Qualification of Coronary Artery Atherosclerotic Burden and Muscle Mass: Comparison of Two Freely Available Software Programs

Aim/Introduction: Sarcopenia and coronary calcification may have a relevant prognostic impact in oncological and nononcological patients. The use of freeware software is promising for quantitative evaluation of these parameters after whole-body positron emission tomography (PET)/computed tomography (CT) and might be useful for one-stop shop risk stratification without additional radiation ionizing burden and health care. In this study, we compared the assessment of CAC score and muscle mass in patients undergoing whole-body PET/CT by two semiautomatic freeware software, Horos and LIFEx for the evaluation of coronary artery calcium (CAC) score and muscle mass in patients undergoing whole-body PET/CT. Materials and Methods: The study population included 40 patients without known coronary artery diseaserandomly selected from those undergoing from January 2019 to December 2021 unenhanced free breathing CT scan as part of whole-body co-registered PET/CT with 2-deoxy-2-[fluorine-18] fluoro-D-glucose or [fluorine-18] fluoromethylcholine PET/CT scans for oncological or non-oncological indications. The total calcium burden in the coronary arteries was quantified based on the Agatston method. The cross-sectional area of all skeletal muscles, was calculated on a single axial image at the level of the third lumbar vertebra, using the open-source Horos and LIFEx software. Once muscles borders were manually traced, a range of attenuation was then set to include tissue between -29 and +150 HU. Statistical difference between data distributions obtained by Horos and LIFEx software was evaluated using the Wilcoxon matched-pairs signedrank test. To assess the agreement between the measurements obtained by the two software we used the Bland-Altman analysis.Results: The muscle areas obtained by the two software programs were comparable, showing high correlation by Lin's concordancecoefficient (0.9997; 95% confidence intervals, CI: 0.9995-0.9999), and a very good agreement by the Bland-Altman analysis (meandifference = 0.41 cm2, lower limit = -1.06 cm2, upper limit = 1.89). For CAC score the Lin's concordance correlation coefficient was0.9976 (95% CI: 0.9965-0.9984) and at Bland-Altman analysis there was a mean difference of 8 (lower limit = -32, upper limit = 48), with a slight overestimation of Horos CAC score as compared to LIFEx. Conclusion: Our results demonstrated that off-line analysis performed with freeware software may allow a comprehensive evaluation of patients undergoing PET/CT studies, allowing the evaluation of muscle mass and calcium score, that may be relevant for the staging and prognostic stratification, beside standard data obtained by PET/CT imaging. For this purpose, the Horos and LIFEx software seem to be interchangeable.