The added value of ECG-gating for the diagnosis of myocardial infarction using myocardial perfusion scintigraphy and artificial neural networks

Validation of an automated method to quantify stress-induced ischemia and infarction in rest-stress myocardial perfusion SPECT

BACKGROUND

Myocardial perfusion SPECT (MPS) is one of the frequently used methods for quantification of perfusion defects in patients with known or suspected coronary artery disease. This article describes open access software for automated quantification in MPS of stress-induced ischemia and infarction and provides phantom and in vivo validation.

METHODS AND RESULTS

A total of 492 patients with known or suspected coronary artery disease underwent both stress and rest MPS. The proposed perfusion analysis algorithm (Segment) was trained in 140 patients and validated in the remaining 352 patients using visual scoring in MPS by an expert reader as reference standard. Furthermore, validation was performed with simulated perfusion defects in an anthropomorphic computer model. Total perfusion deficit (TPD, range 0-100), including both extent and severity of the perfusion defect, was used as the global measurement of the perfusion defects. Mean bias ± SD between TPD by Segment and the simulated TPD was 3.6 ± 3.8 (R(2) = 0.92). Mean bias ± SD between TPD by Segment and the visual scoring in the patients was 1.2 ± 2.9 (R (2)= 0.64) for stress-induced ischemia and -0.3 ± 3.1 (R(2) = 0.86) for infarction.

CONCLUSION

The proposed algorithm can detect and quantify perfusion defects in MPS with good agreement to expert readers and to simulated values in a computer phantom.

Clinical data do not improve artificial neural network interpretation of myocardial perfusion scintigraphy

Artificial neural networks interpretation of myocardial perfusion scintigraphy (MPS) has so far been based on image data alone. Physicians reporting MPS often combine image and clinical data. The aim was to evaluate whether neural network interpretation would be improved by adding clinical data to image data. Four hundred and eighteen patients were used for training and 532 patients for testing the neural networks. First, the network was trained with image data alone and thereafter with image data in combination with clinical parameters (age, gender, previous infarction, percutaneous coronary intervention, coronary artery bypass grafting, typical chest pain, present smoker, hypertension, hyperlipidaemia, diabetes, peripheral vascular disease and positive family history). Expert interpretation was used as gold standard. Receiver operating characteristic (ROC) curves were calculated, and the ROC areas for the networks trained with and without clinical data were compared for the diagnosis of myocardial infarction and ischaemia. There was no statistically significant difference in ROC area for the diagnosis of myocardial infarction between the neural network trained with the combination of clinical and image data (95.8%) and with image data alone (95.2%). For the diagnosis of ischaemia, there was no statistically significant difference in ROC area between the neural network trained with the combination of clinical and image data (87.9%) and with image data alone (88.0%). Neural network interpretation of MPS is not improved when clinical data are added to perfusion and functional data. One reason for this could be that experts base their interpretations of MPS mainly on the images and to a lesser degree on clinical data.

Evaluation of a decision support system for interpretation of myocardial perfusion gated SPECT

PURPOSE

We have recently presented a decision support system for interpreting myocardial perfusion scintigraphy (MPS). In this study, we wanted to evaluate the system in a separate hospital from where it was trained and to compare it with a quantification software package.

METHODS

A completely automated method based on neural networks was trained for the interpretation of MPS regarding myocardial ischaemia and infarction using 418 MPS from one hospital. Features from each examination describing rest and stress perfusion, regional and global function were used as inputs to different neural networks. After the training session, the system was evaluated using 532 MPS from another hospital. The test images were also processed with the quantification software package Emory Cardiac Toolbox (ECTb). The images were interpreted by experienced clinicians at both the training and the test hospital, regarding the presence or absence of myocardial ischaemia and/or infarction and these interpretations were used as gold standard.

RESULTS

The neural network showed a sensitivity of 90% and a specificity of 85% for myocardial ischaemia. The specificity for the ECTb was 46% (p < 0.001), measured at the same sensitivity. The neural network sensitivity for myocardial infarction was 89% and the specificity 96%. The corresponding specificity for the ECTb was 54% (p < 0.001).

CONCLUSION

A decision support system based on neural networks presents interpretations more similar to experienced clinicians compared to a conventional automated quantification software package. This study shows the feasibility of disseminating the expertise of experienced clinicians to less experienced physicians by the use of neural networks.