AI Predicts Low Risk DCIS Features

Researchers developed a deep learning pipeline to predict key biomarkers in ductal carcinoma in situ (DCIS) from digitized pathology slides.
The study focused on oestrogen receptor-positive, human epidermal growth factor receptor 2-negative, grade 1 or 2 DCIS diagnosed through screening.
Artificial intelligence models were trained and tested on data from 887 patients in a Dutch multicentre dataset, followed by external validation on 259 patients in the UK.
Deep learning models accurately predicted tumour grade, oestrogen receptor status, and human epidermal growth factor receptor 2 status directly from pathology images.
The artificial intelligence methods reliably stratified patients according to their eligibility for active surveillance, helping to identify those unlikely to require intensive treatment.
The findings suggest that these AI-assisted pathology tools have the potential to reduce overtreatment in patients with DCIS by guiding less aggressive management strategies.

You may have missed