1
|
Gu Z, Dai W, Chen J, Jiang Q, Lin W, Wang Q, Chen J, Gu C, Li J, Ying G, Zhu Y. Convolutional neural network-based magnetic resonance image differentiation of filum terminale ependymomas from schwannomas. BMC Cancer 2024; 24:350. [PMID: 38504164 PMCID: PMC10949807 DOI: 10.1186/s12885-024-12023-0] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/18/2023] [Accepted: 02/20/2024] [Indexed: 03/21/2024] Open
Abstract
PURPOSE Preoperative diagnosis of filum terminale ependymomas (FTEs) versus schwannomas is difficult but essential for surgical planning and prognostic assessment. With the advancement of deep-learning approaches based on convolutional neural networks (CNNs), the aim of this study was to determine whether CNN-based interpretation of magnetic resonance (MR) images of these two tumours could be achieved. METHODS Contrast-enhanced MRI data from 50 patients with primary FTE and 50 schwannomas in the lumbosacral spinal canal were retrospectively collected and used as training and internal validation datasets. The diagnostic accuracy of MRI was determined by consistency with postoperative histopathological examination. T1-weighted (T1-WI), T2-weighted (T2-WI) and contrast-enhanced T1-weighted (CE-T1) MR images of the sagittal plane containing the tumour mass were selected for analysis. For each sequence, patient MRI data were randomly allocated to 5 groups that further underwent fivefold cross-validation to evaluate the diagnostic efficacy of the CNN models. An additional 34 pairs of cases were used as an external test dataset to validate the CNN classifiers. RESULTS After comparing multiple backbone CNN models, we developed a diagnostic system using Inception-v3. In the external test dataset, the per-examination combined sensitivities were 0.78 (0.71-0.84, 95% CI) based on T1-weighted images, 0.79 (0.72-0.84, 95% CI) for T2-weighted images, 0.88 (0.83-0.92, 95% CI) for CE-T1 images, and 0.88 (0.83-0.92, 95% CI) for all weighted images. The combined specificities were 0.72 based on T1-WI (0.66-0.78, 95% CI), 0.84 (0.78-0.89, 95% CI) based on T2-WI, 0.74 (0.67-0.80, 95% CI) for CE-T1, and 0.81 (0.76-0.86, 95% CI) for all weighted images. After all three MRI modalities were merged, the receiver operating characteristic (ROC) curve was calculated, and the area under the curve (AUC) was 0.93, with an accuracy of 0.87. CONCLUSIONS CNN based MRI analysis has the potential to accurately differentiate ependymomas from schwannomas in the lumbar segment.
Collapse
Affiliation(s)
- Zhaowen Gu
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Wenli Dai
- Zhejiang University School of Mathematical Sciences, Hangzhou, Zhejiang, China
| | - Jiarui Chen
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Qixuan Jiang
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Weiwei Lin
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Qiangwei Wang
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Jingyin Chen
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Chi Gu
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China
| | - Jia Li
- Ningbo Medical Center Lihuili Hospital, Department of Neurosurgery, Ningbo University, 1111, Jiangnan Road, Ningbo, Zhejiang, China.
| | - Guangyu Ying
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China.
| | - Yongjian Zhu
- Department of Neurosurgery, The Second Affiliated Hospital, School of Medicine, Zhejiang University, 88, Jiefang Road, Hangzhou, China.
- Clinical Research Center for Neurological Diseases of Zhejiang Province, Hangzhou, China.
