MSCT tackles acute chest pain in emergency room

February 7, 2005

As a one-stop scanning tool for everything from abdominal trauma to chest pain, CT is fast becoming an emergency room's best friend. Medical centers save trauma treatment costs by taking less staff to support a patient until a diagnosis is made, and they improve emergency care by cutting the time patients spend being imaged before being stabilized and acutely managed.

As a one-stop scanning tool for everything from abdominal trauma to chest pain, CT is fast becoming an emergency room's best friend. Medical centers save trauma treatment costs by taking less staff to support a patient until a diagnosis is made, and they improve emergency care by cutting the time patients spend being imaged before being stabilized and acutely managed.

CT provides exceptional detail. Tiny, scattered opacities that might have suggested pneumonia on older scanners have been supplanted by "tree and bud" signs throughout the lungs that show the true extent of the inflammatory disease process.

CT is fast replacing some standard emergency imaging procedures. Because a chest CT scan can rule out an aortic injury in eight seconds, it leads to a diagnosis and triage decision far more quickly and less invasively than a two-hour angiographic exam. CT is less cumbersome than arteriography or V/Q scans, which often require transporting patients to the catheterization lab or the nuclear medicine department and waiting for the vascular or imaging team to assemble.

The introduction of 64-slice scanners may revolutionize vascular and cardiac imaging in the emergency department. Dr. Theodore Dubinsky, an assistant professor of radiology at the University of Washington, Seattle, is convinced that radiologists can see myocardial infarcts on current CT scans. In the future, when scans are gated, CT could be used to identify or rule out a coronary cause for acute chest pain. Multislice CT will open the door to imaging the heart in diastole with high resolution of the muscle and coronary vessels. Common acute conditions of the chest, such as pulmonary embolism and aortic dissection, can be checked in one scan.

"The CT study for acute chest pain patients is going to be a fundamental change in how we manage these patients," Dubinsky said. "As it is now, clinicians have to pick: Do they think a patient has an aortic dissection, a pulmonary embolism, or a myocardial infarction? If they suspect an MI, they put an electrocardiogram on the patient and draw bloods, but ECGs are only 60% to 65% sensitive for MI. I have a feeling CT is going to be better. We've seen MIs in patients who have negative ECGs and normal serologies."


Cardiovascular CT is routine for evaluating both nontraumatic and traumatic vascular emergencies, including abdominal aortic aneurysm, aortic dissection, inflammatory disease of the blood vessels, thromboembolism, and aortic injury. Today's scanners can scour the lungs for pulmonary emboli and move right down the body to the pelvis and legs to search for venous thromboemboli in the pelvic veins or extremities. Sixteen-slice machines can traverse the body from the top of the head to the femur in two minutes, acquire 500 or so slices, and produce 3D models of the blood vessels in the head and neck, chest, and abdomen. The machines can clearly and accurately evaluate up to 90% of all aortic traumas, said Dr. Robert Novelline, a professor of radiology and director of emergency radiology at Massachusetts General Hospital.

Evaluation of chest pain by directly imaging the coronary arteries is just beginning. In one case report, Dubinsky and his colleagues at Harborview Medical Center wrote that MSCT revealed a distinct area of hypoperfusion in the lateral ventricular myocardial wall of a 55-year-old man who had an indeterminate ECG after being brought to the emergency department because of left-side chest pain.

Even older scanners can play a role. In a study of 28 patients who had been admitted to an emergency department in Japan because of chest pain consistent with acute myocardial ischemia but nondiagnostic ECG changes and normal serum troponin T levels, four-slice CT detected all 19 cases of angiography-confirmed acute coronary syndrome. Scans performed within six hours of the onset of chest pain correctly identified 50% or greater stenosis in major coronary arteries in the presence of noncalcified plaque in all 19 patients. It also detected nontransmural perfusion defects in four patients and produced only one false-positive result. Authors of the study, which was presented at the 2004 RSNA meeting, concluded that four-slice CT was 100% sensitive and 89% specific for identifying and triaging patients with acute coronary syndrome in the emergency department.

