# Application and performance of deep learning models for the automated diagnosis of cervical central spinal stenosis on MRI: a systematic review

## TL;DR

This paper reviews how deep learning models can automatically diagnose cervical spinal stenosis from MRI scans, showing high accuracy but needing more validation.

## Contribution

The study systematically evaluates deep learning models for diagnosing cervical central spinal stenosis using MRI, highlighting performance and limitations.

## Key findings

- Most models achieved AUC ≥0.90 and accuracy ≥0.85 in diagnosing CCSS.
- Only one study reported true external validation, raising concerns about generalizability.
- Reporting inconsistencies hinder comparability across studies.

## Abstract

Cervical central spinal stenosis (CCSS) is a leading cause of adult spinal cord dysfunction. Magnetic resonance imaging (MRI) is the reference standard, but reporting is time-consuming and subject to inter-observer variability. Artificial intelligence (AI)—especially deep-learning—may enable automated, consistent assessment.

Evaluation of performance metrics of AI models for diagnosing CCSS.

Following PRISMA 2020, we searched PubMed, Cochrane, Embase, IEEE Xplore, and Web of Science (2015–July 2025) for studies training and evaluating AI models using MRI to diagnose or grade CCSS. We excluded studies limited to foraminal stenosis, non-MRI modalities, thoracic/lumbar levels, segmentation-only or image-enhancement tools without diagnostic output, and studies focused solely on non-stenotic cervical pathologies. Data were extracted on MRI protocol, model type, data splits and external validation, stenosis classification, and diagnostic performance.

Ten studies (2019–2025) met inclusion criteria, predominantly single-centre and retrospective. Most models used T2-weighted axial and/or sagittal MRI; CNNs (e.g., ResNet-50, EfficientNet) and Transformer-based architectures were common. Sensitivities ranged roughly 0.67–1.00 and specificities 0.42–0.97 across models, with many reporting AUCs ≥0.90 and accuracies ≥0.85. Only one study reported true external test performance. Reporting of confidence intervals, processing time, and explainability (e.g., Grad-CAM) was inconsistent.

Deep-learning shows promising diagnostic performance for automated CCSS assessment on MRI and could reduce variability and reporting time. However, generalisability remains uncertain due to small, retrospective, largely single-centre cohorts and scarce external validation. Standardized reporting (e.g., CLAIM) and prospective, multi-centre validation is needed before routine clinical deployment.

•Systematic review of application of AI deep learning on MRI for the diagnosis of cervical central spinal stenosis.•Inclusion of CNN and Transformer-based models.•Most models achieved high diagnostic performance (AUC ≥0.90; accuracy ≥0.85).•External validation was rare; limited comparability due to reporting heterogeneity.•Standardized reporting and multicentre prospective validation are essential.

Systematic review of application of AI deep learning on MRI for the diagnosis of cervical central spinal stenosis.

Inclusion of CNN and Transformer-based models.

Most models achieved high diagnostic performance (AUC ≥0.90; accuracy ≥0.85).

External validation was rare; limited comparability due to reporting heterogeneity.

Standardized reporting and multicentre prospective validation are essential.