How does AI help with pancreatic cancer?
AI is being studied at almost every step of pancreatic cancer care. For detection, the PANDA model found pancreatic lesions on routine CT scans in a large Chinese study, with 92.9% sensitivity. For treatment, researchers are testing AI that outlines tumours on scans, judges whether surgery is possible, and predicts how a tumour may respond to chemotherapy. Most treatment studies are small and early, so these tools support specialists rather than make decisions for patients. Knowing what each tool does helps you ask better questions.
What the treatment research shows
- Outlining tumours: a deep-learning model measured tumour size on CT almost as experts did in 230 test patients.
- Surgery planning: an early study used AI to check whether a tumour touches nearby blood vessels, which affects whether it can be removed.
- Chemotherapy: small studies predict response to chemotherapy before surgery from CT scans and the CA19-9 blood marker.
- Radiotherapy: the SMART trial re-planned radiotherapy daily using MRI in 136 patients; this is adaptive image guidance rather than AI itself.
The limits
Nearly all treatment studies are small, look back at past patients and come from single centres. None has yet shown that AI improves survival. Many tools are tested only in large research centres, so availability in India is limited. Until larger trials report, the best place for these tools is inside a specialist pancreatic cancer team, where surgeons, oncologists and radiologists review each scan together.
Questions to ask your doctor
Take these to your next appointment:
- Can my scans be reviewed by a pancreatic cancer team to see if surgery is possible?
- Would chemotherapy before surgery help in my case?
- Is adaptive or image-guided radiotherapy available, and would it suit me?
- Is there a clinical trial I could join?
Where OncoVOIS fits
OncoVOIS, a CancerFax cancer-navigation app, explains CT, pathology and blood test reports in plain words and helps you prepare questions about surgery, chemotherapy and trials for your oncologist. It is available on Google Play for Android phones, and the CancerFax team can help with expert opinions and trials.
Sources
1. Cao K, et al. Large-scale pancreatic cancer detection via non-contrast CT and deep learning. Nature Medicine, 2023. https://www.nature.com/articles/s41591-023-02640-w 2. Deep-learning segmentation of pancreatic tumours on CT. PubMed, 2023. https://pubmed.ncbi.nlm.nih.gov/37296006/ 3. Viviers CGA, et al. Deep learning for vascular involvement in pancreatic cancer. arXiv (ICCV workshop), 2023. https://arxiv.org/abs/2310.00639 4. Prediction of neoadjuvant chemotherapy response in pancreatic cancer using CT and neural networks. The American Surgeon, 2021. https://pubmed.ncbi.nlm.nih.gov/33381979/ 5. Parikh PJ, Chuong MD, et al. Multi-institutional phase 2 trial of ablative 5-fraction stereotactic MR-guided on-table adaptive radiation therapy for pancreatic cancer (SMART). https://baptisthealth.elsevierpure.com/en/publications/a-multi-institutional-phase-2-trial-of-ablative-5-fraction-stereo/