Integrating Radiomics and Genomics to Improve the Clinical Assessment of Pancreatic Cysts and Early Detection of Pancreatic Cancer
Abstract
With the growing use of sophisticated radiographic imaging, incidental pancreatic cysts are becoming increasingly encountered in clinical practice. In fact, it is estimated that over 6 million Americans have a pancreatic cyst. Considering a subset of pancreatic cysts can harbor or transform into pancreatic cancer, the detection of a pancreatic cyst can lead to significant anxiety for both the patient and the treating physician. This in turn can lead to potentially invasive, expensive, and even harmful procedures and surgical management options for patients. However, only a subset of pancreatic cysts are precursors to pancreatic cancer and, among these precursors, a minority will progress into pancreatic cancer. Thus, upon identification of a pancreatic cyst, there is a dire need for diagnostic tests that can identify patients that harbor cystic precursors to pancreatic cancer, detect early pancreatic cancers arising from a pancreatic cyst, and/or determine the likelihood of developing pancreatic cancer within the detected pancreatic cyst. Recently, we have reported that radiomic biomarkers from standard abdominal imaging and genomic biomarkers from aspirated pancreatic cyst contents can be a useful tool in the evaluation of pancreatic cysts. For instance, through quantitative imaging of abdominal computed tomography (CT) scans, we identified a set of highly sensitive and highly specific radiomic biomarkers for the detection of early pancreatic cancers. Similarly, we published one of the largest prospective clinical studies to evaluate genomic biomarkers within aspirated pancreatic cyst contents. This study was performed within a U.S. government-accredited clinical laboratory and evaluated 595 patients. Through genomic testing of pancreatic cyst contents, we found specific genomic alterations that were highly accurate in identifying cystic precursors to pancreatic cancer and detecting early pancreatic cancers within a pancreatic cyst. Both sets of genomic biomarkers have been integrated into the clinical evaluation of pancreatic cyst patients at not only our institution, but at medical institutions across the United States. However, while the individual performance of radiomic and genomic biomarkers is encouraging, their combination could be used to further improve their clinical utility. Moreover, it is also important to note that radiomic and genomic biomarkers have demonstrated significant promise in being able to predict whether a pancreatic cyst patient will develop early pancreatic cancer in the future. Hence, we hypothesize that a comprehensive set of radiomic and genomic biomarkers will improve the clinical evaluation and predictive assessment of patients with pancreatic cysts. This proposal represents the formation of a diverse team of investigators with a strong background in early detection of pancreatic cancer and will address the Pancreatic Cancer Research Program (PCARP) focus area of “integration of biologic and imaging biomarkers to drive more precise and earlier detection and prognosis.” The Specific Aims are as follows. Specific Aims: Aim 1. To assess whether the combination of radiomic and genomic biomarkers are superior in performance compared to each assay alone in distinguishing mucinous cysts from non-mucinous cysts and detecting advanced neoplasia within a mucinous cyst. For Aim 1, we will evaluate individual and combined performance of our radiomic and genomic biomarkers within two comprehensive studies to (1) determine the accuracy of detecting those cysts with the potential for developing into pancreatic cancer (Aim 1A) and (2) identify the presence of early pancreatic cancer within a pancreatic cyst (Aim 1B). Paired CT scans and pancreatic cyst contents for Aim 1A will include 230 patients that have been well characterized. Similarly, for Aim 1B, a separate group of 265 patients will be analyzed. For both studies, investigators performing individual biomarker testing will
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 05, 2021
- Source ID
- W81XWH2110710
Entities
People
- Eugene J Koay
Organizations
- The University of Texas MD Anderson Cancer Center
- United States Army