Role of Positron Emission Tomography in Lymphoma

Abstract

Positron emission tomography (PET) scanning, particularly using [F]fluorodeoxyglucose (FDG), has recently emerged as a powerful functional imaging tool for assessment of patients with Hodgkin’s disease and nonHodgkin’s lymphoma (NHL). Numerous investigations have been reported on its use for staging, restaging, and monitoring tumor response of lymphoma, with only a few exploring its value in predicting the tumor’s malignancy grade. In this issue of the Journal of Clinical Oncology, Schoder et al seek to determine whether the intensity of FDG uptake measured using the standardized uptake value (SUV) could differentiate between indolent and aggressive NHL. The SUV is a semiquantitative measure of the degree of FDG uptake, which may be conceptualized as the ratio between the tumoral concentration of tracer and its concentration in the entire body if the tracer were evenly distributed throughout. Because of its simplicity, this parameter is widely utilized both in the research setting and in clinical practice. In their study of 97 patients with NHL, Schoder et al concluded that the SUV is lower in indolent than in aggressive NHL, that patients with an SUV 10 have a high likelihood of aggressive NHL, and that this information may be helpful if there is discordance between biopsy and clinical behavior. These findings clearly have implications in treatment planning and prognosis. While the study by Schoder et al is the largest to date to attempt to distinguish between indolent and aggressive NHLs based on the SUV, the findings are not very different from those of previously reported smaller studies. Similar to those previous studies, a considerable overlap in the SUVs was found between indolent and aggressive NHL, with a relatively wide range of SUVs observed even within the same histologic subtype. This variability could have been even greater had the authors determined the SUVs for all well-delineated tumor lesions in each patient rather than only for the most intense lesion because of the known intrapatient variability in SUV in patients with NHL and other cancer types. Despite this overlap, two conclusions may be drawn reasonably from the data. First, an SUV 13 in the most intense lesion indicates a high likelihood of aggressive histology, while an SUV of 6 is very likely associated with indolent histology. Second, these upper and lower cutoff SUVs provide relatively high accuracy in grading NHL in only approximately 55% of patients, because only approximately 58% of patients with biopsy-proven aggressive histology had an SUV 13, and slightly less than 50% of those with indolent histology had an SUV 6. Consequently, approximately 45% of patients remain in a gray zone between the SUVs of 6 and 13, where the overlap is so pronounced that it practically precludes a justifiable degree of confidence in grading NHL. The adverse effect of this overlap becomes apparent when examining the results of the receiver operating characteristic (ROC) analysis used to determine optimum cutoff SUV that distinguishes indolent from aggressive histologies: this optimum cutoff SUV of 10, although associated with a fair balance of sensitivity and specificity, resulted in a 29% misclassification rate for aggressive NHL and a 19% misclassification rate for indolent NHL. The latter may result in overtreatment of a significant fraction of patients with indolent lymphoma. Conversely, misclassification of almost one third of patients with aggressive lymphoma may result in their undertreatment, which cannot be justified considering the potentially curable nature of their disease. The moderate sensitivity and specificity of the cutoff SUV of 10 needs to be further evaluated in light of the fact that this value was derived post hoc, potentially resulting in an overestimation of its diagnostic accuracy. However, the findings in the 22 patients in whom the biopsy was taken from a site other than that with the most intense FDG uptake could serve as an initial test for the performance of this cutoff in an independent, albeit small, cohort of JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 23 NUMBER 21 JULY 2