Development and validation of an endoscopic images‐based deep learning model for detection with nasopharyngeal malignancies

Abstract

BackgroundDue to the occult anatomic location of the nasopharynx and frequent presence of adenoid hyperplasia, the positive rate for malignancy identification during biopsy is low, thus leading to delayed or missed diagnosis for nasopharyngeal malignancies upon initial attempt. Here, we aimed to develop an artificial intelligence tool to detect nasopharyngeal malignancies under endoscopic examination based on deep learning.MethodsAn endoscopic images-based nasopharyngeal malignancy detection model (eNPM-DM) consisting of a fully convolutional network based on the inception architecture was developed and fine-tuned using separate training and validation sets for both classification and segmentation. Briefly, a total of 28,966 qualified images were collected. Among these images, 27,536 biopsy-proven images from 7951 individuals obtained from January 1st, 2008, to December 31st, 2016, were split into the training, validation and test sets at a ratio of 7:1:2 using simple randomization. Additionally, 1430 images obtained from January 1st, 2017, to March 31st, 2017, were used as a prospective test set to compare the performance of the established model against oncologist evaluation. The dice similarity coefficient (DSC) was used to evaluate the efficiency of eNPM-DM in automatic segmentation of malignant area from the background of nasopharyngeal endoscopic images, by comparing automatic segmentation with manual segmentation performed by the experts.ResultsAll images were histopathologically confirmed, and included 5713 (19.7%) normal control, 19,107 (66.0%) nasopharyngeal carcinoma (NPC), 335 (1.2%) NPC and 3811 (13.2%) benign diseases. The eNPM-DM attained an overall accuracy of 88.7% (95% confidence interval (CI) 87.8%–89.5%) in detecting malignancies in the test set. In the prospective comparison phase, eNPM-DM outperformed the experts: the overall accuracy was 88.0% (95% CI 86.1%–89.6%) vs. 80.5% (95% CI 77.0%–84.0%). The eNPM-DM required less time (40 s vs. 110.0 ± 5.8 min) and exhibited encouraging performance in automatic segmentation of nasopharyngeal malignant area from the background, with an average DSC of 0.78 ± 0.24 and 0.75 ± 0.26 in the test and prospective test sets, respectively.ConclusionsThe eNPM-DM outperformed oncologist evaluation in diagnostic classification of nasopharyngeal mass into benign versus malignant, and realized automatic segmentation of malignant area from the background of nasopharyngeal endoscopic images.

A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis

“…Of the 69 studies, 25 studies did an out-of-sample external validation and were therefore included in a meta-analysis. 21,28,30,34,[36][37][38][39]43,[53][54][55][56]61,[65][66][67]70,73,74,79,81,90,91,99 In line with the aims of this review, all eligible studies were included regardless of the target condition. The meta-analysis therefore included diagnostic classifications in multiple specialty areas, including ophthalmology (six studies), breast cancer (three studies), lung cancer (two studies), dermatological cancer (three studies), trauma and orthopaedics (two studies), respiratory disease (two studies), Of these 25 studies, only 14 used the same sample for the out-of-sample validation to compare performance between deep learning algorithms and health-care professionals, with 31 contingency tables for deep learning algorithm performance and 54 tables for healthcare professionals (figure 3).…”

Artificial intelligence‐based detection of pharyngeal cancer using convolutional neural networks

“… † We use the TNM classification of malignant tumors from the union for international cancer control 22…”

“…Thus, in this study, we demonstrate the diagnostic ability of an AI‐based diagnosing system to detect pharyngeal cancer from EGD images. Although there are two studies about AI diagnosis of pharyngeal cancer, in which otolaryngologists demonstrated the use of AI with images in per‐nasal endoscopy,21,22 no published study has demonstrated the usefulness of AI diagnosis to detect SPC by EGD.…”

Emerging radiotherapy technologies and trends in nasopharyngeal cancer

“…To make this happen, one needs to build convolutional neural networks (CNN), identify the features of the image to be extracted, and finally identify the tumor using deep learning algorithms. Li et al [80] developed an endoscopic image-based deep learning model that had 88% accuracy. This was achieved using 28,966 images from 7951 subjects over 8 years to form the test set for an endoscopic image-based NPC detection model.…”