Deep learning FDG uptake classification enables total metabolic tumor volume estimation in diffuse large B-cell lymphoma

Abstract

Total metabolic tumor volume (TMTV), calculated from ¹⁸F-labeled fluoro-2-deoxyglucose (¹⁸F-FDG) positron-emission tomography–computed tomography (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 (AI)-based method could estimate TMTV with a comparable prognostic value to TMTV measured by experts. Methods: Baseline ¹⁸F-FDG PET/CT scans of 301 DLBCL patients from the REMARC trial (NCT01122472) were retrospectively analyzed. An automated whole-body high-uptake segmentation algorithm identified all three-dimensional regions of interest (ROI) 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 AI-based TMTV was estimated as the sum of the volumes of ROIs classified as suspicious uptake. The reference TMTV was measured by two experienced readers using independent semiautomatic software. The AI-based TMTV was compared to the reference TMTV in terms of prognostic value for progression-free survival (PFS) and overall survival (OS). Results: The AI-based TMTV was significantly correlated with the reference TMTV (ρ=0.76; p<0.001). Using the AI-based approach, an average of 24 regions per subject with increased tracer uptake were identified, and an average of 20 regions per subject were correctly identified as nonsuspicious or suspicious, yielding 85% classification accuracy, 80% sensitivity, 88% specificity, compared to the reference TMTV region. Both TMTV results were predictive of PFS (hazard ratio: 2.4 and 2.6 for AI-based and reference TMTVs, respectively; p<0.001) and OS (hazard ratio: 2.8 and 3.7 for AI-based and reference TMTVs, respectively; p<0.001). Conclusion: TMTV estimated fully automatically using an AI-based approach 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 physiological uptake may considerably simplify TMTV estimation, reduce observer variability and facilitate the use of TMTV as a predictive factor in DLBCL patients.