An estimated 64,000 people in the U.S. will be diagnosed with pancreatic cancer in 2023, according to the American Cancer Society, and about 90 percent of these cases will be pancreatic ductal adenocarcinoma (PDAC). With only about 10 percent of patients surviving past 5 years after diagnosis, PDAC has the lowest 5-year survival rate of any major cancer and is difficult to detect in its early stages, leading to a high mortality rate.
"Even when pancreas cancer is detected at its earliest stage, it almost always has shed cells throughout the body and the cancer returns," noted Peter Allen, MD, Chief of the Division of Surgical Oncology in the Department of Surgery at Duke University School of Medicine. "That's why we shifted our focus to these precancerous cysts, known as intraductal papillary mucinous neoplasms (IPMNs). Most IPMNs will never progress to pancreas cancer, but by distinguishing which ones will progress, we are creating an opportunity to prevent an incurable disease from developing,"
IPMNs represent a detectable precursor to PDAC. However, identifying IPMNs that are at high risk of progressing to PDAC is challenging. Current management strategies for IPMNs are limited due to the difficulty of accurately predicting the timing of progression. Consensus guidelines have been developed to predict the risk of high-grade dysplasia (HGD) and early cancer through various parameters, such as clinical, radiographic, laboratory, endoscopic, and cytologic factors. However, the overall accuracy of these guidelines is around 60 percent (Ann Surg Oncol 2019; https://doi.org/10.1245/s10434-019-07921-8). Multiple studies have proposed biomarkers for high-risk IPMNs, but their accuracy is limited due to intralesional heterogeneity. Consequently, there is an urgent need for more accurate biomarkers for high-risk IPMNs (Biochim Biophys Acta Rev Cancer 2019; doi: 10.1016/j.bbcan.2019.188318).
Multiplex digital spatial profiling is a new technology that can help distinguish disease phenotypes in heterogeneous tissues. In a recent study, researchers at Duke Health utilized digital spatial RNA profiling to identify markers of dysplasia and biological processes associated with malignant progression in IPMNs. The study was published in the journal Science Advances (2023; doi: 10.1126/sciadv.ade4582).
Previous studies have shown that IPMN can have multiple dysplastic epithelial subtypes, including pancreaticobiliary (PB), intestinal (INT), and gastric foveolar (GF), among others. The PB and INT subtypes are associated with a higher risk of invasive cancer, while GF is considered a less aggressive subtype. Patients with the PB subtype have a poor prognosis, and studies suggest that patients with invasive lesions from the PB subtype have similar outcomes as patients with conventional PDAC.
The current investigation led by Allen involved profiling dysplastic epithelium (83 regions) in surgically resected IPMN tissues (12 patients) to differentiate subtypes and predict genes associated with malignancy. The technology allowed the researchers to precisely characterize regions of IPMN histopathology and produce robust gene expression patterns.
The findings revealed that PB and GF lesions may share a common neoplastic cell lineage, while the INT subtype appears to be distinct. The study used unsupervised dimensionality reduction analysis to differentiate subtypes of IPMN and found that transcriptional dysregulation within PB lesions was similar to pancreatic cancer, while INT and GF foci did not show the same dysregulation. This suggests that GF lesions could be considered a precursor to PB epithelium.
The team also explored the underlying biology of IPMN by combining supervised and unsupervised analyses. They found that neoplastic progression in IPMN is driven by two transcriptional programs: inflammatory signaling (TNF-NF[kappa]B) and cell proliferation (S and G2-M phases). Activation of cell proliferation in INT subtype than PB lesions was greater, whereas inflammatory signaling was more prominent in PB than INT. Moreover, the team identified a 46-gene cluster associated with PB epithelium, inflammatory signaling, and PDAC genes, which could potentially be used as a measure of malignancy risk in IPMN.
This study's findings may be limited because the probe panel used only measured a small portion of human protein-coding genes. The small sample size of 12 specimens may also not reflect the full range of disease biology across common epithelial subtypes and may omit rare entities such as oncocytic IPMN. In addition, the requirement for every specimen to have regions of HGD may also have limited the study's design.
In summary, the study identified subtype-specific and high-risk marker genes that could help in the development of a more accurate risk assessment tool for IPMN.
"We found very distinct markers for high-grade cell abnormalities, as well as for slow-growing subtypes," Allen concluded. "Our work now is focusing on finding it in the cyst fluid. If we can identify these unique markers in cyst fluid, it could provide the basis for a protein biopsy that would guide whether we should remove the cyst before cancer develops and spreads."
Dibash Kumar Das is a contributing writer.