Preoperative Psoas Muscle Index Predicts Long-Term Survival but Not Postoperative Morbidity After Curative-Intent...

Sarcopenia is increasingly recognized as a host-related prognostic factor in gastric cancer; however, its relationship with early postoperative outcomes remains inconsistent. The psoas muscle index (PMI) is a practical CT-based surrogate of skeletal muscle reserves.

This retrospective cohort from two university-based tertiary referral centers included patients who underwent curative-intent upfront gastrectomy for gastric cancer between 2015 and 2020. Patients who received neoadjuvant chemotherapy, or who presented with radiologically bulky lymph nodes or distant metastasis, were excluded. Bilateral psoas muscle areas at L3 were summed and normalized to height squared (mm²/m²). Sex-specific Turkish cut-offs defined low PMI (<530 mm²/m² men; <360 mm²/m² women). Early postoperative outcomes and time-specific OS (1-, 3-, 5-year) were compared across PMI-defined categories.

A total of 184 patients were included (106 low-PMI; 57.6% and 78 high-PMI; 42.4%). Low PMI was not associated with higher postoperative morbidity: anastomotic leak 4.72% vs 3.85% (p=1.000), surgical-site infection 11.3% vs 10.3% (p=1.000), transfusion requirement 31.1% vs 30.8% (p=1.000), median hospital stay 13 vs 13 days (p=0.379). In contrast, survival clearly diverged: 1-year OS was not significant (80.19% vs 91.03%; p=0.060), whereas 3-year OS (54.72% vs 79.49%; p<0.001) and 5-year OS (43.40% vs 74.36%; p<0.001) were markedly inferior in low-PMI patients. In multivariable Cox regression, PMI independently predicted mortality (HR 0.996; p<0.001).

Low PMI was not associated with early morbidity, but it independently identified patients at substantially higher long-term mortality risk, supporting the concept that diminished muscle mass reflects impaired long-term biological reserve rather than short-term surgical vulnerability, and may be incorporated into routine preoperative risk stratification.