SVM Based Automatic Classification of Human Stomach Cancer with Optical Coherence Tomography Images

Site Luo,Yingwei Fan,Hao Liu,Xin An,Huikai Xie,Peng Li,Hongen Liao,Li Huo

DOI: https://doi.org/10.1364/cleo_at.2018.jtu2a.99

2018-01-01

Abstract:Classification is carried out by SVM and human with 1539 normal and 1567 cancerous OCT images. Accuracy, sensitivity and specificity of human are 82.8%, 73.9%, 93.5%, and that of SVM are 91.3%, 95.3%, 87.2%.

What problem does this paper attempt to address?

Classification of Human Stomach Cancer Using Morphological Feature Analysis from Optical Coherence Tomography Images

Site Luo,Yingwei Fan,Wei Chang,Hongen Liao,Hongxiang Kang,Li Huo

DOI: https://doi.org/10.1088/1612-202x/ab3638

IF: 1.704

2019-01-01

Laser Physics Letters

Abstract:Optical coherence tomography is radiation-free, and it is considered a tool of optical biopsy. Classification of normal and cancerous tissues is very important for the guidance of surgeons. Here, we develop the morphological feature analysis-based classification (MFAC) method, combining it with machine learning to identify cancerous tissues. We extract five quantitative morphological features from one OCT image through the structured analysis. Five classifiers are involved to make a classification: the support vector machine, the K-nearest neighbor, the random forest, logic regression, and the conventional threshold method. Sensitivity, specificity, and accuracy are used to evaluate these classifiers and are compared with each other. We launched the experimental research of the imaging of ex vivo patients’ stomach cancerous tissue with the OCT system. The results showed the three additional features specially designed for stomach cancer are remarkably better than the traditional image feature. The best feature demonstrated over 95% accuracy under all five classifiers. The designed feature based on the layer structure of the stomach tissue is significantly effective. This MFAC method will be used to image the in vivo tissue in clinical applications in the future.
Classification of in Vivo Autofluorescence Spectra Using Support Vector Machines

WM Lin,X Yuan,P Yuen,WI Wei,J Sham,PC Shi,J Qu

DOI: https://doi.org/10.1117/1.1628244

IF: 3.758

2004-01-01

Journal of Biomedical Optics

Abstract:An algorithm based on support vector machines (SVM), the most recent advance in pattern recognition, is presented for use in classifying light-induced autofluorescence collected from cancerous and normal tissues. The in vivo autofluorescence spectra used for development and evaluation of SVM diagnostic algorithms were measured from 85 nasopharyngeal carcinoma (NPC) lesions and 131 normal tissue sites from 59 subjects during routine nasal endoscopy. Leave-one-out cross-validation was used to evaluate the performance of the algorithms. An overall diagnostic accuracy of 96%, a sensitivity of 94%, and a specificity of 97% for discriminating nasopharyngeal carcinomas from normal tissues were achieved using a linear SVM algorithm. A diagnostic accuracy of 98%, a sensitivity of 95%, and a specificity of 99% for detecting NPC were achieved with a nonlinear SVM algorithm. In a comparison with previously developed algorithms using the same dataset and the principal component analysis (PCA) technique, the SVM algorithms produced better diagnostic accuracy in all instances. In addition, we investigated a method combining PCA and SVM techniques for reducing the complexity of the SVM algorithms.
Classification of Hepatic Tissues from CT Images Based on Texture Features and Multiclass Support Vector Machines

Luyao Wang,Zhi Zhang,Jingjing Liu,Bo Jiang,Xiyao Duan,Qingguo Xie,Daoyu Hu,Zhen Li

DOI: https://doi.org/10.1007/978-3-642-01510-6_43

2009-01-01

Abstract:A computer-aided diagnosis (CAD) of X-ray Computed Tomography (CT) liver images with contrast agent injection is presented. Regions of interests (ROIs) on CT liver images are defined by experienced radiologists. For each ROI, texture features based on first order statistics (FOS), spatial gray level dependence matrix (SGLDM), gray level run length matrix (GLRLM) and gray level difference matrix (GLDM) are extracted. Support vector machine (SVM) is originally for binary classification. In order to classify hepatic tissues from CT images into primary hepatic carcinoma, hemangioma and normal liver, we utilize two methods to construct multiclass SVMs: one-against-all (OAA), one-against-one (OAO) and compare their performance. The result shows that a total accuracy rate of 97.78% is obtained with the multiclass SVM using the OAO method. Our study has some practical significance for clinical diagnosis.
Classification of Oral Salivary Gland Tumors Based on Texture Features in Optical Coherence Tomography Images

Yang Zihan,Shang Jianwei,Liu Chenlu,Zhang Jun,Liang Yanmei

DOI: https://doi.org/10.1007/s10103-021-03365-3

IF: 1.2

2022-01-01

Laser Physics

Abstract:Currently, the diagnoses of oral diseases primarily depend on the visual recognition of experienced clinicians. It has been proven that automatic recognition based on images can support clinical decision-making by extracting and analyzing objective hidden information. In recent years, optical coherence tomography (OCT) has become a powerful optical imaging technique with the advantages of high resolution and non-invasion. In our study, a dataset composed of four kinds of oral salivary gland tumors (SGTs) was obtained from a homemade swept-source OCT, including two benign and two malignant tumors. Seventy-six texture features were extracted from OCT images to create computational models of diseases. It was demonstrated that the artificial neural network (ANN) based on principal component analysis (PCA) can obtain high diagnostic sensitivity and specificity (higher than 99%) for these four kinds of tumors. The classification accuracy of each tumor is larger than 99%. In addition, the performances of two classifiers (ANN and support vector machine) were quantitatively evaluated based on SGTs. It was proven that the texture features in OCT images provided objective information to classify oral tumors.
Classification of Gastric Cancerous Tissues by a Residual Network Based on Optical Coherence Tomography Images

Luo Site,Ran Yuchen,Liu Lifei,Huang Huihui,Tang Xiaoying,Fan Yingwei

DOI: https://doi.org/10.1007/s10103-022-03546-8

2022-01-01

Lasers in Medical Science

Abstract:Optical coherence tomography (OCT) is a noninvasive, radiation-free, and high-resolution imaging technology. The intraoperative classification of normal and cancerous tissue is critical for surgeons to guide surgical operations. Accurate classification of gastric cancerous OCT images is beneficial to improve the effect of surgical treatment based on the deep learning method. The OCT system was used to collect images of cancerous tissues removed from patients. An intelligent classification method of gastric cancerous tissues based on the residual network is proposed in this study and optimized with the ResNet18 model. Four residual blocks are used to reset the model structure of ResNet18 and reduce the number of network layers to identify cancerous tissues. The model performance of different residual networks is evaluated by accuracy, precision, recall, specificity, F1 value, ROC curve, and model parameters. The classification accuracies of the proposed method and ResNet18 both reach 99.90%. Also, the model parameters of the proposed method are 44% of ResNet18, which occupies fewer system resources and is more efficient. In this study, the proposed deep learning method was used to automatically recognize OCT images of gastric cancerous tissue. This artificial intelligence method could help promote the clinical application of gastric cancerous tissue classification in the future.
Visualization and Tissue Classification of Human Breast Cancer Images Using Ultrahigh-Resolution Oct

Xinwen Yao,Yu Gan,Ernest Chang,Hanina Hibshoosh,Sheldon Feldman,Christine Hendon

DOI: https://doi.org/10.1002/lsm.22654

2017-01-01

Lasers in Surgery and Medicine

Abstract:Background and ObjectiveBreast cancer is one of the most common cancers, and recognized as the third leading cause of mortality in women. Optical coherence tomography (OCT) enables three dimensional visualization of biological tissue with micrometer level resolution at high speed, and can play an important role in early diagnosis and treatment guidance of breast cancer. In particular, ultra-high resolution (UHR) OCT provides images with better histological correlation. This paper compared UHR OCT performance with standard OCT in breast cancer imaging qualitatively and quantitatively. Automatic tissue classification algorithms were used to automatically detect invasive ductal carcinoma in ex vivo human breast tissue.Study Design/Materials and MethodsHuman breast tissues, including non-neoplastic/normal tissues from breast reduction and tumor samples from mastectomy specimens, were excised from patients at Columbia University Medical Center. The tissue specimens were imaged by two spectral domain OCT systems at different wavelengths: a home-built ultra-high resolution (UHR) OCT system at 800nm (measured as 2.72m axial and 5.52m lateral) and a commercial OCT system at 1,300nm with standard resolution (measured as 6.5m axial and 15m lateral), and their imaging performances were analyzed qualitatively. Using regional features derived from OCT images produced by the two systems, we developed an automated classification algorithm based on relevance vector machine (RVM) to differentiate hollow-structured adipose tissue against solid tissue. We further developed B-scan based features for RVM to classify invasive ductal carcinoma (IDC) against normal fibrous stroma tissue among OCT datasets produced by the two systems. For adipose classification, 32 UHR OCT B-scans from 9 normal specimens, and 28 standard OCT B-scans from 6 normal and 4 IDC specimens were employed. For IDC classification, 152 UHR OCT B-scans from 6 normal and 13 IDC specimens, and 104 standard OCT B-scans from 5 normal and 8 IDC specimens were employed.ResultsWe have demonstrated that UHR OCT images can produce images with better feature delineation compared with images produced by 1,300nm OCT system. UHR OCT images of a variety of tissue types found in human breast tissue were presented. With a limited number of datasets, we showed that both OCT systems can achieve a good accuracy in identifying adipose tissue. Classification in UHR OCT images achieved higher sensitivity (94%) and specificity (93%) of adipose tissue than the sensitivity (91%) and specificity (76%) in 1,300nm OCT images. In IDC classification, similarly, we achieved better results with UHR OCT images, featured an overall accuracy of 84%, sensitivity of 89% and specificity of 71% in this preliminary study.ConclusionIn this study, we provided UHR OCT images of different normal and malignant breast tissue types, and qualitatively and quantitatively studied the texture and optical features from OCT images of human breast tissue at different resolutions. We developed an automated approach to differentiate adipose tissue, fibrous stroma, and IDC within human breast tissues. Our work may open the door toward automatic intraoperative OCT evaluation of early-stage breast cancer. (C) 2017 Wiley Periodicals, Inc.
Identifying gastric cancer by optical coherence tomography based on an optimized depth-resolved attenuation estimation

Chao Zhao,Yukang Shu,Lin Cui,Xiang He,Weizheng Mao,Xiaofeng Shi,Hang Zhao,Jun Ma

DOI: https://doi.org/10.1088/2040-8986/ad5456

IF: 2.1

2024-06-07

Journal of Optics

Abstract:Optical Coherence Tomography (OCT) is a reliable technique for cancer detection, with the potential to improve accuracy in identifying cancerous tissue through effective use of OCT image data. We proposed an optimized depth resolution estimation (Opt-DRE) based optical attenuation coefficient (OAC) mapping method to reduce the error effect caused by discretization without considering whether the light is completely attenuated or not. The attenuation coefficient maps of gastric tissue were reconstructed using this method. Features were extracted from these maps using gray-level co-occurrence matrix (GLCM) and gray histogram (GH), and a support vector machine (SVM) was employed as a classifier for identifying gastric cancer tissues. The recognition accuracy achieved was 98.60%, which was higher than the 94.30% accuracy in the control group without OACs. This approach enhances the utilization of OCT image information and improves its practical application ability by increasing the recognition accuracy.

optics
Gastric and Colon Cancer Imaging with Swept Source Optical Coherence Tomography

Site Luo,Hao Liu,Hui Zhao,Xin An,Huikai Xie,Li Huo

DOI: https://doi.org/10.1109/cleopr.2017.8118726

2017-01-01

Abstract:Swept source Optical coherence tomography (SS-OCT) is used to image the cancer samples of 12 gastric and colon cancer patients, including 5 gastric cancer and 7 colon cancer patients. The OCT images and histopathology sections show obvious difference between the normal tissue and cancer tissue. The regular layered structure can be seen at OCT images of normal tissue, while the disordered textures and destroyed layers' architectures preserve as the common features of cancer tissues' images. The variation of texture matched between the OCT images and ordinal pathological sections.
Identification of Oral Squamous Cell Carcinoma in Optical Coherence Tomography Images Based on Texture Features

Zihan Yang,Jianwei Shang,Chenlu Liu,Jun Zhang,Yanmei Liang

DOI: https://doi.org/10.1142/s1793545821400010

IF: 2.396

2021-01-01

Journal of Innovative Optical Health Sciences

Abstract:Surgical excision is an effective treatment for oral squamous cell carcinoma (OSCC), but exact intraoperative differentiation OSCC from the normal tissue is the first premise. As a noninvasive imaging technique, optical coherence tomography (OCT) has the nearly same resolution as the histopathological examination, whose images contain rich information to assist surgeons to make clinical decisions. We extracted kinds of texture features from OCT images obtained by a home-made swept-source OCT system in this paper, and studied the identification of OSCC based on different combinations of texture features and machine learning classifiers. It was demonstrated that different combinations had different accuracies, among which the combination of texture features, gray level co-occurrence matrix (GLCM), Laws' texture measures (LM), and center symmetric auto-correlation (CSAC), and SVM as the classifier, had the optimal comprehensive identification effect, whose accuracy was 94.1%. It was proven that it is feasible to distinguish OSCC based on texture features in OCT images, and it has great potential in helping surgeons make rapid and accurate decisions in oral clinical practice.
Boundary Region of Stomach Mucinous Carcinoma with Swept Source Optical Coherence Tomography

