Development of a microsimulation of melanoma mortality for evaluating the effectiveness of population-based skin cancer screening

The effectiveness of a population-based skin cancer screening program: evidence from Germany

In this paper, we analyze how a nationwide population-based skin cancer screening program (SCS) implemented in Germany in 2008 has impacted the number of hospital discharges following malignant skin neoplasm diagnosis and the malignant melanoma mortality rate per 100,000 inhabitants. Our panel data, drawn from the Eurostat database, cover subregions in 22 European countries, measured at the lowest nomenclature of territorial units for statistics (NUTS) level for 2000-2013. Applying fixed effects methods, we find a significantly positive and robust effect of the German SCS on the number of patients diagnosed with malignant skin neoplasm. However, the program does not significantly influence the melanoma mortality rate. This finding conflicts with the decreased melanoma mortality rate found for the pilot SCS program in northern Germany. Our results indicate that Germany's nationwide SCS program is effective in terms of a higher diagnosis rate for malignant skin neoplasms and thus may contribute to an improvement in the early detection of skin cancer.

Incidence, Mortality, and Trends of Nonmelanoma Skin Cancer in Germany

Increasing incidence rates (IRs) of nonmelanoma skin cancer (NMSC) in white populations have been described worldwide. Cancer registry data from the Saarland and Schleswig-Holstein federal states were used to analyze incidence and mortality trends in Germany. Age-standardized rates were compared with crude rates to assess disease burden. Joinpoint regression models were used to estimate annual percentage changes and 95% confidence intervals, allowing us to assess temporal trends between 1970 and 2012. Incidence predictions until 2030 were based on age-period-cohort models and linear extrapolation techniques. In the Saarland federal state, between 1970 and 2012, NMSC age-standardized and crude IRs increased 10- to 22-fold, respectively. In Schleswig-Holstein, between 1999 and 2012 crude IRs doubled, reaching 250 cases/100,000 persons per year in 2012, with age-standardized IRs increasing 1.5-fold. During this period, NMSC mortality remained stable or decreased. For 2030, the predicted age-standardized IRs are as follows: males, 230 cases; females, 180-200 cases. The predicted crude IRs for the same year are males, 450-500 cases; females, 380-430 cases. There is a continuous long-term increase of NMSC incidence with no tendency for leveling off. By 2030, the current NMSC IR in Germany is expected to double.

Skin Cancer Screening in Germany. Documenting Melanoma Incidence and Mortality From 2008 to 2013

BACKGROUND

Nationwide skin cancer screening was introduced in Germany in 2008. The positive results of a pilot project carried out in 2003-4 in the federal state of Schleswig-Holstein had implied that screening would lower the mortality from melanoma.

METHODS

Data on the incidence of invasive malignant melanoma of the skin (MM; ICD-10: C43) were extracted from the databases of the Association of Population-based Cancer Registries in Germany (GEKID) and from the Schleswig-Holstein cancer registry. Mortality rates were extracted from the official cause-of-death statistics.

RESULTS

With the beginning of nationwide screening in 2008, the age-standardized incidence rate in Germany increased by approximately 28% to 18.2 cases per 100 000 persons in 2010. In Schleswig-Holstein, the incidence fell after the pilot project ended and has been comparable to the nationwide incidence since 2008. For Germany overall, there has been no downward trend in MM mortality since the introduction of nationwide screening; in 2013, the mortality rate was 2.3 deaths per 100 000 persons per year. In the area of the pilot study, mortality declined to a level of 1.0/100 000/year until 2008 and then began to rise again. At present, the mortality due to MM in Schleswig- Holstein is once again the same as that in Germany overall (2.4/100 000/year).

CONCLUSION

The introduction of nationwide skin cancer screening in 2008 has not yet led to any measurable decline in mortality due to melanoma. The current method of screening seems to be less thorough than that used in the pilot project; this may explain the absence of a decline in MM-related mortality in Germany overall up to the year 2013, as well as the rising mortality in Schleswig-Holstein since the end of the pilot program. The generation of a robust set of data on how skin cancer screening can be optimized now seems urgently necessary.

Using Active Learning for Speeding up Calibration in Simulation Models

BACKGROUND

Most cancer simulation models include unobservable parameters that determine disease onset and tumor growth. These parameters play an important role in matching key outcomes such as cancer incidence and mortality, and their values are typically estimated via a lengthy calibration procedure, which involves evaluating a large number of combinations of parameter values via simulation. The objective of this study is to demonstrate how machine learning approaches can be used to accelerate the calibration process by reducing the number of parameter combinations that are actually evaluated.

METHODS

Active learning is a popular machine learning method that enables a learning algorithm such as artificial neural networks to interactively choose which parameter combinations to evaluate. We developed an active learning algorithm to expedite the calibration process. Our algorithm determines the parameter combinations that are more likely to produce desired outputs and therefore reduces the number of simulation runs performed during calibration. We demonstrate our method using the previously developed University of Wisconsin breast cancer simulation model (UWBCS).

RESULTS

In a recent study, calibration of the UWBCS required the evaluation of 378 000 input parameter combinations to build a race-specific model, and only 69 of these combinations produced results that closely matched observed data. By using the active learning algorithm in conjunction with standard calibration methods, we identify all 69 parameter combinations by evaluating only 5620 of the 378 000 combinations.

CONCLUSION

Machine learning methods hold potential in guiding model developers in the selection of more promising parameter combinations and hence speeding up the calibration process. Applying our machine learning algorithm to one model shows that evaluating only 1.49% of all parameter combinations would be sufficient for the calibration.

Melanoma mortality following skin cancer screening in Germany

OBJECTIVES

In 2003, a skin cancer screening campaign based on total body skin examination was launched in the federal state of Schleswig-Holstein, Germany. 20% of adults aged 20 and over were screened. In 2008, a 48% decline in melanoma mortality was reported. In the same year, skin screening was extended to the rest of Germany. We evaluated whether melanoma mortality trends decreased in Germany as compared with surrounding countries where skin screening is uncommon. We also evaluated whether the initial decreasing mortality trend observed in Schleswig-Holstein was maintained with a longer follow-up.

SETTING AND PARTICIPANTS

Regional and national melanoma mortality data from 1995 to 2013 were extracted from the GEKID database and the Federal Statistical Office. Mortality data for Germany and surrounding countries from 1980 to 2012 were extracted from the WHO mortality database.

PRIMARY AND SECONDARY OUTCOME MEASURES

Age-adjusted (European Standard Population) mortality rates were computed and joinpoint analysis performed for Schleswig-Holstein, Germany and surrounding countries.

RESULTS

In Schleswig-Holstein, melanoma mortality rates declined by 48% from 2003 to 2008, and from 2009 to 2013 returned to levels observed before screening initiation. During the 5 years of the national programme (2008-2012), melanoma mortality rates increased by 2.6% (95% CI -0.1 to 5.2) in men and 0.02% (95% CI -1.8 to 1.8) in women. No inflexion point in trends was identified after 2008 that could have suggested a decreasing melanoma mortality. Trends of cutaneous melanoma mortality in Germany from 1980 to 2012 did not differ from those observed in surrounding countries.

CONCLUSIONS

The transient decrease mortality in Schleswig-Holstein followed by return to pre-screening levels could reflect a temporal modification in the reporting of death causes. An in-depth evaluation of the screening programme is required.