Emergency Radiology: Protocols for Trauma and Acute Care

Emergency radiology is the subspecialty that provides around-the-clock imaging interpretation for patients in acute distress, and its speed and accuracy directly affect whether people live or die. When a hospital implemented 24/7 radiology coverage at a level-one trauma center, the average time patients spent in the resuscitation unit dropped by about a quarter, translating to hundreds of thousands of dollars in annual savings and, more critically, faster clinical decisions. The field sits at the intersection of nearly every emergency scenario you can imagine, from car crashes and strokes to blood clots in the lungs and ruptured arteries, and the imaging choices made in those first minutes shape everything that follows.

Whole-Body CT in Major Trauma

The single biggest shift in how severely injured patients are imaged over the past two decades has been the move from selective scanning, where clinicians order individual X-rays or targeted CT scans of the body parts they suspect are hurt, to whole-body CT, where the scanner captures everything from the head to the pelvis in one rapid pass. A large retrospective study across German trauma centers found that integrating whole-body CT into early resuscitation was an independent predictor of survival, with a relative mortality reduction of roughly 13 to 25 percent depending on the scoring method used. The number needed to scan, meaning how many patients you’d have to image to save one additional life, ranged from about 17 to 32.¹

A systematic review and meta-analysis pooling data from multiple studies confirmed the trend: overall mortality was lower in patients who received whole-body CT compared with those who received selective scanning, with a pooled odds ratio of 0.75 favoring the whole-body approach.² A separate French multicenter study echoed the finding: crude 30-day mortality was about 16 percent among whole-body CT patients versus 22 percent among those who had selective imaging.³

The advantage is straightforward: injuries that would otherwise be missed or caught late, like a slowly bleeding spleen or a subtle aortic tear, show up immediately on a comprehensive scan. The tradeoff is radiation exposure and occasional overdiagnosis of incidental findings that may not matter in the acute setting. But for patients with polytrauma, the survival benefit has made whole-body CT the standard diagnostic method during the resuscitation phase at most major centers.⁴

Bedside Ultrasound and Its Limits

Before a trauma patient ever reaches the CT scanner, the first imaging they usually get is a bedside ultrasound called FAST, which stands for Focused Assessment with Sonography in Trauma. It takes seconds, requires no radiation, and can detect free fluid in the abdomen or around the heart, a strong clue that something is bleeding internally. The catch is that its sensitivity varies considerably.

One large study found that the overall sensitivity of FAST for detecting abdominal injury was only about a third, though specificity was very high at roughly 99 percent. Sensitivity climbed to nearly half in patients who arrived with low blood pressure, where there tends to be more free fluid to detect.⁵ When emergency physicians performed the exam themselves in a separate study, sensitivity was much higher, over 93 percent, with specificity in a similar range. Radiology residents in the same study achieved even slightly higher sensitivity.⁶ The wide gap between these studies likely reflects differences in injury severity, operator skill, and the timing of the exam relative to the injury.

An interesting finding is that repeating the FAST exam over time can substantially improve its accuracy. One study tracking serial exams showed sensitivity climbing from about 64 percent at four hours to roughly 88 percent at twelve hours after injury, as bleeding accumulated and became easier to see on ultrasound.⁷ The practical implication is that a negative FAST in the first hour after a car accident does not rule out internal bleeding. If the patient’s condition worsens or there is persistent clinical suspicion, a CT scan or repeat ultrasound is warranted.

Stroke Imaging and the Clock

Few emergencies are as time-sensitive as a stroke. The phrase “time is brain” exists because every minute a large artery is blocked, roughly two million neurons die. Emergency radiology’s job in these cases is not just to confirm a stroke but to figure out whether there is brain tissue that can still be saved, which determines whether the patient is a candidate for clot-retrieval procedures.

