Optimisation of metabolic criteria in the prognostic assessment in patients with lymphoma. A multicentre study.

Abstract

To compare sensitivity, specificity and predictive value of Deauville score (DS) vs. ΔSUVmax in interim-treatment PET (iPET) and end-treatment PET (ePET), in patients with diffuse large B cell lymphoma (DLBCL), Hodgkin lymphoma (HL), and follicular lymphoma (FL). Retrospective longitudinal multicentre study including 138 patients (46 DLBCL, 46 HL, 46 FL), on whom 3 18F-FDG PET/CT were performed: baseline, iPET, and ePET. Visual (DS) and semi-quantitative (ΔSUVmax) parameters were determined for iPET and ePET. Predictive value was determined in relation to disease-free interval. Statistical analysis. iPET for DLBCL, HL, and FL: 1) sensitivity of DS: 76.92/83.33/61.53%; specificity: 78.78/85/81.81%; 2) sensitivity of ΔSUVmax: 53.84/83.33/61.53%; specificity: 87.87/87.50/78.78%. ePET for DLBCL, HL and FL: 1) sensitivity of DS: 61.53/83.33/69.23%; specificity: 90.90/85/87.87%; 2) sensitivity of ΔSUVmax: 69.23/83.33/69.23%; specificity: 90.90/87.50/84.84%. Predictive assessment. iPET study: in DLBCL, DS resulted in 10.3% recurrence of negative iPET, and 17.1% in ΔSUVmax at disease-free interval; in HL, both parameters showed a 2.8% recurrence of negative iPET; in FL, DS resulted in 15.6% recurrence of negative iPET, and 16.1% in ΔSUVmax, with no statistical significance. ePET study: in DLBCL, DS resulted in 14.3% recurrence of negative ePET, and 11.8% in ΔSUVmax at disease-free interval; in HL and FL, both methods showed 2.8 and 12.5% recurrence in negative ePET, respectively. DS and ΔSUVmax did not show significant differences in DLBCL, HL and FL. Their predictive value also did not show significant differences in HL and FL. In DLBCL, DS was higher in iPET, and ΔSUVmax in ePET.