Opportunistic prognostication by computerized tomography (CT) in the emergency department: analysis on 1920 patients and creation of a simple and fast scoring system.

PURPOSE

To use simple CT measurements of musculoskeletal and cardiovascular systems to create a CT-based score to predict mortality in patients admitted to the Emergency Department (ED).

METHODS

The study received IRB approval. Non-contrast abdominal CT of > 18 year old patients admitted to the ER between January 2019 and January 2020 were evaluated by a team of twelve radiologists to calculate: (1) diameter of the infrarenal aorta in millimeter; (2) cross sectional area and composition (Hounsfield units) of the psoas muscle at the third lumbar vertebra (LV); (3) bone density, as quantified at the first lumbar vertebra (LV); (4) presence or absence of dilated abdominal aorta. Thirty-day all-cause mortality (ACM) was determined through hospital and electronic records.

RESULTS

N = 1920 unique patients were evaluated. The mean age was 65 ± 19 years and 46% were female. Death occurred in 7.9% of patients by 30 days from admission. The derivation dataset comprised 1462 patients. At multivariable analysis, age (OR 1.02, 95% CI: 1.007-1.04, p = 0.005), psoas cross sectional area (OR 0.99, 95% CI: 0.997-0.999, p < 0.001) and density (OR 0.96, 95% CI: 0.95-0.98, p < 0.001), and dilated infrarenal aorta (OR 1.85, 95% CI: 1-3.28, p = 0.04) were predictors of the outcome. We accordingly derived a 4-item risk score. In the derivation dataset, the score yielded moderate-high discrimination, with an AUC of 0.73 and excellent diagnostic agreement. In the validation dataset (N = 458), discrimination was high (AUC = 0.83).

CONCLUSION

Simple measurements gathered during a standard CT may allow determining the risk of mortality in the heterogeneous patient population admitted to the ED in a cost- and time-effective manner.