Leica Biosystems Clears the Industry's First Standalone AI Quality-Control Software for Digital Pathology 🔗

Leica Biosystems, a Danaher company, secured multiple FDA 510(k) clearances that extend its clinical digital pathology portfolio, headlined by Aperio iQC DX, described as the industry's first standalone AI-assisted quality-control software. The clearances also add the Aperio GT 180 DX scanner and enhance the GT 450 DX. The point of an AI QC layer is unglamorous and structural: catch image-quality problems early, before a slide reaches the pathologist, so laboratories can scale digital workflows without losing consistency under rising case volumes and staffing pressure. Digitisation was the last decade's project. Governing the quality of what gets digitised is this one's.

Coreline Soft and INFINITT Deploy an AI Chest-CT Workflow at a US Health System 🔗

Coreline Soft and INFINITT North America have put a chest-CT AI reading workflow into routine use at St. Joseph's Health in New Jersey, their second North American deployment together. Coreline's AVIEW platform is integrated directly into the existing INFINITT PACS, with AI pre-analysis flagging and quantifying findings such as lung nodules inside the radiologist's normal workflow, no separate application required. The initial focus is lung cancer screening, with prior-image comparison to support tracking of subtle changes over time. This is the version of diagnostic AI that actually sticks: embedded in the system radiologists already use, not bolted on beside it.

Freenome's AI Blood Test for Lung Cancer Wins FDA Breakthrough Device Designation 🔗

The FDA granted Breakthrough Device Designation to SimpleScreen Lung, Freenome's investigational AI-powered multiomic blood test for lung cancer screening. It is aimed at adults aged 50 to 80 with a 20 pack-year smoking history who are not currently being screened, a large and under-served group given how low uptake of guideline lung screening remains. The designation brings prioritised FDA interaction rather than clearance, but it signals where blood-based, AI-driven early detection is heading: toward the cancers where the screening gap, not the science, is the bottleneck.

Craif Raises $33M to Bring an AI Urine Test for Cancer to the US 🔗
Craif, a Japanese bio-AI company, closed an approximately 33 million dollar Series D, taking total funding to around 88 million dollars. Its platform uses AI to analyse microRNA in urine for early detection of cancer and other diseases. The round, co-led by Granite-Integral and diagnostics maker Tauns, funds a reimbursement-driven US launch and a Japanese regulatory filing for a software-based medical device programme targeting pancreatic cancer.

The signal: Capital keeps moving into non-invasive, AI-driven early detection, and it is widening beyond blood. Urine-based liquid biopsy is a harder sell on familiarity but an easier one on patient burden, and Craif is filing its diagnostic explicitly as a software medical device.

Bioptimus Gains a Landmark UK Colorectal Dataset to Train Its Biology Foundation Model 🔗

Bioptimus, the Paris-based AI company founded by a team from Google DeepMind and Owkin, signed a partnership with Cancer Research Horizons for access to the S:CORT colorectal cancer dataset: over 1,750 patient profiles spanning imaging, genomics, transcriptomics and clinical outcomes, with up to six modalities per patient. The data feeds STELA, Bioptimus' initiative to build the largest clinically linked multimodal tissue atlas, and is used to refine M-Optimus, its foundation model for biology. Bioptimus' models are already in use by 16 of the top 20 pharma companies.

The signal: In AI pathology the architecture is increasingly commoditised and the moat is data. Deeply phenotyped, outcome-linked, multimodal patient cohorts are rare, expensive and mostly locked inside institutions. When one of the most comprehensive colorectal datasets in Europe goes to a foundation-model company, the competitive question for every AI diagnostics vendor becomes less about the model and more about who they can get their data from.

Twelve US Health Systems Form a Consortium to Govern Diagnostic AI 🔗

Twelve US health systems, including Cedars-Sinai, Mount Sinai, Northwell and Houston Methodist, formed a Diagnostic AI Consortium with Aidoc. Together they care for nearly twenty million patients a year. The group will co-design AI-enabled diagnostic workflows, measure their effect on safety, quality and speed to diagnosis, and develop shared implementation and governance practices any health system can adopt. Aidoc provides the technical layer through its CARE foundation model and aiOS platform, already running in close to two thousand hospitals. First findings are expected in 2027.

The signal: This is the clearest sign yet that diagnostic AI has moved past the single-algorithm question. The buyers are organising, collectively, around how AI is evaluated and governed at enterprise scale, not whether one model is accurate in a controlled study. For vendors, the bar just moved from FDA clearance to demonstrable governance inside a health system. That is a harder and more durable requirement.