A plain, unenhanced head CT can reliably rule out bleeding (hemorrhagic stroke) but is not great at detecting early ischemic strokes, the kind caused by a clot. One study found that plain CT had a sensitivity of only about 55 percent for acute ischemic stroke. Adding CT angiography, which maps the blood vessels, and perfusion CT, which measures blood flow through the brain tissue, raised that sensitivity to roughly 79 percent.⁸ Combining both techniques gives the most complete picture: the site of the blockage, the core of dead tissue, and the surrounding at-risk tissue that might still be rescued with treatment.⁹

This multimodal approach has become standard at comprehensive stroke centers. It adds only a few minutes to the scan but provides information that fundamentally changes whether a patient gets a clot-busting drug, a catheter-based procedure, or supportive care alone.

Hunting for Pulmonary Embolism

CT pulmonary angiography is the go-to test for suspected blood clots in the lungs, but there’s a persistent problem with overuse. One study found that about a third of patients scanned for suspected pulmonary embolism actually had one.¹⁰ Other studies show the number is far lower. An emergency department analysis found the yield was just 9.4 percent, with potentially avoidable imaging performed in roughly half to three-quarters of patients depending on the criteria applied.¹¹ A retrospective study at another center found an even lower rate of about 12 percent, well below the threshold recommended by professional guidelines, and estimated that up to half of those scans could have been avoided with proper use of clinical scoring tools and a blood test called D-dimer.¹²

The pattern is clear: many emergency departments skip the step of stratifying patients by risk before ordering the scan. D-dimer is a cheap blood test that, when negative in a low-risk patient, effectively rules out pulmonary embolism without any imaging at all. When clinicians skip that step and go straight to CT, the result is unnecessary radiation, contrast dye exposure, and incidental findings that trigger even more testing. This remains one of the most actionable quality-improvement targets in emergency radiology.

Aortic Emergencies

Acute aortic syndromes, including aortic dissection, intramural hematoma, and penetrating aortic ulcer, are among the most lethal conditions that come through an emergency department. CT angiography of the chest has become the gold standard for evaluating these cases because it can capture the entire aorta and its branches in a single, fast acquisition.¹³ The non-contrast phase is critical for detecting acute bleeding within the aortic wall, while the contrast-enhanced phase reveals the tear itself, any separation between the true and false channels of blood flow, and whether branch arteries supplying the brain, kidneys, or gut are compromised.¹⁴

Getting the technical details right matters here more than in almost any other scan. Motion from the heartbeat can mimic a dissection flap on the image, and poorly timed contrast injection can obscure a slow-flowing false lumen. These pitfalls are well recognized but still trip up readers, especially overnight when fatigue sets in.

Spinal Emergencies and When to Skip the X-Ray

After a fall or a car accident, one of the first questions is whether the cervical spine is injured. Two clinical decision rules help clinicians decide who actually needs imaging: the Canadian C-Spine Rule and the NEXUS criteria. A head-to-head comparison found the Canadian rule was more sensitive, catching over 99 percent of significant injuries compared with about 91 percent for NEXUS. It was also more specific, which translated to fewer unnecessary imaging studies.¹⁵ A meta-analysis broadly confirmed the sensitivity gap, finding pooled sensitivity of about 99 percent for the Canadian rule versus 90 percent for NEXUS.¹⁶

Beyond trauma, spinal cord compression from tumors, infections, or disc herniations is a separate emergency where MRI is the key tool. Delay in diagnosis can lead to permanent paralysis, which makes urgent MRI availability a defining feature of how well a hospital handles these cases.¹⁷ A two-center study looking at patients referred for urgent MRI on suspicion of cauda equina syndrome, a condition where the nerve roots at the base of the spine are compressed, found that only about 5 percent of those scanned actually had the condition.¹⁸ That low hit rate reflects appropriate caution: the consequences of missing the diagnosis are severe enough that the threshold for ordering the scan stays low.

