Variations in Surveillance Imaging for Patients with Treated Diffuse Large B-Cell Lymphoma in the National Comprehensive Cancer Network (NCCN) Lymphoma Database.

Utility and optimal timing of surveillance imaging for patients with treated diffuse large B-cell lymphoma (DLBCL) remains uncertain, especially in the era of positron emission tomography.

Data were obtained from the NCCN Non-Hodgkin's Lymphoma Outcomes Database, a prospective cohort study collecting comprehensive clinical, treatment, and outcome data for patients seen at 7 participating NCCN centers. Patients who presented between January 1, 2001 and June 30, 2007 with a new diagnosis of DLBCL in remission for at least three months after the completion of initial therapy were eligible for this study. For those patients with at least two years of follow-up, we assessed overall number and modality of imaging studies performed starting 90 days after the end date of their first-line therapy until documented relapse or end of two-year follow-up window. We then created three subcategories depending upon surveillance strategy: “PET” = positron emission tomography scans or combined positron emission tomography/limited computed tomography scans only; “CT”= computed tomography scans only; or “BOTH PET AND CT” = combination of both “PET” and “CT” at different times. For those who relapsed after being in primary remission for at least 90 days, we assessed the nature of the visit where relapse was first documented, recent accompanying imaging, and the presence of a confirmatory biopsy within one month before or after that relapse.

As of July 1, 2009, 433 newly-diagnosed DLBCL patients had at least two years of follow-up from 6 of the 7 participating NCCN centers; median age was 56 and 57% were male. Initial stage and International Prognostic Index (IPI) were as follows: I = 24%; II= 24%; III= 16%; IV=36%; and low= 47%; low-intermediate= 26%; high-intermediate= 19% and high= 8%. For surveillance imaging as defined above, 14% received PET, 48% received CT, and 38% received BOTH PET AND CT. Range of median frequencies across NCCN centers are reported below, both overall and by test type. Overall median number of tests received per year differed significantly across NCCN centers (p for Kruskal-Wallis < 0.0001).

Of 990 patients in primary remission for at least 90 days, there were 108 (11%) documented relapses. Of those, 20% were found at non-routine visits with symptoms, 41% at routine visits with symptoms, 37% at routine visits without symptoms and 2% were unclear. Of those found at routine visits without symptoms, 73% had undergone imaging within one month before and up to the date of relapse. Finally, 77% of those with relapse underwent confirmatory biopsy within one month before or after relapse.

The frequency of surveillance imaging for patients with treated DLBCL is highly variable, even within the context of tertiary care academic centers. Despite controversy surrounding its use, a high proportion of DLBCL patients in remission are followed with PET modalities. Those being followed exclusively with PET modalities tend to have less frequent imaging, which may have implications for comparative effectiveness. Routine imaging likely identifies relapse in about a quarter of patients with DLBCL; the impact on ultimate outcome remains to be further evaluated.

Friedberg: Genentech: Membership on an entity's Board of Directors or advisory committees. Blayney: American Society of Clinical Oncology: Membership on an entity's Board of Directors or advisory committees; BlueCross Blue Shield of Michigan: Research Funding; NCCN: Honoraria; University of Michigan Health System: Employment.