Bayesian approach to predicting cancer incidence for an area without cancer registration by using cancer incidence data from nearby areas

Projections in Breast and Lung Cancer Mortality among Women: A Bayesian Analysis of 52 Countries Worldwide

Among women, lung cancer mortality rates have surpassed those for breast cancer in several countries. This reflects the breast cancer mortality declines due to access to screening and effective treatment alongside the entrance of certain countries in stages of the tobacco epidemic in which smoking becomes more prevalent in women. In this study, we project lung and breast cancer mortality until 2030 in 52 countries. Cancer mortality data were obtained from the WHO Mortality Database. Age-standardized mortality rates (ASMR), per 100,000, were calculated (direct method) for 2008 to 2014 and projected for the years 2015, 2020, 2025, and 2030 using a Bayesian log-linear Poisson model. In 52 countries studied around the world, between 2015 and 2030, the median ASMR are projected to increase for lung cancer, from 11.2 to 16.0, whereas declines are expected for breast cancer, from 16.1 to 14.7. In the same period, the ASMR will decrease in 36 countries for breast cancer and in 15 countries for lung cancer. In half of the countries analyzed, and in nearly three quarters of those classified as high-income countries, the ASMR for lung cancer has already surpassed or will surpass the breast cancer ASMR before 2030. The mortality for lung and breast cancer is higher in high-income countries than in middle-income countries; lung cancer mortality is lower in the latter because the tobacco epidemic is not yet widespread. Due to these observed characteristics of lung cancer, primary prevention should still be a key factor to decrease lung cancer mortality.Significance: The mortality for lung and breast cancer is projected to be higher in high-income countries than in middle-income countries, where lung cancer mortality is expected to surpass breast cancer mortality before 2030. Cancer Res; 78(15); 4436-42. ©2018 AACR.

Assessing predicted age-specific breast cancer mortality rates in 27 European countries by 2020

BACKGROUND

We assessed differences in predicted breast cancer (BC) mortality rates, across Europe, by 2020, taking into account changes in the time trends of BC mortality rates during the period 2000-2010.

METHODS

BC mortality data, for 27 European Union (EU) countries, were extracted from the World Health Organization mortality database. First, we compared BC mortality data between time periods 2000-2004 and 2006-2010 through standardized mortality ratios (SMRs) and carrying out a graphical assessment of the age-specific rates. Second, making use of the base period 2006-2012, we predicted BC mortality rates by 2020. Finally, making use of the SMRs and the predicted data, we identified a clustering of countries, assessing differences in the time trends between the areas defined in this clustering.

RESULTS

The clustering approach identified two clusters of countries: the first cluster were countries where BC predicted mortality rates, in 2020, might slightly increase among women aged 69 and older compared with 2010 [Greece (SMR 1.01), Croatia (SMR 1.02), Latvia (SMR 1.15), Poland (SMR 1.14), Estonia (SMR 1.16), Bulgaria (SMR 1.13), Lithuania (SMR 1.03), Romania (SMR 1.13) and Slovakia (SMR 1.06)]. The second cluster was those countries where BC mortality rates level off or decrease in all age groups (remaining countries). However, BC mortality rates between these clusters might diminish and converge to similar figures by 2020.

CONCLUSIONS

For the year 2020, our predictions have shown a converging pattern of BC mortality rates between European regions. Reducing disparities, in access to screening and treatment, could have a substantial effect in countries where a non-decreasing trend in age-specific BC mortality rates has been predicted.

Bayesian prediction of lung and breast cancer mortality among women in Spain (2014-2020)

BACKGROUND

Breast cancer (BC) is the main cause of cancer mortality among women, and mortality from lung cancer (LC) is increasing among women. The purpose of the present study was to project the mortality rates of both cancers and predict when LC mortality will exceed BC mortality.

METHODS

The cancer mortality data and female population distribution were obtained from the Spanish National Statistics Institute. Crude rate (CR), age-standardized rate (ASR), and age-specific rate were calculated for the period 1980-2013 and projected for the period 2014-2020 using a Bayesian log-linear Poisson model.

