CTA Utilization to Reduce Radiation Exposure in Low-Risk Pulmonary Embolism Evaluations: A Single-Center Retrospective Audit

Background. Pulmonary embolism (PE) is the obstruction of the pulmonary arteries secondary to a venous thromboembolism (Vyas et al. 2024). Computed tomography pulmonary angiography (CTPA) is the imaging standard for suspected PE, but its value depends on appropriate patient selection. A diagnostic yield of roughly 10% is widely cited as the lower threshold below which overuse should be suspected. United States emergency department (ED) cohorts frequently fall well below this floor, with pooled yields near 3%, and lower D-dimer use has been associated with lower yield. In patients stratified as low-risk by the Wells score, a negative D-dimer can exclude PE and prevent unnecessary imaging, yet adherence to this pathway is inconsistent. Objective. To quantify CTPA diagnostic yield and the number needed to scan (NNS), measure adherence to Well’s Score patient stratification criteria, and estimate the radiation burden delivered to PE-negative patients. Methods. We performed a retrospective chart audit of all chest CTAs ordered between April 1 and April 30, 2025 at a community teaching hospital in the Ohio River Valley(Trinity Medical Center West ). Wells scores were calculated for each patient. D-dimer testing patterns were evaluated against Wells’ Score pathway expectations. Data were analyzed using Fisher's exact test and the Mann–Whitney U test. Effective dose was estimated at approximately 2 mSv per CTA-PE protocol (ICRP reference values). Results. A total of 125 CTAs were ordered for 89 patients; 85 of which were ordered to evaluate for PE. The overall diagnostic yield was 8.3% (7/85; NNS 12), above the typical US ED average but below the 10% appropriateness floor. 34 patients had a Well Score >4 (suggestive of increased likelyhood of PE )scored PE-likely (Wells >4), with a PE rate of 14.7% (NNS 6.8).51 patients had a well score ≤4 (suggestive of decreased likelihood of having PE ) , with a PE rate of 3.9%. Only 23.5% of patients eligible for which a D-dimer-first evaluation was indicated had a D-dimer obtained before CTA, and all eligible patients proceeded to imaging regardless of testing results. “Only 23.5% of eligible patients who were indicated for a D-dimer-first evaluation actually received the test before CTA, and all eligible patients underwent imaging regardless of the results PE as the number-one or equally likely diagnosis” was the only significant predictor of a positive CTA (OR 7.76, p=0.048). Mean Wells score was significantly higher in PE-positive than PE-negative patients (5.36 vs 2.96; p=0.027). Approximately 236 mSv of cumulative effective dose was delivered to PE-negative patients, equivalent to roughly 76 years of US background radiation.; 91.7% of evaluable CTAs were non-diagnostic for PE. Conclusions.While CTA ordering at this institution out-performed many US EDs on yield (Kline et al. 2020), it failed to reach the appropriateness floor. This shortfall was driven by a process gap rather than poor clinical selection: three of four low-risk patients bypassed the D-dimer testing and proceeded directly to imaging. Reducing reliance on clinical gestalt with decision support at the point of CTA order entry, such as D-dimer testing and interpretation in elevated-risk populations such as smokers, chronic lung disease, malignancy, and advanced age, represents a high-yield and inexpensive target to reduce radiation risk and financial burden for patients and hospital systems. Because ordering tracked clinical impression rather than structured scoring, embedding decision support at the point of CTA order entry — paired with attention to D-dimer interpretation in a population enriched for smoking, chronic lung disease, malignancy, and advanced age — represents a high-yield, low-cost target for radiation and utility stewardship.