Background: There are ongoing challenges in revascularizing patients with severe left ventricular (LV) dysfunction, with high rates of short- and long-term complications. Cardiac magnetic resonance (CMR) imaging is widely used for these patients; however, the prognostic value of the sum scar score and microvascular obstruction (MVO) remains unclear. Method: A retrospective study of data from King Abdullah Medical City (KAMC) between 2023 and 2025 was conducted to evaluate revascularization decisions made based on CMR studies. Patients who had normal LV function, no angiogram report, or incomplete CMR imaging were excluded. The primary outcome was hospitalization or death in patients with or without revascularization. Results: The cohort included 145 patients. The patients had a mean age of 56.7 ± 9.7 years and were predominantly male 131 (90.3%), with a high prevalence of cardiovascular risk factors: diabetes 102 (70.8%), hypertension 84 (58.7%), known dyslipidemia 46 (32%), and smoking 65 (45.5%). Most patients underwent a viability study within 7 days of a myocardial infarction (MI), 58 (40%), and a further 42 (28.9%) did so within 7–15 days. The CMR LVEF was 29.5 ± 8, and the sum scar score was 38.8 ± 21.9; MVO was observed in 34 patients (23.4%). The number of patients who underwent revascularization at the index admission was 66 (45.5%), of whom 48 (33%) received percutaneous intervention (PCI) and 18 (12.4%) underwent coronary artery bypass surgery (CABG). The patients lost to follow-up numbered 51 (35%). There were no significant differences between the two groups for the primary endpoints, including hospitalization (p-value: 0.61) and mortality (p-value: 0.31). Conclusion: In this retrospective study, CMR did not have an independent effect on hospitalization or mortality in patients who had revascularization compared to medical therapy groups. MVO and the LGE scar score were not significantly associated with cardiovascular outcomes. Our cohort was underpowered, with a high proportion of patients lost to follow-up, limiting the generalizability of the data.
Shalaby et al. (Wed,) studied this question.