RESUMEN
BACKGROUND: Methods for grading and localization of lumbar disc herniation (LDH) on MRI are complex, time-consuming, and subjective. Utilizing deep learning (DL) models as assistance would mitigate such complexities. PURPOSE: To develop an interpretable DL model capable of grading and localizing LDH. STUDY TYPE: Retrospective. SUBJECTS: 1496 patients (M/F: 783/713) were evaluated, and randomly divided into training (70%), validation (10%), and test (20%) sets. FIELD STRENGTH/SEQUENCE: 1.5T MRI for axial T2-weighted sequences (spin echo). ASSESSMENT: The training set was annotated by three spinal surgeons using the Michigan State University classification to train the DL model. The test set was annotated by a spinal surgery expert (as ground truth labels), and two spinal surgeons (comparison with the trained model). An external test set was employed to evaluate the generalizability of the DL model. STATISTICAL TESTS: Calculated intersection over union (IoU) for detection consistency, utilized Gwet's AC1 to assess interobserver agreement, and evaluated model performance based on sensitivity and specificity, with statistical significance set at P < 0.05. RESULTS: The DL model achieved high detection consistency in both the internal test dataset (grading: mean IoU 0.84, recall 99.6%; localization: IoU 0.82, recall 99.5%) and external test dataset (grading: 0.72, 98.0%; localization: 0.71, 97.6%). For internal testing, the DL model (grading: 0.81; localization: 0.76), Rater 1 (0.88; 0.82), and Rater 2 (0.86; 0.83) demonstrated results highly consistent with the ground truth labels. The overall sensitivity of the DL model was 87.0% for grading and 84.0% for localization, while the specificity was 95.5% and 94.4%. For external testing, the DL model showed an appreciable decrease in consistency (grading: 0.69; localization: 0.66), sensitivity (77.2%; 76.7%), and specificity (92.3%; 91.8%). DATA CONCLUSION: The classification capabilities of the DL model closely resemble those of spinal surgeons. For future improvement, enriching the diversity of cases could enhance the model's generalization. TECHNICAL EFFICACY: Stage 2.
RESUMEN
BACKGROUND: ACDF treatment of CSM is currently recognized as a surgical method with reliable efficacy. However, the cervical radiographic findings in a certain group of patients showed that the symptoms were not completely relieved. This study will investigate the relationship between cervical parameters and prognoses after ACDF surgery. METHODS: This study collected cases of CSM treated with ACDF in Zhongda Hospital from May 2014 to June 2018. The investigators recorded gender, age, cervical sagittal parameters, fusion segment, BMI, symptom duration, and NDI score. To compare the changes of parameters after surgery and explore the correlation between each factor and NDI score. RESULTS: Generally, cervical lordosis increased and TS-CL decreased after surgery and during follow-up. Postoperative T1S, SVA and SCA decreased significantly compared to preoperative. T1S was positively correlated with CL (r = 0.245), SVA (r = 0.184), and negatively correlated with SCA (r = - 0.314) and NT (r = - 0.222). The last follow-up NDI score was positively correlated with T1S (r = 0.689), SVA (r = 0.155), TS-CL (r = 0.496), and age (r = 0.194), while negatively correlated with SCA (r = - 0.142). A linear regression model was established with the following formula: NDI = 0.809 × (T1S) - 0.152 × (CL) + 1.962 × (Sex) + 0.110 × (Age). T1S (B = 0.205, P < 0.001), CL (B = - 0.094, P = 0.041), and NT (B = 0.142, P = 0.023) were independent risk factors that affected whether the last follow-up NDI score was greater than preoperative. CONCLUSIONS: In ACDF treatment of CSM, there exists a close correlation between cervical sagittal parameters and NDI scores. T1S, CL, sex, and age were linearly dependent on NDI scores. The increase of T1S, NT, and the decrease of CL were risk factors that affected follow-up NDI score greater than preoperative. Reducing T1S is beneficial to clinical recovery.