What is the clinical evidence from this study?

Study design: Cohort. Population: Ischemic heart disease / Acute coronary syndrome (n=29521). Intervention: Machine learning (ML) models. Primary outcome: In-hospital, 30-day, and 1-year mortality (ROC-AUC 0.927 to 0.943 (in-hospital, TEST1)).

What does this research mean for the field?

Machine learning models demonstrate high discrimination and clinical utility for predicting in-hospital, 30-day, and 1-year mortality after percutaneous coronary intervention, identifying age, haemodynamic status, and renal function as consistent key predictors. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

July 8, 2026Journal of Cardiovascular Translational ResearchOpen Access

Predicting Mortality After Percutaneous Coronary Intervention in a Multiethnic Southeast Asian Population: Insights From Machine Learning

View Full Paper

Ask AI

Bookmark

Why the study?

Can machine learning models accurately predict in-hospital, 30-day, and 1-year mortality in patients undergoing percutaneous coronary intervention?

Population

29,521 patients undergoing percutaneous coronary intervention, including high-risk acute coronary syndrome…

Design

Cohort

Follow-up

1 year

Key result

Machine learning models demonstrated high discrimination for predicting in-hospital mortality after non-elective PCI, with ROC-AUC ranging from 0.927 to 0.943 in external validation.

Authors

YLYih Miin Liew YCYin Kia Chiam PNPei Ling Ngo

Discussion

Member takes

Overview

ML models may aid ACS PCI mortality risk stratification; leaves open clinical utility pending prospective validation.

Key Points

This research aims to identify predictors of mortality after percutaneous coronary intervention in high-risk patients using machine learning models.
Analyzed nationwide registry data from 2007 to 2020, involving 29,521 patients.
Compared seven machine learning models for predicting mortality at different time points: in-hospital, 30-day, and 1-year.
Performed logistic recalibration and validated results using external cohorts (TEST1 and TEST2).
In-hospital mortality discrimination ranged from 0.927 to 0.943 in TEST1 and 0.865 to 0.884 in TEST2.
For 30-day mortality, ROC-AUC ranged from 0.902 to 0.923 in TEST1 and 0.753 to 0.838 in TEST2.
Key predictors consistently identified were age, haemodynamic status, and renal function.

Study Design

Type

Cohort (n=29,521)

Multicenter

Yes

Structured PICO

Can machine learning models accurately predict in-hospital, 30-day, and 1-year mortality in patients undergoing percutaneous coronary intervention?

Population

29,521 patients undergoing non-elective percutaneous coronary intervention for acute coronary syndrome from a nationwide registry in Malaysia (2007–2020).

Exposure

Seven machine learning (ML) models for predicting mortality

Outcome

In-hospital, 30-day, and 1-year mortalityhard clinical

Main Result

Effect estimate: ROC-AUC 0.927 to 0.943 (in-hospital, TEST1)

Machine learning models can accurately predict short- and long-term mortality following PCI in a multiethnic Southeast Asian population, identifying age, hemodynamic status, and renal function as key predictors.

Cite This Study

Liew et al. (2026) conducted a cohort in Ischemic heart disease / Acute coronary syndrome (n=29,521). Machine learning (ML) models was evaluated on In-hospital, 30-day, and 1-year mortality (ROC-AUC 0.927 to 0.943 (in-hospital, TEST1)). Machine learning models demonstrated high discrimination for predicting in-hospital mortality after non-elective PCI, with ROC-AUC ranging from 0.927 to 0.943 in external validation.

synapsesocial.com/papers/6a4de835d2ea289ef6282fbe https://doi.org/10.1007/s12265-026-10812-5

View Full Paper

Ask AI

Bookmark

July 8, 2026Journal of Cardiovascular Translational ResearchOpen Access

Predicting Mortality After Percutaneous Coronary Intervention in a Multiethnic Southeast Asian Population: Insights From Machine Learning

View Full Paper

Ask AI

Bookmark

Why the study?

Can machine learning models accurately predict in-hospital, 30-day, and 1-year mortality in patients undergoing percutaneous coronary intervention?

Population

29,521 patients undergoing percutaneous coronary intervention, including high-risk acute coronary syndrome…

Design

Cohort

Follow-up

1 year

Key result

Machine learning models demonstrated high discrimination for predicting in-hospital mortality after non-elective PCI, with ROC-AUC ranging from 0.927 to 0.943 in external validation.

Authors

YLYih Miin Liew YCYin Kia Chiam PNPei Ling Ngo

Discussion

Member takes

Overview

ML models may aid ACS PCI mortality risk stratification; leaves open clinical utility pending prospective validation.

Key Points

This research aims to identify predictors of mortality after percutaneous coronary intervention in high-risk patients using machine learning models.
Analyzed nationwide registry data from 2007 to 2020, involving 29,521 patients.
Compared seven machine learning models for predicting mortality at different time points: in-hospital, 30-day, and 1-year.
Performed logistic recalibration and validated results using external cohorts (TEST1 and TEST2).
In-hospital mortality discrimination ranged from 0.927 to 0.943 in TEST1 and 0.865 to 0.884 in TEST2.
For 30-day mortality, ROC-AUC ranged from 0.902 to 0.923 in TEST1 and 0.753 to 0.838 in TEST2.
Key predictors consistently identified were age, haemodynamic status, and renal function.

Study Design

Type

Cohort (n=29,521)

Multicenter

Yes

Structured PICO

Can machine learning models accurately predict in-hospital, 30-day, and 1-year mortality in patients undergoing percutaneous coronary intervention?

Population

29,521 patients undergoing non-elective percutaneous coronary intervention for acute coronary syndrome from a nationwide registry in Malaysia (2007–2020).

Exposure

Seven machine learning (ML) models for predicting mortality

Outcome

In-hospital, 30-day, and 1-year mortalityhard clinical

Main Result

Effect estimate: ROC-AUC 0.927 to 0.943 (in-hospital, TEST1)

Cite This Study

synapsesocial.com/papers/6a4de835d2ea289ef6282fbe https://doi.org/10.1007/s12265-026-10812-5

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark