Deep learning–based endoscopic image recognition for detection of early gastric cancer: a Chinese perspective
2018; Elsevier BV; Volume: 88; Issue: 1 Linguagem: Inglês
10.1016/j.gie.2018.01.029
ISSN1097-6779
AutoresZhijie Wang, Qianqian Meng, Shuling Wang, Zhao‐Shen Li, Yu Bai, Dong Wang,
Tópico(s)Lung Cancer Diagnosis and Treatment
ResumoWe read with great interest the recent article in which Kanesaka et al1Kanesaka T. Lee T.C. Uedo N. et al.Computer-aided diagnosis for identifying and delineating early gastric cancers in magnifying narrow-band imaging.Gastrointest Endosc. 2018; 87: 1339-1344Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar reported a computer-aided system for identifying early gastric cancers (EGC). The diagnostic performance (accuracy of 96.3%) suggests the great potential of computer-aided diagnosis for EGC. This is especially true in countries such as China that have a high incidence of gastric cancer but a low EGC detection rate.2Zong L. Abe M. Seto Y. et al.The challenge of screening for early gastric cancer in China.Lancet. 2016; 388: 2606Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar Recent reports3Chen W. Zheng R. Baade P.D. et al.Cancer statistics in China, 2015.CA Cancer J Clin. 2016; 66: 115-132Crossref PubMed Scopus (14103) Google Scholar have estimated that about 679,100 new cases of gastric cancer were confirmed in China each year and that more than 80% of patients received their diagnoses at an advanced stage with poor prognosis. Thus, computer-aided methods are expected to play an important role in the detection of EGC. However, we regretfully found that most current studies1Kanesaka T. Lee T.C. Uedo N. et al.Computer-aided diagnosis for identifying and delineating early gastric cancers in magnifying narrow-band imaging.Gastrointest Endosc. 2018; 87: 1339-1344Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar, 4Miyaki R. Yoshida S. Tanaka S. et al.A computer system to be used with laser-based endoscopy for quantitative diagnosis of early gastric cancer.J Clin Gastroenterol. 2015; 49: 108-115Crossref PubMed Scopus (51) Google Scholar, 5Liu X. Wang C. Bai J. et al.Hue-texture-embedded region-based model for magnifying endoscopy with narrow-band imaging image segmentation based on visual features.Comput Methods Programs Biomed. 2017; 145: 53-66Crossref PubMed Scopus (5) Google Scholar required high-quality, narrow-band imaging (or laser-based) and magnified images for algorithm training and for diagnosis. Meanwhile, advanced magnifying endoscopes and endoscopists with advanced skills in early detection are not always available in many countries,6Chen P.J. Lin M.C. Lai M.J. et al.Accurate classification of diminutive colorectal polyps using computer-aided analysis.Gastroenterology. 2017; 154: 568-575Abstract Full Text Full Text PDF PubMed Scopus (227) Google Scholar including China. Given these problems, we believe that training with an algorithm using white light and nonmagnifying images is especially important for China and other countries with limited access to advanced imaging endoscopes. Human eyes may hardly capture minute lesions during initial nonmagnified gastroscopy, but this problem can potentially be solved with the use of computer vision technologies. Other than traditional image recognition methods, deep learning could enable computer to learn adaptive image features directly from the data sets. The larger data sets that are given, the better performance may be achieved with algorithms. Then China's huge number of patients with gastric cancer (that means mass image data), in a sense, is of great advantage. The government has made artificial intelligence a national strategy since 2017, and we are confident that deep learning–based endoscopic image recognition would make a big difference in China for screening of EGC, as it has done in skin cancer and diabetic retinopathy.7Esteva A. Kuprel B. Novoa R.A. et al.Dermatologist-level classification of skin cancer with deep neural networks.Nature. 2017; 542: 115-118Crossref PubMed Scopus (54) Google Scholar, 8Wong T.Y. Bressler N.M. Artificial intelligence with deep learning technology looks into diabetic retinopathy screening.JAMA. 2016; 316: 2366-2367Crossref PubMed Scopus (157) Google Scholar Computer-aided diagnosis for identifying and delineating early gastric cancers in magnifying narrow-band imagingGastrointestinal EndoscopyVol. 87Issue 5PreviewMagnifying narrow-band imaging (M-NBI) is important in the diagnosis of early gastric cancers (EGCs) but requires expertise to master. We developed a computer-aided diagnosis (CADx) system to assist endoscopists in identifying and delineating EGCs. Full-Text PDF Response:Gastrointestinal EndoscopyVol. 88Issue 1PreviewWe thank Dr Zhijie Wang1 for his interest in and comments to our work.2 We appreciate and concur with his kind comments. Magnifying narrow-band imaging (M-NBI) has been shown to enable the evaluation of microsurface and microvasculature of the GI mucosa and has significantly improved the diagnostic accuracy3; however, the interpretation of M-NBI images is usually difficult for beginners. In 2010, Dr Tsung-Chun Lee attended Dr Noriya Uedo's presentation about M-NBI diagnosis in gastric mucosal lesions at the Asian Pacific Digestive Week. Full-Text PDF
Referência(s)