In a Houston Methodist Young-ACS PCI registry study of 452 adults age 50 or younger treated with PCI for ACS, unsupervised machine learning identified 3 phenotypic clusters with sharply different outcomes. The high-risk cluster (n=78) had more diabetes, heart failure, prior MI/PCI, lower LVEF, and a 4.5-fold higher all-cause mortality vs the low-risk cluster (HR 4.50, 95% CI 1.68-12.1).

Why It Matters To Your Practice

• Young ACS patients are not a uniform low-risk group, even when all undergo PCI.

• Simple phenotype patterns may help clinicians spot patients with heavier comorbidity burden who warrant closer follow-up and more aggressive secondary prevention.

• Diabetes mellitus (DM), HF, PAD, prior MI/PCI, and depressed LVEF clustered with worse prognosis.

Clinical Implications

• After PCI in younger ACS patients, consider risk beyond age alone when planning discharge, surveillance, and counseling.

• Patients resembling the high-risk phenotype may merit tighter control of cardiometabolic risk factors, medication adherence review, and earlier outpatient reassessment.

• The moderate-risk and low-risk clusters did not differ significantly in mortality, suggesting the sickest subgroup may drive most excess risk.

Insights

• The registry included patients from 2010-2022 and used agglomerative clustering on demographic, clinical, and angiographic variables.

• Low-risk patients (n=273) had less HF and PAD, preserved median LVEF of 0.60, and shorter culprit-vessel stent length.

• High-risk patients had DM in 67.9%, hypertension in 93.6%, HF in 65.4%, prior MI in 41.0%, prior PCI in 46.2%, median LVEF of 0.3, and more mechanical circulatory support use.

• A 2-cluster sensitivity analysis showed similar outcome separation, supporting the robustness of the findings.

The Bottom Line

• In young adults undergoing PCI for ACS, unsupervised ML separated 3 clinically recognizable phenotypes, and the highest-risk cluster had markedly worse survival.

• For clinicians, the takeaway is practical: do not let younger age obscure major comorbidity burden when estimating prognosis after ACS PCI.