Establishment and verification of nomogram prediction model for risk of coronary microvascular injury in patients with ST-segment elevation myocardial infarction after percutaneous coronary intervention

Li Jingchang; Hu Manman; Zhou Qian; Liu Lin; Chai Zhiyong

doi:10.3969/j.issn.1674-4055.2025.03.23

Abstract：Objective To discuss the establishment and verification of nomogram prediction model for risk of coronary microvascular injury(CMI)in patients with ST-segment elevation myocardial infarction(STEMI)after percutaneous coronary intervention(PCI).Methods The clinical data of STEMI patients underwent PCI(n=410)were collected in Central Hospital of Nanyang City from June 2021 to June 2024.In a 7:3 ratio,the patients were divided into modeling group and verification group by using simple random sampling method.According to the results of echocardiography,the modeling group was divided into CMI subgroup(n=129)and non-CMI subgroup(n=158).The influence factors were examined by using binary Logistic regression analysis in modeling group after PCI.A nomogram model was established and verified externally with data of verification group.The decision curve of predicting CMI by different factor was analyzed in STEMI patients after PCI.Results The difference in general data had no statistical significance between modeling group and verification group(P>0.05).The percentages of smoking and coronary slow flow(CSF),time from onset to PCI,pre-dilation time and levels of D-dimer(D-D),glycated hemoglobin(HbAlc)and homocysteine(Hcy)were higher,and left ventricular ejection fraction(LVEF)was lower in CMI subgroup than those in non-CMI subgroup(P<0.05).The results of binary Logistic regression analysis showed that smoking history(OR=2.950,95%CI:1.464～5.942),CSF(OR=2.270,95%CI:1.183～4.358),longer time from onset to PCI(OR=2.013,95%CI:1.468-2.760),longer pre-dilation time(OR=1.887,95%CI:1.369～2.466),and increased D-D(OR=9.845,95%CI:3.438～28.195),HbA1c(OR=1.971,95%CI:1.440～2.695)and Hcy(OR=1.201,95%CI:1.091～1.323)were independent risk factors,and LVEF(OR=0.861,95%CI:0.797～0.930)was a protective factor of CMI occurrence.The nomogram prediction model established based on above influence was verified by using Bootstrap,and verified internally through modeling group and externally through.The results showed that C-index was 0.885 in modeling group and 0.890 in verification group,and calibration curves of both groups showed good agreement with ideal curve.The results of discrimination evaluation showed that AUC was 0.885(95%CI:0.846～0.923,P<0.001),sensitivity was 0.760,specificity was 0.880 and Youden index was 0.640 in modeling group,and AUC was 0.890(95%CI:0.852-0.927,P<0.001),sensitivity was 0.791,specificity was 0.861 and Youden index was 0.652 in verification group.The results of decision curve analysis showed that,within the threshold range of 0.00～1.000,the combined prediction of smoking history,CSF,time from onset to PCI,pre-dilation time,D-D,HbA1c,Hcy and LVEF levels for CMI occurrence in STEMI patients after PCI demonstrated a superior net reclassification improvement compared to individual predictions alone.Conclusion Smoking history,CSF,time from onset to PCI,pre-dilation time,LVEF,HbA1c,D-D and Hcy levels are influence factors for CMI occurrence in STEMI patients after PCI,and the nomogram prediction model established based on these factors has a higher predictive efficacy.