Digital Architecture of the Croatian National Lung Cancer Screening Program: Observational Implementation Study (Preprint)

Read full text
Authors: Antun Aboud, Kristina Krpina, Stjepan Orešković, Miroslav Samaržija
Cited by: 0
References: 7

Abstract

BACKGROUND Scaling lung cancer screening from controlled trials to nationwide implementation requires interoperable digital infrastructure capable of coordinating primary care, radiology, pulmonology, and centralized governance. Although low-dose computed tomography (LDCT) reduces lung cancer mortality in high-risk populations, few countries have embedded screening programs directly within national health information systems to enable standardized workflows, real-time monitoring, and data-driven quality control. OBJECTIVE To describe the digital architecture, interoperability framework, and real-world performance of the Croatian National Lung Cancer Screening Program (CNLCSP), implemented as a native extension of the Central Health Information System of the Republic of Croatia (CEZIH). METHODS This retrospective observational implementation study analyzed structured program data collected between October 2020 and December 2025. The CNLCSP targets individuals aged 50–75 years with ≥30 pack-years of smoking history who are current smokers or former smokers who quit within 15 years. The program operates entirely within CEZIH through role-specific modules for general practitioners (GPs), radiologists, pulmonologists, and national coordinators. Core digital functionalities include electronic eligibility verification, paperless referral and scheduling, structured radiology and pulmonology reporting based on modified I-ELCAP guidelines, AI-assisted volumetric nodule analysis integrated into the reporting workflow with mandatory radiologist second reading, secure DICOM-based telemedicine image transfer, and a centralized analytics module providing real-time dashboards of predefined quality indicators, including radiation dose metrics. RESULTS From October 2020 to December 2025, over 54,000 individuals were screened, generating more than 80,000 LDCT examinations across 27 radiology centers and 6 pulmonology centers, involving more than 2,000 GPs. Positive radiological findings were reported in 4.45% of examinations. Continuous digital monitoring supported a mean effective radiation dose of 0.85 mSv, below the program limit of 1.5 mSv. The interoperable CEZIH-based infrastructure enabled expansion from 16 to 27 radiology centers while maintaining standardized reporting and centralized oversight. CONCLUSIONS Embedding lung cancer screening as a native component of a national health information system enables scalable implementation, structured data capture, AI-supported clinical workflows with human oversight, and real-time governance. The Croatian model illustrates how digital integration within existing health infrastructure can support population-level screening and may serve as a transferable informatics framework for other health systems.

Fields of Study & Subjects

Lung Cancer Diagnosis and Treatment Electronic Health Records Systems AI in cancer detection

Publication details

Publication date
26 February 2026
Language
EN
Document type
Preprint