THE ROLE OF NUCLEAR MEDICINE IN LYMPHOMAS – A FOCUS ON PROGNOSTIC EVALUATION OF DIFFUSE LARGE B-CELL LYMPHOMA WITH POSITRON EMISSION TOMOGRAPHY
Czibor Sándor
Rácz Károly Klinikai Orvostudományi
Dr. Fekete Andrea
SE Szülészeti és Nőgyógyászati Klinika, Baross utcai részleg
2025-09-19 15:00:00
Klinikai haematológia
Dr. Masszi Tamás
Dr. Györke Tamás
Dr. Varga Gergely
Dr. Ritter Zsombor
Dr. Ács Nándor
Dr. Erdélyi Dániel János
Dr. Garai Ildikó
Objective: To investigate potential roles of existing [18F]FDG-PET/CT prognostic parameters in DLBCL and to establish new or modified methods which are both easy to implement in routine workflows and yield more accurate results than the current ones.
Methods: First, data of 107 DLBCL NOS patients from a multicentre, prospective trial were used for analysis of baseline volumetric values (MTV and TLG, also normalised for body weight) segmented with three different methods and interim parameters. Secondly, retrospective analysis of 50 baseline PET/CT scans from one centre was performed to investigate textural PET features and the newly defined MTVrate value, alongside MTV and clinical data, including prognostic model-building utilizing a machine learning algorithm. 24-month PFS was the clinical endpoint at both studies and ROC analyses were performed to define optimal cut-off points for continuous parameters. The PFS of low- and high-risk groups were compared with log-rank and Cox-regression analysis.
Results: In the analysis of the multicentre data, MTV and TLG calculations showed good correlation among the three segmentation methods, however, optimal cut-off points were markedly different. Body weight-adjusted MTV and TLG did not provide markedly better prognostic values. Highest hazard ratio was shown for rPET (HR=9.09) and it was also shown to be an independent predictor of PFS (p=0.041; HR=9.15) in a multivariate Cox-regression model. A combined analysis showed that ΔSUVmax positive patients with high MTV formed a group with distinctly poor PFS (35.3%). Among the single-centre data, individual analysis showed the highest AUC on ROC analysis for MTVrate at 0.74, followed by LDH, MTV, and skewness, with AUCs of 0.68, 0.63, and 0.55, respectively, and these parameters were also able to differentiate the PFS. A combined analysis including MTV and MTVrate identified a subgroup with particularly low PFS at 38%. The machine learning-based model had an AUC of 0.83 and the highest relative importance was attributed to five textural features and both MTV and MTVrate.
Conclusion: Results from the analysis of both datasets underline the importance of FDG-PET/CT in the prognostic evaluation of DLBCL patients and indicate PET-based markers that are highly effective and easily implementable in routine clinical practice: MTV and the novel MTVrate feature at baseline and rPET at interim PET.
