Purpose: Pancreatic cancer remains one of the most deadly cancers and carries a dismal 5-year survival rate of <5%. Therefore, there is urgent need to develop a highly accurate and minimally invasive (e.g., without instrumentation of the pancreatic duct given high rate of complications) method of detection. Our group has developed a collection of novel light-scattering technologies that provide unprecedented quantitative assessment of the nanoscale architecture of the epithelium. We propose a novel approach to predict pancreatic cancer through the assessment of the adjacent periampullary duodenal mucosa without any interrogation of the pancreatic duct or imaging of the pancreas.
Experimental design: Endoscopically and histologically normal-appearing periampullary duodenal biopsies obtained from 19 pancreatic cancer patients were compared with those obtained at endoscopy from 32 controls. Biopsies were analyzed using our newly developed optical technologies, four-dimensional elastic light-scattering fingerprinting (4D-ELF) and low-coherence enhanced backscattering (LEBS) spectroscopy.
Results: 4D-ELF- and LEBS-derived optical markers from normal-appearing periampullary duodenal mucosa can discriminate between pancreatic cancer patients and normal controls with 95% sensitivity and 91% specificity. Moreover, the diagnostic performance of these optical markers was not compromised by confounding factors such as tumor location and stage.
Conclusions: Here, we showed, for the first time, that optical analysis of histologically normal duodenal mucosa can predict the presence of pancreatic cancer without direct visualization of the pancreas.