MOOSY-32 e-Nose Dataset for Prostate Cancer Assessment

This Zenodo release formally preserves and versions the dataset originally made publicly available through GitHub in association with the paper “Prostate cancer detection using e-nose and AI for high probability assessment”. The dataset contains processed measurements acquired with the MOOSY-32 electronic nose from urine samples obtained from patients diagnosed with prostate cancer (CaP) and benign prostatic hyperplasia (HBP/BPH). The released data consist of two patient-disjoint HDF5 tables: a training set and a held-out test set, each containing 12,800 rows and 34 columns. Each row represents one sensor response curve summarized through 32 input variables, together with the corresponding class labels. The study cohort included 40 patients, divided into independent training and test groups. The dataset does not contain direct patient identifiers and must be used exclusively for research and methodological evaluation. It is not a validated medical device or clinical decision-support system. The associated publication is: Talens, J. B., Pelegri-Sebastia, J., Sogorb, T., and Ruiz, J. L. (2023). Prostate cancer detection using e-nose and AI for high probability assessment. BMC Medical Informatics and Decision Making, 23, 205. https://doi.org/10.1186/s12911-023-02312-2