Studies of 2026

Evaluation patient-derived serum samples for breast cancer biomarker detection

Nonappa Nonappa, Tampere University

The project aims to develop a rapid and reliable method for detecting biomarkers associated with the early-stage metastasis of primary breast cancer. By identifying and validating key molecular indicators of metastatic potential, we seek to enhance early detection strategies, which are critical for improving patient prognosis and guiding personalized treatment approaches. The overarching goal is to evaluate and validate the expression, relevance, and potential clinical utility of breast cancer biomarkers in patient-derived samples. The objectives are outlined based on our lab-based experimental validation using various breast cancer cell lines.

Prospective validation of machine learning model predicting clinical laboratory measurements (VALID)

Andrea Ganna ja Leena Viiri, Helsinki University

This medical research study aims to validate a machine learning model we have developed to predict the values of commonly used clinical laboratory measurements. In this study, we focus on clinical laboratory markers that can be easily measured from blood, have a well-established role in clinical diagnostics, and have a clearly characterized genetic background. In this recall-by-biobank study conducted through Finnish biobanks, approximately 2,000 individuals aged 30–70 years will be invited to provide blood samples. The clinical laboratory measurements obtained from these samples will be used to validate (i.e., test the performance of) the machine learning model. Participants will be invited across different predicted levels of the laboratory measurements (within and outside reference ranges) in order to evaluate the performance of the model. Including individuals with a wide range of predicted values allows us to assess how well the model performs in different situations, such as detecting abnormal values and confirming predictions for individuals whose values fall within reference ranges. This comprehensive approach will provide a more reliable assessment of the model’s accuracy and its potential usefulness in practical applications.