Exploring trends in laryngeal cancer incidence, mortality and survival: implications for research and cancer control.

South Australian registry data were used to investigate trends in laryngeal cancer age-standardised incidence, mortality and disease-specific survival from 1977 to 2005. Incidence rates decreased by 32% from 1980-84 to 2000-05, affecting both sexes and ages under 70 years. There were concurrent reductions in mortality, although statistical significance was not achieved with the numbers of deaths examined (p>0.05). More than other cancers, laryngeal cancers presented in: the 50-79 year age range; males, particularly those born in Southern Europe; UK/Irish migrants; and residents of lower socio-economic areas. Compared with other cancers, laryngeal cancers were less common in more recent diagnostic periods. The ratio of glottis to other laryngeal cancers was higher in males, older patients, and those born in Southern Europe, UK/Ireland and Western Europe. A secular increase in this ratio was evident. The five-year survival from laryngeal cancer was 68%, with poorer outcomes applying for older patients, non-metropolitan residents, patients with cancers of laryngeal sub-sites other than glottis, and potentially patients born in Southern Europe. Secular changes in survival were not observed. Reductions in incidence are attributed to decreases in tobacco smoking in males and reductions in per capital alcohol consumption since the 1970s. The higher ratio of glottis to other laryngeal cancer sub-sites in males may indicate a greater contribution made by tobacco, as opposed to alcohol, in males. The lower survival observed in non-metropolitan patients may reflect poorer access to radiation oncology and other specialist services, although delays in diagnosis for other reasons may have contributed.

Comparing cancer profiles and survival of aboriginal and non-aboriginal patients in South Australia: where are the opportunities for improving Aboriginal health?

Data from the South Australian Cancer Registry (SACR) for 1977-2003 were used to calculate expected and actual distributions of cancer sites in Aboriginal versus non-Aboriginal populations. Expected distributions were calculated using indirect standardisation and compared with actual distributions using a global Chi-square test. Individual contributions to the Chi-square statistic (from each cancer site) were examined using a z-test and Bonferroni corrected p-value. The expected figure for each cancer site corresponds to the number of cancers we would have expected in Aboriginal patients if they had the same cancer distribution of site by age as the non-Aboriginal population. Expected 5- and 10-year survivals were also calculated and compared to expected survivals drawn from Statewide survivals adjusted for age at diagnosis. There was an overall significant difference in expected and actual cancer site distributions for South Australian Aboriginal male (c2 (17df) = 202.94) and female (c2 (20df) = 311.93) patients, and all patients collectively (c2 (22df) = 485.43). Aboriginal patients had poorer expected 5- and 10-year survival compared with South Australian non-Aboriginal patients, and even poorer actual 5- and 10-year survival than expected. The differences between the expected and actual cancer site distributions reflect the disparities in risk factor prevalence for largely preventable cancers and the survival results reflect the multitude of obstacles confronting Aboriginal patients with cancer compared with non-Aboriginal cancer patients. This study provides areas of focus for interventions to reduce cancer levels in the Aboriginal population and to improve survival of Aboriginal people diagnosed with cancer.

Myeloid leukaemia treatment and survival--the South Australian experience, 1977 to 2002.

OBJECTIVE: To evaluate trends in survival and treatment for myeloid leukaemia in South Australia during 1977-2002, using population-based survival data plus data on survival and treatment of patients at three teaching hospitals. METHODS: Population data were analysed using relative survival methods and hospital registry data using disease-specific survival. Univariate and multivariable analyses were undertaken. Multiple logistic regression analysis was used to investigate factors associated with first-line chemotherapy. RESULTS: South Australia recorded 1,572 new cases of acute myeloid leukaemia (AML) in 1977-2002, together with 536 cases of chronic myeloid leukaemia (CML). Of these cases, 42.6% were recorded in teaching hospital registries. The five-year survival for AML at the teaching hospitals of 14.5% was similar to the corresponding 12.0% for South Australia as a whole. The five-year survival for CML at these hospitals was higher, however, at 48.1% compared with 37.5% for all South Australian cases. Younger patients had higher survivals, both for AML and CML. An increase in survival was evident for more recently diagnosed cases for both leukaemia types, after adjusting for age. This increase in survival was accompanied by an increase over time in the proportion of patients at teaching hospitals having a primary course of chemotherapy. Cytarabine in combination with other agents was the most common induction therapy for AML. While hydroxyurea was the most common first-line treatment of CML, there were changes in clinical policies towards higher-dose treatments, plus trials of new agents and combination therapies. CONCLUSIONS: Secular gains in survival have occurred from AML and CML in association with an increased use of chemotherapy.

Changes in incidence of in situ and invasive breast cancer by histology type following mammography screening.

