AJR:185, August 2005 533

AJR 2005; 185:533–540

0361–803X/05/1852–533

© American Roentgen Ray Society

White et al.Emergency Department Evaluation of Chest Pain

C T • O r i g i n a l R e s e a rc h

Chest Pain Evaluation in the Emergency Department: Can MDCT Provide a Comprehensive Evaluation?

Charles S. White1

Dick Kuo2

Mark Kelemen3

Vineet Jain1

Amy Musk1

Eram Zaidi1Katrina Read4

Clint Sliker1

Rajnish Prasad3

White CS, Kuo D, Kelemen M, et al.

Received March 1, 2005; accepted after revision April 22, 2005.

Supported by a grant from Philips Medical Systems.

1Department of Diagnostic Radiology, University of Maryland School of Medicine, 22 S Greene St., Baltimore, MD 21201. Address correspondence to C. S. White (cwhite@umm.edu).

2Division of Emergency Medicine, Department of Surgery, University of Maryland School of Medicine, Baltimore, MD 21201.

3Division of Cardiology, Department of Internal Medicine, University of Maryland School of Medicine, Baltimore, MD 21201.

4Clinical Scientist, Philips Medical Systems, Department of Radiology, University of Maryland School of Medicine, Baltimore, MD.

OBJECTIVE. The purpose of our study was to determine whether MDCT can provide acomprehensive assessment of cardiac and noncardiac causes of chest pain in stable emergencydepartment patients.

SUBJECTS AND METHODS. Patients with chest pain who presented to the emer-gency department without definitive findings of acute myocardial infarction based on history,physical examination, and ECG were recruited immediately after the initial clinical assess-ment. For each patient, the emergency department physician was asked whether a CT scanwould normally have been ordered on clinical grounds (e.g., to exclude pulmonary embolism).Each consenting patient underwent enhanced ECG-gated 16-MDCT. Ten cardiac phases werereconstructed. The images were evaluated for cardiac (coronary calcium and stenosis, ejectionfraction, and wall motion and perfusion) and significant noncardiac (pulmonary embolism, dis-section, pneumonia, and so forth) causes of chest pain. Correlation was made between the pres-ence of significant cardiac and noncardiac findings on CT and the final clinical diagnosis basedon history, examination, and any subsequent cardiac workup at the 1-month follow-up by aconsensus of three physicians.

RESULTS. Sixty-nine patients met all criteria for enrollment in the study, of whom 45(65%) would not otherwise have undergone CT. Fifty-two patients (75%) had no significantCT findings and a final diagnosis of clinically insignificant chest pain. Thirteen patients (19%)had significant CT findings (cardiac, 10; noncardiac, 3) concordant with the final diagnosis. CTfailed to suggest a diagnosis in two patients (3%), both of whom proved to have clinically sig-nificant coronary artery stenoses. In two patients (3%), CT overdiagnosed a coronary stenosis.Sensitivity and specificity for the establishment of a cardiac cause of chest pain were 83% and96%, respectively. Overall sensitivity and specificity for all other cardiac and noncardiaccauses were 87% and 96%, respectively.

CONCLUSION. ECG-gated MDCT appears to be logistically feasible and shows prom-ise as a comprehensive method for evaluating cardiac and noncardiac chest pain in stable emer-gency department patients. Further hardware and software improvements will be necessary foradoption of this paradigm in clinical practice.

iagnosing the cause of acute chestpain in the emergency departmentremains a formidable task becauseof extensive etiology that ranges

from benign to potentially lethal. The evalua-tion of many of these conditions, particularlythe conventional assessment for the presenceof cardiac disease in the acute setting, is ofteninconclusive and may require further invasivetesting. MDCT has been shown to be effectivefor the delineation of many causes of chestpain that may be inapparent on initial clinicalor radiographic evaluation, including pneumo-nia, aortic dissection, and pulmonary embo-lism [1, 2]. More recently, promising results

have been obtained with the use of MDCT inthe evaluation of coronary artery stenosis [3,4]. Our study was undertaken for two pur-poses: to determine the logistic feasibility ofusing MDCT in the emergency department set-ting to assess chest pain, and to assess sensitiv-ity and specificity of the MDCT technology inproviding an evaluation of cardiac and noncar-diac causes of acute chest pain in stable emer-gency department patients.

