Search Results

You are looking at 1 - 1 of 1 items for

  • Author or Editor: Ana G. Clavere-Graciette x
  • Refine by Access: All Content x
Clear All Modify Search

Abstract

OBJECTIVE

To validate the performance of a novel, integrated test for canine cancer screening that combines cell-free DNA quantification with next-generation sequencing (NGS) analysis.

SAMPLE

Retrospective data from a total of 1,947 cancer-diagnosed and presumably cancer-free dogs were used to validate test performance for the detection of 7 predefined cancer types (lymphoma, hemangiosarcoma, osteosarcoma, leukemia, histiocytic sarcoma, primary lung tumors, and urothelial carcinoma), using independent training and testing sets.

METHODS

Cell-free DNA quantification data from all samples were analyzed using a proprietary machine learning algorithm to determine a Cancer Probability Index (High, Moderate, or Low). High and Low Probability of Cancer were final result classifications. Moderate cases were additionally analyzed by NGS to arrive at a final classification of High Probability of Cancer (Cancer Signal Detected) or Low Probability of Cancer (Cancer Signal Not Detected).

RESULTS

Of the 595 dogs in the testing set, 89% (n = 530) received a High or Low Probability result based on the machine learning algorithm; 11% (65) were Moderate Probability, and NGS results were used to assign a final classification. Overall, 87 of 122 dogs with the 7 predefined cancer types were classified as High Probability and 467 of 473 presumably cancer-free dogs were classified as Low Probability, corresponding to a sensitivity of 71.3% for the predefined cancer types at a specificity of 98.7%.

CLINICAL RELEVANCE

This integrated test offers a novel option to screen for cancer types that may be difficult to detect by physical examination at a dog’s wellness visit.

Open access
in Journal of the American Veterinary Medical Association