Search

IntroductionInterim PET (iPET) assessment is important for response adaptation in Hodgkin lymphoma (HL). The current standard for iPET assessment is the Deauville score (DS). The aim of our study was to evaluate the causes of interobserver variability in assigning the DS for iPET in HL patients and to make suggestions for improvement.MethodsAll evaluable iPET scans from the RAPID study were re-read by two nuclear physicians, blinded to the results and patient outcomes in the RAPID trial. The iPET scans were assessed visually according to the DS and, thereafter, quantified using the qPET method. All discrepancies of more than one DS level were re-evaluated by both readers to find the reason for the discordant result.ResultsIn 249/441 iPET scans (56%) a concordant visual DS result was achieved. A "minor discrepancy" of one DS level occurred in 144 scans (33%) and a "major discrepancy" of more than one DS level in 48 scans (11%). The main causes for major discrepancies were 1) different interpretation of PET-positive lymph nodes-malignant vs. inflammatory; 2) lesions missed by one reader and 3) different assessment of lesions in activated brown fat tissue. In 51% of the minor discrepancy scans with residual lymphoma uptake, additional quantification resulted in a concordant quantitative DS result.ConclusionDiscordant visual DS assessment occurred in 44% of all iPET scans. The main reason for major discrepancies was the different interpretation of PET positive lymph nodes as malignant or inflammatory. Disagreements in evaluation of the hottest residual lymphoma lesion can be solved by the use of semi-quantitative assessment.

IntroductionThe automatic classification of lymphoma lesions in PET is a main topic of ongoing research. An automatic algorithm would enable the swift evaluation of PET parameters, like texture and heterogeneity markers, concerning their prognostic value for patients outcome in large datasets. Moreover, the determination of the metabolic tumor volume would be facilitated. The aim of our study was the development and evaluation of an automatic algorithm for segmentation and classification of lymphoma lesions in PET.MethodsPre-treatment PET scans from 60 Hodgkin lymphoma patients from the EuroNet-PHL-C1 trial were evaluated. A watershed algorithm was used for segmentation. For standardization of the scan length, an automatic cropping algorithm was developed. All segmented volumes were manually classified into one of 14 categories. The random forest method and a nested cross-validation was used for automatic classification and evaluation.ResultsOverall, 853 volumes were segmented and classified. 203/246 tumor lesions and 554/607 non-tumor volumes were classified correctly by the automatic algorithm, corresponding to a sensitivity, a specificity, a positive and a negative predictive value of 83%, 91%, 79% and 93%. In 44/60 (73%) patients, all tumor lesions were correctly classified. In ten out of the 16 patients with misclassified tumor lesions, only one false-negative tumor lesion occurred. The automatic classification of focal gastrointestinal uptake, brown fat tissue and composed volumes consisting of more than one tissue was challenging.ConclusionOur algorithm, trained on a small number of patients and on PET information only, showed a good performance and is suitable for automatic lymphoma classification.

AimqPET is a quantitative method used to assess FDG-PET response in lymphoma. qPET was developed using 898 scans from children with Hodgkin Lymphoma (HL) in the EuroNet-PHL-C1 (C1) trial. The aim of this study was to determine if qPET could be applied as an alternative response method in adults in the RAPID trial.MethodsPET-CT scans performed after 3 cycles of ABVD in RAPID were re-evaluated by an independent reader, blinded to PET results and outcome in RAPID. All initially involved regions were assessed visually and by qPET. The distribution of qPET measurements was compared for RAPID and C1 patients. Previously published qPET thresholds corresponding to visual DS (vDS) of 1-5 in C1 were used to derive quantitative DS (qDS) for RAPID patients.ResultsPET-CT scans were available for 450 patients from RAPID. vDS were 1 (171 scans), 2 (153 scans), 3 (72 scans), 4 (31 scans) and 5 (23 scans) respectively. The distribution of qPET values was similar to C1 patients, with a unimodal 'normal' distribution and a long tail to the right, suggestive of favorable response in the majority and less favorable response in the minority with outlying values. qPET thresholds from C1 applied in RAPID patients gave 86% concordance for vDS and qDS. There was 97% concordance for complete metabolic response (CMR; DS 1-3) vs. no-CMR using the Lugano classification.ConclusionqPET which was developed in pediatric patients receiving more intensive OEPA chemotherapy, was a suitable quantitative method for assessing response in adult patients treated with ABVD in a response-adapted setting in the RAPID trial.