Development and external validation of a machine learning model based on multimodal dynamic data for…

AI / informaticsFrontiers in neurology· 2026

Development and external validation of a machine learning model based on multimodal dynamic data for predicting shunt-dependent hydrocephalus after aneurysmal subarachnoid hemorrhage: a dual-center retrospective cohort study.

Chengzhang Zheng, Jianhuang Huang, Xinguo Wang … Xiang Gu

CerebrovascularAI / informaticsPrognosisQ2IF 3.1

Machine learning model externally validated for predicting shunt-dependent hydrocephalus after subarachnoid hemorrhage

Development and external validation of a machine learning model based on multimodal dynamic data for predicting shunt-dependent hydrocephalus after aneurysmal subarachnoid hemorrhage: a dual-center retrospective cohort study.

Chengzhang Zheng … Xiang Gu

Frontiers in neurology · 2026 · Q2 in Neurology

Purpose

To develop and externally validate a machine learning model for predicting shunt-dependent hydrocephalus (shunt-dependent hydrocephalus) after aneurysmal subarachnoid hemorrhage (aSAH) using multimodal dynamic data.

Methods

This dual-center retrospective study enrolled aSAH patients from Mindong Hospital (development cohort, 2018-2025) and Putian University Hospital (external validation cohort, 2016-2025).

n = 228 patients

AUC 0.867

Results

AUC 0.867

external testing confirms strong accuracy, 1.0 would be perfect prediction

n = 228 patients

More results

The development cohort included 228 patients (shunt-dependent hydrocephalus 21.9%); external validation included 102 patients (shunt-dependent hydrocephalus 23.5%).

Random forest achieved optimal internal validation performance (AUC 0.894, AUPRC 0.676, Brier 0.112).

At threshold 0.30, sensitivity and specificity were 0.920 and 0.840.

More results

Risk stratification demonstrated shunt-dependent hydrocephalus rates of 5.3, 35.7, and 64.7% in low-, intermediate-, and high-risk groups ( p < 0.001).

“

Conclusion · 1 of 2

The random forest model integrating multimodal dynamic data accurately predicts shunt-dependent hydrocephalus after aSAH with strong external validity; ablation analysis confirmed that incorporating dynamic features significantly improved discrimination (AUC 0.894 vs.

Conclusion · 2 of 2

0.831 for the static-only model, ΔAUC = 0.063, p = 0.028), identifying dynamic monitoring data - rather than algorithm choice alone - as a key driver of predictive performance and facilitating early identification and risk stratification of high-risk patients.

Development and external validation of a machine learning model based on multimodal dynamic data for predicting shunt-dependent hydrocephalus after aneurysmal subarachnoid hemorrhage: a dual-center retrospective cohort study.

Development and external validation of a machine learning model based on multimodal dynamic data for predicting shunt-dependent hydrocephalus after aneurysmal subarachnoid hemorrhage: a dual-center retrospective cohort study.

Related papers