Portuguese National Cancer Registry: from local registries to high-quality nationwide real-world data Authors: TM Garcia, A Sousa, P Leite-Silva, L Lopes-Conceição, O Godinho, MJ Bento Published: 2026 ABSTRACT AbstractBackgroundPortugal’s population-based cancer registries have evolved from fragmented municipal initiatives, starting in 1976, to full national coverage by 2007, culminating in the legislative creation of the National Cancer Registry [Registo Oncológico Nacional (RON)] in 2017.Patients and methodsThis short communication describes RON’s structural evolution, data quality, research utility, and readiness for international reporting.ResultsAlthough regional registries captured 20 core epidemiological variables, RON introduced >30 new clinical variables aligned with European Network of Cancer Registries recommendations, refining tumor characterization and treatment details. Analysis of 2022 data reveals that core incidence variables are 100% complete, with microscopic verification (97.8%), death certificate only (<0.1%), and unknown basis of diagnosis (<0.1%) meeting international benchmarks. However, challenges persist: stage is missing in 40.4% of cases, ill-defined morphology or topography affects 3.5%, unrelated surgical codes affect 14.9%, and poor stage standardization is evidenced by 779 recorded options. Despite partial automations, manual data entry remains substantial, contributing to a 3-year publication lag. Nevertheless, RON has supported 56 national and international studies since 2018, with data now sufficient for consideration in international reports.ConclusionsPortugal has achieved remarkable progress in establishing a comprehensive national cancer registry with robust core data quality and growing international recognition. Strategic investments in automated data integration, real-time validation, targeted training, and phased data release, building on existing European collaborations, will further strengthen RON as a timely, high-impact resource for real-world oncology research and policy.Key words: oncology, population-based cancer registries Introduction Population-based cancer registries (PBCRs) are a cornerstone of oncologic epidemiology and public health surveillance, facilitating systematic data collection on cancer incidence, mortality, and survival within their respective geographic populations. Beyond monitoring disease burden, PBCRs also serve to evaluate cancer control programs, assess treatment patterns and outcomes, and facilitate real-world research that informs clinical practice and health policies. Portugal’s journey toward a unified national PBCR reflects the challenges and achievements of developing robust population-level surveillance. The earliest PBCR in Portugal dates back to 1976 with the registry at the hospital of Viana do Castelo, an initiative that was rapidly followed across the country.1,2 In 1988, a governmental decree established regional registries [Registos Oncológicos Regionais (RORs)] across the northern, central, and southern health administrative regions, wherein each hospital’s local registry submitted cases to its corresponding regional database.3 By 2007, Portugal achieved full population–level coverage of the entire continental territory and the autonomous regions.1 This regional infrastructure provided the essential groundwork upon which a national system would be constructed. The legislative framework for the National Cancer Registry [Registo Oncológico Nacional (RON)] was established in 2018, mandating a registry centralizing the collection and analysis of data from all oncologic patients diagnosed and/or treated in Portugal.4 The objectives include monitoring institutional activity, evaluating organized screening programs, assessing therapeutic effectiveness, supporting epidemiological surveillance, and enabling research, including, in articulation with the Portuguese Medicines Agency [Autoridade Nacional do Medicamento e Produtos de Saúde, I.P. (INFARMED)], monitoring the effectiveness of medications and medical devices.4 The quality of cancer registries has been a topic of ongoing worldwide discussion, particularly as registry data are increasingly used to shape oncology policies.5,6 This concern is especially relevant for the Portuguese reality because, despite its evolution, RON still mainly relies on manual data entry and curation. Eight years after its establishment, this short communication aims to describe the evolution of the structure and quality of the Portuguese cancer registry and assess its readiness to report internationally and support real-world oncology research. Methods We conducted a descriptive analysis of RON’s data framework, comparing it with ROR’s, namely the number and type of variables included in both. Data quality was assessed using cases diagnosed between 2018 and 2022 (N = 330 706) for completeness, validity, comparability, and timeliness assessment. All data were extracted from the RON centralized database. Data quality indicators focused on validity, completeness, comparability, and timeliness as defined by the International Agency for Research on Cancer and the European Network of Cancer Registries (ENCR).7,8 Results Evolution of data collection and framework The transition from ROR to RON did not fundamentally change the hierarchical structure of data collection, with local registries submitting cases manually to RON through a unique electronic platform. Primary data sources remained the hospital’s systems containing clinical information, and secondary data sources include administrative systems for patient demographics and vital status. However, a standardized dataset was introduced (Figure 1), one that responds to ENCR’s recommendations,9 with the addition of new variables for this purpose. This dataset allowed the refinement of treatment and tumor characterization to enhance clinical granularity and allow the National Registry to achieve its objectives besides data reporting. This dataset is systematically collected in its entirety, without a defined minimum dataset with just a few core variables. Change in framework between Regional Cancer Registry and National Cancer Registry. ECOG, Eastern Cooperative Oncology Group; ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; PR, progesterone receptor; RON, Registo Oncológico Nacional; ROR, Registos Oncológicos Regionais; TNM, tumor–node–metastasis. RON also implemented automated integrations extracting core data from administrative systems, hospital information systems (hospitalization and surgery), and pathology reports (Figure 2). Although there is considerable potential for further integrations, these automations represent an important first step toward reducing manual effort and improving data completeness. Still, they are not yet available across all hospitals, with registries retaining a substantial reliance on manual data curation by local registry personnel. Primary and secondary sources of RON. RON, Registo Oncológico Nacional. (Image was made using Canva.com). Quality indicators implemented To support local registries in monitoring their data quality, the RON coordinating team provides monthly indicators on completeness. However, of the 685 public, private, and social sector institutions eligible to register in RON, only 55 had registered at least one case for the 2022 incidence data publication. Table 110 presents the completeness and validity indicators. Completeness of core variables used for incidence reporting is 100%. However, stage is missing in 40.4% of cases. The high microscopic verification (97.8%) and low death-certificate-only and unknown basis of diagnosis (both at <0.1%) meet internationally defined benchmarks. However, ill-defined morphology or topography affects 3.5% of cases, surgical procedure not corresponding to tumor type affects 14.9%, and poor stage standardization affects staging in 1.7%, with 779 different options recorded. Table 1Quality indicators for cancer cases diagnosed between 2018 and 2022 (N = 330 706)IndicatorN (%)Proportion of microscopically verified cases323 354 (97.8%)Proportion of death-certificate-only cases29 (0.0%)Proportion of cases with unknown basis of diagnosis31 (0.0%)Proportion of cases with ill-defined morphology or topography11 510 (3.5%)Proportion of cases with missing or unknown data for: Sex0 Age0 Residence0 Date of diagnosis0 Topography0 Morphology0 Behavior0 Grade0 Stage133 637 (40.4%) Surgery date or procedure (n = 207 315)2 (0.0%) Systemic therapy date or drug (n = 115 992)1153 (1.0 %) Radiotherapy date (n = 64 400)6 (0.0%) Vital status0Proportion of cases with coding errors Lateralitya1484 (0.4%) Date of deathb (n = 87 526)124 (0.04%) Surgical procedurec (n = 207 315)31 008 (14.9%) Staged5630 (1.7%)aInvalid when bilaterality is reported in unilateral organs, and ‘not applicable’ recorded where specification is required.bInvalid when date of death is before date of diagnosis.cInvalid with surgery not applicable to tumor.dInvalid when different from American Joint Committee on Cancer, Eighth Edition, definitions.10 Quality indicators for cancer cases diagnosed between 2018 and 2022 (N = 330 706) Invalid when bilaterality is reported in unilateral organs, and ‘not applicable’ recorded where specification is required. Invalid when date of death is before date of diagnosis. Invalid with surgery not applicable to tumor. Invalid when different from American Joint Committee on Cancer, Eighth Edition, definitions.10 Another type of severely missing information is follow-up other than vital status, namely regarding recurrence of tumor and treatment response. Most institutions do not have the capacity to register this information because of delay in recording primary tumor incidence data. Comparability is 100% because all institutions use International Classification of Diseases for Oncology, Third Edition, coding11 for registering and publications use International Classification of Diseases, 10th Revision, coding.12 The timeliness lag between diagnosis and publication of incidence data is currently 3 years. Previously, some regional registries with higher capacity achieved a shorter lag of 2 years. However, the combination of having to wait for all registries to submit data alongside the rising number of cancer cases in Portugal without a corresponding increase in registry capacity has extended the lag to 3 years for all registries. International reporting and real-world studies RON reports national-level indicators of cancer incidence (crude and age-standardized rates) and survival (5-year overall and net survival). These metrics are published in annual reports as the principal cancer burden indicators for Portugal, with stratification by sex, age group, and region of residence. Although the quality of these core indicators is robust and meets international standards for descriptive epidemiology, persistent missing data for stage currently limit meaningful analysis of stage-specific incidence and survival. Despite this, RON has successfully contributed to European-level collaborative projects and provided data for the European Cancer Information System 2022 data call (submission currently under review). This participation demonstrates the registry’s growing integration into European cancer surveillance networks and its alignment with internationally harmonized data standards. Furthermore, although a call for the Cancer Incidence in Five Continents edition considering data from 2018 onward is yet to be issued, the quality of the variables required for inclusion is now sufficient for RON to be considered for inclusion in the forthcoming edition. Furthermore, RON was able to provide data for many national and international studies since its implementation in 2018. These studies (N = 56) include themes such as incidence trends and environmental risks, treatment patterns, impact of screening programs, and differences in survival outcomes (Table 2). Table 2Distribution of studies across cancer type and themeCancerSpecific populationCOVID-19Incidence/survivalPreventionRegistryTreatmentBreast2153——Germ cell cancer——1———Cervical——21——Pediatric——1———Digestive1—32—1Head and neck—11———Lung——51—1Lymphoma—————1Leukemia——1———Prostate———1——Ovarian——1———Several/all cancers113534Skin———3——COVID-19, coronavirus disease 2019. Distribution of studies across cancer type and theme COVID-19, coronavirus disease 2019. Discussion Portugal’s PBCR has come a long way, from not being included in European cancer incidence estimates in 2006, when Spain’s data were used as proxy for Portugal,1 to becoming an undeniable part of European projects. Further, RON now has annual publications of incidence and survival that inform national health plans on oncology. The evolution from ROR to RON was characterized by the expansion of important clinical domains and the addition of many new variables to characterize both the tumor and its treatment. This transformation positions RON as a comprehensive research resource capable of supporting advanced real-world oncology studies, including comparative effectiveness research, health services disparities analysis, and molecular epidemiology. However, despite the important legislative and organizational advancement, the operational reality of Portuguese cancer registration is facing substantive challenges. Despite advances in automation, RON remains heavily dependent on manual data entry and curation, which introduces inevitable human error, processing delays, and data inconsistencies. Our analysis revealed that a few quality issues persist, with the 40% missing rate for stage being particularly concerning. This limits the ability to conduct stage-specific survival analyses, which are essential for understanding cancer outcomes and evaluating screening effectiveness. The issue of poor standardization, evidenced by >500 different stage options recorded, indicates inadequate data validation during entry. This is compounded by the 15% inappropriate surgery coding rate. A recent study that evaluated radiotherapy utilization patterns among older patients in Portugal showed the usage of RON but also revealed its limitations.13 It compared treatment patterns with European recommendations and identified that, across most cancer sites and age groups, external beam radiation therapy utilization fell below recommended levels, with an estimated 22% of new cancer cases potentially not receiving optimal radiotherapy. However, the results were limited by the lack of follow-up data, which often restricts RON-based studies to characterizing the diagnosis and initial treatment. The 3-year data publication lag14 also limits the registry’s capacity to inform real-time or near-real-time oncology policy and research. In a rapidly evolving field like oncology, such delays reduce the relevance of registry data for monitoring treatment patterns and outcomes. Nevertheless, these limitations can be reframed as opportunities. Future perspectives Building on existing automated integrations, RON could substantially reduce the manual burden and improve timeliness by incorporating additional sources, such as pharmacy and radiology systems, and genetic and molecular pathology databases. This would require substantial informatic intervention because standardized systems are not used by all institutions. Moreover, using natural language processing to extract structured data from unstructured reports and clinical notes and machine leaning to assist in case finding would assist in reducing manual burden. This would mostly improve the completeness of follow-up data. The implementation of real-time data validation rules during data entry (e.g. date logic checks and laterality-organ compatibility controls) alongside closed menus with standardized coding options and automated flagging of inconsistent entries for manual review would further strengthen data quality. Implementing a minimum dataset with a phased approach to data registration could also be adopted. Variables essential for descriptive incidence and basic indicators could be made available on a timely basis for final incidence publications, whereas fully curated data would be reserved for older cases. This would allow annual reports to be released without waiting for complete cohort maturity and at the same time aid some of the institutions that report no cases at all owing to being short-staffed. Some of these are already in discussion within the RON’s coordinating team because they seem to be practical solutions to improve the registry. Conclusions Although RON has made remarkable progress in clinical granularity and research utility, it continues to face operational challenges, such as heavy reliance on manual data entry, disparities, and reporting delays, that are common to many other countries.6 Active participation in European collaborative projects has contributed to improving the quality of the Portuguese cancer registry since its inception because these initiatives are grounded in international standards, guidelines, and best-practice recommendations. Building on this foundation, Portugal should now capitalize on the identified opportunities by pursuing a more innovative pathway, investing in emerging informatics solutions that enhance automation while maintaining essential human validation. This balanced approach will further strengthen the cancer registry as a robust, timely, and high-impact resource for oncology research and public health planning. Acknowledgments Funding None declared. Disclosure The authors have declared no conflicts of interest. References