PET positive, PET negative, or PET peeve?

In this issue of Blood, Horning and colleagues report that there was disagreement by 3 nuclear medicine experts on interpretation of FDG-PET images obtained in the context of a clinical trial in aggressive non-Hodgkin lymphoma 32% of the time.¹  These results emphasize the need for standardized criteria to interpret interim PET scans in lymphoma and should cause physicians to question the practice of changing therapy based on PET imaging outside the context of a trial.

Positron emission tomography (PET) is a functional imaging technique that uses a glucose analog (2-fluoro-2-deoxy-D-glucose [FDG]) radiolabeled with the positron emitter fluorine-18 to evaluate glycolytic activity, which is increased in most histologies of lymphoma.^2,3  Several studies have suggested a role for FDG-PET in the diagnosis and follow-up of patients with lymphoma, and PET is now recommended as part of routine staging and assessment of response in curable lymphomas—particularly diffuse large B-cell lymphoma and Hodgkin lymphoma.⁴ 

More recently, there has been significant interest in performing “interim” PET scans after 2 to 3 cycles of chemotherapy as an early biomarker of resistant disease. A frequently cited trial extolling the benefits of early interim PET enrolled 260 patients with de novo Hodgkin lymphoma and performed a PET scan after 2 cycles of standard ABVD (combination of doxorubicin/bleomycin/vinblastine/dacarbazine) chemotherapy.⁵  No treatment change was permitted on the basis of the interim PET scan. Two-year progression-free survival for patients with positive interim PET (n = 50) was only 12% and for patients with negative interim PET exceeded 95%. Somewhat lost in these exciting results are details regarding interpretation of the PET scans. Two international expert readers were required to reach consensus for each positive scan. Moreover, lesions with “minimal residual uptake,” arbitrarily defined as a standardized uptake value between 2 and 3.5, were considered to be negative scans. These criteria have never been evaluated prospectively. Despite these limitations, as a result of this study, cooperative groups in both Europe and the United States are evaluating treatment algorithms that change therapy based on an interim positive PET scan. Many physicians have already adopted this practice. Indeed, a very common question in our consultation clinic is “what to do with an interim positive PET scan?”

For more than a decade, it has been clear that understanding of physiologic uptake and artifacts associated with FDG is critical to accurate interpretation of PET scans.⁶  False-positive scans can result from brown fat, rebound thymic uptake, and increased diffuse bone marrow and muscle uptake at the completion of therapy, which do not represent refractory disease. Indeed, preliminary results of a trial that incorporated biopsies of PET-positive sites following R-CHOP therapy (ri-tuximab plus cyclophosphamide/doxorubicin/vincristine/prednisone) for diffuse large B-cell lymphoma revealed that only a minority of the biopsies (4 of 36) were positive for lymphoma.⁷ 

However, for the clinician, binary criteria (positive or negative) are easiest to interpret. Two years ago, consensus criteria were developed for interpreting scans after completion of chemotherapy. Mediastinal blood pool activity was recommended as the reference background activity to define PET positivity in lymphoma, and specific recommendations were provided for in terpretation of extranodal sites.⁸  These international criteria were not developed for interim scans during therapy and have never rigorously evaluated prospectively.

The study by Horning et al reveals this major limitation in the use of interim PET scans as part of an algorithm to treat patients with diffuse large B-cell lymphoma.¹  The 3 nuclear medicine physicians involved in these reviews were experts; 2 of them co-authored the International Criteria for interpreting follow-up PET scans in lymphoma.⁸  Of 12 scans, one physician rated 2 scans as positive, the second rated 9 scans as positive, and the third rated 6 scans as positive. Because of nonspecific uptake, PET may have disadvantages in the evaluation of extranodal disease sites, and indeed extranodal disease proved most difficult to interpret in Horning and colleagues' study. Remarkably, after discussion of all scans, the experts were able to reach consensus on only 3 of them. If 3 experts are able to agree only two-thirds of the time, the lack of fidelity of PET interpretation will trump any impact of the alternative therapy being tested in a clinical trial.

The ECOG group is to be congratulated for this important contribution that clearly indicates it is time for a pause. Until criteria for interpretation of PET scans are prospectively validated, and the experts can routinely agree on what is PET positive and what is PET negative, physicians should not change therapy for lymphoma based on an interim PET in practice, and probably should not even routinely perform such scans. Cooperative study groups need to consider these results carefully and include prospective data evaluating agreement in PET interpretation criteria used for any studies incorporating interim PET imaging.