A new AI system identifies brain tumours before surgery with over 97% accuracy, helping surgeons plan care more safely.
The system from Thomas Jefferson University is designed to reduce much of that uncertainty before a patient enters the operating room.
And experts suggest it could become a ‘silent partner’ in everyday diagnosis.
A research team developed the automated machine learning (AutoML) program to distinguish pituitary macroadenomas from parasellar meningiomas using standard MRI scans.
The findings, reported in Otolaryngology–Head and Neck Surgery, represent the first time this technology has been used to distinguish these two growths preoperatively.
Pituitary macroadenomas grow from the hormone-making gland at the base of the brain. Parasellar meningiomas arise from the tissue surrounding that area. Both can press on optic nerves, blood vessels, and vital brain structures.
Because each tumour requires a different surgical approach, an early and accurate diagnosis can guide surgeons’ entry into the skull and the outcomes they can promise.
Gurston G Nyquist, a professor of otolaryngology and neurological surgery, said: ‘Our automated machine learning model achieved over 97% accuracy in distinguishing between two common types of skull base tumours using preoperative MRI scans. This work is significant because it demonstrates that automated machine learning can streamline model development for medical imaging classification, reducing barriers to implementing AI-based diagnostic support in otolaryngology.’
Nyquist added that the system could assist hospitals of all sizes.
He added: ‘While multi-institutional validation and careful integration into clinical workflows are warranted, this study represents an important step in the development of reliable tools that may improve skull base tumour diagnosis in both community and tertiary care settings.’
According to the study authors, accuracy in reading scans for this region ranges from about 83% to nearly 97%, depending on training and experience.
When a scan misleads surgeons, the impact is felt in the operating room. A surgeon might take a longer route than necessary, expect to remove more tumour than is safe, or underestimate the risks associated with blood vessels. Extra tests and second opinions can also delay care and increase stress for families.
The new system aims to sharpen this first judgement. AutoML automates many technical steps that typically require advanced programming skills.
In this project, the program learned to detect subtle image patterns that people struggle to measure consistently. It does not replace a doctor’s eye; rather it adds another perspective.
To train the model, the team used 1,628 MRI images from 116 patients. At standard settings, the system correctly classified tumours 97.55% of the time. For pituitary macroadenomas, it reached 97% sensitivity and 98.96% specificity. Sensitivity reflects how well true cases are found. Specificity shows how often false calls are avoided.
For parasellar meningiomas, the sensitivity was 98.41% and specificity at 95.53%.
The researchers then tested the model on 959 new images that it had not analysed. The accuracy held steady. That external test is essential before any tool can be trusted in clinics.
The team sees the tool supporting early triage, training young doctors to recognise patterns, and helping surgeons prepare for future challenges.
The technology could also shorten the time between a scan and a treatment plan. Patients may receive clearer explanations and fewer last-minute changes in strategy. Hospitals without specialist teams could benefit from expert-level support at the point of care.
Over time, the approach may lower costs by reducing extra tests and repeat operations. The model may also train future doctors to recognise difficult cases earlier.
As it expands to other conditions, AI could become a silent partner in daily diagnosis, experts say.
