In the Journals

‘Major leap’: Test reduces unnecessary surgery for pancreatic cysts

A laboratory test that uses molecular and clinical markers, as well as machine learning, may help improve the management of pancreatic cysts and reduce unnecessary surgeries, according to data published in Science Translational Medicine.

One of the study authors, Christopher L. Wolfgang, MD, PhD, of the Sol Goldman Pancreatic Cancer Research Center at Johns Hopkins University, said that pancreatic cysts present a fundamental problem for treatment because most of them will never actually progress to cancer, but it is difficult to discern between low- and high-risk.

“We currently do not have a test to differentiate the high-risk lesions from the low-risk lesions,” Wolfgang said during a conference call. “In other words, we don’t have an accurate method that reliably tells us whether to immediately remove the cysts with surgery vs. closely observe vs. discharge from further care.”

This exposes more patients to the risks of treatment, including radiation and procedure complications. However, Wolfgang said a new test that combines molecular analysis with conventional clinical and imaging feature — known as CompCyst — could be a “major leap forward” in the management of pancreatic cysts.

Researchers used CompCyst to categorize 862 pancreatic cysts into three groups; cysts with no potential to turn cancerous, mucin-producing cysts with a small risk for cancer progression, and cysts with a high likelihood for progression to cancer. Depending on which category each cyst fell into, researchers could determine if the patients needed no monitoring, periodic monitoring or immediate surgery.

Investigators found that current clinical practice only correctly identified 19% of patients with benign cysts while CompCyst identified 60%. Among patients who should be monitored, CompCyst identified 49% compared with 34% for current clinical practice.

Researchers believe that CompCyst could have helped prevent unnecessary surgery in at least 60% of the patients included in the study.

Bert Vogelstein, MD, Clayton Professor of Oncology and co-director of the Ludwig Center at the Johns Hopkins Kimmel Cancer Center, said that their study shows that artificial intelligence can help bridge the gap between lab results and clinical judgement in a way that benefits patient care.

“There’s a feeling among many that genetic tests, like the one described in this paper, will eventually replace clinical judgement,” he said on the conference call. “We don’t believe that this is true. We believe that tests like this will inform rather than replace clinical judgement.” by Alex Young

Disclosures: Vogelstein reports being a member of the scientific advisory board for CAGE, Eisai-Morphotek, NeoPhore, Nexus (Camden Partners) and Sysmex-Inostics. He is also a founder of Personal Genome Diagnostics and PapGene. Wolfgang reports no relevant financial disclosures. Please see the full study for all other authors’ relevant financial disclosures.

A laboratory test that uses molecular and clinical markers, as well as machine learning, may help improve the management of pancreatic cysts and reduce unnecessary surgeries, according to data published in Science Translational Medicine.

One of the study authors, Christopher L. Wolfgang, MD, PhD, of the Sol Goldman Pancreatic Cancer Research Center at Johns Hopkins University, said that pancreatic cysts present a fundamental problem for treatment because most of them will never actually progress to cancer, but it is difficult to discern between low- and high-risk.

“We currently do not have a test to differentiate the high-risk lesions from the low-risk lesions,” Wolfgang said during a conference call. “In other words, we don’t have an accurate method that reliably tells us whether to immediately remove the cysts with surgery vs. closely observe vs. discharge from further care.”

This exposes more patients to the risks of treatment, including radiation and procedure complications. However, Wolfgang said a new test that combines molecular analysis with conventional clinical and imaging feature — known as CompCyst — could be a “major leap forward” in the management of pancreatic cysts.

Researchers used CompCyst to categorize 862 pancreatic cysts into three groups; cysts with no potential to turn cancerous, mucin-producing cysts with a small risk for cancer progression, and cysts with a high likelihood for progression to cancer. Depending on which category each cyst fell into, researchers could determine if the patients needed no monitoring, periodic monitoring or immediate surgery.

Investigators found that current clinical practice only correctly identified 19% of patients with benign cysts while CompCyst identified 60%. Among patients who should be monitored, CompCyst identified 49% compared with 34% for current clinical practice.

Researchers believe that CompCyst could have helped prevent unnecessary surgery in at least 60% of the patients included in the study.

Bert Vogelstein, MD, Clayton Professor of Oncology and co-director of the Ludwig Center at the Johns Hopkins Kimmel Cancer Center, said that their study shows that artificial intelligence can help bridge the gap between lab results and clinical judgement in a way that benefits patient care.

“There’s a feeling among many that genetic tests, like the one described in this paper, will eventually replace clinical judgement,” he said on the conference call. “We don’t believe that this is true. We believe that tests like this will inform rather than replace clinical judgement.” by Alex Young

Disclosures: Vogelstein reports being a member of the scientific advisory board for CAGE, Eisai-Morphotek, NeoPhore, Nexus (Camden Partners) and Sysmex-Inostics. He is also a founder of Personal Genome Diagnostics and PapGene. Wolfgang reports no relevant financial disclosures. Please see the full study for all other authors’ relevant financial disclosures.