Artificial intelligence and chordoma: A scoping review of the current landscape and future directions
Introduction: Chordomas are rare, locally aggressive tumours that present significant treatment challenges due to their proximity to critical neurovascular structures. Artificial intelligence (AI) methodologies have shown promise in enhancing diagnostic precision, surgical planning, and prognosticat...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-01-01
|
| Series: | Brain and Spine |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772529425000906 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Introduction: Chordomas are rare, locally aggressive tumours that present significant treatment challenges due to their proximity to critical neurovascular structures. Artificial intelligence (AI) methodologies have shown promise in enhancing diagnostic precision, surgical planning, and prognostication in various cancers. Research question: What is the current landscape of AI applications in chordoma management, and what are the key limitations and future directions for integrating AI into clinical practice for this rare malignancy? Materials and methods: We conducted a scoping review following the PRISMA-ScR guidelines and the Arksey and O'Malley framework. A search of five databases with an end date of November 9, 2024, identified peer-reviewed studies assessing AI or machine learning applications in chordoma management. Data extraction focused on study characteristics, methodologies, clinical tasks, and performance metrics. Results: Twenty-one studies published between 2017 and 2024 were included, encompassing 5486 patients. The studies addressed diverse clinical tasks: 7 focused on differentiating chordomas from other tumours or classifying subtypes, 6 on survival prediction, 2 on tumour segmentation, 2 on outcome prediction, and 4 miscellaneous tasks. Common algorithms used included convolutional neural networks, support vector machines, random forests, and clustering algorithms. Limitations identified across studies included small sample sizes, single-center data, reliance on single data modalities, and issues with model interpretability. Discussion and conclusion: AI applications in chordoma management show potential in improving diagnostic accuracy, surgical planning, and prognostication. Future research should focus on collaborative efforts for larger, diverse datasets with external validation cohorts, interpretable multimodal models, and validation through prospective clinical trials. |
|---|---|
| ISSN: | 2772-5294 |