Luo Site,Fan Yingwei,Chang Wei,An Xin,Zhao Hui,Huo Li

DOI: https://doi.org/10.3788/aos201838.0517001

2018-01-01

Acta Optica Sinica

Abstract:To provide a real-time intraoperative feedback and suggest an applied potential to guide stomach cancer surgery for removal of normal tissue as little as possible, we use an optical coherence tomography (OCT) to image normal tissue and tumor tissue of 39 stomach cancerous patients, and compare OCT images and pathological sections. The results show that OCT images of 38 patients show obvious difference between normal tissue and tumor tissue. For the 1 stomach mucinous careinomas, the mucus secreted by tumor of stomach mucinous careinoma destroys the layer structure, texture, and mucosa of stomach tissue. The poorer backscattering of mucus and the greater difference between backscattering of mucosal tissue make the OCT images of stomach mucinous careinoma clearer. The B-mode ultrasonic scanning (B-scan) OCT images can distinguish the boundary of the cancerous tissue and the normal tissue. A volumetric image can be built with B-scan images by the three-dimensional (3D) reconstruction method. The layered section images in tissue-depth direction are obtained after volumetric surface flattening. These images provide another dimensionality to view stomach cancerous boundary. The cross validation with B-scan images can show excellent distinction.
Deep-Learning-Based Automated Identification and Visualization of Oral Cancer in Optical Coherence Tomography Images

Zihan Yang,Hongming Pan,Jianwei Shang,Jun Zhang,Yanmei Liang

DOI: https://doi.org/10.3390/biomedicines11030802

IF: 4.757

2023-03-07

Biomedicines

Abstract:Early detection and diagnosis of oral cancer are critical for a better prognosis, but accurate and automatic identification is difficult using the available technologies. Optical coherence tomography (OCT) can be used as diagnostic aid due to the advantages of high resolution and non-invasion. We aim to evaluate deep-learning-based algorithms for OCT images to assist clinicians in oral cancer screening and diagnosis. An OCT data set was first established, including normal mucosa, precancerous lesion, and oral squamous cell carcinoma. Then, three kinds of convolutional neural networks (CNNs) were trained and evaluated by using four metrics (accuracy, precision, sensitivity, and specificity). Moreover, the CNN-based methods were compared against machine learning approaches through the same dataset. The results show the performance of CNNs, with a classification accuracy of up to 96.76%, is better than the machine-learning-based method with an accuracy of 92.52%. Moreover, visualization of lesions in OCT images was performed and the rationality and interpretability of the model for distinguishing different oral tissues were evaluated. It is proved that the automatic identification algorithm of OCT images based on deep learning has the potential to provide decision support for the effective screening and diagnosis of oral cancer.

biochemistry & molecular biology,medicine, research & experimental,pharmacology & pharmacy
Identification of Oral Precancerous and Cancerous Tissue by Swept Source Optical Coherence Tomography.