Subjects and MethodsPatients

Patients presenting in our emergency depart-ment with acute chest pain between November

D

White et al.

534 AJR:185, August 2005

2003 and July 2004 were approached for recruit-ment in this prospective study. Acute chest pain inthe University of Maryland emergency departmentis typically classified on a scale of 1 to 5 on the ba-sis of the initial clinical impression. These are clin-ical categories based on the patient’s chief com-plaint, symptoms and signs, risk factors, and ECG.Category 1 is an acute myocardial infarction. Cate-gory 2 is considered definite angina with uncer-tainty regarding acute myocardial infarction. Cate-gories 3, 4, and 5 are probable angina, probably notangina, and not angina, respectively. Patients inwhom there is less concern for angina but suspicionof clinically significant noncoronary chest pain(e.g., pulmonary embolism) are usually graded ascategory 3 or 4 rather than category 5, conferring anonanginal component to the classification system.These are not final diagnoses but rather simple,practical, and functional categories along a contin-uum adopted broadly by emergency physicians andadvocated by the Society of Chest Pain Centers [5].

Clinically stable patients with chest pain classi-fied from 2 to 4 were eligible for inclusion based onthe assessment of the emergency department phy-sician and other inclusion and exclusion criteria. In-clusion criteria included age greater than 18 yearsand the ability to understand and sign a consentform. Exclusion criteria were clinical instability(such as arrhythmia, congestive heart failure, andhypotension), compromised renal function, aller-gies to contrast material, and pregnancy. Category1 patients were excluded because of clinical insta-bility. Category 5 patients were deemed unlikely tohave a significant cause of chest pain and were alsoexcluded. The emergency department physicianwas asked in each case before the study whether CTwould have been performed for the conventionalworkup. The CT study was done early in the clini-cal evaluation of the patient in the interval immedi-ately after the ECG was done and blood samplesdrawn and before a decision was made as to furthercare or studies for the patient and before results ofcardiac enzymes were available. Patient demo-graphics are summarized in Table 1.

Seventy-eight patients with chest pain were en-rolled. Nine patients were excluded before com-pleting the protocol. In four patients, the CT wasperformed but the raw projection data were erasedfrom the scanner hard drive before multiple phasesof the cardiac cycle could be reconstructed. In threepatients, no CT was performed because the patientswere undergoing evaluation elsewhere and couldnot be scheduled for CT. The other two patients leftthe emergency department against medical advicebefore CT scanning. Ultimately, 69 patients partic-ipated in the chest pain protocol.

The study protocol was approved by our institu-tional review board (IRB) and informed consent

was given by each patient. All Health InsurancePortability and Accountability Act procedureswere followed. As part of the study protocol, all pa-tients enrolled in the study received a consultationwith a cardiologist.

Image AcquisitionCT scans were acquired on each patient using a

16-MDCT scanner (MX8000IDT, Philips MedicalSystems). The CT scanner is located adjacent to theemergency suite, and thus patients could be moni-tored as necessary by emergency department per-sonnel while undergoing CT. For each patient, ret-rospective ECG-gated images were obtainedthrough the entire chest during a single breath-holdbeginning at the inferior margin of the heart and ex-tending to the top of the lung apices. Patients wereadvised to exhale slowly if they could not maintainbreath-holding throughout the scanning. The scan-ning protocol included collimation of 0.75 × 16 mmwith reconstructed axial image thickness of 1 mm.Scanning technique was 140 kVp and 350–500mAs. A pitch of 0.2–0.3 was used with a scannerrotation time of 0.42 sec. Iodinated contrast mate-rial, 120–150 mL, was injected through an 18- to

20-gauge angiocatheter into an antecubital vein at3–4 mL/sec. Automated bolus timing was per-formed using a threshold value of 150 H and a re-gion of interest placed over the ascending aorta. Af-ter additional IRB approval was obtained, a β-blocker (Lopressor [metoprolol, Novartis], 5 mgIV) was used to control heart rates greater than 70beats per minute. Of the final 32 patients, 10 re-ceived β-blockers.

