Can Nuclear Medicine Technologists Assess Whether a Myocardial Perfusion Rest Study Is Required?

Patients

The study group consisted of patients who underwent MPI between September 15, 2004, and September 14, 2005, at Sahlgrenska University Hospital in Gothenburg, Sweden. Patients with incomplete data and studies with technical problems were excluded. The study population comprised 532 patients, with a mean age of 62 ± 11 y; 49% were men. Only 1 examination per patient was included. MPI was performed for diagnosing coronary artery disease (CAD) in 421 patients, for managing known CAD in 107, and for other reasons in 4. Seventy-nine patients (15%) with previous MI and 98 patients (18%) with a history of revascularization were included. Diabetes was present in 16%, hypertension in 50%, and hyperlipidemia in 44% of the patients; 16% of the patients were current smokers, and 30% had a family history of cardiac disease. The study was approved by the Research Ethics Committee at Gothenburg University.

Radionuclide Imaging

Patients were stressed using either maximal exercise, symptom-limited ergometry testing (53%), or pharmacologic testing with adenosine. The exercise or pharmacologic stress was continued for at least 2 min after tracer injection. The gated SPECT studies were performed using a 2-d nongated stress and gated rest ^99mTc-sestamibi protocol. Stress and rest acquisition began approximately 60 min after injection of ^99mTc-sestamibi (600 MBq). Images were acquired using 2 different dual-head SPECT cameras (Infinia or Millennium VG; GE Healthcare) equipped with a low-energy, high-resolution collimator. Images were acquired with the patient in a supine position, using step-and-shoot mode, a circular acquisition, a 64 × 64 matrix, a zoom factor of 1.28, a pixel size of 6.9 mm, 60 projections over 180°, 40 s per projection, and an energy window of 20%. In patients weighing over 90 kg, the acquisition time per projection was increased to 55 s. During rest acquisition, the patient was monitored using a 3-lead electrocardiogram. The acceptance window was opened to ±20% of the predefined R–R interval, except for a limited number of studies in which a wider acceptance window was used. Other beats were rejected. Each R–R interval was divided into 8 equal time intervals. Gated SPECT acquisition was performed at the same time as ungated routine SPECT acquisition. An automatic motion-correction program was used in studies indicating patient motion during acquisition.

Tomographic reconstruction of nongated data was performed using filtered backprojection with a Butterworth filter having a critical frequency of 0.52 cycles/cm and a power factor of 5. The gated data were reconstructed using filtered backprojection with a Butterworth filter having a critical frequency of 0.40 cycles/cm and a power factor of order 10. No attenuation or scatter correction was used.

Stress–Rest Interpretation

Visual interpretation of the complete stress and rest images by 1 experienced physician at the time of clinical reporting was used as the gold standard. All cases categorized as infarction or ischemia present or probably present were categorized as the group requiring a rest study (i.e., the “rest-study-required group”), whereas all other cases were categorized as the group not requiring a rest study (i.e., the “no-rest-study-required group”). Rest, stress, gated images, clinical information, and electrocardiogram findings were available for this interpretation. The CEqual and QGS software packages were used to aid this visual interpretation (4,5).

Stress-Only Assessment

A total of 3 nuclear medicine technologists and 3 physicians independently classified each of the 532 stress studies as study-required or no-rest-study-required. Two of the technologists had more than 10 y of experience in MPI, and 1 had only 1.5 y of experience. All 3 physicians were experienced in MPI. One of the physicians performed the gold standard interpretations 1 y before he performed the stress-only interpretations. Two technologists and 2 physicians came from one hospital and the other 2 from another hospital.

To assist the technologists in their classifications, the most experienced physician gave an MPI assessment tutorial and presented guidelines for classifying the stress-only images (Table 1). The rules comprised lower-normal count levels for men and women in different parts of the left ventricle. The criteria were not to be applied to patients with sarcoidosis, systemic sclerosis, or systemic lupus erythematosus. These groups of patients have been reported to have abnormal scan findings at rest that may normalize with pharmacologic stress (6,7).

All observers reviewed the stress slice images and bull's-eye plots using the EXINI heart software (EXINI Diagnostics AB). No quantification data based on reference databases were available, but the count values in percentage of the maximal left ventricular activity were displayed on the screen when the pointer was moved to an area of interest. All observers reviewed the stress images, but only 2 of the physicians also had access to clinical information.

Statistical Methods

The χ² test was used to analyze the differences in classification of patients into the rest-study-required and no-rest-study-required groups between the nuclear medicine technologists and the physicians. P values of less than 0.05 were considered statistically significant.

Main Findings

Nuclear medicine technologists can assess that a rest study is not needed. The risk that this assessment is false is no higher for a technologist (1.3%) than for a physician (2.6%). The 2 experienced technologists and 1 of the physicians produced few if any false classifications in the infarction or ischemia group while identifying 19%−26% of the no-rest-study-required cases.

Two physicians had access to clinical data when assessing the stress images. One of these physicians exhibited performance similar to that of the 2 experienced technologists, indicating that clinical data are of minor importance to this classification task. This result is in accordance with the findings of Worsley et al. (1); they, however, proposed that patients with previous MI need to have a rest study, giving a high level of perfusion abnormalities in their patient group. Most important in the stress-only approach is to correctly identify those patients who might have MI or ischemia and therefore require a rest study. Assuming that these patients can be identified accurately, it is of interest to minimize the number of superfluous rest studies. The physicians were significantly better than the technologists at identifying such patients (32% vs. 21%, respectively) but at the cost of a higher number of false classifications in the rest-study-required group. Both physicians and technologists achieved lower values than those (52%) reported by Worsley et al. (1). One factor that may at least partly explain this difference is the higher proportion of women in the patient group of the present study (51% here vs. 31% in the study by Worsley et al. (1)). Perfusion defects in the anterior wall of the left ventricle can be explained by an attenuation artifact, but this is difficult to assess from the stress images only (Fig. 1A).

Clinical Implications

If the nuclear medicine technologist who acquires and reviews the stress images could also make the stress-only assessment, clinic workflow would improve. The results of this study indicate that this is feasible. If this is to be the aim, the technologists must receive relevant education. It was, therefore, of interest to see that our approach of giving the technologists an interpretation tutorial and guidelines for classifying stress-only images (Table 1) resulted in such good performance, even for the technologist with only 1.5 y of MPI experience. The only 2 false classifications in the rest-study-required group made by the technologists were cases meeting the criteria presented in Table 1.

Study Limitation

The gold standard for determining the scintigraphic presence of myocardial infarction or ischemia was the rest–stress interpretation made by 1 experienced physician. Ideally, the gold-standard method should be used independently of the myocardial perfusion images; this is difficult to accomplish, however, especially as it is important to have a large number of subjects who represent cases found in clinical routine. We used consecutive patients from a 12-mo period; that is, we focused on increasing the number of subjects instead of including fewer subjects using, for example, coronary angiography or cardiac events as the gold standard. Therefore, this study was not designed to assess the value of a stress-only or stress–rest MPI study in relation to a cardiac event rate or angiographic results.

The guidelines presented in Table 1 have not been validated in clinical settings with other patient populations, acquisition protocols, γ-cameras, and the like. They should, therefore, be regarded only as examples of criteria that could be used to evaluate a stress-only study. Other semiquantitative methods, such as the summed stress score, could serve as alternative approaches for this type of analysis.