Develop a Prediction Model for Nonmelanoma Skin Cancer Using Deep Learning in EHR Data

  • Chih-Wei Huang
  • Alex P. A. Nguyen
  • Chieh-Chen Wu
  • Hsuan-Chia Yang
  • Yu-Chuan (Jack) LiEmail author
Part of the Studies in Computational Intelligence book series (SCI, volume 899)


We aimed to develop deep learning models for the prediction of the risk of advanced nonmelanoma skin cancer (NMSC) in Taiwanese adults. We collected the data of 9494 patients from Taiwan National Health Insurance data claim from 1999 to 2013. All patients’ diseases and medications were included in the development of the convolution neural network (CNN) model. We used the 3-year medical data of all patients before the diagnosed NMSC as the dimensional time in the model. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were computed to measure the performance of the model. The results showed the mean (SD) of AUC of the model was 0.894 (0.007). The performance of the model observed with the sensitivity of 0.83, specificity of 0.82, and 0.57 for PPV value. Our study utilized CNN to develop a prediction model for NMSC, based on non-image and multi-dimensional medical records.



This research is sponsored in part by Ministry of Science and Technology (MOST) under grant MOST 107-2634-F-038-002, MOST 106-2634-F-038-001-CC2; The Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan (TMU DP2-107-21121-01-A-06); MOST 108-2410-H-038-010-SSS; Taipei Medical University under grant TMU107-AE1-B18.

We would like to thank Mr. Chia-Wei Liang for his assistant in this study.


  1. 1.
    Apalla, Z., Lallas, A., Sotiriou, E., et al.: Epidemiological trends in skin cancer. Dermatol. Pract. Concept. 7, 1–6 (2017)CrossRefGoogle Scholar
  2. 2.
    Lecun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521, 436 (2015)CrossRefGoogle Scholar
  3. 3.
    Loh, T.Y., Ortiz, A., Goldenberg, A., et al.: Prevalence and clinical characteristics of nonmelanoma skin cancers among hispanic and asian patients compared with white patients in the united states: a 5-year, single-institution retrospective review. Dermatol. Surg. 42, 639–645 (2016). Official publication for American Society for Dermatologic Surgery [et al.]CrossRefGoogle Scholar
  4. 4.
    Lomas, A., Leonardi-Bee, J., Bath-Hextall, F.: A systematic review of worldwide incidence of nonmelanoma skin cancer. Br. J. Dermatol. 166, 1069–1080 (2012)CrossRefGoogle Scholar
  5. 5.
    Nguyen, P.A., Jack Li, Y.C.: Artificial intelligence in clinical implications. Comput. Methods Programs Biomed. 166, A1 (2018)CrossRefGoogle Scholar
  6. 6.
    Nguyen, P.A., Syed-Abdul, S., Iqbal, U., et al.: A probabilistic model for reducing medication errors. PLoS ONE 8, e82401 (2013)CrossRefGoogle Scholar
  7. 7.
    Olsen, C.M., Neale, R.E., Green, A.C., et al.: Independent validation of six melanoma risk prediction models. J. Invest. Dermatol. 135, 1377–1384 (2015)CrossRefGoogle Scholar
  8. 8.
    Sng, J., Koh, D., Siong, W.C., et al.: Skin cancer trends among Asians living in Singapore from 1968 to 2006. J. Am. Acad. Dermatol. 61, 426–432 (2009)CrossRefGoogle Scholar
  9. 9.
    Vuong, K., Armstrong, B.K., Weiderpass, E., et al.: Development and external validation of a melanoma risk prediction model based on self-assessed risk factors. JAMA Dermatol. 152, 889–896 (2016)CrossRefGoogle Scholar
  10. 10.
    Vuong, K., Mcgeechan, K., Armstrong, B.K., et al.: Risk prediction models for incident primary cutaneous melanoma: a systematic review. JAMA Dermatol. 150, 434–444 (2014)CrossRefGoogle Scholar

Copyright information

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021

Authors and Affiliations

  • Chih-Wei Huang
    • 1
  • Alex P. A. Nguyen
    • 1
  • Chieh-Chen Wu
    • 1
  • Hsuan-Chia Yang
    • 1
  • Yu-Chuan (Jack) Li
    • 1
    Email author
  1. 1.International Center for Health Information Technology, College of Medical Science and TechnologyTaipei Medical UniversityTaipeiTaiwan

Personalised recommendations