Appendicitis and the Stepwise Approach

Suspected appendicitis is one of the most common reasons for emergency abdominal imaging. A stepwise algorithm, starting with ultrasound and moving to low-dose CT only when ultrasound is inconclusive, has shown excellent accuracy while cutting radiation significantly. In one study of 183 adults, ultrasound alone resolved the diagnosis for nearly half of patients, and the full algorithm achieved sensitivity and specificity both above 96 percent while reducing the need for standard full-dose CT.¹⁹

This stepwise strategy is especially valuable in pregnant patients, for whom radiation and iodinated contrast are concerns. A study applying a similar algorithm in pregnant women with suspected appendicitis, using ultrasound first and low-dose CT with oral contrast as a backup, achieved 100 percent sensitivity and 92 percent specificity for appendicitis while keeping the radiation dose below 2.5 millisieverts per scan.²⁰

Protecting Children from Unnecessary Scans

Children are more vulnerable to radiation than adults because their cells are dividing faster and they have more years ahead in which a radiation-induced cancer could develop. The PECARN rule, a clinical decision tool for pediatric head trauma, was designed to identify which children truly need a CT scan and which can safely be observed. Multiple validations have confirmed it works. One Italian study found that applying the rule would have reduced CT scans by about 29 percent without missing a single case of clinically important traumatic brain injury.²¹

A larger study of over 1,300 children found that only about 4 percent required CT scanning based on the rule. Among the vast majority who did not receive a scan, none returned with delayed intracranial bleeding or neurological deterioration.²² A Dutch study projected that systematic application of the PECARN rules could cut pediatric head CT rates by more than half in children over two years of age, from about 63 percent to 24 percent.²³ Despite this strong evidence, adoption remains uneven. Many emergency departments still default to scanning kids whose mechanism of injury sounds scary, even when the clinical exam suggests the risk is low.

Interventional Radiology in Hemorrhage Control

Emergency radiology is not purely diagnostic. Interventional radiologists can thread a catheter into a bleeding artery and plug it with tiny coils, gel foam, or other materials, a procedure called embolization. In pelvic fractures, which can bleed massively from branches of the internal iliac arteries, transcatheter arterial embolization is often the first-line treatment for arterial hemorrhage.²⁴

Technical success rates range from roughly 74 to 100 percent across published series.²⁵ A systematic review found efficacy rates between 81 and 100 percent, with about 10 percent of patients needing a repeat procedure. Mortality from uncontrolled intrapelvic bleeding after embolization was around 6 percent, and complication rates were very low.²⁶ For patients who are too unstable for open surgery or whose bleeding source is difficult to access surgically, embolization can be lifesaving.

The Overnight Read and How Often It Goes Wrong

At most teaching hospitals, after-hours imaging is initially interpreted by radiology residents, with an attending radiologist reviewing the studies the next morning. The gap between the preliminary report and the final attending read is a well-studied vulnerability. The major discrepancy rate, meaning an error that could change patient management, hovers around 1.7 to 2.6 percent across studies.²⁷²⁸

Discrepancy rates are not uniform. CT scans have higher discrepancy rates than plain X-rays or ultrasound, and musculoskeletal injuries are missed more often than other pathology. More junior residents have higher error rates than their senior peers, with one study showing the rate dropping from about 2 percent in the first two years of training to about 1.2 percent in the fourth year.²⁹ For spinal cord and nerve compression, where the stakes are highest, third- and fourth-year residents performed well, reaching roughly 98 percent sensitivity for detecting compression on MRI.³⁰

Robust systems for communicating critical findings help close this gap. Best practice calls for direct closed-loop communication for high-acuity findings, meaning the radiologist contacts the treating physician directly and confirms the message was received.³¹ Some institutions have developed secure-chat systems embedded in the electronic medical record to make this communication faster and more traceable.³²

Turnaround Time and Why It Matters

How long it takes from the moment a CT is ordered to the moment the clinician sees the report has a direct impact on emergency department flow. One quality-improvement initiative found that total CT order-to-result turnaround time dropped from about 157 minutes to 130 minutes after workflow changes, a reduction of roughly 17 percent. The time from scan completion to report dropped as well, from 84 to about 73 minutes.³³ The downstream effect on total emergency department length of stay was modest, about a 4 percent reduction, because imaging is only one piece of the puzzle. But in a department seeing hundreds of patients a night, even small gains compound.