RESULTS

All calculated rates were greater for BC than for LC in 2013 (CR, 27.3 versus 17.3; ASR, 13.5 versus 9.3), and the CR was not projected to change by 2020 (29.2 versus 27.6). The ASR for LC is expected to surpass that of BC in 2019 (12.9 versus 12.7).

CONCLUSIONS

By 2020 the LC mortality rates may exceed those of BC for ages 55-74 years, possibly because of the prevalence of smoking among women, and the screening for and more effective treatment of BC. BC screening could be a good opportunity to help smokers quit by offering counseling and behavioral intervention.

Estimates of prostate cancer burden in Italy

Age-standardized incidence rates of prostate cancer (PC) sharply increased during the period 1990-2005 in Italian areas covered by cancer registries, while corresponding mortality rates remained nearly constant. The latest observations have reported on a reversal of the incidence trend with decreasing values after 2005. We provided incidence, mortality, and prevalence estimates at national and geographical area levels, together with time projections up to the year 2020. We applied the MIAMOD method, using as input national mortality data for the years 1970-2010 and population-based survival data for the period of diagnosis (1985-2002). We assumed relative survival of prostate cancer remained constant after the year of diagnosis (2005). The age-standardized incidence rates of PC were estimated to increase during the period 1984-2005, from 31 per 100,000 in 1984 to 93 per 100,000 in 2005. From 2005 onwards, the estimated rates declined to 71 in 2015 and to 62 in 2020. Age-standardized mortality rates slightly increased from 1970 up to about 19 per 100,000 in 1999 and then started to decrease with an estimated reduction of about 2.3% per year. Mortality projections indicated a continuing reduction, with a predicted age-standardized rate of about 12 per 100,000 in 2020. Prevalence was estimated to continuously increase up to a crude prevalence value of 1.2% in the year 2020. The results indicate that the epidemic peak of PC was reached around the year 2005 followed by declining incidence rates, while a substantial decrease in mortality, starting during the early 2000s, is expected to continue during the 2010s.

Contribution of changes in demography and in the risk factors to the predicted pattern of cancer mortality among Spanish women by 2022

BACKGROUND

Changes in the burden of cancer mortality are expected to be observed among Spanish women. We predict those changes, in Spain, for breast cancer (BC), colorectal cancer (CRC), lung cancer (LC) and pancreatic cancer (PC) from 2013 to 2022.

METHODS

Bayesian age-period-cohort modeling was used to perform projections of the cancer burden in 2013-2022, extrapolating the trend of cancer mortality data from 1998 to 2012. We assessed the time trends of the crude rates (CRs) during 1998-2012, and compared the number of cancer deaths between the periods 2008-2012 and 2018-2022 to assess the contribution of demographic changes and changes in the risk factors for cancer.

RESULTS

During 1998-2012, CRs of cancer decreased for BC (0.3% per year) and increased for LC (4.7%), PC (2%) and CRC (0.7%). During 2013-2022, CRs might level off for CRC, whereas the time trends for the remaining cancers might continue at a similar pace. During 2018-2022, BC could be surpassed by CRC as the most frequent cause of cancer mortality among Spanish women, whereas LC could be the most common cause of cancer mortality among women aged 50-69 years (N/year=1960 for BC versus N/year=1981 for LC). Comparing 2018-2022 and 1998-2012, changes in the risk factors for cancer could contribute 37.93% and 18.36% to the burden of LC and PC, respectively, and demographic shifts - mainly due to ageing (19.27%) - will drive the burden of CRC.

CONCLUSIONS

During 2018-2022, demographic changes (ageing) and changes in risk factors could have a different impact on the lifetime risk of cancer among Spanish women.

Selecting the minimum prediction base of historical data to perform 5-year predictions of the cancer burden: The GoF-optimal method

BACKGROUND

Predicting the future burden of cancer is a key issue for health services planning, where a method for selecting the predictive model and the prediction base is a challenge. A method, named here Goodness-of-Fit optimal (GoF-optimal), is presented to determine the minimum prediction base of historical data to perform 5-year predictions of the number of new cancer cases or deaths.