Zihan Yang,Jianwei Shang,Chenlu Liu,Jun Zhang,Yanmei Liang

DOI: https://doi.org/10.1002/lsm.23461

2021-01-01

Lasers in Surgery and Medicine

Abstract:BACKGROUND AND OBJECTIVES:Distinguishing cancer from precancerous lesions is critical and challenging in oral medicine. As a noninvasive method, optical coherence tomography (OCT) has the advantages of real-time, in vivo, and large-depth imaging. Texture information hidden in OCT images can provide an important auxiliary effect for improving diagnostic accuracy. The aim of this study is to explore a reliable and accurate OCT-based method for the screening and diagnosis of human oral diseases, especially oral cancer.MATERIALS AND METHODS:Fresh ex vivo oral tissues including normal mucosa, leukoplakia with epithelial hyperplasia (LEH), and oral squamous cell carcinoma (OSCC) were imaged intraoperatively by a homemade OCT system, and 58 texture features were extracted to create computational models of these tissues. A principal component analysis algorithm was employed to optimize the combination of texture feature vectors. The identification based on artificial neural network (ANN) was proposed and the sensitivity/specificity was calculated statistically to evaluate the classification performance.RESULTS:A total of 71 sites of three types of oral tissues were measured, and 5176 OCT images of three types of oral tissues were used in this study. The superior classification result based on ANN was obtained with an average accuracy of 98.17%. The sensitivity and specificity of normal mucosa, LEH, and OSCC are 98.17% / 98.38%, 93.81% / 98.54%, and 98.11% / 99.04%, respectively.CONCLUSION:It is demonstrated from the high accuracies, sensitivities, and specificities that texture-based analysis can be used to identify oral precancerous and cancerous tissue in OCT images, and it has the potential to help surgeons in diseases screening and diagnosis effectively.
Human colorectal cancer tissue assessment using optical coherence tomography catheter and deep learning

Hongbo Luo,Shuying Li,Yifeng Zeng,Hassam Cheema,Ebunoluwa Otegbeye,Safee Ahmed,William C. Chapman,Matthew Mutch,Chao Zhou,Quing Zhu

DOI: https://doi.org/10.1002/jbio.202100349

IF: 3.3899

2022-02-28

Journal of Biophotonics

Abstract:Optical coherence tomography (OCT) can differentiate normal colonic mucosa from neoplasia, potentially offering a new mechanism of endoscopic tissue assessment and biopsy targeting, with a high optical resolution and an imaging depth of ~1 mm. Recent advances in convolutional neural networks (CNN) have enabled application in ophthalmology, cardiology, and gastroenterology malignancy detection with high sensitivity and specificity. Here we describe a miniaturized OCT catheter and a residual neural network (ResNet)-based deep learning model manufactured and trained to perform automatic image processing and real-time diagnosis of the OCT images. The OCT catheter has an outer diameter of 3.8 mm, a lateral resolution of ~7 um, and an axial resolution of ~6 um. A customized ResNet is utilized to classify OCT catheter colorectal images. An area under the receiver operating characteristic (ROC) curve (AUC) of 0.975 is achieved to distinguish between normal and cancerous colorectal tissue images.This article is protected by copyright. All rights reserved.

optics,biochemical research methods,biophysics
Convolutional neural network (CNN) classification of breast cancer in optical coherence tomography (OCT) images

Diana Mojahed,Peter Chang,Yu Gan,Xinwen Yao,Brigid Angelini,Hanina Hibshoosh,Richard Ha,Christine P. Hendon

DOI: https://doi.org/10.1117/12.2504966

2019-01-01

Abstract:The purpose of this study was to develop and evaluate the performance of a convolutional neural network (CNN) that uses a novel A-line based classification approach to detect cancer in OCT images of breast specimens. Deep learning algorithms have been developed for OCT ophthalmology applications using pixel-based classification approaches. In this study, a novel deep learning approach was developed that classifies OCT A-lines of breast tissue. De-identified human breast tissues from mastectomy and breast reduction specimens were excised from patients at Columbia University Medical Center. A total of 82 specimens from 49 patients were imaged with OCT, including both normal tissues and non-neoplastic tissues. The proposed algorithm utilized a hybrid 2D/1D convolutional neural network (CNN) to map each single B-scan to a 1D label vector, which were derived from manual annotation. Each A-line was labelled as one of the following tissue types: ductal carcinoma in situ (DCIS), invasive ductal carcinoma (IDC), adipose, and stroma. Five-fold cross-validation Dice scores across tissue types were: 0.82-0.95 for IDC, 0.54-0.75 for DCIS, 0.67-0.91 for adipose, and 0.61-0.86 for stroma. In a second experiment, IDC and DCIS were combined as a single tissue class (malignancy) while stroma and adipose were combined as a second tissue class (non-malignancy). In this setup, the experiment yielded five-fold cross-validation Dice scores between 0.89-0.93, respectively. Future work includes acquiring more patient samples and to compare the algorithm to previous works, including both deep learning and traditional automatic image processing methods for classification of breast tissue in OCT images.
Application of Hyperosmotic Agent to Determine Gastric Cancer with Optical Coherence Tomography Ex Vivo in Mice.

