Registry-derived stage (RD-Stage) for capturing cancer stage at diagnosis for endometrial cancer: development and validation of business rules and capacity to report at a national level.

Abstract

Abstract Background Registry-derived stage (RD-Stage) provides a mechanism to capture cancer stage at diagnosis from routinely captured data available to population-based cancer registries (PBCRs). In 2021, a project was undertaken to develop business rules to capture RD-Stage for endometrial cancer, understand current capacity to capture this information at a national level, and assess how it compares with stage recorded in clinical notes at diagnosis by clinicians. Methodology Business rules for deriving RD-stage (Endometrial carcinoma) were developed using AJCC Cancer Staging Manual 8th edition and endorsed by a tumour-specific Expert Working group comprising cancer specialists responsible for delivering cancer care and PBCR epidemiologists and medical coders. Baseline completeness of data fields required to calculate RD-Stage, and an overall proportion of cases for whom an RD stage could be assigned was assessed across each Australian jurisdiction. A validation study was undertaken in one jurisdiction to compare RD-Stage (Endometrial cancer) calculated by the Victorian Cancer Registry (VCR) with clinical stage recorded in cases diagnosed in 2018-2020 in the National Gynae-Oncology Registry (NGOR). Results The level of completeness of data to enable RD-Stage (Endometrial carcinoma) to be calculated across jurisdictions ranged from 0 to 89%. Three jurisdictions captured degree of spread of cancer only (instead of TNM data) and therefore captured no data which would enable RD-Stage to be calculated. The RD-Stage (Endometrial carcinoma) validation study found that RD-Stage could not be derived for 64/485 (13%) cases in the VCR and was not captured for 44/485 (9%) cases in NGOR. There was concordance at stage level (I, II, III, IV) in 393/410 (96%) of cases (95.8%, Kendall’s coefficient=0.95). Conclusion A lack of consistency in data captured by, and data sources reporting to, PBCRs meant that it was not possible to provide an accurate national baseline of endometrial carcinoma stage at diagnosis. In a sample of Victorian cases, there was very good concordance between RD-Stage (Endometrial carcinoma) and that recorded by clinicians in medical records and captured by NGOR. However, for 10% of cases RD-Stage could not be calculated because data was not available. RD-Stage provides a useful tool to be used for population epidemiological purposes.