Healthcare AI Daily Briefing

Industry & Products

Clinical knowledge-augmented AI boosts AML data abstraction accuracy to over 95%.

In a UCHealth and Verily pilot using healthcare-native AI architecture, extracting myeloblast percentage accuracy rose from 72% with standard zero-shot models to more than 95% when grounded in clinical ontologies. The pilot projected that an estimated 1,200 hours of manual abstraction could be reduced to just 40 hours for acute myeloid leukemia variable extraction. [4]

Medical Imaging

Echocardiographic RV-PA coupling enhances PAH risk scores.

Journal article

In 401 patients with prevalent pulmonary arterial hypertension, adding echocardiographic variables to ESC/ERS and REVEAL 2.0 scores improved c-index for morbi-mortality prediction, with machine learning-derived RV-PA coupling showing the greatest influence. [1]

Diagnostics & Pathology

AI-Gemini detects oral lesions with 97.1% accuracy.

Journal article

In a prospective multicenter study of 350 patients across twenty university and tertiary head and neck centers in Italy, Belgium, France, Spain, and Israel, AI-Gemini using Gemini 2.5 Advanced achieved 97.1% accuracy for lesion detection and malignancy classification compared with histopathology. [2]

Industry & Products

UCHealth selected Abridge vendor for AI scribe due to EHR integration and validated performance.

Podcast from Healthcare AI Pioneers

The interview describes UCHealth's process of evaluating several AI scribe products, focusing on EHR integration and validation, ultimately choosing Abridge as their vendor because it integrates with their EHR and offers a well-validated product. [5]

Policy & Governance

Clear AI governance vital for clinical integration.

Podcast from The Signal Room

The interview describes how really clear AI governance, especially around clinical pieces, is vitally important in a healthcare system to bring technology into the organization and manage change. [6]

Operations & Workflow

AI CDI tools drive 22-percentage-point chart completion increase at Intermountain.

Intermountain Health reported a 22-percentage-point increase in inpatient chart completion during or immediately after rounds after implementing AI clinical documentation improvement tools, alongside an 11.5-point rise in Net EHR Experience Score. The organization retained physicians for higher-level clinical judgment while automating routine documentation tasks. [3]