Scholars Central

Research articlePublished 24 Feb 2025

Combining Urinary Biomarker Panels and Machine Learning For Earlier Detection of Pancreatic Cancer

Authors: Ruben Berreby, Ecole Jeannine Manuel

Department of Oncology, University of Versailles, Paris, France

Journal of Clinical & Experimental Pathology

Volume 15 · Issue 1 · Pages 1–5

Ask about this article

Abstract

Select any words in the title or abstract, or press this and choose a sentence.

Close to 500,000 people were diagnosed with pancreatic cancer worldwide in 2020, and close to 450,000 people worldwide died from it that same year. Pancreatic cancer is the 12th most prevalent cancer in the world and one of the most deadly due to its aggressive nature as well as its lack of clearly identified symptoms, resulting in late diagnosis. It has the lowest 5-year overall survival rate of any malignancy due to late diagnosis (7.2%), and a survival rate, when detected in the late stages with distant metastatic spreading, of merely 2.4%. The scope of this investigation focuses on the detection of pancreatic cancer with a machine model trained on urinary biomarker data (urine samples). By using various techniques to improve the results of our models, very successful results were obtained. Despite the unequal distribution of data for each stage of cancer of the patients studied, the logistic regression model had a very successful accuracy score, significantly higher than the accuracy scores of previous research done without the use of machine learning. Their results obtained an AUC of 0.85, while AUC using the logistic regression model in this investigation reached 0.98. Results for the random forest classifier were very successful as well, as it performed better than the logistic regression model for predicting the stages in minority, rather than simply the major stages, and its accuracy score reached 89%.

Ask about this article

Answers come only from this article's title, abstract and keywords, each sentence pointing to where it comes from. Your question is sent to OpenAI to write the answer. Scholars Central does not store it.

Try:

Follow

New articles from what you follow go into your email digest, once you agree to alert emails.

  • JournalJournal of Clinical & Experimental Pathology
  • SubjectPathology & Lab Medicine
  • BylineArticles whose byline prints “Ruben Berreby”
  • BylineArticles whose byline prints “Ecole Jeannine Manuel”