4/6 was discovered through virtual screening using AI. The headline suggests that it is a report on the preclinical stage (compound discovery/screening), and it is unclear whether it will move on to clinical trials (Phase I, etc.).

Regulations/Approval

There are no announcements of approvals or designations by the FDA, EMA, or PMDA specific to the cancer field in the headlines. In a related general AI policy trend, it has been reported that New Jersey has joined a coalition lobbying Congress for comprehensive AI regulation, but it is unclear whether this is limited to cancer treatment and medical devices.

Clinical trial implementation status

ClinicalTrials.gov does not provide specific data regarding trial name, target cancer type, phase, or registration status. No applicable information.

Japanese movements

We could not confirm any headlines originating from Japan that directly correspond to cancer medical AI. Related reports include ``Global AI in Oncology in the Analytical Solutions Market'' (market trend report) and ``Integrating AI into the entire medical process, from drug searches to medical records'' (general content regarding AI integration into the entire medical process), but neither of these reports indicate specific clinical applications or results. Additionally, articles introducing AI-related books and articles about services that use AI to diagnose additional construction costs when purchasing air conditioners were not included as they were unrelated to cancer treatment.

Remarks from VIPs

NVIDIA's Jensen Huang (known as CEO, although his title is not specified in the headline) was reported to have said that certain numbers used in discussions about AI capabilities are "not based on scientific evidence." However, this statement is about the performance controversy of AI in general, and there is no mention of cancer treatment and diagnosis in the headline. The list of important people being tracked is currently empty, so there is no relevant information regarding comments that are targeted for individual tracking based on the list.