Introduction:
This measure presents, for the first time, national stage at diagnosis data for the top five incidence cancers (breast (female), colorectal, lung, melanoma, and prostate cancer) in Australia. Stage at diagnosis indicates the extent to which a cancer has spread when first diagnosed and is an important prognostic factor for an individual’s cancer outcomes. It also provides contextual information for interpreting cancer survival and recurrence, and treatment patterns at the population level. The collection of quality national cancer incidence data accompanied by stage at diagnosis enhances our ability to:
- Understand the respective contributions of stage as compared with other factors (e.g. treatment) to variations in survival; and
- Help identify and inform where further research and targeted cancer control strategies may impact on tumour types and population groups where cancers may commonly be diagnosed at an advanced stage.
Using data sources that are routinely accessible to all cancer registries, the following analysis for “capture of stage data” presents the proportion of incident cancer cases in 2011 that were assigned a value for Registry-Derived stage at diagnosis (RD-Stage), henceforth referred to as ‘stage’. For this measure, the proportion of cases that were able to be staged is also referred to as “staging completeness”. The “capture of stage data” is one of two measures reported on the NCCI website that are related to adult cancer stage at diagnosis. The second measure is the “distribution of cancer stage” .
More information about data sources, methods for collection, and guidance for interpreting the data can be found in the ‘About the Data’ tab.
Charts
About this measure
The lack of high quality national cancer staging data is an identified gap in our data knowledge in Australia. As part of the Stage, Treatment and Recurrence (STaR) project, Cancer Australia has worked with population-based cancer registries (PBCRs) and the Australasian Association of Cancer Registries (AACR) to develop nationally-standardised methodologies for collecting stage at diagnosis data. This work has initially focussed on the top five incident cancers (breast (female), colorectal, lung, prostate cancer, and melanoma), for population-level reporting purposes.
Cancer Australia supported the Cancer Council Victoria (CCV) in developing Business Rules for the collection of national cancer stage at diagnosis for invasive tumours based on the Tumour, Node, and Metastases (TNM) staging system as developed and maintained by the American Joint Committee on Cancer (AJCC) in collaboration with the International Union for Cancer Control (UICC). The TNM Staging System is based on the extent of the tumour (T), the extent of spread to the lymph nodes (N), and the presence of metastasis (M). As each cancer type has its own classification system, letters and numbers do not always mean the same thing for every kind of cancer. Once the T, N, and M are determined, they are combined, and an overall stage of 0, I, II, III, IV is assigned1.
Using these Business Rules, all Australian PBCRs have derived cancer stage at diagnosis (RD-Stage) using data sources that are routinely accessible to all cancer registries. RD-Stage at diagnosis is defined as the best estimate of summary TNM stage at diagnosis that can be derived nationally by Australian cancer registries from the data sources available to them. Its purpose will be for population-based analyses, not clinical applications.
RD-Stage has the following values:
- Stage 1 (equivalent to TNM stage I; Early stage)
- Stage 2 (equivalent to TNM stage II; Early stage)
- Stage 3 (equivalent to TNM stage III; Locally advanced)
- Stage 4 (equivalent to TNM stage IV; Metastatic)
- Stage unknown
- Stage not applicable (including tumours with morphologies not eligible for TNM and tumours with no histological confirmation) – These have been excluded for the purpose of this analysis.
This measure presents, for the first time, national stage at diagnosis data for the top five incidence cancers (breast (female), colorectal, lung, melanoma, and prostate) in Australia using data sources that are routinely accessible to all cancer registries. To date, RD-Stage has been collected only for an index year (2011) as this represents the most up-to date cancer incidence data for which 5-year relative survival by stage could be reported in Australia (up to 2016 – data forthcoming). In collaboration with Cancer Australia, the PBCRs and AACR, the Australian Institute of Health and Welfare (AIHW) has combined newly collected information on RD-Stage with information on cancer incidence from the Australian Cancer Database (ACD).
Upcoming releases for stage at diagnosis data:
Cancer Australia, the PBCRs, the AACR and the AIHW are also working together to combine incidence by stage at diagnosis with mortality data from the National Death Index (NDI). Once complete, these data will allow for the reporting of national 5-year relative survival rates by stage at diagnosis collected by registries for the first time in Australia. Currently, the only international data for relative survival (1-year) by cancer stage of which we are aware is published by the National Cancer Registration and Analysis Service (NCRAS) in England – These data can be accessed through the ‘References’ tab.
It is anticipated that Australian national 5-year relative survival data by cancer type and stage at diagnosis for the top five incidence cancers will be released on the NCCI website later in 2018. This will be an important opportunity to benchmark survival by cancer stage at diagnosis across comparable populations.