The average scanning time was 30 sec with anadditional 3–4 min for preprocedure placement andadjustment of the ECG leads. A large field of view(350–400 mm) was used to encompass the entirechest at 75% of the R-R interval. The raw data wereused to reconstruct axial images as well as coronaland sagittal maximum-intensity-projection refor-mations. Axial images were also subsequently re-constructed in a smaller field of view (250 mm) tar-geted to the heart. These reconstructions wereobtained at 10% intervals for a total of 10 phases.

Image InterpretationAn initial assessment was made of noncardiac

disease and for the contrast-enhanced presence ofcoronary artery calcification using the large field-of-view images (ECG-gated 75% R-R interval).Specific noncardiac entities that were evaluated in-cluded, but were not limited to, aortic dissection,pulmonary embolism, pneumonia, pneumothorax,pericardial effusion, and rib fracture. A qualitativeassessment of the presence and extent of coronaryartery calcifications was made. No quantitativesoftware was used because all scans were obtainedafter the administration of IV contrast material.Coronary artery calcification was characterized asnone, mild, moderate, or severe. All of this infor-mation was immediately communicated to the clin-ical team in the emergency department using a pre-liminary report form (Fig. 1A). This abbreviatedinitial evaluation was necessitated by the time re-quired to reconstruct the 10 cardiac phases consist-ing of 2,500–3,500 images, which ranged from 15min if all images were reconstructed immediatelyto several hours, depending on other duties of thetechnologists and the clinical demands made on theCT scanner.

All imaging data were processed by a fellow re-ceiving training in thoracic radiology using a dedi-cated workstation (MxView, Philips Medical Sys-tems). Postprocessing consisted of the calculation ofejection fraction and the generation of curved planarreconstructions of each of the main coronary arter-ies. The 10 cardiac phases were loaded into a cineloop, and wall motion was assessed qualitatively forareas of hypo- or akinesis and to determine the end-diastolic and end-systolic phases for analysis of theejection fraction. The endocardial margins on thesephases were drawn manually on contiguous short-

TABLE 1: Demographics of 69 Patients with Chest Pain

Demographic Variable No. %

Patient age (yr)

Mean 51

Range 33–81

Patient sex

Male 35 51

Female 34 49

Chest pain evaluation category

2 17 25

3 32 46

4 18 26

None assigned 2 3

Smoker (current or former) 48 70

Nonsmoker 21 30

CHD risk factors

0 6

1–2 25

3–4 26

≥ 5 12

Body mass index

< 25 15

25–30 13

> 30 18

Not obtained 23

Note—CHD = coronary heart disease.

Emergency Department Evaluation of Chest Pain

AJR:185, August 2005 535

axis images. Papillary muscles were excluded fromcontouring. Automated software summated the sur-face areas of these images to calculate the ejectionfraction according to the formula:

ejection fraction = (end-diastolic volume –end-systolic volume) / end-diastolic volume.

Short- and long-axis images were assessed for per-fusion defects on each phase and were defined asareas of decreased perfusion that could be visual-ized in two projections.

Construction of curved planar reconstructions ofthe coronary arteries was based on the optimal car-diac phase, defined as that with the least amount ofmotion. The axial and postprocessed images wereassessed for the presence of stenosis, which wasquantified subjectively as being greater (signifi-cant) or less than 50% (nonsignificant). All imag-ing observations were agreed on by consensus ofthe two radiologists. Because the final assessmentwas typically completed after patient triage, coro-nary stenoses or other significant findings identi-fied after postprocessing were communicated to the

cardiology consultant. A final CT assessment of thecause of the chest pain was recorded on a case re-port form (Fig. 1B).

Clinical Follow-Up and AssessmentEmergency department data and all available

medical records were reviewed for each patient be-tween 1 and 2 months after the emergency depart-ment visit. Data were collected as to whether thepatient left against medical advice from the emer-gency department and whether the patient requiredeither hospitalization or a subsequent emergencydepartment visit. In addition to the CT, informationon other relevant diagnostic tests was recorded, in-cluding coronary angiography, stress echocardio-graphy, or radionuclide stress testing obtainedwithin 1 month of presentation to the emergencydepartment.

