Response assessment in lymphoma: time to lower the bar?

Abstract

Positron emission tomography (PET) using 18 fl uoro-2deoxyglucose (FDG) has signifi cantly enhanced imaging of lymphoma in a variety of clinical settings, improving both the sensitivity and, in the appropriate setting, the specifi city of lymphoma detection. While the benefi t of PET scanning depends in part on the histology of the lymphoma in question, it is clear that FDG-PET improves both initial staging and response assessment in aggressive, potentially curable lymphomas, allowing optimization of therapy. Th e setting in which the use of FDG-PET scanning has shown the most utility is in assessment of response following completion of therapy. While in the past, response assessment was based on resolution of size abnormalities in aff ected lymph nodes and tissues as measured by computed tomography (CT), it was clear that residual anatomic abnormalities following lymphoma treatment could be related to scar tissue, and not to active lymphoma. FDG-PET, in measuring the metabolic activity of residual abnormalities, can predict the presence or absence of residual active lymphoma better than CT analysis of size alone. Juweid and colleagues [1] demonstrated that incorporating metabolic data derived from FDG-PET along with standard CT criteria led to better prediction of outcome with post-treatment response assessment than CT alone. Th e International Harmonization Project (IHP) subsequently defi ned criteria for “ positive ” versus “ negative ” FDG-PET scans following completion of therapy [2], and these defi nitions were incorporated into the revised Response Criteria for Lymphoma published by Cheson et al . in 2007 [3]. Specifi cally, the reference uptake intensity for comparison was defi ned to be that of the mediastinal blood pool (MBP), with areas of uptake greater than background or MBP (depending on size and location) defi ned as abnormal, and suggestive of residual lymphoma. While these criteria have been incorporated into posttreatment response defi nitions, it is clear that the use of MBP as a reference is not optimal in all circumstances. FDG-PET has been explored for early response assessment in both Hodgkin (HL) and non-Hodgkin lymphoma (NHL). While the jury is out regarding its utility in NHL, Gallamini and others have shown that PET does predict outcome when used early in therapy for HL [4]. When investigators have studied the optimal reference for defi ning a positive study in the interim setting, however, it is clear that the use of MBP is too restrictive, and that uptake between that of the MBP and liver, referred to as minimal residual uptake or MRU, is associated with outcomes comparable to those less than the MBP. Th erefore, in interim response assessment, liver uptake has been adopted as a more appropriate reference, as defi ned in the Deauville (also known as London) criteria [5]. While the IHP criteria have been incorporated into standard post-treatment imaging, it is clear that many patients with a “ positive ” PET scan will not relapse, and some that do relapse quite late, suggesting that the MBP cut-off for negativity may be too stringent. Th is question was explored in the study by Manohar and colleagues published in this issue. Th ey retrospectively reviewed post-treatment PET scans in a cohort of 69 patients with aggressive non-Hodgkin lymphoma, including diff use large B-cell lymphoma (DLBCL) and anaplastic large T-cell lymphoma (ALCL). Th ey then scored the scans as positive or negative based on three diff erent sets of criteria in order to compare accuracy. Th ese included the IHP criteria, London (Deauville) criteria and so-called “ Gallamini criteria, ” defi ned as standardized uptake value (SUV) 3.5, which was the reference used in the Gallamini study of interim PET in HL. Th ey found that all three sets of criteria had similar excellent negative predictive value (NPV), meaning that patients with a negative scan by any criteria had uniformly excellent outcomes. However, when compared to London or Gallamini criteria, the IHP response criteria had inferior positive predictive value (PPV), with more patients having “ positive ” scans by these rigorous criteria staying in long-term remissions. Th ese fi ndings suggest that, similar to interim response assessment, low grade residual uptake at the time of post-treatment evaluation is not predictive of ultimate relapse, and that the IHP criteria for a negative scan may be too stringent. Several caveats must be mentioned. First, this is a retrospective study in a limited data set. While the fi ndings are