REMOD, a new AI model, detects pancreatic cancer on CT scans much earlier and more accurately than radiologists.
Key Details
- 1The AI model, REDMOD, was trained to identify radiomic signatures of pre-diagnostic pancreatic ductal adenocarcinoma (PDAC) on CT scans.
- 2It was trained with imaging data from multiple institutions and developed a 40-feature radiomic signature.
- 3Tested on over 200 abdominal CT scans from patients later diagnosed with PDAC despite no initial radiologist findings.
- 4REMOD detected pancreatic cancer on average 475 days before clinical diagnosis.
- 5The model achieved nearly 75% sensitivity, about twice as high as radiologists' sensitivity.
- 6At a 24-month prediction window, REDMOD's sensitivity was triple that of radiologists.
Why It Matters
Early detection of pancreatic cancer is vital due to its poor prognosis when caught late. AI advancements like REDMOD could revolutionize detection efficacy, potentially improving survival rates and demonstrating the impact of imaging AI in clinical practice.
Source
Radiology Business
Related News
•AuntMinnie
MRI-Based AI Model Accurately Subtypes Colorectal Cancer
A machine-learning MRI radiomics model shows high accuracy for preoperatively subtyping colorectal cancer, potentially enabling improved treatment planning.
•AuntMinnie
Study Reveals Large Performance Gaps Among AI Tools for Lung Cancer Detection
Seven commercial AI devices for lung cancer detection on chest x-rays show substantial and clinically meaningful performance variability.
•Radiology Business
Radiologist-AI Collaboration Boosts Pulmonary Embolism Detection on CTPA
Combining AI with radiologists substantially improves detection of pulmonary embolism (PE) on CT pulmonary angiography (CTPA) scans, new research shows.