# Grey-to-white matter ratio on computed tomography for predicting neurological outcome in patients with heat stroke: a retrospective cohort study

**Authors:** Hua Wei, Hongling Zhu, Menglong Liu, Xiaodan Zhu, Anyong Yu, Can Luo, Qingbo Zeng, Fating Zhou, Haizhen Duan

PMC · DOI: 10.3389/fneur.2025.1556822 · Frontiers in Neurology · 2025-05-13

## TL;DR

This study shows that a lower grey-to-white matter ratio on CT scans can predict poor neurological outcomes in heat stroke patients.

## Contribution

The study introduces the grey-to-white matter ratio (GWR) as a novel predictor of neurological outcomes in heat stroke patients.

## Key findings

- All GWR parameters were significantly lower in patients with poor neurological outcomes.
- Combining GWR with qSOFA improved prediction accuracy for poor outcomes in heat stroke patients.

## Abstract

Grey-to-white matter ratio (GWR) is an early and sensitive indicator of cerebral oedema in patients with hypoxic-ischaemic brain injury, we aimed to evaluate the prognostic value of GWR for predicting neurological outcome in heat stroke patients.

This multicentre retrospective analysis included 86 patients with heat stroke patients who underwent cranial computed tomography (CT). Patients were stratified by Cerebral Performance Category (CPC) scores at discharge: good outcome (CPC 1–2, n = 65) versus poor outcome (CPC 3–5, n = 21) in the derivation cohort. Seven GWR parameters were calculated from Hounsfield unit measurements at three different regions (basal ganglia, centrum semiovale, high convexity): putamen/corpus callosum (PU/CC), caudate nucleus/posterior limb of internal capsule (CN/PLIC), CN/CC, PU/PLIC, GWRbasal ganglia, GWRcerebrum, and GWRaverage. Prognostic performance of GWR was compared with qSOFA using receiver operating characteristic (ROC) analysis. And a validation cohort was used to verify the reliability.

All GWRs were significantly lower in the poor outcome group than in the good outcome group. ROC analysis showed the following areas under the curve: PU/CC, 0.836; CN/PLIC, 0.815; CN/CC, 0.858; PU/PLIC, 0.814; GWRbasal ganglia, 0.855; GWRcerebrum, 0.803; GWRaverage, 0.837. The cutoff values with 90.77% specificity in predicting poor outcome were as follows: PU/CC, 1.20 (sensitivity, 76.19%); CN/PLIC, 1.17 (sensitivity, 52.38%); CN/CC, 1.20 (sensitivity, 76.19%); PU/PLIC, 1.20 (sensitivity, 61.90%); GWRbasal ganglia, 1.23 (sensitivity, 80.95%); GWRcerebrum, 1.19 (sensitivity, 57.14%); GWRaverage, 1.23 (sensitivity, 71.43%). The sensitivity of GWRbasal ganglia significantly increased when combined with qSOFA in the derivation and validation cohorts.

A low GWR was strongly associated with poor outcome in the heat stroke patients. The GWR may be useful as an objective early predictor of poor neurological outcome in the heat stroke patients. Incorporating the GWR with qSOFA significantly enhanced the prediction performance.

## Full-text entities

- **Diseases:** heat stroke (MESH:D018883), hypoxic-ischaemic brain injury (MESH:D002534), cerebral oedema (MESH:D001929)
- **Species:** Homo sapiens (human, species) [taxon 9606]

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12106023/full.md

## Figures

6 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12106023/full.md

## References

34 references — full list in the complete paper: https://tomesphere.com/paper/PMC12106023/full.md

---
Source: https://tomesphere.com/paper/PMC12106023