A consensus group consisting of one emergencydepartment physician, one cardiologist, and one ra-diologist was convened to determine a final diagno-sis. The consensus group used the following guide-lines to adjudicate each case: For patients who weredischarged from the emergency department and

who did not receive further testing, the diagnosisprovided at the time of discharge from the emer-gency department by the emergency departmentphysician was deemed definitive. For patients whohad coronary angiography, stress echocardio-graphy, or radionuclide stress testing, the results ofthese tests were used to arrive at a final judgment.Positive coronary angiography was defined asshowing a stenosis greater than 50% in a major ves-sel. For clinically significant diagnoses for whichCT is considered a standard reference technique(e.g., pulmonary embolism, aortic dissection), theCT findings were regarded as definitive. Usingthese guidelines based on the best available clinicaland testing information, a final diagnosis for theemergency department visit was determined. Fi-nally, results of invasive coronary angiography,stress echocardiography, and radionuclide stresstesting were correlated with their respective find-ings on CT angiography.

Statistical AnalysisSensitivity, specificity, positive predictive

value, and negative predictive value were calcu-

A B

Fig. 1—Radiology report forms.A and B, Initial (A) and final (B) case report forms. Final form also included a coronary artery scoring sheet (not shown).

White et al.

536 AJR:185, August 2005

lated for the emergency department CT using the fi-nal clinical diagnosis as the reference standard.Separate values were calculated for the diagnosis ofcardiac chest pain only and for the full assessmentof cardiac and noncardiac diagnoses. For differ-ences between ejection fractions as calculated onMDCT and radionuclide testing, an unpaired t testwas used. A p value of less than 0.05 was consid-ered to indicate statistical significance.

ResultsOverall Clinical Assessment

The 69 patients who completed the chestpain protocol included 35 men (51%) and 34women (49%) with a mean age of 51 years(range, 33–81 years). Forty-five patients(65%) would not otherwise have undergoneCT, according to the judgment of the emer-gency physician caring for the patient. Seven-teen patients (25%) presented with chest painthat was classified as category 2, 32 (46%)with category 3, and 18 (26%) with category4. In two patients (3%), no category was as-signed by the emergency physician. Twenty-nine patients (42%) required hospital admis-sion for further evaluation, including 13(76%) of those with category 2 pain, 11(32%) with category 3 pain, and three (17%)with category 4 pain. Both patients for whomno category was assigned were also admitted.One patient was ruled in for an acute myocar-dial infarction and a second was diagnosedwith acute coronary syndrome.

Fifty-two (75%) of the 69 patients had nosignificant CT findings and a final diagnosisof clinically insignificant chest pain (Fig. 2).Thirteen patients (19%) had significant CTfindings concordant with the final diagnosis(cardiac, 10; noncardiac, 3). Each of the 10patients with cardiac disease and a positiveCT diagnosis was deemed to have a diagnosisof angina due to coronary artery disease(Fig. 3). The three noncardiac diagnoses werepericarditis with a moderately large pericar-dial effusion, subtle pneumonia, and pulmo-nary embolism (Figs. 4 and 5). In two patients(3%), CT failed to suggest a clinically signif-icant diagnosis. Both of those were clinicallysignificant coronary artery stenoses identifiedon angiography in the left anterior descendingand first diagonal branches, respectively. Inboth patients, image quality was adversely af-fected by motion (Fig. 6). In two additionalinstances (3%), CT overdiagnosed a coronarystenosis. Both incorrectly identified lesionswere in the mid left anterior descending ar-tery. The decision on final diagnosis wasbased on clinical data in 34 patients (49%),

radionuclide testing in 15 (22%), coronaryangiography in 11(16%), stress echocardio-graphy in six (9%), and CT alone in three(4%). The diagnoses made on CT alone werethose with noncardiac causes, for which CT isa standard reference technique.

