Comparison Between Different PET and CT-Based Imaging Interpretation Criteria at Interim Imaging in Patients With Diffuse Large B-Cell Lymphoma.

Abstract

To evaluate the predictive value of interim PET (iPET) in diffuse large B-cell lymphoma (DLBCL) using 5 different imaging interpretation criteria: Deauville 5-point scale criteria, International Harmonization Project (IHP) criteria, Response Evaluation Criteria In Solid Tumors (RECIST) 1.1, European Organization for Research and Treatment of Cancer, and PET Response Criteria in Solid Tumors (PERCIST) 1.0. We retrospectively reviewed records from 38 patients with DLBCL who underwent baseline and iPET at our institution. Imaging was interpreted according to the previously mentioned criteria. Results were correlated with end-of-treatment response, based on reports at the end of treatment radiological examinations, overall survival (OS), and progression-free survival (PFS) to assess and compare the predictive value of iPET according to each criterion. We also evaluated the concordance between different criteria. The Deauville and PERCIST criteria were the most reliable for predicting end-of-treatment response, reporting an accuracy of 81.6%. They also correlated with OS and PFS (P = 0.0004 and P = 0.0001, and P = 0.0007 and P = 0.0002, for Deauville and PERCIST, respectively). Interim PET according to European Organization for Research and Treatment of Cancer also predicted the end-of-treatment response with an accuracy of 73.7% and had a significant correlation with OS (P = 0.007) and PFS (P = 0.007). In contrast, the IHP criteria and RECIST did not predict outcomes: the accuracy for end-of-treatment response was 34.2% and 36.8%, respectively, with no significant correlation with OS or PFS (P = 0.182 and P = 0.357, and P = 0.341 and P = 0.215, for OS and PFS, respectively). The predictive value of iPET in DLBCL patients is most reliable using the Deauville and PERCIST criteria. Criteria that rely on anatomical characteristics, namely, RECIST and IHP criteria, are less accurate in predicting patient outcomes in DLBCL.