Derivation of Best-Practice Scan Speeds and Excess Scan Durations for CT Pulmonary Angiography: Analysis Using Registry Data for 166,769 Examinations Across 121 Sites

<b>BACKGROUND.</b> Radiology practices' potential use of a fixed scan speed results in some patients being scanned more slowly than is necessary for the clinical scenario. For CT pulmonary angiography (CTPA), use of fixed scan speeds can lead to prolonged breath-hold requirements and potentially lower image quality from motion artifact. <b>OBJECTIVE.</b> The purpose of this study was to develop best-practice scan speeds for CTPA examinations as well as to assess the extent of variation from these speeds and resulting excess scan durations in real-world clinical practice. <b>METHODS.</b> This retrospective study included 192,779 acquisitions from 166,769 CTPA examinations performed in adult patients (97,649 women and 68,925 men; median age, 60 years old) from January 1, 2016, to January 1, 2021. The examinations were performed at 121 sites using 277 physical scanners representing 28 scanner models from four vendors, and they were identified from an international CT dose registry. Acquisition characteristics were extracted from the registry and were used to calculate scan speeds and durations. Acquisitions were stratified into five equally sized radiation dose categories using CTDI_vol values. For each combination of scanner model and dose category, the best-practice scan speed was calculated as the 95th-percentile actual speed, although the best-practice speed of a higher dose category was assigned if faster. Excess scan durations were calculated with respect to hypothetic durations if acquisitions used best-practice speeds. <b>RESULTS.</b> The acquisitions had speeds that, on average, were 30% slower than the best-practice scan speed. A total of 87% of acquisitions were slower than the best-practice speed; 62% were at least 20% slower than the best-practice speed. Acquisitions had a median scan duration of 4.8 seconds; use of best-practice scan speeds would have saved a median of 1.2 seconds. The 95th-percentile excess scan duration was 7.1 seconds, which was exceeded by at least one examination for 58% of sites. <b>CONCLUSION.</b> CTPA commonly uses speeds slower than the proposed best-practice speeds for a given scanner's capabilities, potentially leading to greater motion artifact and suggesting widespread opportunity for improvement. Simple protocol modifications can decrease scan duration while still allowing adequate radiation dose. <b>CLINICAL IMPACT.</b> The findings indicate widespread opportunity for CTPA protocol improvement through simple adjustments designed to shorten scan speeds at a maintained dose output that preserves image quality.