Chen S, Li G, Li F, Wang G, Wang Q. A dynamic nomogram for predicting the probability of irreversible neurological dysfunction after cervical spinal cord injury: research based on clinical features and MRI data.
BMC Musculoskelet Disord 2023;
24:459. [PMID:
37277760 DOI:
10.1186/s12891-023-06570-z]
[Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 12/06/2022] [Accepted: 05/25/2023] [Indexed: 06/07/2023] Open
Abstract
BACKGROUND
Irreversible neurological dysfunction (IND) is an adverse event after cervical spinal cord injury (CSCI). However, there is still a shortage of objective criteria for the early prediction of neurological function. We aimed to screen independent predictors of IND and use these findings to construct a nomogram that could predict the development of neurological function in CSCI patients.
METHODS
Patients with CSCI attending the Affiliated Hospital of Southwest Medical University between January 2014 and March 2021 were included in this study. We divided the patients into two groups: reversible neurological dysfunction (RND) and IND. The independent predictors of IND in CSCI patients were screened using the regularization technique to construct a nomogram, which was finally converted into an online calculator. Concordance index (C-index), calibration curves analysis and decision curve analysis (DCA) evaluated the model's discrimination, calibration, and clinical applicability. We tested the nomogram in an external validation cohort and performed internal validation using the bootstrap method.
RESULTS
We enrolled 193 individuals with CSCI in this study, including IND (n = 75) and RND (n = 118). Six features, including age, American spinal injury association Impairment Scale (AIS) grade, signal of spinal cord (SC), maximum canal compromise (MCC), intramedullary lesion length (IMLL), and specialized institution-based rehabilitation (SIBR), were included in the model. The C-index of 0.882 from the training set and its externally validated value of 0.827 demonstrated the model's prediction accuracy. Meanwhile, the model has satisfactory actual consistency and clinical applicability, verified in the calibration curve and DCA.
CONCLUSION
We constructed a prediction model based on six clinical and MRI features that can be used to assess the probability of developing IND in patients with CSCI.
Collapse