Evaluation of a deep learning segmentation tool to help detect spinal cord lesions from combined T2 and STIR acquisitions in people with multiple sclerosis.
Saved in:
| Title: | Evaluation of a deep learning segmentation tool to help detect spinal cord lesions from combined T2 and STIR acquisitions in people with multiple sclerosis. |
|---|---|
| Authors: | Lodé, Baptiste1,2 (AUTHOR), Hussein, Burhan Rashid2 (AUTHOR), Meurée, Cédric2 (AUTHOR), Walsh, Ricky2 (AUTHOR), Gaubert, Malo2 (AUTHOR), Lassalle, Nicolas3 (AUTHOR), Courbon, Guilhem4 (AUTHOR), Martin, Agathe4 (AUTHOR), Le Bars, Jeanne4 (AUTHOR), Durand-Dubief, Françoise5,6 (AUTHOR), Bourre, Bertrand7 (AUTHOR), Maarouf, Adil8 (AUTHOR), Outteryck, Olivier9 (AUTHOR), Mehier, Clément3 (AUTHOR), Poulin, Alexandre3 (AUTHOR), Cathelineau, Camille3 (AUTHOR), Hong, Jeremy3 (AUTHOR), Criton, Guillaume10,11 (AUTHOR), Motillon-Alonso, Sophie3 (AUTHOR), Lecler, Augustin12,13 (AUTHOR) |
| Source: | European Radiology. Oct2025, Vol. 35 Issue 10, p5954-5964. 11p. |
| Subjects: | Deep learning, Multiple sclerosis, Spinal cord tumors, Sensitivity analysis, Medical equipment, Inter-observer reliability |
| Abstract: | Objective: To develop a deep learning (DL) model for the detection of spinal cord (SC) multiple sclerosis (MS) lesions from both sagittal T2 and short tau inversion recovery (STIR) sequences and to investigate whether such a model could improve the performance of clinicians in detecting SC lesions. Materials and methods: A DL tool was developed based on SC sagittal T2 and STIR acquisitions from the imaging database of the French MS registry (OFSEP), including retrospective data from 40 different scanners. A multi-reader study based on retrospective data was performed between December 2023 and June 2024 to compare the performance of 20 clinicians in interpreting upper and lower SC acquisitions with and without the use of the tool. A ground truth was established by three experts. Sensitivity, precision, and inter-reader variability were evaluated. Results: We included 50 patients (39 females, median age: 41 years [range: 15–67]) with SC MRI acquired between February 2017 and December 2022. When reading with the tool, the clinicians' mean sensitivity to detect SC lesions improved (from 74.3% [95% CI = 67.8–80.6%] to 79.2% [95% CI: 73.5–85.0%]; p < 0.0001), with no evidence of difference in the mean precision: (69.0% [95% CI: 62.8–75.2%] vs 70.1% [95% CI: 64.3–75.9%]; p = 0.08). Inter-reader variability in lesion detection was slightly improved with the tool (Light's kappa = 0.55 vs 0.60), but without statistical difference (p = 0.056). Conclusion: The use of an automatic tool can help clinicians detect SC lesions in pwMS by increasing their sensitivity. Key Points: QuestionNo tool to help detect MS SC lesions is used in clinical practice despite their frequency and prognostic value. FindingsThis DL-based tool led to improvement in clinicians' sensitivity in detecting SC lesions from both sagittal T2 and STIR sequences, without decreasing precision. Clinical relevanceOur study indicated the potential of a DL-based tool to assist clinicians in the challenging task of detecting SC lesions in people with MS on a combination of sequences commonly acquired in clinical practice. [ABSTRACT FROM AUTHOR] |
| Copyright of European Radiology is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Objective: To develop a deep learning (DL) model for the detection of spinal cord (SC) multiple sclerosis (MS) lesions from both sagittal T2 and short tau inversion recovery (STIR) sequences and to investigate whether such a model could improve the performance of clinicians in detecting SC lesions. Materials and methods: A DL tool was developed based on SC sagittal T2 and STIR acquisitions from the imaging database of the French MS registry (OFSEP), including retrospective data from 40 different scanners. A multi-reader study based on retrospective data was performed between December 2023 and June 2024 to compare the performance of 20 clinicians in interpreting upper and lower SC acquisitions with and without the use of the tool. A ground truth was established by three experts. Sensitivity, precision, and inter-reader variability were evaluated. Results: We included 50 patients (39 females, median age: 41 years [range: 15–67]) with SC MRI acquired between February 2017 and December 2022. When reading with the tool, the clinicians' mean sensitivity to detect SC lesions improved (from 74.3% [95% CI = 67.8–80.6%] to 79.2% [95% CI: 73.5–85.0%]; p < 0.0001), with no evidence of difference in the mean precision: (69.0% [95% CI: 62.8–75.2%] vs 70.1% [95% CI: 64.3–75.9%]; p = 0.08). Inter-reader variability in lesion detection was slightly improved with the tool (Light's kappa = 0.55 vs 0.60), but without statistical difference (p = 0.056). Conclusion: The use of an automatic tool can help clinicians detect SC lesions in pwMS by increasing their sensitivity. Key Points: QuestionNo tool to help detect MS SC lesions is used in clinical practice despite their frequency and prognostic value. FindingsThis DL-based tool led to improvement in clinicians' sensitivity in detecting SC lesions from both sagittal T2 and STIR sequences, without decreasing precision. Clinical relevanceOur study indicated the potential of a DL-based tool to assist clinicians in the challenging task of detecting SC lesions in people with MS on a combination of sequences commonly acquired in clinical practice. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 09387994 |
| DOI: | 10.1007/s00330-025-11541-0 |
