Qadir's project, "Radiologist-AI Collaboration for Early Detection of Pancreatic Cancer: Optimizing Interaction in IPMN Management," builds on his ongoing research using artificial intelligence to identify patients at higher risk of developing pancreatic cancer.

Pancreatic cancer is one of the deadliest forms of cancer, with a five-year survival rate of less than 15 percent. Because the disease often causes few symptoms in its early stages, it is frequently diagnosed after it has already advanced.

"That's what makes precursor lesions like IPMNs so valuable," Qadir said. "They give us a chance to intervene before cancer develops, or while it's still potentially curable. If we can tell which patients are truly at high risk, we can catch the disease in the narrow window where treatment can change the outcome."

An intraductal papillary mucinous neoplasm (IPMN) is a type of pancreatic cyst that can develop into pancreatic cancer over time. While most IPMNs are harmless, identifying the patients whose cysts are likely to become cancerous is essential for determining who would benefit from surgery while avoiding unnecessary procedures for others.

Unlike many AI studies that focus solely on prediction accuracy, Qadir's research examines how AI can best support radiologists during clinical decision-making.

"Most AI research in medicine stops at one question: can the model make an accurate prediction?" he said. "A model that performs well in isolation can still be harmful if it's introduced into the workflow in the wrong way. This work focuses on how radiologists and AI should work together."

The project will evaluate different approaches to integrating AI-based decision support into radiologists' workflows to determine which methods provide the best balance between detecting high-risk lesions and minimizing unnecessary interventions.

"Right now, the guidelines clinicians use are good at catching high-risk cysts, but they're not specific," Qadir explained. "That can lead to patients undergoing major pancreatic surgery for cysts that potentially were never going to become cancer. The goal is fewer unnecessary operations without missing the high-risk lesions."

The competitive IUSCCC Trainee Pilot Award provides up to $10,000 to graduate students and postdoctoral researchers conducting innovative cancer research. The program supports projects spanning cancer biology, prevention, early drug development, and clinical research.

For Qadir, the award represents more than financial support.

"Beyond the funding, this recognition is a signal that the field sees clinical-AI collaboration, not just algorithm accuracy, as the direction that matters," he said. "Most of all, patients with pancreatic cancer or at risk of developing pancreatic cancer need better tools, and this award lets me contribute to that."

Looking ahead, Qadir hopes the project will help establish best practices for integrating AI into clinical workflows and lay the foundation for bringing these tools into routine patient care.

"I hope that this grant lets us pinpoint the best way to deploy the developed tool and lays the groundwork for eventually bringing it into real clinical practice," he said.

Qadir’s principal investigator is Fiona Kolbinger, Research Assistant Professor of Biomedical Engineering.