The sensitivity of CT as compared with thefinal diagnosis for coronary artery diseasewas 83%. Specificity, negative predictivevalue, and positive predictive value were96%, 96%, and 83%, respectively. Overall,the sensitivity of cardiac and noncardiac con-ditions (including CT-based diagnoses) was87%. Specificity, negative predictive value,and positive predictive value were 96%, 96%,and 87%, respectively.

Coronary Artery CalcificationQualitative analysis of the extent of coro-

nary artery calcification on the contrast-en-hanced study revealed no calcification in 42patients. Mild, moderate, or severe amountsof calcification were present in 19, four, andfour patients, respectively. Of the 10 patientswith significant coronary artery disease foundon CT, coronary artery calcification wasgraded qualitatively as mild in one, moderatein six, and severe in three. Of the two patientswith negative CT findings who proved tohave coronary artery disease, one had no vis-ible coronary artery calcification and the sec-ond had mild coronary artery calcification.

Functional AssessmentTwenty-one patients underwent stress nu-

clear medicine testing with calculation ofejection fraction within 1 week of the MDCT.MDCT yielded a significantly higher ejectionfraction (mean, 63%; range, 47–82%) thanthe radionuclide study (mean, 52%; range,36–63%) in these patients (p < 0.01). Twelvepatients had a difference of 10% or less be-tween the radionuclide stress and MDCTejection fraction, five and four patients haddiscrepancies of 11–20% and 21–30%, re-spectively. One patient showed a focal perfu-sion defect at the cardiac apex on CT (Fig. 7).A wall motion abnormality was identified ina similar location on stress echocardiography.

ComplicationsNon-life-threatening complications that

were ascribed to the trial protocol occurred intwo patients. One 40-year-old woman devel-oped urticaria shortly after discharge from theemergency department that was presumablycaused by IV contrast material. She returnedto the emergency department and was given

oral diphenhydramine, with prompt resolu-tion of symptoms. A 54-year-old man devel-oped renal insufficiency after the contrast-en-hanced CT and required 2 days of additionalhospitalization. He was discharged with base-line renal function.

DiscussionSubstantial advances have occurred in pa-

tient evaluation and triage over the past de-cade, but the assessment of chest pain in theemergency department remains a significantchallenge. Although a cardiac or noncardiacdiagnosis may be immediately apparent, ini-tial clinical evaluation is often equivocal, re-sulting in a high proportion of hospital admis-sions [6]. Nevertheless, it is estimated that4–8% of patients are inappropriately dis-charged from the emergency department andultimately prove to have a myocardial infarc-tion, the most important cause of acute chestpain [7, 8].

In the emergency department, the diagno-sis of acute cardiac ischemia, which includesacute myocardial infarction and unstable an-gina, remains primarily clinical, and is guidedby history, risk factors, and ECG results. Thisdiagnostic pathway is known to lack sensitiv-ity [9, 10]. Serum markers of myocardial in-jury (CK-MB [creatininekinase myoglobin]troponin, and myoglobin) also are a criticalpart of chest pain assessment but typically donot show elevations until more than 6 hr afterthe onset of the chest pain [11]. Thus, theymay not be useful in the hyperacute settingwhen the administration of thrombolyticsmight lead to maximal preservation of myo-cardial tissue. Moreover, these markers donot allow rapid exclusion of myocardial is-chemia, which would permit early dischargefrom the emergency department.

Other imaging-based diagnostic strategieshave been attempted to assess cardiac causesof chest pain. Echocardiography with thepatient at rest and after stress can show wallmotion abnormalities due to ischemia and canassess valvular, pulmonary artery, and peri-cardial disease [12–14]. However, the tech-nique is operator dependent and requiresconsiderable experience. Moreover, pooracoustic windows may limit the study [12].Nuclear scintigraphy using thallium ortechnetium-99m sestamibi may detect abnor-malities of myocardial perfusion, but this mayreflect remote infarction. These techniqueshave a high sensitivity and moderate specific-ity [15]. An important additional limitation isthe need to transport the patient to a gamma

Emergency Department Evaluation of Chest Pain

AJR:185, August 2005 537

A B

Fig. 2—48-year-old man who presented to emergency department with chest pain and normal coronary arteries by CT angiography.A and B, Curved planar reformations from MDCT show right coronary artery (A) and left anterior descending artery (B).

