Search

Background: We analyzed the prognostic value of a new baseline positron emission tomography (PET) parameter reflecting the spread of the disease, the largest distance between two lesions (Dmax). We tested its complementarity to metabolic tumor volume (MTV) in a large cohort of diffuse large B-cell lymphoma (DLBCL) patients from the REMARC trial (NCT01122472)., Patients and Methods: MTVs were defined using the 41% maximum standardized uptake value threshold. From the three-dimensional coordinates, the centroid of each lesion was automatically obtained and considered as the lesion location. The distances between all pairs were calculated. Dmax was obtained for each patient and normalized with the body surface area [standardized Dmax (SDmax)]., Results: From the REMARC trial, 290 patients aged 60-80 years were included: 91% had an advanced stage and 71% International Prognostic Index (IPI) ≥3. High versus low SDmax significantly impacted progression-free survival (PFS) (P < 0.0001) and overall survival (OS) (P = 0.0027). Patients with SDmax > 0.32 m -1 (n = 82) had a 4-year PFS and OS of 46% and 71%, respectively, against 77% and 87%, respectively, for patients with low SDmax. High SDmax and high MTV were independent prognostic factors of PFS (P = 0.0001 and P = 0.0010, respectively) and OS (P = 0.0028 and P = 0.0004, respectively). Combining MTV and SDmax yielded three risk groups with no (n = 109), one (n = 122) or two (n = 59) factors (P < 0.0001 for both PFS and OS). The 4-year PFS were 90%, 63%, 41%, respectively, and the 4-year OS were 95%, 79%, 66%, respectively. In addition, patients with at least two of the three factors including high SDmax, high MTV, Eastern Cooperative Oncology Group (ECOG) ≥2 had a higher number of central nervous system relapse (P = 0.017)., Conclusions: SDmax is a simple feature that captures lymphoma dissemination, independent from MTV. These two PET metrics, SDmax and MTV, are complementary to characterize the disease, reflecting the tumor burden and its spread. This score appeared promising for DLBCL baseline risk stratification., Competing Interests: Disclosure NC is an employee at Siemens Healthcare GmbH. OC declares research funding from Roche, Takeda, Gilead, advisory board and honoraria from Celgene, Roche, Takeda, Gilead, BMS, Merck, AbbVie, Janssen outside the submitted work. CT declares honoraria from Roche, Amgen, Janssen, Celgene, Gilead Science/Kyte; consulting/advisory role from Roche, Gilead Sciences, Janssen, Celgene, Novartis; research funding and travel, accommodations, expenses from Roche, Novartis. All other authors have declared no conflicts of interest., (Copyright © 2020 European Society for Medical Oncology. Published by Elsevier Ltd. All rights reserved.)

Total metabolic tumor volume (TMTV), calculated from 18 F-FDG PET/CT baseline studies, is a prognostic factor in diffuse large B-cell lymphoma (DLBCL) whose measurement requires the segmentation of all malignant foci throughout the body. No consensus currently exists regarding the most accurate approach for such segmentation. Further, all methods still require extensive manual input from an experienced reader. We examined whether an artificial intelligence-based method could estimate TMTV with a comparable prognostic value to TMTV measured by experts. Methods: Baseline 18 F-FDG PET/CT scans of 301 DLBCL patients from the REMARC trial (NCT01122472) were retrospectively analyzed using a prototype software (PET Assisted Reporting System [PARS]). An automated whole-body high-uptake segmentation algorithm identified all 3-dimensional regions of interest (ROIs) with increased tracer uptake. The resulting ROIs were processed using a convolutional neural network trained on an independent cohort and classified as nonsuspicious or suspicious uptake. The PARS-based TMTV (TMTV PARS ) was estimated as the sum of the volumes of ROIs classified as suspicious uptake. The reference TMTV (TMTV REF ) was measured by 2 experienced readers using independent semiautomatic software. The TMTV PARS was compared with the TMTV REF in terms of prognostic value for progression-free survival (PFS) and overall survival (OS). Results: TMTV PARS was significantly correlated with the TMTV REF (ρ = 0.76; P < 0.001). Using PARS, an average of 24 regions per subject with increased tracer uptake was identified, and an average of 20 regions per subject was correctly identified as nonsuspicious or suspicious, yielding 85% classification accuracy, 80% sensitivity, and 88% specificity, compared with the TMTV REF region. Both TMTV results were predictive of PFS (hazard ratio, 2.3 and 2.6 for TMTV PARS and TMTV REF , respectively; P < 0.001) and OS (hazard ratio, 2.8 and 3.7 for TMTV PARS and TMTV REF , respectively; P < 0.001). Conclusion: TMTV PARS was consistent with that obtained by experts and displayed a significant prognostic value for PFS and OS in DLBCL patients. Classification of high-uptake regions using deep learning for rapidly discarding physiologic uptake may considerably simplify TMTV estimation, reduce observer variability, and facilitate the use of TMTV as a predictive factor in DLBCL patients., (© 2021 by the Society of Nuclear Medicine and Molecular Imaging.)