Related Experiment Video
Updated: May 9, 2025
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
Evaluating the Diagnostic Potential of Biomarker Panels in Breast Cancer and Prostate Adenocarcinoma
Kldiashvili Ekaterina1, Iordanishvili Saba1,2, Adamia Sophia2
1Petre Shotadze Tbilisi Medical Academy Tbilisi Georgia.
Background:
Noninvasive diagnostic methods are essential for early cancer detection and improved patient outcomes. Circulating biomarkers, measurable indicators of pathological processes, offer a promising avenue, yet optimal panels for reliable cancer diagnosis remain undefined. This study evaluates the diagnostic performance of selected plasma biomarkers in distinguishing breast cancer and prostate adenocarcinoma patients from healthy individuals, using statistical analysis and machine learning.
Materials And Methods:
We analyzed blood samples from 162 participants (73 cancer patients: 51 with breast cancer and 22 with prostate adenocarcinoma; 89 healthy controls). Levels of 12 cancer-associated biomarkers-including Ki67, DNMT1, BRCA1, and MPO-were quantified using enzyme-linked immunosorbent assays (ELISA). Statistical analyses, including the Mann-Whitney U test and machine learning models (random forest), were employed to assess the predictive accuracy of these biomarkers in distinguishing between cancerous and healthy states.
Results:
Biomarkers such as Ki67, DNMT1, and MPO were significantly elevated in cancer groups. Random forest models using selected combinations (e.g., BRCA1-CTA-TP53) achieved perfect classification accuracy (AUC = 1.00). However, high inter-marker correlations suggested potential redundancy, underscoring the need for biomarker panel optimization.
Conclusion:
Our findings support the potential of biomarker panels for accurate, noninvasive cancer diagnostics. Further validation in larger, more diverse cohorts is warranted to establish clinical utility and generalizability.
