Deep Learning Prediction of Cancer Prevalence from Satellite Imagery
Open Access
Abstract
The worldwide growth of cancer incidence can be explained in part by changes in the prevalence and distribution of risk factors. There are geographical gaps in the estimates of cancer prevalence, which could be filled with innovative methods. We used deep learning (DL) features extracted from satellite images to predict cancer prevalence at the census tract level in seven cities in the United States. We trained the model using detailed cancer prevalence estimates from 2018 available in the CDC (Center for Disease Control) 500 Cities project. Data from 3500 census tracts covering 14,483,366 inhabitants were included. Features were extracted from 170,210 satellite images with deep learning. This method explained up to 64.37% (median = 43.53%) of the variation of cancer prevalence. Satellite features are highly correlated with individual socioeconomic and health measures that are linked to cancer prevalence (age, smoking and drinking status, and obesity). A higher similarity between two environments is associated with better generalization of the model (p = 1.10–6). This method can be used to accurately estimate cancer prevalence at a high spatial resolution without using surveys at a fraction of the cost.Keywords
This publication has 17 references indexed in Scilit:
- Combining satellite imagery and machine learning to predict povertyScience, 2016
- Cancer incidence trends using American Community Survey estimates are not consistent with SEER for small populationsCancer Epidemiology, 2016
- The NumPy Array: A Structure for Efficient Numerical ComputationComputing in Science & Engineering, 2011
- Matplotlib: A 2D Graphics EnvironmentComputing in Science & Engineering, 2007
- Environmental and Heritable Factors in the Causation of Cancer — Analyses of Cohorts of Twins from Sweden, Denmark, and FinlandNew England Journal of Medicine, 2000
- Incidence of cancer in the electronics industry: using the new Swedish Cancer Environment Registry as a screening instrument.Occupational and Environmental Medicine, 1983
- A Swedish cancer-environment register available for research.Scandinavian Journal of Work, Environment & Health, 1981
- Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practicesInternational Journal of Cancer, 1975
- Some Neglected Leads to Cancer CausationJNCI: Journal of the National Cancer Institute, 1971
- THE RELATIONSHIP BETWEEN HUMAN SMOKING HABITS AND DEATH RATESJAMA, 1954