Comparison of four different imaging response criteria in patients with Hodgkin and non-Hodgkin lymphoma using PET/CT

Objectives Various criteria are available to assess treatment response in oncology patients using CT and/or metabolic information from PET. Four such criteria are commonly used, but there is limited understanding of the inter-criteria variability. We hypothesized that PET/CT evaluation of treatment response may differ between these criteria.

Methods Forty pre and post treatment PET/CT images of 20 patients with Hodgkin (HL) and Non-Hodgkin lymphoma (NHL) seen at Stanford Medical Center from 2006 to 2010 were reviewed by GD. Images were classified as complete response (CR), partial response (PR), stable disease (SD) and disease progress (DP) according to four response criteria: PERCIST 1.01, IWG2, EORTC3, and RECIST 1.14. Medical charts were reviewed to assess clinical response.

Results Of the 10 patients with HL there was only 40% congruence among the 4 response criteria, all classified as PR. EORTC, PERCIST and IWG had matching results in only in 1 patient classified as CR. Seventy percent of patients with HL were clinically classified as CR and 30% as PR. PERCIST was the most consistent with clinical outcomes in 70% of HL patients (Fig 1a). Of the 10 patients with NHL, there was congruence among the 4 response criteria in only 30%. Of these, 1 was classified as SD and the others were classified as PR. In addition, 40% of patients were classified as CR by EORTC, PERCIST and IWG. The remaining patients had mixed response by the 4 criteria. Sixty percent of patients with NHL were clinically classified as CR, 30% as PR and 10% as PD. PERCIST was the most consistent with clinical outcomes in 90% of patients (Fig 1b).

Conclusions There was low correlation among the 4 criteria used for treatment response evaluation in patients with HL and NHL using PET/CT. This may have significant implications in the management of these patients. In this cohort, PERCIST was most consistent with clinical outcomes in patients with HL and NHL. Further studies are needed to confirm these results in larger cohorts and in different malignancies