(I-628) Machine learning-based detection of sarcopenic obesity and adverse outcome prediction in patients undergoing surgical treatment for spinal metastases

Introduction: The distribution and proportions of lean and fat tissues may dictate prognosis and outcomes of patients undergoing oncologic treatment. Specifically, in obese patients, sarcopenia may be easily overlooked as an additive prognostic factor when compared to obesity alone. The role of this body phenotype, known as sarcopenic obesity (SO), has not yet been studied in the context of postoperative outcomes among patients undergoing surgical treatment for spinal metastases.

Methods: Records of patients undergoing surgical treatment for spinal metastases between 2010 and 2019 were retrospectively examined. Sarcopenia and adiposity were assessed on routinely available 3-month preoperative CT images by using a validated deep learning methodology. Based on these composition analyses, patients were classified with SO and non-sarcopenic obesity. Following Nearest neighbor propensity matching, rates of postoperative complications, length of stay, 30-day readmission, and 90-day and 1-year all-cause mortality were evaluated.

Results: A total of 62 patients with obesity underwent surgical treatment for spinal metastases during the study period. Of the 62 , 37 patients had non-sarcopenic obesity, while 25 had sarcopenic obesity. After propensity matching, a total of 50 records were evaluated, composed equally f patients with non-sarcopenic obesity and SO (n=25). Patients with SO were noted to have increased rates of discharge to acute rehab vs. home (80% vs. 40%, p=.004), 30-day readmissions (72% vs. 44%, p=.045), 90-day mortality (48% vs. 16%, p=.015), 1-year mortality (64% vs. 32%, p=.024), and months to mortality following surgery (12.60±19.84 vs. 37.16±35.19 p=.002). No significant differences were noted in length of stay or postoperative complications when comparing the two groups (p>.05).

Conclusion : The SO phenotype is associated with increased rates of non-home discharge, readmission, and postoperative mortality. Our study suggests that SO is an important prognostic factor to consider when planning the course of care for patients with spinal metastases.

How to Improve Patient Care: