Build an Open Standard Comprehensive Cancer Cohort Research Database

This study presents a framework that leverages the healthcare interoperability standards, mainly regarding Fast Healthcare Interoperability Resources (FHIR) and Digital Imaging and Communications in Medicine (DICOM), to create a comprehensive cancer cohort research database. We developed a robust data collection pipeline to enable seamless data exchange across multiple Electronic Health Records (EHR) systems, encompassing patient demographic data, radiology and pathology reports, CT images, whole slide images in DICOM format, and genomic data. The framework supports the collection and analysis of data from 625 cases of non-small cell lung cancer and 373 cases of head and neck cancer. Collected data were converted into FHIR format by mapping vocabulary items, utilizing international terminologies such as ICD, LOINC, and SNOMED CT. For imaging reports, we adopted a FHIR imaging report implementation guide based on the Taiwan Core Implementation Guide to integrate clinical data with pathology images relevant to pathology reports. Genomic data were stored in cBioPortal, and a centralized portal was implemented to integrate these diverse data sources using open-source solutions.