The implementation of round-the-clock radiology coverage at an academic trauma center demonstrated the larger version of this effect: average resuscitation unit length of stay fell from about 11 hours to about 8 hours, a 26 percent drop, once attending radiologists were available 24/7 rather than relying solely on preliminary reads.³⁴

Cumulative Radiation in Frequent Visitors

One concern that rarely gets discussed with patients is the radiation burden that accumulates in people who visit the emergency department repeatedly. A study tracking frequent ED visitors over about eight years found that the median cumulative CT dose was 91 millisieverts, with some patients receiving as much as 579 millisieverts across dozens of scans. The estimated lifetime attributable cancer risk for the most heavily scanned patients was roughly one in 17.³⁵ A separate study focused on young adults with recurrent abdominal pain found that about 41 percent of those who had multiple CT scans exceeded a cumulative dose of 50 millisieverts, with most of that dose accumulating within a single year.³⁶

These numbers do not mean that any individual CT scan is dangerous. The risk from a single emergency scan is extremely small and almost always justified by the clinical question. The concern is systemic: no one is tracking the total across hospitals, and patients with chronic conditions, sickle cell crises, kidney stones, or recurrent pancreatitis can accumulate doses without any single provider realizing how high the total has gotten. Dose-tracking registries and clinical decision support that flags high cumulative exposure are gaining traction but are far from universal.

Imaging Pregnant Patients

Imaging during pregnancy creates an understandable tension. Physicians worry about fetal radiation exposure and tend to delay or avoid CT, sometimes at the cost of missing serious maternal diagnoses. A multi-institutional survey found that while most physicians rated themselves as fairly or very confident in their knowledge of emergency imaging protocols for pregnant patients, the actual correct response rate on a knowledge assessment was only about 69 percent, with several questions scoring below 50 percent. Radiologists outperformed other specialties, but gaps existed across the board.³⁷

The reality is that most diagnostic CT exams deliver fetal doses well below the threshold for deterministic harm, and major professional societies endorse the position that a medically indicated scan should not be withheld solely because of pregnancy. The bigger risk is often the missed diagnosis. The stepwise appendicitis algorithm described earlier, starting with ultrasound and using low-dose CT only when needed, is a good model of how to balance these concerns without leaving a serious condition undiagnosed.

AI Tools at the Reading Station

Artificial intelligence is increasingly present in emergency radiology, most commonly as triage software that flags studies likely to contain critical findings so they move to the top of the reading queue. For detecting intracranial hemorrhage on brain CT, one study found that AI assistance significantly improved the sensitivity of inexperienced readers, from about 59 percent to 73 percent. But that gain came with a tradeoff: specificity dropped, meaning more false alarms. For experienced radiologists, the algorithm made no measurable difference in sensitivity, specificity, or overall accuracy.³⁸

That pattern, AI helping novices more than experts, has shown up repeatedly in medical imaging research. The technology is genuinely useful as a safety net for off-hours coverage and for ensuring that an obvious finding does not sit in a queue for an hour before someone looks at it. But the idea that AI will replace the emergency radiologist anytime soon does not match the evidence. Experienced readers already perform at a level the algorithms have trouble exceeding, and the clinical judgment required to integrate imaging findings with patient history and physical exam remains firmly human territory.

Contrast Reactions and Preparedness

Many emergency CT scans require intravenous contrast dye, and while allergic-like reactions are uncommon, severe ones can be life-threatening. The key preparedness measures include making sure staff are trained to recognize anaphylaxis, that intramuscular epinephrine autoinjectors are immediately available at every scanner, and that visual aids for managing reactions are posted where technologists and nurses can reference them quickly. An important nuance is distinguishing allergic-like reactions, which require epinephrine, from physiologic reactions like vasovagal episodes or anxiety responses, which usually need only supportive care.³⁹ The most common pitfall is hesitating to give epinephrine when a reaction is clearly escalating. In the context of emergency radiology, where patients are already acutely ill and the scan cannot wait, getting contrast reactions right the first time is non-negotiable.