General purpose deep learning attenuation correction improves diagnostic accuracy of SPECT MPI: multicenter study

作者：Aakash D. Shanbhag, Robert J. H. Miller, M. Lemley, P. Kavanagh, Joanna X. Liang, Anna M. Marcinkiewicz, V. Builoff, S. V. Van Kriekinge, Terrence D. Ruddy, M. Fish, Andrew J. Einstein, Monica Martins, J. Halcox, Philipp A. Kaufmann, C. Buckley, T. Bateman, Daniel S. Berman, D. Dey, P. Slomka · 发表于：JACC Cardiovascular Imaging · 年份：2025 · DOI：10.1016/j.jcmg.2025.06.010 · 被引用次数：6 · 研究领域：Medicine

Background Single photon emission computed tomography(SPECT) myocardial perfusion imaging(MPI) uses computed tomography(CT)–based attenuation correction(AC) to improve diagnostic accuracy. Deep-learning (DL) has the potential to generate synthetic AC images, as an alternative to CT-based AC. Objectives This study evaluates whether DL-generated synthetic SPECT images could enhance accuracy of conventional SPECT MPI. Methods We developed a DL model in a multicenter cohort of 4894 patients from 4 sites to generate simulated SPECT AC images(DeepAC). The model was externally validated in 746 patients from 72 sites in a clinical trial (NCT01347710) and 320 patients from another external site. In the first external population, we assessed the diagnostic accuracy for obstructive coronary artery disease (CAD)—defined as left main stenosis ≥50% or ≥70% in other vessels—for total perfusion deficit (TPD). In the latter, we completed change analysis and compared quantitative scores for AC, DeepAC, and non-attenuation correction (NC) with clinical scores. Results In the first external cohort (mean age 63 ± 9.5, 69.0% male), 206 patients(27.6%) had obstructive CAD. The area under the receiver operating characteristic curve (AUC) of DeepAC TPD (0.77, 95% confidence interval [CI] 0.73 – 0.81) was higher than NC TPD (AUC 0.73, 95% CI 0.69 – 0.77, p<0.001). In the second external cohort, DeepAC quantitative scores had closer agreement with actual AC scores compared to NC. Conclusion In a multic...