Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Apr 14;13(8):1417.
doi: 10.3390/diagnostics13081417.

NAMSTCD: A Novel Augmented Model for Spinal Cord Segmentation and Tumor Classification Using Deep Nets

Ricky Mohanty  1 , Sarah Allabun  2 , Sandeep Singh Solanki  3 , Subhendu Kumar Pani  4 , Mohammed S Alqahtani  5   6 , Mohamed Abbas  7 , Ben Othman Soufiene  8
Affiliations

NAMSTCD: A Novel Augmented Model for Spinal Cord Segmentation and Tumor Classification Using Deep Nets

Ricky Mohanty et al. Diagnostics (Basel). .

Abstract

Spinal cord segmentation is the process of identifying and delineating the boundaries of the spinal cord in medical images such as magnetic resonance imaging (MRI) or computed tomography (CT) scans. This process is important for many medical applications, including the diagnosis, treatment planning, and monitoring of spinal cord injuries and diseases. The segmentation process involves using image processing techniques to identify the spinal cord in the medical image and differentiate it from other structures, such as the vertebrae, cerebrospinal fluid, and tumors. There are several approaches to spinal cord segmentation, including manual segmentation by a trained expert, semi-automated segmentation using software tools that require some user input, and fully automated segmentation using deep learning algorithms. Researchers have proposed a wide range of system models for segmentation and tumor classification in spinal cord scans, but the majority of these models are designed for a specific segment of the spine. As a result, their performance is limited when applied to the entire lead, limiting their deployment scalability. This paper proposes a novel augmented model for spinal cord segmentation and tumor classification using deep nets to overcome this limitation. The model initially segments all five spinal cord regions and stores them as separate datasets. These datasets are manually tagged with cancer status and stage based on observations from multiple radiologist experts. Multiple Mask Regional Convolutional Neural Networks (MRCNNs) were trained on various datasets for region segmentation. The results of these segmentations were combined using a combination of VGGNet 19, YoLo V2, ResNet 101, and GoogLeNet models. These models were selected via performance validation on each segment. It was observed that VGGNet-19 was capable of classifying the thoracic and cervical regions, while YoLo V2 was able to efficiently classify the lumbar region, ResNet 101 exhibited better accuracy for sacral-region classification, and GoogLeNet was able to classify the coccygeal region with high performance accuracy. Due to use of specialized CNN models for different spinal cord segments, the proposed model was able to achieve a 14.5% better segmentation efficiency, 98.9% tumor classification accuracy, and a 15.6% higher speed performance when averaged over the entire dataset and compared with various state-of-the art models. This performance was observed to be better, due to which it can be used for various clinical deployments. Moreover, this performance was observed to be consistent across multiple tumor types and spinal cord regions, which makes the model highly scalable for a wide variety of spinal cord tumor classification scenarios.

Keywords: classification; convolutional; cord; mask regions; neural network; segmentation; segments; spinal; tumor.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Spinal cord tumor classification model flowchart.
Figure 2
Figure 2
Overall flow of the proposed model.
Figure 3
Figure 3
Input spinal cord image, its Saliency Mask, and final Saliency Map image.
Figure 4
Figure 4
Masked Region CNN model for pixel-level classification for segment extraction.
Figure 5
Figure 5
Design of the Inception Model.
Figure 6
Figure 6
A typical CNN model used for classification of spinal cord regions.
Figure 7
Figure 7
Accuracy of classification for different models.
Figure 8
Figure 8
PSNR of classification for different models.
Figure 9
Figure 9
Classification delay of different models.
See this image and copyright information in PMC

References

    1. Sabaghian S., Dehghani H., Batouli S.A.H., Khatibi A., Oghabian M.A. Fully Automatic 3D Segmentation of the Thoracolumbar Spinal Cord and the Vertebral Canal from T2-Weighted MRI Using K-Means Clustering Algorithm. Spinal Cord. 2020;58:811–820. doi: 10.1038/s41393-020-0429-3. - DOI - PubMed
    1. Liao C.C., Ting H.W., Xiao F. Atlas-free cervical spinal cord segmentation on midsagittal T2-weighted magnetic resonance images. J. Healthc. Eng. 2017;2017:1–12. doi: 10.1155/2017/8691505. - DOI - PMC - PubMed
    1. Ahammad S.H., Rahman M.Z.U., Lay-Ekuakille A., Giannoccaro N.I. An Efficient optimal threshold-based segmentation and classification model for multi-level spinal cord Injury detection; Proceedings of the 2020 IEEE International Symposium on Medical Measurements and Applications (MeMeA); Bari, Italy. 1 June–1 July 2020.
    1. Mnassri B., Sahnoun M., Hamida A.B. Comparison Study for Spinal Cord Segmentation Methods aiming to detect SC Atrophy in MRI images: Case of Multiple Sclerosis; Proceedings of the 2020 5th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP); Sousse, Tunisia. 2–5 September 2020.
    1. Le Couedic T., Caillon R., Rossant F., Joutel A., Urien H., Rajani R.M. Deep-learning based segmentation of challenging myelin sheaths; Proceedings of the 2020 Tenth International Conference on Image Processing; Paris, France. 9–12 November 2020.