The impact of large lesions on the RECIST assessment criteria

Abstract

17560 Background: Since 2000 the Response Evaluation Criteria in Solid Tumors (RECIST) has been the standard measurement criteria for calculating tumor response. Although RECIST includes a minimun target lesion size to minimize errors in estimating the change in size of small lesions, RECIST does not provide a maximum size for target lesion selection. Previous research has shown that a tumor's rate of growth decreases at is approaches maximum volume. Furthermore, large lesions have the ability to greatly impact the overall assessment, which could result in the misclassification of response. The purpose of this study was to determine if the inclusion of a maximum target lesion size in the RECIST guidelines would significantly impact the response assessment. Methods: We reviewed 29 patients from the Tumor Imaging Metrics Core (TIMC) database who had 1 or more lesions > 7 cm at baseline and had a minimum of 3 time points. Previous research has shown that as a tumor approaches its maximum size its growth rate decreases. Therefore, we chose a diameter of 7 cm to study the impact of large lesions on response assessment. In order to test our hypothesis, were reviewed all scans using both standard and modified RECIST guidelines. The modified criteria included a maximum lesion diameter of 7 cm for selecting target lesions at baseline. We tracked tumor response using both criteria and determined if there was a difference in response assessment. Results: For standard RECIST the mean lesion diameter was 4.5 cm (n = 150) versus 2.5 cm (n = 121) for modified RECIST. Overall, response rates were significantly different when comparing the two criteria (p=.019). Additionally, according to the modified RECIST criteria, 20.7% of the patients would have had partial response (n = 3) or progressive disease (n = 3) one time point earlier. Conclusions: Our results suggest that the inclusion of a maximum lesion size in baseline target selection criteria has the ability to reduce the impact of a single large lesion on the overall response. Furthermore, in some cases, the implementation of a maximum target lesion size will improve RECIST's diagnostic ability. No significant financial relationships to disclose.