Survival prediction models in motor neuron disease

Area of Science:

• Neuroimaging
• Neurology
• Oncology

Background:

• Motor neuron disease (MND) prognosis is challenging.
• Predictive models for MND survival need enhancement.
• Multimodal data integration is crucial for accurate prognostication.

Purpose of the Study:

• To evaluate the predictive value of multimodal brain MRI in motor neuron disease (MND) survival.
• To assess the added prognostic accuracy of MRI metrics beyond clinical and cognitive data.
• To investigate survival prediction in amyotrophic lateral sclerosis (ALS) subtypes.

Main Methods:

• Prospective follow-up of 200 MND patients for a median of 4.13 years.
• Baseline brain MRI (grey matter volumes, diffusion tensor imaging) and clinical/cognitive assessments.
• Multivariable survival modeling (Royston-Parmar) to compare predictive models.

Main Results:

• Clinical model predicted 4-year survival with AUC 0.79.
• Combined clinical and MRI model achieved AUC 0.89.
• For ALS patients, combined model AUC increased from 0.62 to 0.77.

Conclusions:

• Brain MRI measures of structural damage improve MND survival prediction.
• Multimodal approach combining MRI with clinical/cognitive data offers superior prognostication.
• This is particularly relevant for amyotrophic lateral sclerosis patients.