New research presented at the European Congress of Radiology (ECR) 2026 indicates that AI-assisted CT scans can streamline cancer follow-up imaging. The technology reduces reading time while improving consistency in tumor measurement, potentially standardizing radiology assessments and minimizing disagreements in treatment response evaluation.
Accurate tumor measurement is critical for evaluating treatment response. While radiologists use RECIST 1.1 criteria, reader variability can impact clinical decisions. AI-assisted CT tools automatically identify and measure target lesions, reducing workload and subjective differences.
A retrospective multi-center study involving 23 readers evaluated follow-up CT scans from 212 oncology patients. AI-assisted CT significantly reduced reading time per patient by over 35 seconds. At the patient level, clinically meaningful disagreement between radiologists, defined as a 20% difference in the Sum of Longest Diameters, decreased from 43.4% in unassisted cases to 28.3% with AI assistance.
These findings suggest AI-assisted CT can reduce clinically significant disagreements in assessing treatment response, improving overall consistency and efficiency in radiology workflows. Further validation across diverse imaging systems and patient populations is anticipated.