Truace
Subscribe

Distinct clinical patterns of preterm infants vulnerable to intestinal injury identified by unsupervised clustering

Background Current necrotizing enterocolitis (NEC) diagnostic criteria fail to capture the full spectrum of intestinal injury. We aimed to identify diagnosis-independent clinical patterns of intestinal injury among vulnerable very low birth weight (VLBW) infants using unsupervised hierarchical clustering. Methods Using prospectively collected data from 20,352 VLBW infants in the Korean Neonatal Network (2013-2022), unsupervised clustering was performed after excluding variables reflecting prematurity and NEC dia…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1201· cite & verify
Distinct clinical patterns of preterm infants vulnerable to intestinal injury identified by unsupervised clustering
"A very low birth weight baby in HRB Clinic" by Bhakua is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/.

In brief

Using data from 20,352 very low birth weight infants in the Korean Neonatal Network between 2013 and 2022, researchers performed unsupervised hierarchical clustering without NEC-related labels to uncover patterns of intestinal injury vulnerability. They found eight primary clusters, with Cluster 8 showing the highest prevalence of previously diagnosed NEC >= stage 2 and further dividing into five subclusters linked to antenatal antibiotic exposure, assisted reproduction, pregnancy-induced hypertension, chorioamnionitis/PPROM, and severe postnatal morbidities.

The work matters because current NEC diagnostic criteria fail to capture the full spectrum of intestinal injury, and a diagnosis-independent phenotypic framework could enable more precise classification and targeted prevention in this high-risk population. What remains uncertain is whether these computationally derived clusters translate into actionable clinical criteria, improved outcomes, or generalizability beyond the Korean cohort, as the study reports structural validation but not prospective clinical implementation.

Main points

  1. Analysis used prospectively collected data from 20,352 VLBW infants in the Korean Neonatal Network from 2013-2022, excluding variables reflecting prematurity and NEC diagnosis.
  2. Cluster 8 had the highest prevalence of previously diagnosed NEC >= stage 2 and was subdivided into SC23-SC27: antenatal antibiotic exposure, assisted reproduction, pregnancy-induced hypertension, chorioamnionitis/PPROM, and severe postnatal morbidities.
  3. Hierarchical structure validation showed strong concordance between Kaplan-Meier survival estimates and supervised survival model predictions across clusters.

The gain

Unsupervised hierarchical clustering of 20,352 VLBW infants identified eight diagnosis-independent clusters of intestinal injury vulnerability, with high-risk Cluster 8 further split into five etiologic subclusters, creating a framework for a more precise taxonomy beyond traditional NEC staging.

The rundown

Researchers applied unsupervised hierarchical clustering to prospectively collected Korean Neonatal Network data from 2013-2022, deliberately excluding prematurity and NEC diagnosis variables to find diagnosis-independent structure.

The resulting taxonomy identified eight primary clusters and five subclusters within the highest-risk group defined by antenatal and postnatal features, validated by concordance between Kaplan-Meier estimates and supervised survival model predictions.

Sources

  1. Peer-reviewedPediatric Research2026-09-23

The debate