Deep Learning for Soft Tissue Sarcoma Management

Soft tissue sarcomas (STSs) represent a diverse group of tumors that pose significant diagnostic and therapeutic challenges. In a recent review published in the KeAi journal Meta-Radiology, a team of researchers from The Second Xiangya Hospital at Central South University in Changsha, China, explored the potential application of deep learning (DL) in revolutionizing the management of these complex tumors.

"Deep learning has shown remarkable promise in various medical fields, and its application in STSs is no exception. Our review has synthesized the most recent advancements and highlighted how deep learning can improve the accuracy of diagnosis, personalize treatment plans, and predict patient outcomes more effectively and efficiently,” shares Zhihong Li, senior and co-corresponding author of the study.

The review covers several key areas where deep learning is making an impact:

1. Data Acquisition and Processing: The integration of multi-modal data, including radiographic images and histopathological slides, enhances the diagnostic process.
2. Algorithm Development: Advanced deep learningmodels such as convolutional neural networks (CNNs) and generative adversarial networks (GANs) have been developed to improve image analysis and data augmentation.
3. Clinical Applications: Deep learningmodels have been successfully used to automate the contouring of gross tumor volumes (GTVs) for radiation therapy, predict treatment responses, and stratify patients based on risk.
4. Pathological Diagnosis: Automation of diagnostic systems using deep learningalgorithms can assist pathologists in accurately classifying STS subtypes and identifying prognostic biomarkers.

Prof. Chao Tu, who led the study alongside Li, emphasized the importance of high-quality data and optimized algorithms. "The success of deep learning in clinical applications depends heavily on the quality of the input data and the robustness of the algorithms. Our review has underscored the need for well-annotated datasets and continuous algorithm refinement."

This research was supported by several grants, including the National Natural Foundation of China and the Hunan Provincial Natural Science Foundation. The authors hope that their review will encourage further research and the adoption of deep learning technologies in clinical practice, ultimately improving outcomes for patients with soft tissue sarcomas.

Contact author name, affiliation, email address:

Zhihong Li, Department of Orthopaedics, The Second Xiangya Hospital of Central South University, Changsha, China.

Email: lizhihong@csu.edu.cn.

Funder:

This work was supported by the National Natural Foundation of China (82272664, 81902745), Hunan Provincial Natural Science Foundation of China (2022JJ30843), and the Science and Technology Innovation Program of Hunan Province (2023RC3085).

Conflict of interest: 

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The Author Zhihong Li is the Editor-in Chief of the journal, but was not involved in the peer review procedure. This paper was handled by another Editor Board member.

See the article: 

Xu, R., Tang, J., Li, C., Wang, H., Li, L., He, Y., Tu, C., & Li, Z. (2024). Deep learning-based artificial intelligence for assisting diagnosis, assessment, and treatment in soft tissue sarcomas. Meta-Radiology, 2, 100069. https://doi.org/10.1016/j.metrad.2024.100069