PP182 - COMPUTED TOMOGRAPHY CONTRAST PHASE ALTERS SKELETAL MUSCLE QUALITY BUT NOT QUANTITY: PRELIMINARY RESULTS FROM AN INTRA-SUBJECT VALIDATION STUDY

PP182

COMPUTED TOMOGRAPHY CONTRAST PHASE ALTERS SKELETAL MUSCLE QUALITY BUT NOT QUANTITY: PRELIMINARY RESULTS FROM AN INTRA-SUBJECT VALIDATION STUDY

A. Jiménez-Sánchez^1,*, J. M. Rubio-García², V. Gumhalter³, E. Barrera-Guadalupe³, F. X. Palmas-Candia⁴

¹UGC Endocrinología y Nutrición, Hospital Universitario Virgen del Rocío, Sevilla, ²Department of Radiology, Complejo Hospitalario Universitario Insular Materno Infantil de Canarias, ³ARTIS Development, Las Palmas de Gran Canaria, ⁴Endocrinology and Nutrition Department, Hospital Universitari Vall d’Hebron, Barcelona, Spain

Rationale: Computed tomography (CT) is a gold standard to assess skeletal muscle (SM) quantity (by area) and quality (by density, a surrogate for myosteatosis). However, protocol heterogeneity—specifically the presence and timing of intravenous contrast—may hinder its validity and clinical implementation.

Methods: Routinely acquired three-phase abdominal CT scans consisting of non-contrast (NC), arterial (A), and venous (V) phases were included. A blinded single operator analyzed DICOM images with FocusedON-BC software. Two adjacent axial slices at the L3 vertebra were analyzed per phase. SM area (cm²) and density (Hounsfield units, HU) were segmented with a -29 to +150 HU threshold, combining AI and manual corrections. An adjusted error was calculated to isolate the contrast effect in both SM biomarkers, by subtracting baseline intra-phase topographic noise (slice-to-slice physiological variation) from the net inter-phase variation. As a quality control, phases were validated via aorta and inferior vena cava density changes. Statistics included median (IQR), Friedman test, Kendall’s W, and Bonferroni-corrected Wilcoxon post-hoc tests (p < 0.05).

Results: N = 129 patients (774 images) were measured. Phase validation was adequate: aortic density peaked in A (+197.5 HU vs. baseline; p < 0.001), and cava density peaked in V (+66.5 HU; p < 0.001). SM density increased progressively from baseline (16.4 HU) to +2.4 HU in A (p < 0.001), and +7.3 HU in V (p < 0.001). SM area remained stable during the NC-A transition (139.39 vs. 139.54 cm²; p = 1.00), with marginal reductions in A-V (139.54 vs. 138.54 cm²; p < 0.001) and NC-V (139.39 vs. 138.54 cm²; p < 0.001).

Conclusion: CT acquisition phases were interchangeable for SM quantity (area), but not for SM quality (density). Due to the prognostic importance of myosteatosis, methodological solutions for this limitation are needed.

Disclosure of Interest: A. Jiménez-Sánchez: None declared, J. M. Rubio-García: None declared, V. Gumhalter Other: Employed by ARTIS Development, the company that developed and owns the FocusedON-BC software evaluated in this study, E. Barrera-Guadalupe Other: Employed by ARTIS Development, the company that developed and owns the FocusedON-BC software evaluated in this study, F. X. Palmas-Candia: None declared