Influence of artificial intelligence on the performance of endosonographers in detection of solid pancreatic tumors: a tandem randomized video study.

The CAPS Consortium recommends pancreatic surveillance for selected high-risk individuals to detect stage I (T1N0M0, margin-negative resectable) pancreatic cancer and high-grade precursors (PanIN-3, IPMN with high-grade dysplasia); it should ideally be performed in a research setting by a multidisciplinary team at a center with appropriate expertise. Candidate groups and their affected-relative requirements: Peutz-Jeghers/STK11 carriers regardless of family history; CDKN2A carriers regardless of family history (surveillance regardless of family history reached 77% agreement, still meeting the guideline's ≥75% consensus threshold as a Grade 1 'definitely do it' recommendation, while surveillance in those with at least one affected first-degree relative reached stronger 99% agreement); BRCA2 carriers with at least one affected FDR or at least two affected relatives of any degree; ATM, PALB2, and Lynch (MLH1/MSH2/MSH6) carriers only if at least one affected FDR; BRCA1 with an affected FDR did NOT reach consensus; hereditary pancreatitis was an area of disagreement (no consensus on whether or how to surveil); and familial pancreatic cancer kindreds, an individual with at least one affected FDR who in turn has an affected FDR, qualify on pedigree alone without an identified germline variant. Germline testing and genetic counseling should be considered for those eligible for surveillance but are not an absolute prerequisite. Start ages (exact age often did not reach consensus): CDKN2A at age 40; Peutz-Jeghers at least by age 40, or 10 years younger than the youngest affected relative; hereditary pancreatitis/PRSS1 at age 40 or 20 years after the first attack; BRCA2/ATM/PALB2/BRCA1/Lynch carriers at age 45 or 50, or 10 years younger than the youngest affected relative; familial pancreatic cancer kindreds at age 50 or 55, or 10 years younger than the youngest affected relative (experts split between 50 and 55). New-onset diabetes in a high-risk individual should trigger screening regardless of age. Modalities: both baseline and follow-up surveillance use EUS and MRI/MRCP, alternated (consensus that both are used, but NO consensus on if or how to alternate them); CT is reserved for individuals who cannot undergo MRI/EUS or when a solid lesion is detected; CA19-9 is an additional test only when imaging shows worrisome features; routine fasting glucose and/or HbA1c testing is advised. Interval: in the absence of pancreatic abnormalities, or with only low-risk findings such as pancreatic lobulation, cysts without worrisome features, or a <1 cm non-functioning neuroendocrine tumor, continue 12-monthly imaging. Concerning findings: a newly detected solid lesion of uncertain significance should be re-imaged at 3 months (3-6 months for CDKN2A carriers); EUS-FNA is indicated for solid lesions ≥5 mm, cystic lesions with worrisome features (mural nodule, solid component, duct dilation), or a main-pancreatic-duct stricture without a mass. Surgical resection, performed at a specialty center as an oncologic/radical resection, is indicated for a solid lesion (≥5 mm; resection regardless of size did not reach consensus), or an IPMN with a mural nodule, an enhancing solid component, symptoms (pancreatitis/jaundice/pain), or a main pancreatic duct ≥10 mm. Surveillance of average-risk individuals is not recommended.