A B

Fig. 3—52-year-old man who presented to emergency department with chest pain.A, Curved planar reformation from CT angiogram from MDCT shows area of proximal left anterior descending artery (LAD) stenosis (arrow).B, Coronary angiogram confirms LAD stenosis (arrow).

White et al.

538 AJR:185, August 2005

camera, which is often in a remote location[13]. MRI is often impractical because of theneed for specialized equipment that may notbe available in the emergency department anda substantial prevalence of claustrophobia[16]. Each of these imaging techniques is alsolimited in its capability to detect extracardiaccauses of chest pain.

CT, in particular electron beam CT, hasbeen used to risk-stratify patients with acutechest pain by revealing coronary calcium [17].The absence of calcification is associated witha very low likelihood of acute cardiac is-chemia. The latest generation of MDCT scan-ners features ECG-gating, submillimeter spa-tial resolution, and relatively good temporalresolution that permit adequate assessment ofcoronary artery anatomy [18, 19].

Our pilot study shows that MDCT is a fea-sible approach to provide a comprehensivechest pain evaluation in the emergency depart-ment. We selected patients with chest pain lev-els of 2 to 4 who were deemed to be clinicallystable. It is this group of patients without defi-nite evidence of myocardial infarction inwhom MDCT may have its greatest impact.None of our patients ultimately proved to havemyocardial infarction or unstable angina.Presumably this is because of our restrictive in-clusion criteria and the low prevalence of my-ocardial infarction (~ 5%) that has been docu-mented in our emergency department. This

low prevalence indicates the potential impactof the MDCT protocol, if completed expedi-tiously, to rapidly triage patients, particularlythose with a negative study.

In our study, several patients had chest paincaused by coronary artery disease, andMDCT showed coronary artery calcificationor areas of coronary stenosis in most of thesepatients. However, in two patients, an area ofsignificant coronary artery narrowing on cor-onary arteriography was not detected on CT.In part, this deficiency may have been due tothe way the CT scan was acquired. Our proto-col entailed a global evaluation of chest painand thus represented a necessary compromisebetween evaluating the coronary arteries andthe remainder of the chest. We used a largerfocal spot and larger field of view than aretypical for a dedicated coronary artery evalu-ation, and we subsequently reconstructed asmaller field of view centered around theheart. In addition, scanning was initiated atthe bottom of the heart and extended cephaladfor an uninterrupted acquisition through theentire chest, in contrast to the typical coro-nary artery protocol that progresses inferiorlyfrom a level just above the coronary arteries.These modifications undoubtedly led to somedegradation of coronary artery images. More-over, β-blockers were not used for the initialportion of the study. Despite the limitations ofthe technique, MDCT was able to diagnose

clinically significant cardiac and noncardiaccauses of chest pain in most cases.

Other aspects of the MDCT cardiac evalu-ation were less valuable. The assessment ofcoronary artery perfusion after enhancementwas quite subjective. No quantitative thresh-old has been established to define such a de-fect. When a defect was conclusively presentin the judgment of the two observers, its chro-nicity could not be determined. Similar con-siderations apply to the evaluation of wallmotion abnormalities. Ejection fraction wasdepressed in a minority of patients, with a rea-sonable correlation with results from nuclearmedicine testing when available.

Although the number of clinically signifi-cant abnormalities was low in this pilot studyof an acute chest pain imaging protocol,MDCT showed the potential to be a valuablemethod for excluding significant cardiogeniccauses of chest pain, including coronary ar-tery stenoses greater than 50%, as evidencedby a high negative predictive value. Com-pletely normal or not significantly abnormalMDCT findings was the most common resultin our series and was confirmed in most casesby the final diagnosis. In addition, MDCTwas valuable for suggesting noncardiac diag-noses such as pneumonia and pulmonary em-bolism. In this respect, MDCT has an advan-tage over other imaging techniques, such asperfusion radionuclide scintigraphy and

Fig. 4—56-year-old woman who presented to emergency department with chest pain. Left lower lobe pneumonia was found on MDCT. No coronary stenosis was identified.

