Deep Learning Classifier with Patient's Metadata of Dermoscopic Images in Malignant Melanoma Detection

Dina Nur Anggraini Ningrum; Sheng-Po Yuan; Woon-Man Kung; Chieh-Chen Wu; I-Shiang Tzeng; Chu-Ya Huang; Jack Yu-Chuan Li; Yao-Chin Wang

doi:10.2147/JMDH.S306284

Deep Learning Classifier with Patient's Metadata of Dermoscopic Images in Malignant Melanoma Detection

J Multidiscip Healthc. 2021 Apr 21:14:877-885. doi: 10.2147/JMDH.S306284. eCollection 2021.

Authors

Dina Nur Anggraini Ningrum^{1

2}, Sheng-Po Yuan^{1

3}, Woon-Man Kung⁴, Chieh-Chen Wu⁴, I-Shiang Tzeng^{4

5

6}, Chu-Ya Huang⁷, Jack Yu-Chuan Li^#^{1

8

9}, Yao-Chin Wang^#^{10

11}

Affiliations

¹ Graduate Institute of Biomedical Informatics, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan.
² Public Health Department, Universitas Negeri Semarang, Semarang City, Indonesia.
³ Department of Otorhinolaryngology, Shuang-Ho Hospital, Taipei Medical University, New Taipei City, Taiwan.
⁴ Department of Exercise and Health Promotion, College of Kinesiology and Health, Chinese Culture University, Taipei, Taiwan.
⁵ Department of Research, Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation, New Taipei City, Taiwan.
⁶ Department of Statistics, National Taipei University, Taipei, Taiwan.
⁷ Taiwan College of Healthcare Executives, Taipei, Taiwan.
⁸ Department Dermatology, Wan Fang Hospital, Taipei, Taiwan.
⁹ Taipei Medical University Research Center of Cancer Translational Medicine, Taipei, Taiwan.
¹⁰ Graduate Institute of Injury Prevention and Control, College of Public Health, Taipei Medical University, Taipei, Taiwan.
¹¹ Department of Emergency Medicine, Min-Sheng General Hospital, Taoyuan, Taiwan.

^# Contributed equally.

Abstract

Background: Incidence of skin cancer is one of the global burdens of malignancies that increase each year, with melanoma being the deadliest one. Imaging-based automated skin cancer detection still remains challenging owing to variability in the skin lesions and limited standard dataset availability. Recent research indicates the potential of deep convolutional neural networks (CNN) in predicting outcomes from simple as well as highly complicated images. However, its implementation requires high-class computational facility, that is not feasible in low resource and remote areas of health care. There is potential in combining image and patient's metadata, but the study is still lacking.

Objective: We want to develop malignant melanoma detection based on dermoscopic images and patient's metadata using an artificial intelligence (AI) model that will work on low-resource devices.

Methods: We used an open-access dermatology repository of International Skin Imaging Collaboration (ISIC) Archive dataset consist of 23,801 biopsy-proven dermoscopic images. We tested performance for binary classification malignant melanomas vs nonmalignant melanomas. From 1200 sample images, we split the data for training (72%), validation (18%), and testing (10%). We compared CNN with image data only (CNN model) vs CNN for image data combined with an artificial neural network (ANN) for patient's metadata (CNN+ANN model).

Results: The balanced accuracy for CNN+ANN model was higher (92.34%) than the CNN model (73.69%). Combination of the patient's metadata using ANN prevents the overfitting that occurs in the CNN model using dermoscopic images only. This small size (24 MB) of this model made it possible to run on a medium class computer without the need of cloud computing, suitable for deployment on devices with limited resources.

Conclusion: The CNN+ANN model can increase the accuracy of classification in malignant melanoma detection even with limited data and is promising for development as a screening device in remote and low resources health care.

Keywords: artificial neural network; convolutional neural network; embedded artificial intelligence; skin cancer.