In the past few years, the lymphoma literature has featured an increasing number of studies — some well-designed, most not so — designed to investigate the potential for “response-adapted” therapy.
Many of these studies have explored the use of functional imaging, especially with 18FDG-PET, either during or after therapy. The conclusions of these studies have been inconsistent and confusing. Recent reports from the ASCO Annual Meeting only added to that confusion.
The rapid adoption of PET scanning into routine clinical care and its use to guide treatment decisions in the absence of solid, prospective data has had many consequences. I have seen examples of patients who have been switched to intensive salvage regimens for Hodgkin’s lymphoma or diffuse large B-cell lymphoma (DLBCL) on the basis of an “interim” PET scan thought to show a suboptimal response. I have been aware of patients embarking on salvage regimens for DLBCL on the basis of a “positive” PET performed within 2 to 3 weeks of the completion of an induction regimen, and of patients diagnosed with relapsed disease — without biopsy confirmation — on the basis of a positive PET scan performed as part of post-treatment surveillance imaging.
Current data do not support any of these approaches. The overuse of functional imaging has serious implications for patients. Besides the potential for overtreatment, there also is the added anxiety of a false-positive scan, the need for unnecessary and invasive biopsies to evaluate sites that are positive on PET scanning, and the added expense to the patient — and to our health care system — from unnecessary scans and subsequent follow-up evaluations.
So, do we have enough confidence in current data to introduce response-adapted therapy, based on functional imaging, into routine practice? Not yet.
Role of CT/PET
The role of FDG-PET for staging and end-of-treatment response assessment of patients with DLBCL and Hodgkin’s lymphoma is well established. It has been shown to be the most sensitive and specific imaging test available for staging most types of lymphoma — including indolent disease, such as follicular lymphoma — and recent data suggest that routine use of bone marrow biopsies probably can be abandoned for certain types of lymphoma based on data available from CT/PET.
Several studies have shown that results of the end-of-treatment PET scan are highly predictive of subsequent relapse and survival. For example, the technique has been shown to have a very high negative predictive value of 85% to 100% in DLBCL, although the positive predictive value is somewhat lower and more variable (from 40% to 90%). Similar data have been reported in Hodgkin’s lymphoma, and there are emerging data in follicular lymphoma demonstrating correlation between end-of-treatment CT/PET and outcome.
In a study presented at ASCO, Judith Trotman, MBChB, and colleagues analyzed outcomes for patients with follicular lymphoma treated on three large randomized trials in Europe — PRIMA, PET Folliculaire and FOLL05. The analysis included data from 246 scans performed at the end of therapy and assessed centrally by an expert review panel using the Deauville criteria, a standardized 5-point scale.
PET status was shown to be highly predictive of subsequent PFS and OS using cut-off scores of 3 or 4. For example, among those with a cut-off score of 4, the 4-year PFS was 23.2% for PET-positive patients vs. 63.4% for PET-negative patients. The corresponding 4-year OS rates were 87.2% vs. 97.1%. These results were much more strongly predictive than conventional CT. The researchers concluded that CT/PET should be adopted as the standard approach to response assessment in follicular lymphoma.
These are informative data — particularly for the conduct of clinical trials — suggesting that early assessment of post-induction response by PET may serve as a useful surrogate for PFS or OS and provide an early read in patients with a very indolent disease, where survival differences may take several years to emerge.
What do these results mean for routine clinical practice? Probably not much. There are no data to suggest that patients have improved PFS or OS because they had a CT/PET rather than CT assessment of response, and it is also important to consider the context of this study. The scale is not in widespread use and is poorly understood by many oncologists, and the inter-observer variability in interpreting this test outside a clinical trial setting is unknown.
Overall, these data tell us that CT/PET probably is a better biomarker of response than CT alone. Whether we can use those data clinically — especially in an indolent disease with multiple second-, third- and fourth-line treatment options — is unknown. At a time when comparative effectiveness research is growing in importance, no available data suggest that CT/PET has affected patient outcomes.