On the way to developing a one-stop MSCT protocol for determining cardiac, pulmonary, and musculoskeletal causes of acute chest pain, Dr. Charles White is working on what he calls a hybrid plan for assessing patients with intermediate levels of chest pain. These patients experience something between the two extremes of chest pain-severe, clearly heart-related pain and troubling but minor or musculoskeletal discomfort.

The protocol is a compromise between scanning for pulmonary embolism and conducting a specific study of the coronary arteries. A coronary artery study is best achieved with a small field-of-view that is honed down to the heart, with volume coverage extending from the top of the heart to the bottom and cardiac gating, according to White, director of thoracic imaging at the University of Maryland, Baltimore. This approach does not include much of the lungs, however. A pulmonary embolism study uses higher pitch to traverse the chest more quickly and is not gated.

White's compromise is to perform a gated study that begins at the bottom of the cardiac cavity so scanning will cover the entire heart during a breath-hold. Patients then can begin to breathe and exhale slowly while scanning covers the top of the lungs, where images are less vulnerable to motion. He uses a field-of-view large enough to include the periphery of the lungs and reconstructs images only of the heart.

Although scan acquisition according to this protocol is fairly long, on the order of 35 seconds, it is feasible with a 16-slice scanner, White said. A study reported at the 2004 RSNA meeting showed that even though the protocol does not fully optimize the coronary arteries, it produces a good enough view of the blood vessels to rule out significant vascular disease. In the study, which included 69 clinically stable patients, CT identified a significant cause of chest pain in 13 (19%) patients, 10 cardiac and three noncardiac in nature. The 16-slice technique directly visualized the coronary arteries and had a high specificity and negative predictive value for atypical chest pain. CT had a sensitivity of 87%, specificity of 96%, positive predictive value of 87%, and negative predictive value of 96%.

"This approach holds out the prospect that if the CT is negative, the suspicion of cardiac chest pain would be low, and workup could stop instead of waiting the extra time and doing extra tests. It could have a huge impact," White said.

The proportion of myocardial infarctions that are "missed" in an emergency department ranges from 2% to 8%, and the death rate for these patients after hospital discharge is about 25%, said Dr. Mel Clouse, head of interventional radiology at Beth Israel Medical Center in Boston. These patients most likely come in with ruptured plaque and a small amount of clot in the coronary vasculature that either causes chest symptoms or blocks blood flow downstream but does not produce a sufficient degree of ischemia to elevate cardiac enzymes. The patients therefore lie around in the emergency department for several hours, receive hydration and pain medication, and leave when they seem better, only to suffer MI when a coronary vessel later completely clots off.

"Rapid scanners have made it possible to get images of the pulmonary arteries, aorta, and coronaries. There's no question where the technology is going and where it will find one of its important uses-imaging individuals who have chest pain symptoms at the outset. But it will take a little doing to get the protocol worked out so that your timing and the scanner don't produce artifacts," Clouse said.

Clouse is still tinkering with the specifics of his protocol. He has tried setting Hounsfield units for the main pulmonary artery at 180, moving to just above the aortic arch, and gating the study all the way down. This takes about 24 seconds and clearly demarcates the vasculature but fails to anatomically detail lung parenchyma. He is experimenting with other alternatives, such as beginning at the top of the chest and performing nongated imaging through the chest, then moving to the coronary root and switching to gated imaging through the heart. He considers slice thickness critical.

"To the uninitiated eye, the difference between a 0.5- and 1-mm slice is not all that great. But if you do a 0.5-mm slice scan on an individual with a bypass graft in the leg, the difference between 0.5 and 1 mm is really enormous because you can see the actual ribbing on the graft," Clouse said.

