Predicting primary treatment failure using interim FDG-PET scanning in diffuse large B-cell lymphoma.

Interim FDG-PET (iPET) in diffuse large B-cell lymphoma (DLBCL) is increasingly practised and used in clinical trials to adapt further therapy. However, the optimum timing and methodology of iPET remains controversial. We retrospectively analysed the iPET results and outcomes of 200 DLBCL patients where FDG-PET was routinely performed at baseline, after 2 cycles (iPET2) and at completion of chemoimmunotherapy. iPET was also performed after 4 cycles (iPET4) where at iPET2, Deauville score (DS) was ≥4. Scans were assessed by blinded expert lymphoma PET physicians for DS, maximum standard uptake value (SUVmax), total metabolic tumour volume (TMTV) and total lesion glycolysis (TLG). Treatment failure was defined as death, progression or refractory disease. 95.5% of patients received R-CHOP. No baseline PET parameter was predicted for EFS or OS independent of the NCCN-IPI. The multivariable analysis at iPET2 showed DS5 (19.5% of cases) predicted treatment failure (HR 6.29, 95% CI 3.01-13.17, P < .001), but DS4 was equivalent to DS1-3. At iPET4, ΔSUVmax < 66% predicted treatment failure (HR 5.49, 95% CI 3.03-9.99, P < .001). By multivariable analysis of all time points, high NCCN-IPI and DS5 at iPET2 were negative predictors of survival. These findings were independent of novel prognostic markers.