METHODS

An empirical ex-post evaluation exercise for cancer mortality data in Spain and cancer incidence in Finland using simple linear and log-linear Poisson models was performed. Prediction bases were considered within the time periods 1951-2006 in Spain and 1975-2007 in Finland, and then predictions were made for 37 and 33 single years in these periods, respectively. The performance of three fixed different prediction bases (last 5, 10, and 20 years of historical data) was compared to that of the prediction base determined by the GoF-optimal method. The coverage (COV) of the 95% prediction interval and the discrepancy ratio (DR) were calculated to assess the success of the prediction.

RESULTS

The results showed that (i) models using the prediction base selected through GoF-optimal method reached the highest COV and the lowest DR and (ii) the best alternative strategy to GoF-optimal was the one using the base of prediction of 5-years.

CONCLUSIONS

The GoF-optimal approach can be used as a selection criterion in order to find an adequate base of prediction.

Cancer burden in China: a Bayesian approach

Background

Cancer is a serious health issue in China, but accurate national counts for cancer incidence are not currently available. Knowledge of the cancer burden is necessary for national cancer control planning. In this study, national death survey data and cancer registration data were used to calculate the cancer burden in China using a Bayesian approach.

Methods

Cancer mortality and incidence rates for 2004–2005 were obtained from the National Cancer Registration database. The third National Death Survey (NDS), 2004–2005 database provided nationally representative cancer mortality rates. Bayesian modeling methods were used to estimate mortality to incidence (MI) ratios from the registry data and national incidence from the NDS for specific cancer types by age, sex and urban or rural location.

Results

The total estimated incident cancer cases in 2005 were 2,956,300 (1,762,000 males, 1,194,300 females). World age standardized incidence rates were 236.2 per 100,000 in males and 168.9 per 100,000 in females in urban areas and 203.7 per 100,000 and 121.8 per 100,000 in rural areas.

Conclusions

MI ratios are useful for estimating national cancer incidence in the absence of representative incidence or survival data. Bayesian methods provide a flexible framework for smoothing rates and representing statistical uncertainty in the MI ratios. Expansion of China’s cancer registration network to be more representative of the country would improve the accuracy of cancer burden estimates.

Time trends of cancer incidence and mortality in Catalonia during 1993-2007

PURPOSE

To describe time trends of cancer in Catalonia, Spain, during the period 1993-2007.

METHODS/PATIENTS

Data have been provided by two population-based cancer registries, Girona and Tarragona, and the Catalan mortality registry. Cancer incidence in Catalonia has been estimated through modeling methods using data from these health structures.

RESULTS

During 2003-2007, there were 20,042 cancer cases and 9,842 deaths per year among men and 13,673 new cancer cases and 5,882 deaths among women. The most frequent incident cancers among men were prostate (N = 4,258), lung (N = 3,021), colorectal (N = 3,007) and bladder (N = 2,238), whereas among women they were breast (N = 3,907), colorectal (N = 2,088), corpus uteri (N = 734) and lung (N = 527). During 1993-2007, age-standardized incidence rates (ASIRs) rose 1.2 % per year among men [prostate (6.3 %), testis (5.7 %), kidney (2.9 %), liver (2.2 %) and colorectal (2.1 %)]. ASIRs decreased for stomach (-2.9 %), oral cavity and pharynx (-2.8 %), larynx (-2.7 %) and esophagus (-2 %). Among women, ASIRs only rose for lung (5.2 %), kidney (3.1 %), oral cavity and pharynx (2.6 %) and thyroid (1.6 %). ASIRs decreased for corpus uteri (-2.3 %), stomach (-1.7 %) and ovary (-1.6 %). Cancer mortality decreased -1.3 % per year among men and -2.1 % among women during the same period.

CONCLUSION

Among men, the decrease of incidence/mortality of tobacco-related tumors was related to a reduction of smoking prevalence. Among women, the stabilization of breast cancer incidence and the rise of lung cancer incidence are similar to that observed in most European regions. These results allow assessing the effectiveness of public health strategies and they pose new frontiers for cancer control in Catalonia.