Background

FDG PET/CT (PET) is commonly used in the management of patients with lymphomas and is recommended for both initial staging and response assessment after treatment in patients with diffuse large B-cell lymphoma and Hodgkin lymphoma. In follicular lymphoma (FL), PET proved to be more accurate than conventional imaging for initial staging and for restaging and a positive PET after induction treatment (piPET) has been recently confirmed as a negative prognostic factor for progression Free Survival (PFS). So far however a clear definition and standardization of how the assessment of piPET must be done is lacking. In this work we tested the 5 point Deauville scale (5PS), applied by an expert panel to piPET to determine if it could be a reliable and accurate tool to support the prognostic role of piPET in patients with FL.

Patients and Methods

This study was conducted on patients with advanced stage untreated FL enrolled in the FOLL05 randomized trial for whom pi- and baseline PET were available in DICOM format. Scans were then interpreted using visual assessment with the Deauville 5-point scale by an international expert panel (EP). Two EP readers scored piPET independently and a third one was asked to resolve the discrepancies. Binary (+ve vs –ve) and overall concordance rates were calculated using k Cohen’s and Fleiss’s kappa coefficients, respectively using the scores obtained on independent reading. Central review results were correlated with patients features, and with patients outcome using PFS as principal endpoint.

Results

Among 534 patients randomized to the FOLL05 trial we could retrieve a digital PET-DICOM file in 121 patients; 31 patients were excluded for poor quality of the images with a total of 90 pts available for the central review. Median age was 54 (range 36 to 75), 80 (89%) had stage III-IV, 33.3% had high risk FLIPI (score 3-5). After review two different cutoffs were used to classify piPET as +ve or –ve. Using a cut off at 3, rate of +ve piPET was (18%) with a low Cohen’s K index of 0.23. Using a cutoff at 4 +ve piPET rate was 11% but Cohen’s K index improved to 0.62. Using the cut-off at 3 and at 4 concordance rate was 81%, and 94.5% respectively. When 5PS results were correlated with patients features no correlation was shown. Overall at time of current analysis and with a median follow-up of 25.9 months 56 pts (62%) were in continuous CR and 34 (38%) had progressed or relapsed. Overall Results of the 5PS were confirmed to be predictive of PFS (Log rank = 9, P<0.05). The 3-year PFS of piPET -ve and piPET +ve pts with score 3-5 as positive was 38% and 72%, respectively; PPV and NPV were 0.29 and 0.68, respectively. The 3-year PFS of piPET -ve and piPET +ve pts with score 4-5 as positive was 20% and 71%, respectively; PPV and NPV were 0.24 and 0.68, respectively. Looking at single 5PS scores the best discrimination for PFS was observed for scores 4 and 5.

Conclusions

The central review of FDG-PET scans using the 5PS confirmed that a qualitative assessment of response is feasible also for patients with FL. With the qualitative assessment of response the prognostic role of piPET had the best ratio between reproducibility and clinical informative power using a cut off at 4. These results confirm piPET-response as relevant prognostic factor for patients with FL, and warrants the adoption of the 5PS in future studies.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution