Our PhD student to speak at Data Science Summit 2026
Our PhD student Anna Mrukwa will speak at Data Science Summit 2026 with a presentation entitled “Don’t Make the Network Rediscover Anatomy: Computer Vision for Very Early-Stage Lung Cancer in LDCT.”
The work starts from a practical problem in lung cancer screening: AI models do not always transfer reliably between screening programmes. Differences between cohorts and imaging data can affect model behaviour in a new population, motivating the search for approaches that make greater use of stable anatomical information.
The presented method focuses on very early lung lesions and their relationship with the bronchovascular bundle. RONALD is used for patient-specific anatomy extraction. In the evaluated datasets, this step retained 1270 of 1271 (99.92%) annotated lesions from the Pomeranian screening cohort and 166 of 166 (100%) lesions from the DLCS cohort.
The talk will also place this lesion-level work in the broader context of transferring risk models, including SYBIL, to data from the Polish Lung Cancer Screening programme.
The presentation is based on joint work by Anna Mrukwa, MSc Eng.; Prof. Joanna Polańska; Marek Socha, PhD Eng.; Paweł Bożek, MD; Prof. Witold Rzyman, MD, PhD; Prof. Edyta Szurowska, MD, PhD; Katarzyna Dziadziuszko, MD, PhD; Prof. Rafał Dziadziuszko, MD, PhD; Małgorzata Jelitto, MD, PhD; Joanna Bidzińska, PhD Eng.; and Agata Durawa, MD.
Data Science Summit 2026 will take place online on 26 November and at PGE Narodowy in Warsaw on 27 November.
The work was supported by the Silesian University of Technology [02/070/BKM_26/0085] and the Pomeranian Interdisciplinary Centre of Digital Medicine [2023/ABM/02/00018].
