Improving the Classification of Pancreatic Cysts
Pancreatic cysts, pockets of fluid found on or in the pancreas, pose a difficult challenge for clinicians. They are often found incidentally when patients are undergoing imaging for other ailments and may or may not become cancerous. In fact, one study found that as many as 20% of cysts were surgically removed unnecessarily, which can result in dangerous complications. A team of researchers, including Hopkins inHealth member Rachel Karchin, sought a new diagnostic approach to better distinguish which pancreatic cysts may become cancerous, and thus will benefit from surgical removal.
The team used cyst fluid and specialized gene sequencing technology to identify the molecular features typically associated with pancreatic cysts, including mutations of specific genes and abnormal chromosome numbers. A computer algorithm was then run to determine which of these features, in combination with traditional clinical markers (e.g., age, symptoms, and imaging findings), could be used to best determine the patients who would most benefit from surgery. The researchers were able to pinpoint a panel of molecular and clinical markers that could predict who should receive surgery with 92% sensitivity, substantially higher than the 75% sensitivity found when using molecular or clinical markers separately.
Though this research has yet to be independently validated, it is an important step in discovering an optimal panel of molecular and clinical markers to better classify pancreatic cysts. With refinement, it may ultimately lead to fewer unnecessary surgeries.
More about this research can be found here.