AI Tool Predicts Certain GI Cancers Years in Advance

TOPLINE:

The Kettles Esophageal and Cardia Adenocarcinoma predictioN (K-ECAN) tool predicts esophageal adenocarcinoma (EAC) and gastric cardia adenocarcinoma (GCA) using data from the electronic health record (EHR) and is more accurate than other tools, a large study suggests.

METHODOLOGY:

Researchers performed a case-control study using data from the Veterans Health Administration (VHA) Central Cancer Registry.
They identified 8430 patients with EAC and 2965 patients GCA; these patients were compared with more than 10 million control patients.
K-ECAN uses basic information in the EHR to determine an individual's future risk of developing EAC or GCA.

TAKEAWAY:

With an area under the receiver operating characteristic (AUROC) of 0.77, K-ECAN demonstrated better discrimination than previously validated models and published guidelines.
Using only data from 3 to 5 years prior to diagnosis only slightly diminished its accuracy (AUROC, 0.75).
K-ECAN remained the most accurate tool when undersampling men to simulate a non-VHA population (AUROC, 0.85).
Although gastroesophageal reflux disease (GERD) was strongly associated with EAC, it only contributed a small proportion of gain in information for prediction.

IN PRACTICE:

Because K-ECAN does not rely heavily on GERD symptoms to assess risk, it has the “potential to guide providers to increase appropriate uptake of screening. De-emphasizing GERD in decisions to offer screening could paradoxically increase appropriate uptake of screening for EAC and GCA," the authors write.

SOURCE:

The study, with first author Joel H. Rubenstein, MD, with the LTC Charles S. Kettles VA Medical Center, Ann Arbor, Michigan, was published online August 17 in Gastroenterology.

LIMITATIONS:

K-ECAN was developed and validated among US veterans and needs to be validated in other populations.

DISCLOSURES:

Funding for the study was provided by the US Department of Defense. Rubenstein has received research support from Lucid Diagnostics.

For more news, follow Medscape on Facebook, X, Instagram, and YouTube.

Credits:

Lead image: Magicmine/Dreamstime

Send comments and news tips to news@medscape.net.

Cite this: AI Tool Predicts Certain GI Cancers Years in Advance - Medscape - Aug 29, 2023.

Commenting is limited to medical professionals. To comment please Log-in.

Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.

My Alerts