Cancer Australia is also supporting the development of a set of complementary Business Rules for deriving cancer stage at diagnosis for 16 paediatric (childhood) cancer types as well as scoping the development and testing of new business rules for additional cancer types.
Current status
Derivation of RD-Stage values is dependent upon the availability to population-based cancer registries (PBCRs) of the required source records for each incident cancer. The proportion of cancers where an RD-Stage value could be assigned varied with the records available to the PBCRs. This is referred to as staging “completeness”.
Completeness of RD-Stage capture varied by cancer type in 2011
In 2011, RD-Stage “completeness” was high overall for each of the top five incident cancers. At least 94% of incidence cases were able to be assigned a value for RD-stage at diagnosis for melanoma (97%), prostate (97%), and breast (female) cancer (94%). Lower staging completeness was found for colorectal cancer (88%), and for lung cancer (72%). Among colorectal cancers, staging “completeness” was higher for colon (90%) than rectal (83%) cases.
Completeness of RD-Stage by sex and age in 2011
The proportions of top five incidence cancer cases that were able to be staged were similar for both sexes but varied across age groups:
- The proportion of cancer cases that were staged was the same for both sexes for melanoma (97%) and colorectal cancer (88%) and similar for lung cancer (71%-72%).
- For persons less than 50 years of age, a slightly higher proportion of cases were staged than for persons aged 50 years and over for:
- Breast (female) cancer (96% compared to 93%)
- Colorectal cancer (90% compared to 88%)
- Lung cancer (80% compared to 71%).
- Similar proportions of cancers were staged for both <50 years and ≥ 50 years age groups for melanoma (97% for both) and prostate cancer (98% and 97%, respectively).
Examination of staging completeness for each cancer type showed a similar staging completeness for each 5-year age group up to at least 74 years of age.
- For breast (female) cancer, almost all cancers (95% to 98%) were staged for 5-year age groups up to 74 years of age, after which staging completeness decreased for older age groups (from 92% to 67%).
- For colorectal cancer, the majority of cancers (88% to 91%) were staged for 5-year age groups up to 79 years of age, after which staging completeness decreased for older age groups (from 87% to 78%).
- For lung cancer, the majority of cancers (75% to 84%) were staged for 5-year age groups up to 74 years of age, after which staging completeness decreased for older age groups (from 66% to 55%).
- For melanoma, almost all cases (96% to 98%) were staged across all age groups.
- For prostate cancer, almost all cancers (98% to 100%) were staged for 5-year age groups up to 79 years of age, after which staging completeness decreased for older age groups (from 92% to 84%).
- Similar patterns of completeness were apparent among males and females when comparing the proportion of cancers that were staged across 5-year age groups.
Completeness of RD-Stage for Aboriginal and Torres Strait Islander persons in 2011
Due to the small number of incident cancer cases in a single year (2011), results of staging completeness analysis by Aboriginal and Torres Strait Islander (henceforth referred to as Indigenous) status should be interpreted with caution. Indigenous status data are only available for cancer incidence data in this report for New South Wales, Victoria, Queensland, Western Australia and Northern Territory as past investigations indicated that these jurisdictions had a higher completeness of Indigenous status recording.
- A similar proportion of cancers were staged for both Indigenous and non-Indigenous persons, for each cancer type. The following proportions apply to completeness of cancer staging for Indigenous and non-Indigenous persons, respectively:
- Breast (female) cancer (95% compared to 96%)
- Colorectal cancer (87% compared to 89%)
- Lung cancer (75% compared to 73%).
- Melanoma (95% compared to 96%).
- Prostate (98% compared to 97%).
- Similar patterns of staging completeness were apparent for both sexes when comparing the proportion of cancers staged by Indigenous status, except for lung cancer. For lung cancer there were:
- Slightly higher proportions of cancers staged for Indigenous females (77%) than non-Indigenous females (73%).
- Similar proportions of cancers staged for Indigenous males (72%) and non-Indigenous males (74%).
Completeness of RD-Stage by remoteness area of residence in 2011
The proportion of incident cancer cases that were staged was similar for most cancer types when comparing Major Cities to Inner and Outer Regional areas, but the proportion was slightly lower for Remote and Very Remote areas.
- For breast (female) cancer, staging completeness was similar for Major Cities (94%), Inner and Outer regional areas (95%) and Remote Very Remote areas (92%)
- For colorectal, lung and prostate cancer, staging completeness was slightly lower in Remote and Very Remote areas.
- For colorectal cancer, staging completeness was 88% for both Major Cities and Inner and Outer Regional areas, and slightly lower for Remote and Very Remote areas (83%).
- For lung cancer, staging completeness was 73% for Major Cities, 69% for Inner and Outer Regional areas (69%), and slightly lower for Remote and Very Remote areas (67%).
- For prostate cancer, staging completeness was 98% for both Major Cities and Inner and Outer Regional areas, and slightly lower for Remote and Very Remote areas (95%).
- Melanoma staging completeness by remoteness area of residence could not be examined for Remote and Very Remote areas due to very small numbers for unknown stage at diagnosis. Staging completeness was similar when comparing Major Cities and Other Remoteness areas combined (97% compared to 98%)
Completeness of RD-Stage by socioeconomic status in 2011
The proportion of cancers that were staged was similar across socioeconomic status (SES) areas for each cancer type.
- Similar proportions of cancer cases were staged across SES area for:
- Breast (female) cancer (ranging from 93% to 95%).
- Colorectal cancer (ranging from 87% to 89%).
- Melanoma (ranging from 96% to 98%).
- Prostate cancer (ranging from 97% to 98%).
- Lung cancer (ranging from 69% to 74%).
- Similar patterns of staging completeness were apparent for both sexes when also comparing the proportion of cancers staged across SES areas.
International benchmarks for completeness of stage at diagnosis
Due to differences in the scope of data collection, methodology and availability of information for staging, the completeness of international stage at diagnosis varies considerably. Comparisons of staging completeness however, are provided here as context for assessing completeness of staging in Australia. More detailed information on stage data for Canada (Canadian Partnership Against Cancer) and England (National Cancer Intelligence Network) is available the ‘References’ tab. In this section, data are presented for Canada using 2013 data, England using 2012 data, and Australia using 2011 data.
In this section, data are presented for currently available staging completeness in Canada (2013), England (2012), and Australia (2011).
The proportion of staged cancer cases varied for each cancer type internationally. Australia had a comparable or higher level of completeness for stage at diagnosis data for each cancer type compared to Canada and England, except for lung cancer where staging completeness Australia was lower.
About the data
Unit of analysis:
The unadjusted crude proportion of cancer cases for which stage data are available for cases with a principal diagnosis of:
|
Cancer type* |
ICD-10-AM codes |
|
Breast (female) |
C50 |
|
Colorectal |
C18.0, C18.2–C20 |
|
C18** |
|
C19-C20 |
|
Lung |
C34 |
|
Melanoma*** |
C43 |
|
Prostate |
C61 |
*The top 5 incident cancers that were eligible for staging comprise breast (female) cancer (ICD-10 code C50), colorectal cancer excluding appendix (C18.0, C18.2–C20), lung cancer excluding trachea (C34), melanoma of the skin excluding skin of genitals (C43) and prostate cancer (C61). Certain morphology codes that were not eligible for staging are excluded, such as sarcomas, lymphomas or carcinoid tumours.
**Colon cancer (C18) excludes cancer of the appendix (C18.1)
***Excludes melanoma of “unknown primary site”
Numerator: Incident cancer cases for a selected RD-Stage at diagnosis value (staged or unknown) for a selected cancer type.
Denominator: All eligible RD-Stage records that were able to be matched to an incident cancer case in the ACD for the relevant cancer type. The denominator includes cases with an "Unknown" stage at diagnosis for which the registry did not have sufficient information to derive stage.
Scope:
RD-Stage
RD-Stage at diagnosis is defined as the best estimate of summary TNM stage of diagnosis as derived by cancer registries from data sources available to them. These data will be used for statistical purposes as opposed to clinical management and supporting individual patient care. Clinical requirements for prognostic precision differ from epidemiological requirements for comparability and statistical completeness.2 Specifically, the collection of RD-Stage:
- Is intended for epidemiological population-based analyses only – in particular, this information stage at diagnosis will assist in understanding the severity of disease across tumour types and different and sociodemographic groups as well as inform us of patterns of incidence and mortality.
The Business Rules have been tested and reviewed by all states and territories to ensure applicability across all Australian population based cancer registries. The Business Rules have also been endorsed as a national standard for the collection of stage data by the AACR.
Australian Cancer Database3
Cancer incidence indicates the number of new cancers diagnosed during a specified time period (usually one year). The major source of national cancer incidence data is the ACD which contains records of all primary, malignant cancers (except basal cell and squamous cell carcinomas of the skin) diagnosed in Australia since 1982.
All Australian states and territories have legislation that makes cancer a notifiable disease. Various designated bodies, i.e., institutions such as hospitals, pathology laboratories and registries of births, deaths and marriages, are required to report cancer cases and deaths to their jurisdictional cancer registries.
Each registry supplies incidence data annually to the AIHW under an agreement between the registries and the AIHW. These data are compiled into the ACD, the only repository of national cancer incidence data.
Linkage of RD-Stage and the ACD
The data used for reporting this measure have been created by linkage of data from RD-Stage collection and the ACD. These data are therefore limited to records that have been matched across these two collections. For this analysis, 3.7% records in the RD-Stage collection (approximately 2,500 out of 72,200 cases) have been excluded from these analyses for the following reasons:
- RD-stage record did not link to the ACD.
- The RD-Stage record linked to an ACD record that was out of scope.
- RD-Stage record was ineligible for stage (such as sarcomas, lymphomas or carcinoid tumours)
- RD-Stage record was a duplicate
A relatively small number of records (approximately 600, less than 1%) records in the ACD were in scope but did not link to the RD-Stage collection. These records did not link because:
- They had been altered since being submitted to the ACD and were now out of scope; or
- Applied to melanomas of unknown primary site but not coded as such in the ACD; or
- Had been staged at a point after diagnosis but not at diagnosis.
International data on stage at diagnosis – England and Canada
Data presented for the international comparisons of stage at diagnosis have been sourced from the Canadian Partnership Against Cancer (Canada) and the National Cancer Intelligence Network (England). Data for Australia and England are available at the national level. For Canada, data are not available for Quebec. Further information on the scope, and methodology for collection, and relevant caveats are available through the links under the ‘References’ tab.
The Canadian data are presented in greater detail than the Australian and English data, the following changes have been made to the publicly available Canadian data for ease of comparison:
- Stage 0 cancers have been excluded from totals and proportions;
- Sub-stages (e.g. Stage IA, Stage IIB) have been aggregated to Stage 1, Stage 2, Stage 3 and Stage 4, accordingly.
- For lung cancer, small cell and non-small cell lung cancers have been grouped. Small cell lung cancers represented 13% of staged lung cancers in Canada.
- Cases where stage data were “not available” have been combined with “Unknown” stage.
Data caveats
- This analysis presents crude proportions that have not been adjusted.
- Collection of these data has provided an insight into differences in the availability, extent and accessibility of information that is required to derive RD-Stage across Population-based Cancer Registries (PBCRs). Notably, the availability and quality of data accessible to PBCRs was found to improve during the study period.
- Remoteness area of the patient's usual place of residence was defined using the ABS Australian Statistical Geography Standard (ASGS) remoteness structure classification, 2011. The process for calculating remoteness areas results in some records being split across areas, e.g. a record might be 0.6 Major Cities and 0.4 Inner Regional. The number of incident cases by remoteness, expressed as decimals and aggregated for split records, has been rounded to the nearest whole number. Totals may differ from other demographic breakdowns due to rounding.
- Socioeconomic group of the patient's usual place of residence defined using the ABS SEIFA Index of Relative Socioeconomic Disadvantage, 2011.
- Due to the small number of incident cancer cases in a single year (2011), results of analysis by Indigenous status should be interpreted with caution.
- Analyses by Indigenous status are only available in this report for New South Wales, Victoria, Queensland, Western Australia and Northern Territory, where higher completeness of reporting Indigenous status has been determined by the AIHW in past analyses.
References
Cancer Australia, 2008. A National Cancer Data Strategy for Australia. (https://canceraustralia.gov.au/sites/default/files/publications/ncds_final_web1_504af02093a68.pdf).
Cancer Australia. The Stage, Treatment, and Recurrence project. (https://canceraustralia.gov.au/research-data/cancer-data/improving-cancer-data).
Data
Australian Institute of Health and Welfare.
Australian Cancer Incidence and Mortality (ACIM) books provide incidence and mortality by cancer type and selected demographic groups. (https://www.aihw.gov.au/reports/cancer/acim-books/contents/acim-books).
Canadian Partnership Against Cancer.
Provides information on the capture and distribution of stage data for selected provinces.
(http://www.systemperformance.ca/cancer-control-domain/diagnosis/capture-of-stage/).
(http://www.systemperformance.ca/cancer-control-domain/diagnosis/stage-distribution/).
National Cancer Intelligence Network.
The National Cancer Registration and Analysis Service provides information on cancer survival by stage at diagnosis in England.
(http://www.ncin.org.uk/publications/survival_by_stage).
Thursfield V, Farrugia H. Cancer in Victoria: Statistics & Trends 2014. Cancer Council Victoria, Melbourne 2015.
Provides information on prostate cancer stage at diagnosis in Victoria.
http://www.cancervic.org.au/downloads/cec/cancer-in-vic/CCV-statistics-trends-2014.pdf
References
-
American Joint Committee on Cancer. What is cancer staging? Accessed 26 March 2018; https://cancerstaging.org/references-tools/Pages/What-is-Cancer-Staging.aspx
-
Walters S, Maringe C, Butler J et al. 2013. Comparability of stage data in cancer registries in six countries: lessons from the International Cancer Benchmarking Partnership. Int. J. Cancer: 132; 676-685.
-
Australian Institute of Health and Welfare. METeOR: Australian Cancer Database 2014; Quality Statement. Accessed 16 February 2018; http://meteor.aihw.gov.au/content/index.phtml/itemId/687104
