Pancreatic Cancer Early Detection: From High-Risk Surveillance to AI and Liquid Biopsy: RadioGraphics | October 2026

What Does the Review Highlight?

This comprehensive review examines current pancreatic cancer screening strategies and emerging technologies for detecting pancreatic ductal adenocarcinoma (PDAC) at a potentially curable stage.

• Early detection is critical: Five-year survival is approximately 44% for localized disease versus 3% for metastatic disease.
• Screening should target high-risk individuals, particularly those with familial pancreatic cancer, pathogenic germline mutations and hereditary syndromes—not the average-risk population.
• EUS and MRI/MRCP remain central to surveillance. EUS is particularly effective for detecting small solid lesions, while MRI/MRCP better characterizes pancreatic cysts.
• High-risk surveillance can improve early detection: CAPS studies reported five-year survival exceeding 70% among surveillance-detected pancreatic cancers, although observational biases remain possible.
• New-onset diabetes after age 50 is an important warning signal. Approximately 0.85% develop PDAC within three years; END-PAC scoring may help identify individuals needing further evaluation.
• Subtle imaging abnormalities may precede diagnosis by years, including focal pancreatic atrophy, abrupt duct cutoff, unexplained duct dilatation and altered parenchymal enhancement.
• Pancreatic cyst surveillance requires risk stratification. Enhancing mural nodules ≥5 mm, main pancreatic duct ≥10 mm and obstructive jaundice are important high-risk features of IPMN.
• AI-assisted CT and MRI analysis may identify visually occult cancer. Some retrospective radiomics models achieved AUC >0.95 for detecting cancer up to one year before clinical diagnosis.
• Liquid biopsy is an emerging complementary strategy. Circulating DNA/RNA, microRNA, circular RNA, proteomic markers and PAC-MANN assays show promise but lack sufficient validation for routine screening.
• Future screening will likely be multimodal, integrating clinical risk, electronic health records, AI-assisted imaging and molecular biomarkers.

Why Does It Matter?

Conventional pancreatic imaging often detects cancer only after a visible mass develops. Earlier recognition of subtle secondary signs and better identification of high-risk individuals could create opportunities for curative intervention.

The greatest potential lies in combining imaging, clinical risk prediction and molecular biomarkers, rather than relying on any single test.

GastroAGI Takeaway

The future of pancreatic cancer screening is not simply better imaging—it is identifying the right patient, at the right time, before a visible tumour develops.

Currently, risk-based EUS and MRI/MRCP surveillance in experienced centres remains the established approach. New-onset diabetes, suspicious pancreatic duct changes and high-risk cyst features deserve particular attention.