Geospatial patterns by cancer stage across Australia for three common cancers.

Background: Monitoring cancer stage is vital to interpret cancer incidence and survival patterns, yet there are currently no cancer stage estimates by small areas across Australia, despite demonstrated large disparities in cancer incidence and survival. While cancer stage data is not routinely collected in Australia, a pilot project collected stage information nationwide in 2011.

Methods: Data on all primary invasive melanoma, female breast and prostate cancers (stages 1-4) diagnosed during 2011 in Australia were categorised into early and intermediate/advanced stage at diagnosis. Bayesian spatial models were used to estimate standardised incidence rates (SIRs) and proportions of cancer stage across 2148 statistical areas level 2. The correlation between early and more advanced cancer rates was explored using exceedance probabilities.

Results: Both melanoma and prostate cancer had mainly early stage diagnoses. There was large variation in rates across the nation, and also substantial correlation between SIRs of early and more advanced stage for melanoma and prostate cancer. In contrast, breast cancer had a higher proportion of advanced cancers diagnosed, less pronounced variation in rates and limited correlation between early and more advanced stage SIRs. The proportion of cases diagnosed as early stage varied across Australia by type of cancer.

Conclusion: This study uncovered important spatial patterns in the diagnosis of cancer by stage across the country, which varied by cancer type and location. There is an urgent need to have contemporary information about stage at diagnosis routinely included in population-based cancer registries across the country.