AI cuts lesion measurement time by 34%
Automated lesion measurements increased agreement on treatment response, although expert-generated proposals were accepted more often and required fewer adjustments.
By
Andrea Surnit
September 3, 2026
Clinical Scorecard: AI cuts lesion measurement time by 34%
At a Glance
Category Detail
Condition Lesion measurement in cancer follow-up Key Mechanisms Artificial intelligence assistance reduces reading time and increases classification agreement among readers. Target Population Patients with cancer undergoing follow-up CT examinations Care Setting Radiology departments using computed tomography
Key Highlights
AI assistance reduced reading time by 34% compared to unassisted assessment. Interobserver agreement on RECIST response classification improved with AI assistance. AI-assisted measurements showed greater variability compared to unassisted measurements. Expert assistance provided the lowest measurement error. Readers accepted 61% of AI-generated proposals with modifications. Guideline-Based Recommendations
Diagnosis
Use AI-assisted measurements for faster lesion assessment. Management
Consider expert assistance for higher accuracy in lesion measurements. Monitoring & Follow-up
Monitor interobserver variability when using AI-assisted assessments. Risks
AI assistance may introduce greater variability in individual lesion measurements. Patient & Prescribing Data
Patients with predefined target lesions undergoing follow-up CT.
AI assistance may improve workflow but requires careful consideration of measurement variability.
Clinical Best Practices
Utilize AI assistance to enhance efficiency in RECIST assessments. Incorporate expert review to minimize measurement errors. Related Resources & Content
