Artificial Intelligence to Guide Management of Pancreatic Cysts
In a proof-of-concept study, an international scientific team has shown that a laboratory test using artificial intelligence tools has the potential to more accurately sort out which people with pancreatic cysts will eventually develop pancreatic cancers. Their findings were published by Springer et al in Science Translational Medicine.
The test, dubbed CompCyst (for comprehensive cyst analysis), incorporates measures of molecular and clinical markers in cyst fluids, and appears to be on track to significantly improve on conventional clinical and imaging tests, the research team said.
Using information from more than 800 patients with pancreatic cysts who had cyst fluid analysis and cyst removal surgery at 16 medical centers around the world, investigators said CompCyst more often correctly identified which patients needed and likely had a chance to benefit from surgery, and which were unlikely to benefit from surgery or needed further monitoring only, than standard current methods. Specifically, they found that using the test would have recommended against surgery for more than half of patients who underwent cyst removal later deemed unnecessary because the cysts were unlikely to have caused cancer.
“Our study demonstrates the potential role of CompCyst as a complement to existing clinical and imaging criteria when evaluating pancreatic cysts,” said study author Anne Marie Lennon, MBBCh, PhD, of the Johns Hopkins Multidisciplinary Pancreatic Cyst Clinic. “It could provide a greater degree of confidence for physicians when they advise patients that they do not require follow-up and can be discharged from surveillance. Although we still need to prospectively validate this test, our results are exciting because they document a new and more objective way to manage the many patients with this disease.”
Pancreatic Cysts
Pancreatic cysts are common. They are found in 4% of people in their 60s and 8% of people over age 70, according to other published research. That means some 800,000 people with a pancreatic cyst are identified each year in the United States alone. By contrast, only a small fraction of cysts progress to cancer. “The dilemma facing patients and their physicians is the ability to distinguish precancerous cysts from cysts that will not progress to cancer,” said Dr. Lennon.
“Currently available clinical and imaging tests often fail to distinguish precancerous cysts from cysts that have little or no potential to turn cancerous, which makes it difficult to determine which patients will not require follow-up and which patients will need long-term follow-up or immediate surgical resection,” said study investigator Christopher Wolfgang, MD, PhD, MS, of Johns Hopkins Kimmel Cancer Center and the Johns Hopkins Precision Medicine Center of Excellence for Pancreatic Cancer. “As a result, essentially all people diagnosed with a cyst are followed long-term. Surgeons are faced with making recommendations to patients based on the risks and benefits of surgery with limited information. We seldom miss a cancer, but it is at the expense of performing an operation that in hindsight may not have been necessary. This study directly addresses these fundamental problems in management of pancreatic cysts.”
Methods
In the study, the precise nature of the cysts examined was confirmed through histopathologic analysis of resected surgical specimens. The cysts were then classified into three groups: those with no potential to turn cancerous, for which patients would not require periodic monitoring; mucin-producing cysts that have a small risk of progressing to cancer, for which patients can receive periodic monitoring for progression to possible cancer; and cysts for which surgery is recommended because there is a high likelihood of progression to cancer.
The CompCyst test was created with patient data, including clinical impressions and symptoms, images from computed tomography scans, and molecular features such as DNA alterations within cyst fluid.
In the study, the researchers evaluated the molecular profiles, including DNA mutations and chromosome changes, of 862 pancreatic cysts. They then fed the molecular information, along with clinical and radiologic data, into a computer-based program that used artificial intelligence to classify patients into the three groups noted previously.
Results
Based on histopathologic analysis of the surgically resected cysts, the researchers found that surgery was not needed for 45% of the patients that underwent surgery for their cysts. This unnecessary surgery was performed because the clinicians could not determine if the cysts were dangerous. In these patients, if CompCyst had been used, the researchers estimated that 60% to 74% of the patients (depending on the cyst type) could have avoided these unnecessary surgeries.
The study had several limitations, the researchers note, including that pancreatic cyst fluid was obtained at the time of surgery, and that the cysts evaluated are more atypical than those seen in routine clinical practice.
Disclosure: The work was supported by the Lustgarten Foundation—Pancreatic Cancer Research, the Virginia and D.K. Ludwig Fund for Cancer Research, the Sol Goldman Pancreatic Cancer Research Center, the Rolfe Pancreatic Cancer Foundation, the Benjamin Baker Scholarship, and the National Institutes of Health. For full disclosures of the study authors, visit stm.sciencemag.org.
The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.
