CT-Derived L3 Skeletal Muscle Index, Handgrip Strength, and Nutritional Behavior as Complementary Dimensions of Frailty in Older Oncology Inpatients: A Cross-Sectional Observational Study

Background: In older oncology patients, sarcopenia and frailty are highly prevalent but may reflect distinct vulnerability domains. Whether morphologic muscle loss, functional impairment, and nutritional behavior converge in this population remains unclear. We investigated whether CT-derived L3 skeletal muscle index, handgrip strength, G8-based frailty-related vulnerability, and self-reported nutritional behavior represent overlapping or distinct vulnerability dimensions in older oncology inpatients. Methods: This single-center substudy was derived from a prospectively collected screening cohort of hospitalized patients aged ≥65 years. Skeletal muscle index (SMI) was retrospectively quantified at the L3 level from routine CT scans acquired within ±1 month of screening, using sex-specific Prado cut-offs to classify CT-defined low skeletal muscle mass. Handgrip strength was measured by digital dynamometry and categorized using age- and sex-adjusted thresholds. Frailty was assessed with the G8 questionnaire, including item C on mobility. Nutritional behavior, comprising food avoidance, supplement use, and cancer-specific diets, was evaluated by structured self-report. Principal component analysis and correlation heat mapping were used to identify behavioral patterns. Results: Among 116 patients, including 68 with hematologic malignancies and 48 with solid tumors, CT-defined low muscle mass was present in 55.2% and was more frequent in men (p < 0.01). SMI was not associated with handgrip strength (rho = 0.03; p = 0.784). Reduced handgrip strength was associated with lower G8 scores (p < 0.001) and lower G8 item C mobility scores (p = 0.001). Self-reported nutritional behaviors were common and showed no significant association with CT-defined low skeletal muscle mass status. Conclusions: CT-derived muscle mass, muscle strength, and nutritional behavior appear to capture partly non-overlapping vulnerability dimensions in older oncology inpatients, supporting integrated functional and nutritional assessment.