Mia Levy MD, Presenter: Nothing to Disclose

Daniel L. Rubin MD, Abstract Co-Author: Nothing to Disclose

Objective criteria such as Response Evaluation Criteria in Solid Tumors (RECIST) are becoming crucial for assessing the response to treatment. However, applying such criteria could be thwarted by variability in reporting the required data. Our goal was to determine whether the information routinely recorded in radiology reports and image annotations is sufficient for reliably applying RECIST.

Reports and images were obtained of 42 CT scans (13 baseline, 29 follow-up) of the chest, abdomen and pelvis from 13 patients enrolled in clinical trials at two geographically-separate cancer centers where RECIST is used to evaluate treatment response. Radiology reports and image annotations were evaluated. The clinical flow sheets created by oncologists were used as the source of information to determine the gold standard for this study (identity of “target lesions” and measurements of the longest diameter of each target lesion). Radiology reports and CT image annotations created by the interpreting radiologist were evaluated to identify the target lesions and record their measurements. We evaluated whether the radiology information was sufficient and consistent (in baseline and follow up studies) for applying RECIST criteria to evaluate response to treatment.

The 42 CT scans were interpreted by 16 different radiologists. On baseline scans, 71% of the target lesions in the gold standard were identified in radiology reports and 73% in image annotations, while the longest diameter was reported 55% of the time and annotated 50% of the time. On follow-up scans, 38% of target lesions were identified in reports and 70% by image annotation, while the longest diameter was reported only 28% of the time and annotated in images only 26% of the time. In only 26% of the studies were the report and image annotation data sufficient to apply RECIST.

The radiology reporting and image annotations examined were insufficient in most cases for practitioners to reliably apply RECIST criteria. Our results illustrate the need for new training techniques and informatics tools to improve radiologist-oncologist communication and to enable more effective use of image-based quantitative evaluation criteria.

Current radiology reporting focuses on summarizing findings, which is insufficient for consistent application of quantitative criteria to evaluate treatment response, such as RECIST.

Levy, M, Rubin, D, Is Current Radiology Reporting Sufficient for Evaluating the Clinical Response to Treatment? An Analysis of Information Content in Radiology Images and Reports.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6013650.html