Poster Presentation Clinical Oncology Society of Australia Annual Scientific Meeting 2018

Evaluation of a simplified Tumour-Node-Metastasis staging system in a population-based cancer registry (#152)

Sheena Lawrance 1 , Vidur Mahindra 1 , Chau Bui 1 , Claire Cooke-Yarborough 1 , Maria Arcorace 1
  1. NSW Cancer Institute, Eveleigh, NSW, Australia

Aim: In 2017, the New South Wales Cancer Registry (NSWCR) participated in a project, supported by Cancer Australia, aimed at providing national stage data for high-incident cancers, including colorectal cancer, diagnosed in 2011. Simplified business rules  based on American Joint Committee for Cancer (AJCC) Tumour-Node-Metastasis (TNM) staging were applied to obtain Registry-Derived (RD) stage, defined as the best estimate of TNM stage at diagnosis using routine notifications available within cancer registries. RD-stage was compared with summary stage Degree of Spread (DoS), which has been recorded for all applicable cancers in NSWCR at a population-based level since 1972, and AJCC-TNM, which has been collected variably since 2006. We determined the value-add of RD-staging in terms of staging completeness and alignment to gold-standard AJCC-TNM.

Methods: Colorectal stage data was extracted from NSWCR pre- and post- RD-staging to compare data completeness across all three staging systems. The alignment between DoS/RD-stage and AJCC-TNM was compared, as were the expected and observed cross-tabulated frequency distributions using a subset of NSWCR data. To determine differences between use of DoS, RD-stage and AJCC-TNM in research, we compared 3-year cox proportional-hazards survival models developed from each of the three stage variables.

Results: Compared to DoS, RD-staging did not improve stage data completeness for colorectal cancer and only provided small improvements in accuracy (concordance/kappa scores of 99%/99% and 81%/76% for RD-stage and DoS respectively). Both DoS and RD-stage survival models showed highly similar hazard ratios to gold-standard AJCC-SG models.

Conclusion: Although stage data is not mandated for reporting in Australia, it is important for epidemiological analyses. This study provides an evidence-based approach to inform researchers and epidemiologists of the comparability between three different cancer staging systems within a population-based cancer registry. For colorectal cancer, DoS and RD-Stage are shown to be comparable surrogate staging systems for AJCC-TNM stage.