Next Article in Journal
The YAP/TEAD Axis as a New Therapeutic Target in Osteosarcoma: Effect of Verteporfin and CA3 on Primary Tumor Growth
Next Article in Special Issue
Transfer Learning in Breast Cancer Diagnoses via Ultrasound Imaging
Previous Article in Journal
G1 Cell Cycle Arrest and Extrinsic Apoptotic Mechanisms Underlying the Anti-Leukemic Activity of CDK7 Inhibitor BS-181
Previous Article in Special Issue
Machine Learning Algorithms to Predict Recurrence within 10 Years after Breast Cancer Surgery: A Prospective Cohort Study

Deep Learning Prediction of Cancer Prevalence from Satellite Imagery

Laboratory of Artificial Intelligence in Medicine and Biomedical Physics, Stanford University School of Medicine, Stanford, CA 94304, USA
Radiation Oncology Department, Hôpital Européen Georges Pompidou, Assistance Publique–Hôpitaux de Paris, 75015 Paris, France
Authors to whom correspondence should be addressed.
Cancers 2020, 12(12), 3844;
Received: 19 October 2020 / Revised: 15 December 2020 / Accepted: 16 December 2020 / Published: 19 December 2020
(This article belongs to the Special Issue Machine Learning Techniques in Cancer)
Cancer prevalence estimates are used to guide policymaking, from prevention to screening programs. However, these data are only available for 28% of the U.S. population. We used deep learning to analyze satellite imagery in order to predict cancer prevalence with a high spatial resolution. This method explained up to 64.37% of the variation of cancer prevalence. It could potentially be used to map cancer prevalence in entire regions for which these estimates are currently unavailable.
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. View Full-Text
Keywords: cancer; epidemiology; deep learning cancer; epidemiology; deep learning
Show Figures

Graphical abstract

MDPI and ACS Style

Bibault, J.-E.; Bassenne, M.; Ren, H.; Xing, L. Deep Learning Prediction of Cancer Prevalence from Satellite Imagery. Cancers 2020, 12, 3844.

AMA Style

Bibault J-E, Bassenne M, Ren H, Xing L. Deep Learning Prediction of Cancer Prevalence from Satellite Imagery. Cancers. 2020; 12(12):3844.

Chicago/Turabian Style

Bibault, Jean-Emmanuel, Maxime Bassenne, Hongyi Ren, and Lei Xing. 2020. "Deep Learning Prediction of Cancer Prevalence from Satellite Imagery" Cancers 12, no. 12: 3844.

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

Back to TopTop