OBJECTIVE: To investigate secular trends and correlates of incidence of breast cancer by histology type following the introduction of population-based mammography screening. METHODS: Analysis of age-standardised incidence rates for 1,423 in situ and 16,157 invasive carcinomas recorded on the South Australian population-based cancer registry for the 1985-2004 diagnostic period. Multiple logistic regression was undertaken to compare socio-demographic characteristics by histology. Progression from in situ disease was investigated using the Kaplan-Meier method. RESULTS: The incidence of in situ lesions increased approximately seven-fold over the 20-year period, compared with an increase of about 40% for invasive cancers. The increase for in situ lesions was due to increases for ductal carcinomas, with little change for lobular lesions. By comparison, the percentage increase in incidence for invasive cancer was greater for lobular than ductal cancers. Both for in situ and invasive cancers, percentage increases were greatest for the screening target age range of 50-69 years. One in 14 in situ cases was found to progress to invasive cancer within seven years of diagnosis, but insufficient detail was available to determine whether the invasive cancers were a progression of the in situ lesions or whether they originated separately. These invasive cancers were smaller than generally applying for other invasive cancers of the female breast. CONCLUSIONS: The larger secular increases in incidence for in situ than invasive cancers would reflect the dominant role of mammography in the detection of ductal carcinoma in situ. The lack of an increase for lobular in situ lesions may have resulted from their poorer radiological visibility. The greater percentage increase for lobular than ductal invasive lesions may have been due to an increase in imaging sensitivity for these lesions, plus real increases in incidence. The smaller sizes of invasive cancers found in women with a prior in situ diagnosis may have resulted from more intensive medical surveillance, although the possibility of biological differences cannot be discounted.

Reasons for lower survival from non-Hodgkin's lymphoma among older patients.

Cancer-registry data for 710 patients, treated for non-Hodgkin's lymphoma (NHL) at a South Australian teaching hospital between 1977 and 2000, gave a five-year disease-specific survival of 53%, which was similar to population based estimates for Australia, the USA, and Europe. This figure reduced with age at diagnosis from 69% for patients less than 40 years at diagnosis to 30% for those aged 80 years or more. Multivariable analysis indicated that older age was predictive of lower survival (p<0.001), after adjusting for grade (Working Formulation), Ann Arbor stage,bulk disease, B symptoms (weight loss, unexplained fever, night sweats), extra-lymphatic site involvement, and diagnostic period. No other clinical variable, when included in the model, affected the risk coefficient for age. Even among patients gaining complete remission following chemotherapy, the relative risk of death from NHL was 2.11(95% CL: 1.24, 3.57) for patients aged 70 years or more at diagnosis when compared with younger patients. We conclude that older patients have lower survivals not explained by established risk factors and that this also applies to patients who achieve complete remission following chemotherapy.

Are cutaneous melanomas of specified thickness showing deeper levels of invasion at diagnosis?

Secular trends in Clark level were investigated by Breslow category for 8,432 invasive cutaneous melanomas diagnosed in South Australia in 1980-2000. More recently diagnosed lesions were found to have deeper levels. After adjusting for age at diagnosis, tumour site, histology, and thickness measured in half millimetres, the relative odds (95% confidence limits) of penetration to the reticular dermis or subcutaneous fat were 1.99 (1.59, 2.50) for the 1987-93 diagnostic period, and 2.82 (2.25, 3.54) for 1994-2000, when compared with 1980-86. After adjusting for melanoma thickness, the secular trends for deeper lesions applied to a broad cross-section of socio-demographic sub-groups, tumour sites, and histological types. While this similarity in trend would be consistent with a measurement effect, a real change cannot be ruled out and increased emphasis on earlier detection may be warranted. The prognostic implications of changes in inter-relationships between measures of thickness and level require periodic re-evaluation.

Comparative epidemiological characteristics of oesophageal adenocarcinoma and other cancers of the oesophagus and gastric cardia.

Secular trends and epidemiological characteristics of 1,581 oesophageal cancers, diagnosed in South Australian residents in 1977-2000, were analysed by histological type and diagnostic period, using multivariable Poisson regression and logistic regression. The age-adjusted incidence of squamous cell carcinoma did not vary significantly by diagnostic period, either in males (p = 0.195) or females (p = 0.087). By comparison, variations were observed for adenocarcinomas in males (p<0.001) and females (p = 0.015), with an increase in age-adjusted incidence of 169% for males and 150% for females between 1977-81 and 1997-2000. Most of these increases occurred in the 1990s. Secular differences were not evident for tumours of other or unknown histological type. The ratio of adenocarcinomas to squamous cell carcinomas was higher in patients who were aged 80 years or more, male, residents of high socio-economic areas, and those born in the United Kingdom/Ireland. Conversely, relatively low ratios presented for patients born in Southern and other parts of Europe. These differences by country of origin accord with differences between the national incidence rates for these countries, as indicated by international data. Differences in secular trend and country of birth between adenocarcinomas of the oesophagus and gastric cardia suggest that they are not expressions of the same disease. Preventive implications of these results are discussed.