Automatic reorientation algorithm for myocardial perfusion SPECT using segmentation

• Published in: European journal of clinical investigation Vol. 55; no. S1
• Main Authors: Vijande, Ezequiel, Campisi, Roxana, Juarez‐Orozco, Luis Eduardo, Agüero, Roberto, Geronazzo, Ricardo, Namías, Mauro
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.04.2025
• Subjects: Algorithms; automation; Computed tomography; Datasets; Deep learning; Heart; Image processing; Image segmentation; Machine learning; myocardial perfusion imaging; Operators; Perfusion; Positron emission; reorientation; Segmentation; Single photon emission computed tomography; SPECT; Standard deviation; Workflow
• ISSN: 0014-2972; 1365-2362
• DOI: 10.1111/eci.70016

Background Cardiac reorientation is a necessary step in processing myocardial perfusion images. This task usually requires manual intervention and thus introduces intra‐ and inter‐operator variability in the processing workflow that may lead to reduced reproducibility of the results. Methods A deep learning model was trained to perform segmentation of cardiac structures from SPECT images simulated from a real PET/CT dataset. Labels used for training were automatically generated in a semi‐supervised fashion by using TotalSegmentator on CT images. Segmentation results from the trained model were used to calculate cardiac landmarks from which the cardiac axes were defined, and reorientation was performed. Automatic reorientation was compared against the manual reorientation defined by three expert nuclear cardiologists. Results The average rotation difference between cardiac axes calculated from predicted segmentations and ground‐truth segmentations was 5.3°±3.1° on the simulated SPECT test dataset. In real SPECT images, the standard deviation of the angle difference between the automatic method and human experts was lower in all axes and operators compared to the maximum inter‐operator standard deviation. Conclusions The proposed deep learning‐based algorithm provides an automatic method to perform cardiac reorientation in myocardial perfusion SPECT images with an error range like the variability between operators and with the advantage of using objective anatomical landmarks for the definition of cardiac axes. Our results suggest that the proposed algorithm performs cardiac reorientation automatically with a similar variability with respect to human experts as the inter‐operator variability between human experts. Additionally, expert cardiologists' criteria for visually reorienting cardiac studies may differ from the standardized cardiac axis definition. The feasibility of our automatic method may translate to faster and more reproducible workflows in the clinic than the currently used manual and semi‐automatic methods.