Even less certain is the role of so-called “interim” or “early” functional imaging. Several studies in Hodgkin’s lymphoma and DLBCL have investigated the role of functional imaging — typically after two or four cycles of induction chemotherapy — in predicting subsequent PFS and OS. In general, these studies have demonstrated that negative “early” PET predicts improved PFS and OS compared with a positive PET and have raised the possibility that subsequent therapy can be adapted based on this result, reducing the potential for over-treating PET-negative patients who may already be cured and exploring alternative strategies for those who are PET-positive and less likely to be cured without a change in therapy.
This concept is being explored in several well-designed prospective studies in Hodgkin’s lymphoma and DLBCL, addressing the use of PET scanning for assigning patients to consolidation strategies such as involved field radiation therapy or early consideration of intensified chemotherapy. A randomized phase 2 study from the GELA/LYSA groups in France, presented at ASCO, attempted to address this issue. The study design was complex, based upon results of PET scans performed after two and four cycles of R-CHOP or rituximab, doxorubicin, cyclophosphamide, vindesine, bleomycin and prednisone (R-ACVBP). It also was based on the assumption that R-ACVBP plus autologous stem cell transplantation should be the standard of care for poor-risk DLBCL and that anything less than this is “backing off” on treatment.
In this study, even those patients who were PET negative after two cycles of therapy went on to consolidation treatment but not autologous stem cell transplant (ASCT). From this difficult-to-interpret study, using only visual criteria (not the 5-point scale) for PET interpretation, the researchers concluded that — based on negative PET after two and four cycles of therapy — “25% of patients do not need ASCT.” Many of the assumptions in the design of this study are flawed, particularly the use of consolidative ASCT as a “standard” therapy, but this demonstrates how poorly conceived trials are confusing this issue with no suggestion of patient benefit. The 4-year PFS and OS in this study were 75% and 83%, respectively — pretty much what you would expect in advanced DLBCL patients receiving R-CHOP.
A third study presented at ASCO this year explored the role of single-agent brentuximab vedotin (Adcetris, Seattle Genetics) prior to doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) in patients with untreated, limited-stage Hodgkin’s lymphoma to assess its single-agent activity. CT/PET response to two cycles of brentuximab vedotin served as the primary endpoint. Ten of the 12 patients achieved complete metabolic response after two cycles of brentuximab vedotin.
Although not a criticism of this specific study, this highlights another important question about the use of PET-based response-adapted therapy: The predictive value of PET may be related to the specific therapy. Assumptions that all PET-negative responses are the same are probably wrong. Some data in DLBCL suggest PET response may be less predictive of outcome for patients who receive infusional vs. other immunochemotherapy regimens.
Prospective data needed
So what do we really know about response-adapted therapy using PET in lymphoma? We know it is a useful biomarker and that it is better to be PET negative than PET positive at any time during or immediately after therapy. We don’t know much about the reproducibility of PET interpretation outside well-controlled trial settings, and we have no prospective data demonstrating that changing therapy based on PET results improves outcome. Several important, well-designed clinical trials are either recently completed or in progress. We need to wait for mature results of these studies and, in the meantime, be very circumspect in our use of PET scanning as anything other than a prognostic biomarker.
Is there anything out there other than functional imaging showing promise as a marker for response-adapted therapy? A study from Stanford University, presented at ASCO, described the use of acellular tumor DNA from immunoglobulin heavy chain genes from the blood of patients with DLBCL — a “liquid biopsy” method with potential clinical utility in assessing response. These are preliminary data but offer the potential for response assessment free of subjectivity — and hopefully with a reliable, quantitative read out. Until such tests are available, functional imaging may be the best we have, but we need to recognize its limitations and avoid changes in treatment that have no proven benefit to our patients.
Cheson BD. J Clin Oncol. 2011;29:1844-1854.
Cheson BD. Oncology (Williston Park). 2013;27:1289-1292.
Safar V. J Clin Oncol. 2012;30:184-190.
The following were presented at the ASCO Annual Meeting; May 30-June 3, 2014; Chicago:
Casasnovas RO. Abstract #8503.
Federico M. Abstract #8507.
Kurtz DM. Abstract #8504.
Trotman J. Abstract #8502.
For more information:
John Sweetenham, MD, FRCP, FACP, is HemOnc Today’s Chief Medical Editor, Hematology. He can be reached at email@example.com.
Disclosure: Sweetenham reports no relevant financial disclosures.