| |
Collapse
|
2
|
Dauleac C, Boulogne S, Barrey CY, Guyotat J, Jouanneau E, Mertens P, Berhouma M, Jung J, André-Obadia N. Predictors of functional outcome after spinal cord surgery: Relevance of intraoperative neurophysiological monitoring combined with preoperative neurophysiological and MRI assessments. Neurophysiol Clin 2022; 52:242-251. [PMID: 35396150 DOI: 10.1016/j.neucli.2022.03.004] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/13/2021] [Revised: 03/11/2022] [Accepted: 03/11/2022] [Indexed: 10/18/2022] Open
Abstract
OBJECTIVES To assess the accuracy of intraoperative neurophysiological monitoring (IONM) in predicting immediate and 3-month postoperative neurological new deficit (or deterioration) in patients benefiting from spinal cord (SC) surgery; and to identify factors associated with a higher risk of postoperative clinical worsening. METHODS Consecutive patients who underwent SC surgery with IONM were included. Pre and postoperative clinical (modified McCormick scale), radiological (lesion-occupying area ratio), and electrophysiological features were collected. RESULTS A total of 99 patients were included: 14 (14.1%) underwent extradural surgery, 50 (50.5%) intradural extramedullary surgery, and 35 (35.4%) intramedullary surgery. Cumulatively, multimodal IONM (motor and somatosensory evoked potentials, D-wave whenever possible) significantly predicted postoperative deficits (p<0.001), with a sensitivity, specificity, positive predictive value, and negative predictive value of 0.81, 0.93, 0.83, and 0.92, respectively. Sixty (60.6%) patients displayed no IONM change, whereas 39 (39.4%) displayed IONM worsening. In multivariate analysis, predictors for postoperative clinical worsening were: abnormal preoperative electrophysiological assessment (p=0.03), intramedullary tumor (p<0.001), lesion-occupying area ratio ≥0.7 (p<0.001), and IONM alterations (p<0.001). Three months after the surgical procedure, in patients presenting at least one of the risk factors described above, 45/81 (55.6%) and 19/81 (23.5%) were clinically and electrophysiologically improved, respectively; while 13/81 (16.0%) and 10/81 (12.3%) were clinically and electrophysiologically worsened. CONCLUSION Multimodal IONM is an essential tool to guide SC surgery, and enables the accurate prediction of postoperative neurological outcome. Specific attention should be given to patients presenting with preoperative electrophysiological abnormalities, large tumor volume, and intramedullary tumor location.
Collapse
Affiliation(s)
- Corentin Dauleac
- Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurologie Fonctionnelle et Epileptologie, Lyon, France; Université Lyon I, Université Claude Bernard, Lyon, France; Laboratoire CREATIS, CNRS UMR5220, Inserm U1206, INSA-Lyon; Université de Lyon I, Lyon, France.
| | - Sébastien Boulogne
- Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurologie Fonctionnelle et Epileptologie, Lyon, France; Université Lyon I, Université Claude Bernard, Lyon, France; Centre de Recherche de Neurosciences de Lyon, INSERM UMRS 1028, CNRS UMR 5292, Université Claude Bernard Lyon 1, Université de Lyon, F-69000 Lyon, France
| | - Cédric Y Barrey
- Université Lyon I, Université Claude Bernard, Lyon, France; Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurochirurgie C, Lyon, France
| | - Jacques Guyotat
- Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurochirurgie D, Lyon, France
| | - Emmanuel Jouanneau
- Université Lyon I, Université Claude Bernard, Lyon, France; Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurochirurgie B, Lyon, France
| | - Patrick Mertens
- Université Lyon I, Université Claude Bernard, Lyon, France; Centre de Recherche de Neurosciences de Lyon, INSERM UMRS 1028, CNRS UMR 5292, Université Claude Bernard Lyon 1, Université de Lyon, F-69000 Lyon, France; Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurochirurgie A, Lyon, France
| | - Moncef Berhouma
- Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurologie Fonctionnelle et Epileptologie, Lyon, France; Laboratoire CREATIS, CNRS UMR5220, Inserm U1206, INSA-Lyon; Université de Lyon I, Lyon, France; Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurochirurgie D, Lyon, France
| | - Julien Jung
- Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurologie Fonctionnelle et Epileptologie, Lyon, France; Université Lyon I, Université Claude Bernard, Lyon, France; Centre de Recherche de Neurosciences de Lyon, INSERM UMRS 1028, CNRS UMR 5292, Université Claude Bernard Lyon 1, Université de Lyon, F-69000 Lyon, France
| | - Nathalie André-Obadia
- Hospices Civils de Lyon, Hôpital neurologique Pierre Wertheimer, Service de Neurologie Fonctionnelle et Epileptologie, Lyon, France; Université Lyon I, Université Claude Bernard, Lyon, France; Centre de Recherche de Neurosciences de Lyon, INSERM UMRS 1028, CNRS UMR 5292, Université Claude Bernard Lyon 1, Université de Lyon, F-69000 Lyon, France
| |
Collapse
|