[Myocardial perfusion scintigraphy with Tc-99m MIBI in patients with left bundle branch block: Visual quantification of the anteroseptal perfusion imaging for the diagnosis of left anterior descending artery stenosis].

INTRODUCTION: The non-invasive detection of myocardial ischaemia in patients with left bundle branch block (LBBB) remains a challenge. It is often associated with coronary artery disease or hypertension, but frequently there is no indication of cardiovascular pathology at presentation. Exercise-induced electrocardiographic ST segment changes are non-diagnostic. Confirming coronary artery disease has obvious implications for management. Several studies have shown greater cardiac mortality in the presence of LBBB. Generally, a good prognosis has been found in patients with LBBB and normal or near-normal myocardial perfusion scintigraphy (MPS). Various investigators report frequent anteroseptal defects with MPS in patients with LBBB in the absence of significant left anterior descending (LAD) coronary artery disease. Several mechanisms have been proposed to explain this false-positive phenomenon. Various interpretative methods and stress techniques have been evaluated in an attempt to improve the specificity of noninvasive studies for detecting LAD disease. A number of software packages for quantifying myocardial perfusion are commercially available. Quantification is recommended to improve diagnostic accuracy and intra- and inter-observer reproducibility.41 METHODS: Patients with LBBB on ECG, who were referred to our institution (February 2002 to September 2003) for myocardial perfusion scintigraphy, were included in the study. Patients with previous myocardial infarction were excluded, unless the location was confirmed to be not anteroseptal before the onset of LBBB. Patients who did not undergo coronary angiography within six months were also excluded, unless a LAD lesion of > or = 50% was diagnosed more than six months prior to MPS without subsequent intervention, or angiography more than six months later showed a LAD lesion of < or = 50%. Treadmill exercise, dipyridamole or dobutamine infusion were used according to standard protocols and imaging commenced 15-60 minutes later. QPS quantitative software, used to reconstruct the images and quantify perfusion, is described in detail elsewhere. Three experienced nuclear physicians interpreted the studies. Stress and rest perfusion, as well as reversibility, to the anteroseptal wall (excluding the apex), anteroseptal wall and apex, and apex only, were graded on a scale of 0 (normal) to 4 (absent perfusion), where 1 represents mild, 2 moderate, and 3 severe impairment of perfusion. A final decision was made by consensus. Using QPS, summed stress, rest and difference scores were obtained for the same regions. Angiographic correlation was obtained by reviewing the patients' records. Stenosis of the LAD or graft vessel to the LAD of > or =50% was regarded as significant. The Kruskal-Wallace non-parametric test was used to compare the groups with and without significant LAD stenosis. A Bonferroni correction was applied to make provision for multiple testing. Receiver operating characteristic (ROC) analysis was utilised to determine the optimal threshold of the significant measurements to distinguish between the two groups; for this threshold, the sensitivity and specificity were calculated.

RESULTS: Nine men and nine women (42-78 years) satisfied the inclusion criteria and were included in the study. Dipyridamole was used in nine patients, exercise in seven, dobutamine in one, and one patient was injected during a period of typical chest pain. Ten patients had a LAD stenosis of < 50% and eight > or = 50%. The only measurement that yielded a significant difference between the groups was visual improvement in perfusion to the anteroseptal wall and apex between the stress and rest study (p < 0.0096). Even after applying a Bonferroni correction, the value tended towards significance (p = 0.16). A ROC curve was calculated and an optimal threshold of 0.5 determined, which in turn had a sensitivity of 88% and specificity of 67%.

DISCUSSION: Our findings suggest that visual reversibility in the anteroseptal wall and apex gives an indication of significant LAD stenosis in patients with LBBB. This finding agrees with that of Mairesse et al. Wackers argues that cardiomyopathic changes cause anteroseptal perfusion defects in LBBB. It is possible that irreversible perfusion defects in the anteroseptal wall and apex are caused by a constant, stress-independent mechanism, whereas reversible defects indicate underlying ischaemia. Interestingly, quantitative analysis was not helpful in predicting LAD disease. The quantitative software we used is well validated. On the other hand, Svenssson et al. compared three myocardial perfusion quantification software packages and found considerable variation, especially in the presence of perfusion defects. The state of perfusion to the apex was not helpful to detect significant LAD disease. It is known that the LAD usually supplies the apex. Matzer et al. found that requiring the presence of an apical defect improved specificity. This could not be confirmed by Lebtahi et al. or Vaduganathan et al.

LIMITATIONS: A definite limitation of our study was that treadmill stress testing was performed in seven patients. It is currently recommended by most authors that pharmacological stress be performed in patients with LBBB Selection bias is also a limitation because only patients who also had angiography were included in the study (18 out of 91).

CONCLUSION: A visual improvement in anteroseptal and apical myocardial perfusion between stress and rest with Tc-99m MIBI in patients with LBBB probably indicates significant LAD stenosis. In our hands, quantitative software did not aid in the diagnosis. A well-designed, prospective study using a standardized stress protocol (probably dipyridamole or adenosine), which specifically evaluates visual reversibility in the anteroseptal wall and apex, will obviate the need for a Bonferroni correction, and could confirm these findings.