Fig. 5—60-year-old man who presented to emergency department with chest pain. Axial MDCT scan shows pulmonary embolism in right middle lobe pulmonary artery (arrow).

Emergency Department Evaluation of Chest Pain

AJR:185, August 2005 539

A

Fig. 6—60-year-old woman who presented to emergency department with chest pain.A, Curved planar reformation from CT angiogram on MDCT was interpreted as neg-ative. Arrow points to narrowed branch that was not identified prospectively.B, Coronary angiogram shows a 50–60% stenosis (arrow) of diagonal branch.C, Repeat curved planar reformation produced after discrepancy was reported sug-gests presence of stenotic area (arrow) in retrospect.

B C

Fig. 7—56-year-old man who presented to emergency department with history of myocardial infarction (MI) and chest pain. MDCT scan shows endocardial apical perfusion defect (arrows). Patient was asymptomatic at time of scanning with neg-ative acute MI evaluation, and defect was deemed to be chronic.

White et al.

540 AJR:185, August 2005

echocardiography, that are sometimes used inthe emergency department setting.

Several limitations of this study must beemphasized. First, because this was a feasibil-ity study, the number of enrolled patients wassmall. Second, as described earlier, the goalof providing a complete thoracic assessmentthat included pulmonary embolism and aorticdissection necessitated a compromise proto-col that was not optimized to the coronaryarteries. Another technology-related short-coming was that the time required for post-processing necessitated an initial generalevaluation followed by a more detailed car-diac evaluation, often separated by severalhours or more. Thus, a real-time coronary ar-tery assessment proved difficult or impossi-ble. With the advent of the latest generation of64-MDCT scanners, these limitations may bemitigated as a result of faster image acquisi-tion, better spatial and temporal resolution,and more rapid postprocessing.

A fourth limitation was the lack of a stan-dard end point to assess final diagnosis. Asnoted, the patients had variable clinical eval-uations and some had no documented follow-up at our institution after their index emer-gency department visit. In particular, only aminority of patients ultimately underwentcoronary angiography, the anatomic standardof reference. Thus, determination of a finaldiagnosis was necessarily subjective in manyinstances. This shortcoming is common instudies involving emergency department pa-tients [20]. Finally, the design of the study didnot permit assessment of high-risk or less sta-ble patients, thereby limiting the number ofsubjects with clinically significant coronaryartery stenoses. Further studies of this tech-nique will need to assess its performancewhen evaluating patients with a higher likeli-hood of coronary artery disease.

An additional consideration is that CT maynot be the best technique to diagnose certaincauses of chest pain. For instance, muscu-loskeletal chest pain is often best evaluated onphysical examination, gastrointestinal refluxdisease may be assessed with manometrictesting, and pneumonia is often better seen onchest radiography. Thus, although CT has po-tential value in the diagnosis of many life-threatening causes of chest pain, not everycause of chest pain can be diagnosed.

In summary, our pilot study suggests thata comprehensive evaluation of many cardiac

and noncardiac causes of acute chest pain inthe emergency department with MDCT is lo-gistically feasible and may provide clini-cally meaningful data. The greatest potentialimpact appears to be in the exclusion of sig-nificant cardiac disease to supplement as-sessment of established indications in life-threatening noncardiac disease such as pul-monary embolism and aortic dissection.However, in patients with suspected coro-nary disease based on CT, further study isneeded to more clearly elucidate its sensitiv-ity for clinically significant coronary arterystenoses. Moreover, routine implementationof this technique will require further techno-logic advances, such as more rapid scanningand reconstruction, and greater ease of im-age postprocessing. The potential exists thatthese advances will lead to greater reliabilityof delineation of coronary artery anatomyand cardiac function, permitting contempo-raneous clinical feedback of the entireMDCT examination to the emergencydepartment team, thereby facilitating a moreexpeditious triage of the patient withchest pain.

