Nearly perfect reproducibility degree of computed tomography in the evaluation of subcutaneous, visceral, and epicardial adipose volumes and radiodensities in lymphoma patients

Obesity increases cardiovascular risk and the occurrence of certain types of cancer. Additionally, oncological diseases and their anticancer therapies can lead to adipose tissue remodeling, worsening chronic inflammation, insulin resistance, and cardiovascular risk. Therefore, measuring the volume and distribution of body fat can help characterize and monitor the cardiometabolic status of patients undergoing cancer treatment. Positron emission tomography (PET) combined with computed tomography (CT) is a routine hybrid imaging method in oncology practice. PET/CT can provide information on the quality, volume, and distribution of body fat. However, the literature regarding the accuracy and reproducibility of these analyses by PET/CT still needs to be improved.

To evaluate the inter- and intra-observer reproducibility of CT in analyzing measurements of subcutaneous (SC), visceral (VS), and epicardial (EP) fat volumes in lymphoma patients who underwent oncological PET/CT examinations at a Brazilian university hospital.

A blind, comparative, observational, and analytical study. Data collection was performed using an oncological PET/CT examination database. The volumes and attenuation densities of SC, VS, and EP fat were measured by two independent evaluators using the free software LIFEx at two different moments on PET/CT images of 21 lymphoma patients. The data were grouped by evaluator and analysis moment and compared among them. Intra-observer reproducibility was evaluated using the intraclass correlation coefficient (ICC), and inter-observer reproducibility was evaluated using the kappa coefficient (k).

Regarding intra-observer reproducibility, the ICC was greater than 0.99 for assessments of SC and VS fat attenuation density and greater than 0.96 for EP fat attenuation density for both observers. For fat volume assessments, an ICC greater than 0.99 was obtained for SC and VS and greater than 0.97 for both observers. Concerning inter-observer reproducibility, the k coefficient ranged from 0.9 to 0.95 for EP attenuation and volume assessments and was >0.99 for both measures in the other sites.

PET/CT showed nearly perfect intra- and inter-observers reproducibility in evaluating the density and volume of SC, VS, and EP fat in lymphoma patients.