Applied Radiology

Noninvasive tests to evaluate patients for suspected coronary artery disease play an increasing role in our diagnostic armamentarium. There are several excellent ways to assess patients noninvasively including exercise treadmill tests, ultrafast CT scans, echocardiography, nuclear imaging, and MRI. The ideal use for each modality is still being defined, but depends upon availability, physician expertise, and the patient's pre-test probability of disease. PET imaging has emerged as an excellent nuclear imaging technique for the evaluation of patients with suspected or known coronary artery disease. It is the gold standard for detecting viable myocardium, but many physicians may not be aware that PET imaging has several advantages over SPECT imaging for accurate detection of coronary artery disease.

This article focuses on detection of coronary artery disease in women. The potential advantages of PET imaging in this area are outlined. Primarily because of the ability of PET to eliminate attenuation artifacts that are problematic in SPECT imaging, PET appears to be more accurate in women. Perhaps the time has come to rekindle our enthusiasm for PET rubidium imaging as the non-invasive study of choice for the diagnosis of coronary artery disease. --Byron Williams, MD

Coronary artery disease remains the leading cause of death in the western world. This statement is true for both men and women. ¹ Although breast cancer is an important medical issue in females, diseases of the heart and circulatory system are the number one killers. It is estimated that up to 12 to 14 million people are currently diagnosed with coronary artery disease in the United States and another equally large group of people are undiagnosed but have clinically "silent" disease. In fact, silent myocardial ischemia is more common than most physicians appreciate and is estimated to occur in more than 30% of females over age 65. ² Furthermore, approximately half of all individuals with coronary disease have their initial presentation as either acute myocardial infarction or sudden cardiac death. Data from the Framingham Study demonstrated that approximately one-half to three-quarters of individuals identified to be at increased risk for coronary artery disease remain healthy up to a 20-year follow-up period. ³ Thus, coronary risk factor assessment alone is not sufficient to accurately identify most individuals with significant disease. In order to evaluate asymptomatic or minimally symptomatic individuals for the presence of significant coronary artery disease, some type of provocative testing is necessary to produce abnormalities (evidence of myocardial ischemia) that are not present at rest. In current practice, exercise or pharmacological stress is performed and is often combined with nuclear imaging or ultrasound studies. To further confuse this issue, there are important gender differences that exist when dealing with diagnostic testing and presenting symptoms of coronary artery disease in women as compared with men.

Gender Differences: Clinical symptoms and diagnostic testing

Death from cardiovascular disease poses a greater threat to women than all cancers combined. ¹ Nevertheless, misperceptions persist that females are not at risk for heart disease or stroke, especially when compared with their male counterparts. The presentation of disease is one important gender difference. There seems to be at least a decade lag in the average age of onset of coronary artery disease in women compared with men, such that the prevalence of disease is higher in men than women until age 65. ⁴ At this point, women catch up with men and the prevalence of coronary artery disease becomes higher in women. Most females are post-menopausal before clinical disease is evident, which suggests a protective effect from estrogen. Women often present with symptoms of angina, whereas men are more likely to present with either an acute myocardial infarction or sudden death as the initial manifestation. Only two-thirds of females with typical exertional angina will have documented obstructive coronary artery disease when coronary angiograms are performed. Women appear to be more prone than men to have other causes of chest pain unrelated to coronary artery disease. ^5,6 In women below age 60 with non-specific chest pain, coronary artery disease is very unlikely to be the explanation for the pain, especially in a pre-menopausal woman. It has also been demonstrated that women have a higher fatality rate after their first heart attack than men. Diabetes and dyslipidemia appear to be more prevalent in the female population. ^7,8 Some of these gender differences may be related to more diffuse disease in women, smaller vessels, more vasospasm, or perhaps a gender bias in diagnostic testing and management. ^9,10

The diagnostic accuracy of noninvasive stress testing for the diagnosis and evaluation of suspected coronary artery disease has been well studied in males, but less well studied in women; such that concern about their reliability in women remains a real clinical problem. Disappointing results in accuracy have been reported for both stress EKG and stress nuclear imaging studies in women. ^11-14 Stress echocardiography appears to have similar accuracy to stress nuclear imaging in women. ^15,16 False positive stress EKG and stress thallium studies are more common in women than men. All of the reasons for this discrepancy are not known but may be explained in part to breast attenuation artifact, a higher prevalence of vasospasm, mitral valve prolapse, and ST segment abnormalities (perhaps a digitalis-like estrogen effect on the EKG) as well as Bayes' theorem. In addition, women have a greater incidence of false negative stress EKG and stress perfusion imaging studies due to lower workloads achieved, thus failing to provoke ischemia; and perhaps more diffuse disease than in males resulting in "balanced" ischemia. Breast attenuation artifacts can balance out perfusion defects in areas not affected by the attenuation making the imaging study appear relatively normal. Based on work done at our institution and at other centers in this country, PET imaging appears to be more accurate than SPECT imaging for the appropriate detection of coronary artery disease in both men and women. PET has several physical advantages over SPECT that result in improved accuracy. The main advantages of PET imaging have to do with improved image quality because of better resolution and higher count rates (by a factor of 30-fold when compared with SPECT) and attenuation correction. In addition, PET has the ability to quantify coronary blood flow, which is not possible with SPECT.

