New perspectives on the role of PET imaging in diffuse large B-cell lymphoma

This thesis evaluated the potential of quantitative PET biomarkers at baseline before treatment and response during treatment using the interim PET in DLBCL patients. The aim is to facilitate selection of high risk patients to ultimately improve outcome due to early treatment modifications and identification of low risk patients to allow treatment de escalation. The main findings were that radiomics features at baseline were predictive of 2-year survival without progression, regardless of the applied segmentation method to delineate the lesions and the applied radiomics feature and/or lesion selection. Moreover, adding radiomics features significantly improves prediction of outcome in DLBCL patients compared to currently used risk scores such as the IPI score, leading to better upfront selection of high risk and low risk patients. An interim PET after 2-4 cycles of treatment is able to differentiate between good and poor responders, and especially the negative predictive value of an I PET is high. It is cost effective to reduce treatment based on a negative interim PET after 2 cycles using the Deauville score. With respect to the optimal response criterion, the ΔSUVmax criterion seems to be the best trade off between sensitivity and positive predictive value. Concluding, PET biomarkers can play a pivotal role in personalizing treatment and management of DLBCL patients and in future lymphoma research.