Trauma is a leading cause of death in the US and around the world [1–4]. To maximize the likelihood of positive outcomes, contemporary trauma care involves rapid transfer of severely injured patients by emergency medical services from the site of injury to dedicated trauma centers with appropriate sub-specialization at hand [35–37]. Trauma WBCT is a ubiquitous and rapid exam used to identify and characterize injuries [10], and may confer a survival advantage [15, 16].
Factors that have been reported to influence radiology RTATs include reader experience level, round-the-clock coverage, and the practice setting (e.g., teaching hospital vs teleradiology) [27–29]. We are not aware of prior work that explored the relationship between trauma injury acuity and severity and RTATs for trauma WBCT exams. This study retrospectively examined the association between RTATs and relevant clinical data corresponding with injury acuity and severity in a large cohort of 11251 trauma patients who underwent admission trauma WBCTs.
Life threatening traumatic injuries may stand to benefit the most from rapid diagnosis [11–16], however, we find greater injury acuity and severity are associated with longer RTATs. Two primary correlates of trauma severity were used to stratify patients: polytrauma (ISS ≥ 16) and need for blood transfusion. Hemorrhage is recognized as the leading reversible cause of death in the first hour of admission to a trauma center [7–9]. Patients who required massive transfusion (≥ 10 PRBCs) had almost double the RTAT of those who did not (47.5 vs 24 minutes, p < 0.0001). RTAT was increased 1.7-fold (39 vs 23 minutes, p < 0.0001) in patients requiring any blood transfusion, 1.6-fold (34 vs 21 minutes, p < 0.0001) in patients with polytrauma, and 1.8-fold in those who expired (41 minutes [IQR: 13-72.5] vs 23 minutes [IQR: 4–47]). RTAT was also longer in patients with Shock Index > 1, penetrating injury mechanisms, and ≥ 3 seriously injured body regions.
Among the more severely injured cohort of patients with polytrauma for whom trauma WBCT may be instrumental in surgical planning, age; ISS; number of PRBCs transfused; penetrating trauma mechanism; and serious injuries involving the neck, thorax, and pelvis or lower extremities were independently predictive of RTATs that exceeded approximately one hour. The findings in this study highlight a fundamental paradox in trauma imaging – that the patients who stand to benefit the most from a rapid interpretation often have the most complex and time-consuming findings to report.
Our study does not establish any causality between longer RTAT and poor outcomes, as trauma surgeons may receive wet reads, independently review studies on their own for critical injury, or may seek informal scanner console or reading room consultations. We discourage radiologists from making process-oriented changes that may involve trade-offs between reading time and accuracy based on these results.
Radiology is a forward-thinking field that has consistently embraced advancements in digital technology, leveraging data-driven approaches to remain at the forefront of innovation. RTAT has served as a means of objectively benchmarking improvements in radiologist efficiency with implementation of technologies including PACS, voice recognition, and more recently, scalable deep learning-based AI CAD solutions [17, 18, 20]. AI CAD tools may help facilitate a faster workflow in patients with known or suspected polytrauma. Davis et al., found that RTATs for emergency head CTs decreased following deployment of a commercial ICH detection tool [23], and in other work by O’Neill et al., ICH CAD was shown to have a greater effect on efficiency in ICH-positive patients compared to patients without ICH [24]. Other commercially available FDA-approved products that could have a significant impact on trauma WBCT RTAT include those for spine and rib fractures [38–41]. Counting of rib fractures is an example of a disproportionately time-consuming task that could benefit from automation, and we speculate that this might play a role in our result of thoracic AIS as an independent predictor of longer RTATs in patients with polytrauma.
Commercial CT-based tools are available for triage or detection and grading of intracranial hemorrhage and spine fractures. Proof-of-concept tools in earlier stages of the R&D pipeline include those for detection of active hemorrhage and aortic injury; detection and grading of pelvic fractures; detection, quantification, and grading of solid organ lacerations; as well as detection and quantification of traumatic extraperitoneal pelvic hematoma, hemoperitoneum, hemothorax, and pulmonary contusion [22, 42–56].
Greater institutional, society, governing body, and regional or national-level funding agency awareness and resource allocation is needed to accelerate advancements in this area. Trauma WBCT interpretation represents a long-tailed problem with a very broad range of encountered injury types and severities, and outside of a few use cases with large public datasets [57, 58], data scarcity currently remains a major obstacle.
Ultimately, clinical use and reimbursement by CMS will be based on studies showing improvement in patient outcomes, however outcome studies supporting FDA-approved tools are currently few in number [20]. Since RTAT is routinely collected by radiology departments, this objective parameter can be used as a preliminary indicator of potential efficacy before undertaking time, resource, and capital-intensive clinical trials and prospective outcome studies. Trauma WBCT serves a dual purpose as a screening and surgical planning modality. Patients with negative scans or minor injuries may ultimately have shorter stays that lower hospital costs. Those with the most serious hemorrhage-related injuries and studies that take the longest to read could have more dramatic reductions in RTAT when the radiologist’s workflow is augmented by suites of tools for the spectrum of potentially surgically important injuries. Future work would need to explore potential associated reductions in time to intervention or mortality.
Our large-scale cohort study has several limitations; it is retrospective with data collected from two trauma centers in a single medical system. We do not have comparison data from other institutions using different practice models which limits generalizability. Given prior reports of turn-around times ranging from 30–87 minutes [25, 26], the RTATs of our radiology group are well within the standard of care. We currently use an ICH early notification tool on all patients with head CT sent to PACS. Evaluation of trends related to CAD implementation are outside of the scope of our work. Effects of injury acuity or severity on RTAT may vary for teleradiology practices covering multiple trauma centers and non-traumatic acute care at a range of other sites, community practices without experienced staff in this domain, and academic practices with resident overnight coverage who are faced with a different set of considerations than in our practice. AI CAD tools for torso cavitary hemorrhage, organ injury, cervical spine fracture, and cerebrovascular injury, may be particularly beneficial in settings where trauma/ER subspecialized radiologists are unavailable [21].