JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

A novel cross-validated machine learning based Alertix-Cancer Risk Index for early detection of canine malignancies

  • 0Alertix Veterinary Diagnostics, Stockholm, Sweden.
Frontiers in Veterinary Science +

|

May 12, 2025

|

May 12, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

A new machine learning model, Alertix-Cancer Risk Index (Alertix-CRI), combines canine Thymidine kinase 1 (TK1) and C-reactive protein (cCRP) levels for early tumor detection. This non-invasive biomarker approach significantly improves diagnostic accuracy in dogs.

Area Of Science

  • Veterinary Medicine
  • Biomarker Discovery
  • Machine Learning in Oncology

Background

  • Growing demand for non-invasive tumor biomarkers in veterinary medicine.
  • Thymidine kinase 1 (TK1) is a known proliferation biomarker for canine malignancies.
  • Combining TK1 with inflammatory biomarkers like canine C-reactive protein (cCRP) can enhance early tumor detection sensitivity.

Purpose Of The Study

  • To develop and validate a machine learning model, Alertix-Cancer Risk Index (Alertix-CRI), for early canine tumor detection.
  • To integrate canine TK1 protein, cCRP levels, and age into a predictive model.
  • To assess the diagnostic performance of Alertix-CRI compared to individual biomarkers.

Main Methods

  • Utilized 287 serum samples from healthy dogs and dogs with various tumors.
  • Measured serum TK1 and cCRP levels using ELISA techniques.
  • Developed Alertix-CRI using a generalized boosted regression model (GBM) with 70% training and 30% validation data.

Main Results

  • Both TK1 and cCRP levels were significantly higher in tumor-bearing dogs (p < 0.0001).
  • TK1 and cCRP showed similar sensitivity (54% vs. 51%) at 95% specificity for overall tumors.
  • Alertix-CRI demonstrated high discriminatory capacity with an AUC of 0.98, achieving 90% sensitivity and 97% specificity.

Conclusions

  • Alertix-CRI serves as a valuable decision-support tool for clinicians to differentiate malignant diseases in dogs.
  • The model facilitates advancements in precise and dependable diagnostic tools for early cancer detection and therapy monitoring in veterinary medicine.

Keywords: