GliomaAI-GBM: Non-Invasive MRI-Based Detection of IDH Wildtype Glioblastoma Using Artificial Intelligence
Trial Snapshot
- Phase
- Not Applicable
- Status
- Completed
- Sponsor
- Enrollment
- 1,372
- Locations
- 1
- Primary Endpoint
- Diagnostic performance of GliomaAI-GBM for identification of IDH wildtype glioblastoma from MRI, measured by accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the ROC.
Study Overview
Brief Summary
The goal of this observational study is to learn whether an artificial intelligence system called GliomaAI-GBM can help detect a specific molecular type of brain tumour called IDH wildtype glioblastoma using routine MRI scans. The study uses previously collected and fully anonymised MRI data from 1,372 patients from 13 institutions in the Cancer Imaging Archive (TCIA).
The main questions it aims to answer are:
- How accurately can GliomaAI-GBM identify IDH wildtype glioblastoma from MRI scans?
- How well does the system perform across data from different hospitals and patient groups?
Researchers will use existing MRI scans and clinical information to train and test the AI system. No new scans, treatments, or hospital visits are required for participants, and all data used is fully anonymised and obtained from an existing research database.
Participants will not be asked to do anything, as this study only uses previously collected imaging data.
Study Design
- Study Type
- Observational
- Observational Model
- Cohort
- Time Perspective
- Retrospective
Eligibility Criteria
- Ages
- 18 Years to — (Adult, Older Adult)
- Sex
- All
- Accepts Healthy Volunteers
- No
Inclusion Criteria
- •Adult (>=18 years of age)
- •Having pre op MRI scan
- •Having biopsy / surgery
- •Having post biopsy/ surgery histology diagnosis and genetic analysis.
Exclusion Criteria
- •MRI scan significantly degraded by motion or other artefact
- •Incomplete genetic analysis
- •Prior treatment (e.g., radiotherapy or chemotherapy) before baseline MRI
Outcomes
Primary Outcomes
Diagnostic performance of GliomaAI-GBM for identification of IDH wildtype glioblastoma from MRI, measured by accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the ROC.
Time Frame: Perioperative
The diagnostic performance of the GliomaAI-GBM artificial intelligence model will be assessed by comparing pre-operative MRI-based predictions of IDH wildtype glioblastoma status against post-operative (biopsy or surgery) molecular/genetic profiling results as the reference standard. Performance metrics including accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and AUC will be calculated.
Secondary Outcomes
No secondary outcomes reported