Before CT chest pain imaging becomes routine, improvements also will be needed in reconstruction and postprocessing. White's protocol reconstructs 10 phases from the heart to capture the best phase of imaging of the coronary arteries and perform some functional images.

Although reconstruction algorithms have been streamlined, reconstructing a study consisting of 2000 or more slices can take 20 minutes or longer, which ties up the scanner if reconstructions are done on the operator's console and makes emergency department physicians wait for a diagnostic answer. Hardware changes that speed up reconstruction or shift the process offline may allow reconstructions to be done on an adjacent workstation without interrupting work flow, White said.

Ideally, emergency cardiovascular CT would incorporate one-touch postprocessing so that the 3000 images entering a workstation will generate data that allow complete assessment of the heart, including a reconstruction of the coronary arteries, measurement of ejection fraction and perhaps left ventricular perfusion, a cine loop assessment of wall motion, and even a calcium score.

Postprocessing can take a half-hour or longer, however, just to get images of the coronary arteries and pick out the proper phase while evaluating whether the arteries are stenotic. Measurement of ejection fraction takes time because radiologists have to manually segment the endocardial portion of the endocardium in systole and diastole. Postprocessing remains so slow it is a barrier to a full CT cardiac workup, White said.

Other limitations include the sensitivity of scanners to motion, especially for finding small aortic intimal injuries and peripheral emboli, Dubinsky said. Because radiologists are seeing the interior of the human body in more and more detail, they are now plagued by subtle artifacts that could be ignored in the past as well as any kind of motion-related misregistration.


Although 32-slice CT scanners can cut through the chest in less than 30 seconds and provide high-quality images of the coronary arteries, the increased speed of 64-slice scanners will counteract motion artifacts from patients' breathing and allow gating of all chest studies. With 64-slice scanners, image acquisition will be so fast radiologists should be able to do a scan while the pulmonary arteries as well as the aorta are opacified and the heart and coronary arteries are well visualized.

"Many of our emergency patients are hyperventilating and have a rapid heart rate, so things are moving. The 64-slice scanners should be able to decrease motion artifact from cardiac, aortic pulsatile, and respiratory movement," Novelline said.

In addition to scanning through tissue volume more quickly, 64-slice scanners may increase temporal resolution by decreasing the window of data. Because 64-slice scanners have a gantry rotation of 0.4 seconds or less, they will push the limit and reduce the window of data to 0.05 seconds, which will reduce motion effects and perhaps eliminate the need for beta blockers, White said.

"When the heart rate is faster, it keeps going from systole to diastole at a fast rate, so you need better temporal resolution. The goal of the CT scanner is to decrease the temporal window so there is very little blurring. One way to overcome current scanners' limitations is by giving beta blockers to slow the heart rate down and make it more tolerant to a wider window of data. The other way is to make the temporal window smaller. Some of the reconstruction advantages of the 64-slice machines will help get that smaller window, and so will a faster gantry speed," White said.

The scanners will provide more options for image manipulation, Novelline said. With 64-slice machines, radiologists will be able to sandwich tissue sections and look at two or three sections together or stack slices in coronal or sagittal reformations, all in postprocessing when the patient is off the table.

The thin slices possible on advanced CT scanners should enhance visualization of the distal coronary arteries. The size of the proximal coronary arteries is less than 4 mm, and the diameter of the distal arteries is around 2 mm. A slice thickness of 0.5 mm may make all the difference in finding the smaller distal vessels, White said.

The advantages of 64-slice CT would appear to set the stage for a single noninvasive study of cardiac causes of acute chest pain. It is not clear whether MSCT will define ischemia from infarction. If it does, it may serve as a triage point for sending individuals with a single lesion in the left anterior descending artery to angioplasty and those with a stent or multivessel disease to surgical bypass.

"At some point, we may collect enough data to say that CT works and to show how we can use it for chest pain evaluation. Now it's still the big frontier," Dubinsky said.