PET Imaging with Rubidium-82--A Better Way?

Attenuation correction by PET allows interpretation of perfusion images to be unhindered by soft tissue artifacts (primarily breast attenuation in women, diaphragmatic attenuation in men, and obesity), which enhances the diagnostic quality and accuracy of the imaging study. PET appears to be very accurate for detection of significant coronary artery disease in women and may be the noninvasive test of choice for this population in terms of accuracy and cost effectiveness (especially when the pre-test likelihood of disease is of relatively low to moderate probability). Figure 1 is an example of a very abnormal SPECT thallium image in a woman with atypical chest pain that resulted in coronary arteriograms being performed. Figure 2 shows the PET scan in the same patient performed soon after the coronary arteriograms.

Previously, we studied and reported the results of 57 female patients who were referred to the cardiac catheter lab for arteriography to diagnose suspected coronary artery disease. The sensitivity and specifity of PET imaging detecting greater than 50% luminal diameter narrowing in this selected group of women was 94% and100%, respectively. In this small cohort of selected women patients, PET demonstrated excellent accuracy and correlate well with the "gold standard" of coronary arteriography. ¹⁷ The overall sensitivity and specificity for PET imaging reported from our center is 95% and 95% for males and females.

The relative lack of accuracy of diagnostic testing in women as compared with men cannot be explained solely on the basis of Bayes' theorem alone. In fact, a recently published meta-analysis demonstrated the accuracy of diagnostic testing is lower for women than men. ¹⁸ Recently, we reviewed retrospectively all women who underwent SPECT thallium and PET imaging within 3 months of each other from January 1998 to December 1999. There were 42 patients. Of this group, 12 patients underwent coronary arteriograms, 2 patients who did not have arteriograms had documented prior myocardial infarctions, and 30 patients had atypical symptoms and an otherwise low pre-test likelihood of disease. These women were assigned to the diagnosis of coronary disease if angiography demonstrated >=50% diameter stenosis of one or more major branches of the coronary circulation or documented prior infarction (10 patients were in this group). A total of 32 women were "normal" by coronary angiograms or by virtue of a low pre-test likelihood of disease. Abnormal scans were present in 26 SPECT images and 11 PET images. Normal studies were found in 16 SPECT and 31 PET studies. When sensitivity and specificity are applied to these groups based on our definition of coronary disease, SPECT imaging was 80% and 44%, respectively, compared with 90% and 94% for PET. Admittedly there is a referral bias since many of the PET studies were done because the SPECT was abnormal but the clinical suspicion was relatively low. Nevertheless, it does suggest that PET Rb-82 imaging is a more accurate diagnostic study for appropriate detection of significant coronary artery disease than SPECT thallium imaging (B. Williams, unpublished data, July 2000).

A recent meta-analysis demonstrated a lower sensitivity and specificity for SPECT thallium in women as compared with men. The specificity for thallium was 64% in women and 85% for in. ¹⁸ In theory, Tc-99m sestamibi SPECT imaging should improve this number and some early studies do suggest this, especially if gated wall motion analysis is used in conjunction with the perfusion study. ¹⁹ In our laboratory, using PET Rb-82 perfusion imaging the diagnostic accuracy for both men and women is very similar. The sensitivity and specificity for women is 98% and 95% and for men is 96% and 93%, respectively.

Summary

Noninvasive imaging has many advantages in the assessment and evaluation of the patient with suspected coronary artery disease, not the least of which is relative safety and comfort for the patient. Both PET and SPECT imaging are excellent modalities for noninvasive testing in humans. The physical properties of PET allow for improved accuracy for detection of coronary artery disease when compared with SPECT. This article focuses on the gender differences that are present in the evaluation of patients with suspected coronary disease and points out the improved accuracy of PET imaging in female patients. Perhaps it is time to strongly consider using PET as the preferred diagnostic noninvasive study in women if the technology is available.