Scientific Session 17 — Gastrointestinal ImagingWednesday, April 17, 2013
110. Predicting Survival in Metastatic Melanoma Patients on Bevacizumab Using MASS Criteria and Baseline LDH: Evaluation of a Prospective Phase II Trial
Smith A1, Gray M1*, Martin Del Campo S2, Garner K1, Zhang H1, Elci O1, Zhang X1, Carson W2 1. University of Mississippi Medical Center, Jackson, MS; 2. The Ohio State University, Columbus, OH
Address correspondence to A. Smith (firstname.lastname@example.org)
Objective: Metastatic melanoma survival is devastatingly low, and a predictive biomarker could be used to guide therapy, reduce unnecessary drug-related toxicity, and improve patient counseling. Our objective is to identify clinical and CT parameters that are predictive of progression-free survival (PFS) and overall survival (OS) in metastatic melanoma patients on bevacizumab therapy.
Materials and Methods: An institutional review board–approved HIPAA-compliant secondary analysis of CT and clinical data from 46 patients on bevacizumab therapy from a prospective phase II trial was performed. Up to five target lesions were identified and objective imaging response according to Response Evaluation Criteria in Solid Tumors 1.1, Choi criteria, modified Choi criteria, and Morphology, Attenuation, Size, and Structure (MASS) criteria were made on the first posttherapy CT study compared with the baseline images by reviewers blinded to survival outcome data. Cox proportional hazards models were used to identify associations of survival with baseline clinical variables (gender, age, Eastern Cooperative Oncology Group performance, time from diagnosis, number and location of metastatic sites, lactate dehydrogenase [LDH], hemoglobin, platelets, absolute neutrophils, corrected calcium, and alkaline phosphatase) and imaging findings on the first posttherapy CT study (absolute and percent change in target lesion size and attenuation and objective imaging responses). Log-rank test and receiver operator characteristics (ROC) curve with area under the curve (AUC) were used to evaluate predictive accuracy.
Results: Median PFS and OS for the entire cohort (n = 46) were 4 months and 14 months, respectively. In multivariate analysis, MASS criteria response on the first posttherapy CT strongly predicts PFS (p < 0.001) and OS (p < 0.001). Favorable response and unfavorable response by MASS criteria were significantly different from indeterminate response for predicting PFS (hazards ratio [HR], 0.34, p = 0.032; HR, 5.73, p < 0.001). High baseline LDH is associated with decreased PFS (HR = 1.32 for each increment of 100, p = 0.001) and OS (HR = 1.42 for each increment of 100, p = 0.001). Other clinical variables were not associated with survival. ROC analysis verified that an index combining baseline LDH and MASS criteria response on the first posttherapy CT has a high degree of accuracy in predicting PFS of more than 9 months (AUC = 0.97) and OS of more than 2 years (AUC = 0.87).
Conclusion: An index combining serum LDH and MASS criteria imaging response on the first posttherapy CT study has high accuracy for predicting PFS and OS in patients with metastatic melanoma who are undergoing bevacizumab therapy.