Healthcare AI Daily Briefing

Medical Imaging

CIPHER shows promise for ICI-P risk assessment but needs prospective validation.

Journal article

CIPHER, a CT-based deep learning foundation model, was pretrained on 590,284 CT slices from 2,500 NSCLC patients and fine-tuned on an internal MD Anderson cohort of 347 patients (33 ICI-P cases). In head-to-head benchmarking, it achieved an AUC of 0.83, outperforming clinical, radiomics and ensemble models. The model requires prospective validation before clinical use. [3]

Operations & Workflow

Trust key for PerfectServe to grow and retain customers in consolidating niche.

In the niche that PerfectServe occupies—which Castel says is consolidating—trust is a key trait to grow and to retain customers, and any AI they use must be able to show its reasoning to maintain that trust. [6]

Diagnostics & Pathology

TRIGS and SAM-TIL matched gold standard in pCR and event-free survival.

Preprint

In the PARTNER randomized trial with 285 patients, TRIGS and foundation model-based SAM-TIL showed no substantial difference from gold standard clinical assessment in predicting pathological complete response (AUC 0.60-0.64) or event free survival (Integrated Brier score approximately 0.10). [2]

Industry & Products

Aidoc gets third FDA Breakthrough Device Designation for CARE CXR Triage.

Aidoc received its third FDA Breakthrough Device Designation in just over a year for CARE Chest X-Ray Triage, which prioritizes pneumomediastinum and lobar/lung collapse. The solution is pending 510(k) clearance and is not yet available for sale in the United States. [4]

Industry & Products

Over 900 clinicians using ambient AI platform at oncology network.

Podcast from Beyond the Chart: Exploring the AI Frontier for Oncology

The interview describes over nine hundred clinicians actively using an ambient AI platform at one oncology network, calling it 'a new muscle' that the technology company and practices had to learn to deploy. [1]

Industry & Products

Enveda raises $311M Series E at $2B valuation to advance AI-discovered nature drugs into clinical trials.

Enveda secured $311 million in Series E funding at a $2 billion valuation, doubling its value from a year ago. The biotech startup uses AI to discover drugs from plants and microbes, advancing candidates into human clinical trials for severe skin conditions and weight-loss maintenance after GLP-1s, though no AI-discovered drug has yet received FDA approval. [5]