ISHL10 Abstract T021

Semi-quantitative parameters to improve the interim FDG-PET/CT positive predictive value in Hodgkin Lymphoma

Interim FDG-PET/CT (iPET), according to the 5-point Deauville score (5p-DS) is the strongest predictor for outcome in Hodgkin Lymphoma (HL). In PET-guided treatment approaches, iPET-positive patients are candidates for more intensive and potentially more toxic treatments. Nevertheless, recent studies have shown that the positive predictive value (PPV) of iPET visual analysis with 5p-DS has still some limitations to optimally identify patients at different prognosis. Semi-quantitative parameters could improve PPV of iPET. Aim of this retrospective study is to compare PPV of visual and semi-quantitative analysis of iPET in HL patients treated with ABVD.

We studied 67 HL patients (median age 39 years; 30 females, 37 males; 38 limited Stage, 29 advanced) at our Institution between 2007 and 2013, who were treated with ABVD. iPET was performed after 2 cycles of ABVD. For visual analysis, the 5p-DS was used, setting different cut-points at 5p-DS>3 and 5p-DS>4 as positive. For semi-quantitative analysis, we evaluated the ratio between lesion and liver SUVmax (rPET); the ratio between lesion and mediastinal blood pool SUVmax (mPET); the ratio between lesion SUVpeak and liver SUVmean (qPET). Primary endpoint was two-year progression-free survival (PFS). ROC analysis was used to determine the best cut-point of semi-quantitative parameters to identify treatment failures.

In visual analysis, 25/67 patients had 5p-DS >3, and 5p-DS was >4 in 14/67 patients. PFS according to 5p-DS>3 and 5p-DS>4 was 53% and 27%, respectively and their PPV was 40% and 57%, respectively. The semi-quantitative parameters between residual lesion and the different backgrounds rPET, mPET and qPET were prognostic factors in our population (p<0.01). The most accurate cut-point in predicting adverse events for rPET, mPET and qPET were 1.14, 2 and 1.46, respectively. For values higher than these cut-point, PFS were 15%, 25% and 20% and the PPV were 70%, 63% and 80%, respectively.

iPET semi-quantitative parameters appear to perform better than visual analysis for outcome prediction in HL. In particular, ratios between residual lesion and background SUV (liver or mediastinal blood pool) could improve the predictive value of relapse or progression.