Although surgical resection improves overall survival in patients with diffuse Low-grade gliomas (DLGG), it can also result in deterioration of neurocognitive function, which are poorly understood and lack effective predictive models. This study aims to develop a model using whole-brain tumor burden metrics, inflammatory and molecular markers for predicting high risk of neurocognitive decline (ND) postoperatively. The study involved 192 patients with left frontal DLGG. MRI data were analyzed to derive whole-brain tumor burden metrics, including tumor radiomics, whole-brain cortical thickness, myelin content, and network characteristics. postoperative inflammatory and molecular markers were collected. Postoperative follow-up neurocognitive function was assessed using the Montreal Cognitive Assessment at 3 months and 1 year. Machine learning models were constructed using Pycaret.

Inclusion Criteria:

1. Histopathologically proven DLGG based on 2021 WHO criteria
2. Age ≥ 18 years,
3. Karnofsky Performance Status ≥ 70%,
4. Tumor involved unilateral left frontal lobe and did not reach the central sulcus.

Exclusion Criteria:

1. With previous treatment before image acquisition,
2. Presenting cognitive impairment pre-operatively based on The Montreal Cognitive Assessment (MoCA)
3. Survival of less than 3 months post-surgery
4. Lack of compliance with neurocognitive assessments in 3-month and 1-year after surgery

192 glioma patients