Artificial intelligence to detect malignant eyelid tumors from photographic images

Zhongwen Li,Xiaomei Yu,Weiwei Xie,Mengjie Pei,Jiewei Jiang,Guohai Wu,Xuefang Wu,Layi Wang,Hongyun Chen,Zhen Li,Wei Qiang

doi:10.1038/s41746-022-00571-3

Abstract

Malignant eyelid tumors can invade adjacent structures and pose a threat to vision and even life. Early identification of malignant eyelid tumors is crucial to avoiding substantial morbidity and mortality. However, differentiating malignant eyelid tumors from benign ones can be challenging for primary care physicians and even some ophthalmologists. Here, based on 1,417 photographic images from 851 patients across three hospitals, we developed an artificial intelligence system using a faster region-based convolutional neural network and deep learning classification networks to automatically locate eyelid tumors and then distinguish between malignant and benign eyelid tumors. The system performed well in both internal and external test sets (AUCs ranged from 0.899 to 0.955). The performance of the system is comparable to that of a senior ophthalmologist, indicating that this system has the potential to be used at the screening stage for promoting the early detection and treatment of malignant eyelid tumors.

Highlights

Eyelid tumors are the most common neoplasm encountered in daily ophthalmology practice[1,2]
After excluding 150 photographic images without histopathological diagnoses, a total of 1,417 images with 1,533 eyelid tumors delineated by tight bounding boxes were used to establish and evaluate an eyelid tumor detection system (ETDS)
Deep learning system versus ophthalmologists For differentiating malignant eyelid tumors from benign ones based on the external test set, the junior ophthalmologist achieved an accuracy of 72.3% with a

Summary

Introduction

Eyelid tumors are the most common neoplasm encountered in daily ophthalmology practice[1,2]. As eyelids have many tissue types, various benign and malignant tumors can develop[3]. Malignant eyelid tumors pose a great threat because of their proximity to the eyeballs, brain, and paranasal sinuses, which may cause cosmetic disfigurement and severe morbidity[4,5]. Recognition and treatment of malignant eyelid tumors can result in the most cosmetically and functionally satisfactory outcomes[4–6]. Melanoma and sebaceous gland carcinoma (SGC) of the eyelid are rare lesions, they have high mortality[7,8]. The estimated 5-year survival rate of these malignant eyelid tumors can be over 99% if they could be detected in their earliest stages (depth of skin invasion ≤0.76 mm)[8]. Early detection of these malignant eyelid tumors is considerably critical

Methods

Results

Conclusion