Role of Advanced MRI-Based Perfusion and Diffusion Imaging in Predicting Recovery After the Surgical Management of Cervical Myelopathy: A Retrospective Study

TL;DR

This study shows that advanced MRI techniques can predict how well patients with cervical myelopathy will recover after surgery, using measures like fractional anisotropy and perfusion fraction.

Contribution

The study introduces a novel use of combined perfusion and diffusion MRI metrics to predict recovery outcomes in cervical myelopathy patients.

Findings

Fractional anisotropy (FA) and perfusion fraction (f) strongly correlate with postoperative neurological recovery.

A combined FA + f model achieved the highest predictive accuracy (AUC=0.90) for recovery outcomes.

Advanced MRI metrics outperform conventional MRI in predicting recovery after cervical myelopathy surgery.

Abstract

Background: Cervical myelopathy is a common cause of non-traumatic spinal cord dysfunction, where both structural compression and microvascular compromise contribute to neurological deficits.

Objective: This study aimed to evaluate the prognostic value of advanced MRI-based perfusion and diffusion parameters in predicting neurological recovery following surgical decompression in patients with cervical myelopathy.

Methodology: This retrospective observational study was conducted at the Department of Neurosurgery, Punjab Institute of Neurosciences, Lahore, Pakistan, from September 2024 to September 2025 and included 105 patients with clinically and radiologically confirmed cervical myelopathy who underwent decompressive surgery. Preoperative MRI included diffusion tensor imaging (DTI)-derived parameters fractional anisotropy (FA), apparent diffusion coefficient (ADC), and intravoxel incoherent motion (IVIM) perfusion metrics true diffusion coefficient (D), pseudo-diffusion coefficient (D*), and perfusion fraction (f).

Results: The mean preoperative and postoperative Modified Japanese Orthopaedic Association (mJOA) scores were 10.4±2.3 and 14.9±1.8, respectively (p<0.001). FA showed a strong positive correlation with recovery rate (r=0.68; p<0.001), while ADC demonstrated a significant negative correlation (r=-0.59; p<0.001). Perfusion fraction also correlated positively with recovery (r=0.53; p<0.01). The receiver operating characteristic (ROC) analysis identified FA ≥0.41 and f ≥9.5% as optimal thresholds for predicting good recovery (area under the curve (AUC)=0.86 and 0.78, respectively). A combined FA+f model achieved the highest predictive accuracy (AUC=0.90).

Conclusion: It is concluded that advanced MRI-based diffusion and perfusion imaging provides a reliable, quantitative method for predicting postoperative recovery in cervical myelopathy. FA and perfusion fraction are the most effective predictors of neurological improvement, offering superior prognostic value compared to conventional MRI.