References1. Lawler LP, Fishman EK. Multidetector row com-

puted tomography of the aorta and peripheral arter-

ies. Cardiol Clin 2003; 21:607–629

2. Schoepf UJ, Costello P. Multidetector-row CT im-

aging of pulmonary embolism. Semin Roentgenol

2003; 38:106–114

3. Ropers D, Baum U, Pohle K, et al. Detection of cor-

onary artery stenoses with thin-slice multi-detector

row spiral computed tomography and multiplanar

reconstruction. Circulation 2003; 107:664–666

4. Achenbach S, Ropers D, Pohle K, et al. Clinical re-

sults of minimally invasive coronary angiography

using computed tomography. Cardiol Clin 2003;

21:549–559

5. Tatum JL, Jesse RL, Kontos MC, et al. Comprehen-

sive strategy for the evaluation and triage of the

chest pain patient. Ann Emerg Med 1997;

29:116–125

6. Pozen MW, D’Agostino RB, Selker HP, Sytkowski

PA, Hood WB Jr. Predictive instrument to improve

coronary-care-unit admission practices in acute is-

chemic heart disease: a prospective multicenter

clinical trial. N Engl J Med 1984; 310:1273–1278

7. Lee TH, Rouan GW, Weisberg MC, et al. Clinical

characteristics and natural history of patients with

acute myocardial infarction sent home from the

emergency room. Am J Cardiol 1987; 60:219–224

8. McCarthy BD, Beshansky JR, D’Agostino RB,

Selker HP. Missed diagnoses of acute myocardial

infarction in the emergency department: results

from a multicenter study. Ann Emerg Med 1993;

22:579–582

9. Brush JE Jr, Brand DA, Acampora D, Chalmer B,

Wackers FJ. Use of the initial electrocardiogram

to predict in-hospital complications of acute my-

ocardial infarction. N Engl J Med 1985;

312:1137–1141

10. Lee TH, Rouan GW, Weisberg MC, et al. Sensitiv-

ity of routine clinical criteria for diagnosing myo-

cardial infarction within 24 hours of hospitalization.

Ann Intern Med 1987; 106:181–186

11. Lau J, Ioannidis JP, Balk EM, et al. Diagnosing

acute cardiac ischemia in the emergency depart-

ment: a systematic review of the accuracy and clin-

ical effect of current technologies. Ann Emerg Med

2001; 37:453–460

12. Bholasingh R, Cornel JH, Kamp O, et al. Prognostic

value of predischarge dobutamine stress echocar-

diography in chest pain patients with a negative car-

diac troponin T. J Am Coll Cardiol 2003;

41:596–602

13. Soman P, Bokor D, Lahiri A. Why cardiac magnetic

resonance imaging will not make it. J Comput Assist

Tomogr 1999; 23[suppl 1]:S143–S149

14. Gibler WB, Runyon JP, Levy RC, et al. A rapid di-

agnostic and treatment center for patients with chest

pain in the emergency department. Ann Emerg Med

1995; 25:1–8

15. Kontos MC, Tatum JL. Imaging in the evaluation of

the patient with suspected acute coronary syn-

drome. Semin Nucl Med 2003; 33:246–258

16. Kwong RY, Schussheim AE, Rekhraj S, et al. De-

tecting acute coronary syndrome in the emergency

department with cardiac magnetic resonance imag-

ing. Circulation 2003; 107:531–537

17. Georgiou D, Budoff MJ, Kaufer E, Kennedy JM, Lu

B, Brundage BH. Screening patients with chest pain

in the emergency department using electron beam

tomography: a follow-up study. J Am Coll Cardiol

2001; 38:105–110

18. Schoenhagen P, Halliburton SS, Stillman AE, et al.

Noninvasive imaging of coronary arteries: current

and future role of multi-detector row CT. Radiology

2004; 232:7–17

19. Schoepf UJ, Becker CR, Ohnesorge BM, Yucel EK.

CT of coronary artery disease. Radiology 2004;

232:18–37

20. Ioannidis JP, Salem D, Chew PW, Lau J. Accuracy

of imaging technologies in the diagnosis of acute

cardiac ischemia in the emergency department: a

meta-analysis. Ann Emerg Med 2001; 37:471–477