Honglian Xiong,Zhouyi Guo,Changchun Zeng,Like Wang,Yonghong He,Songhao Liu

DOI: https://doi.org/10.1117/1.3103341

IF: 3.758

2009-01-01

Journal of Biomedical Optics

Abstract:Noninvasive tumor imaging could lead to the early detection and timely treatment of cancer. Optical coherence tomography (OCT) has been reported as an ideal diagnostic tool for distinguishing tumor tissues from normal tissues based on structural imaging. In this study, the capability of OCT for functional imaging of normal and tumor tissues based on time- and depth-resolved quantification of the permeability of biomolecules through these tissues is investigated. The orthotopic graft model of gastric cancer in nude mice is used, normal and tumor tissues from the gastric wall are imaged, and a diffusion of 20% aqueous solution of glucose in normal stomach tissues and gastric tumor tissues is monitored and quantified as a function of time and tissue depth by an OCT system. Our results show that the permeability coefficient is (0.94+/-0.04)x10(-5) cms in stomach tissues and (5.32+/-0.17)x10(-5) cms in tumor tissues, respectively, and that tumor tissues have a higher permeability coefficient compared to normal tissues in optical coherence tomographic images. From the results, it is found that the accurate and sensitive assessment of the permeability coefficients of normal and tumor tissues offers an effective OCT image method for detection of tumor tissues and clinical diagnosis.
Method for Optical Coherence Tomography Image Classification Using Local Features and Earth Mover's Distance.

Yankui Sun,Ming Lei

DOI: https://doi.org/10.1117/1.3251059

IF: 3.758

2009-01-01

Journal of Biomedical Optics

Abstract:Optical coherence tomography (OCT) is a recent imaging method that allows high-resolution, cross-sectional imaging through tissues and materials. Over the past 18 years, OCT has been successfully used in disease diagnosis, biomedical research, material evaluation, and many other domains. As OCT is a recent imaging method, until now surgeons have limited experience using it. In addition, the number of images obtained from the imaging device is too large, so we need an automated method to analyze them. We propose a novel method for automated classification of OCT images based on local features and earth mover's distance (EMD). We evaluated our algorithm using an OCT image set which contains two kinds of skin images, normal skin and nevus flammeus. Experimental results demonstrate the effectiveness of our method, which achieved classification accuracy of 0.97 for an EMD+KNN scheme and 0.99 for an EMD+SVM (support vector machine) scheme, much higher than the previous method. Our approach is especially suitable for nonhomogeneous images and could be applied to a wide range of OCT images.
Machine-Learning Classification of Port Wine Stain with Quantitative Features of Optical Coherence Tomography Image

Shengnan Ai,Chengming Wang,Wenxin Zhang,Wenchao Liao,Juicheng Hsieh,Zhenyu Chen,Bin He,Xiao Zhang,Ning Zhang,Ying Gu,Ping Xue

DOI: https://doi.org/10.1109/jphot.2019.2952903

IF: 2.4

2019-01-01

IEEE Photonics Journal

Abstract:Port wine stain (PWS) is the benign congenital capillary malformation of skin, occurring in 0.3% to 0.5% of the population. In this paper, we build two automated support vector machine (SVM) based classifiers by extracting quantitative features from normal and PWS tissue images recorded by optical coherence tomography (OCT). We use both full feature set and simplified feature set for training. Accuracy of 92.7%, sensitivity of 92.3% and specificity of 93.8% were obtained for classifier with full feature set. Accuracy of 92.7%, sensitivity of 94.9% and specificity of 87.5% were obtained for classifier with simplified feature set. Our results suggest that extracting quantitative features from optical coherence tomographic images could be a potentially powerful method for accurately and automatically identifying PWS margins during laser therapy.
Fully Automated Postlumpectomy Breast Margin Assessment Utilizing Convolutional Neural Network Based Optical Coherence Tomography Image Classification Method

Diana Mojahed,Richard S. Ha,Peter Chang,Yu Gan,Xinwen Yao,Brigid Angelini,Hanina Hibshoosh,Bret Taback,Christine P. Hendon

DOI: https://doi.org/10.1016/j.acra.2019.06.018

IF: 5.482

2020-01-01

Academic Radiology

Abstract:BACKGROUND:The purpose of this study was to develop a deep learning classification approach to distinguish cancerous from noncancerous regions within optical coherence tomography (OCT) images of breast tissue for potential use in an intraoperative setting for margin assessment. METHODS:A custom ultrahigh-resolution OCT (UHR-OCT) system with an axial resolution of 2.7 μm and a lateral resolution of 5.5 μm was used in this study. The algorithm used an A-scan-based classification scheme and the convolutional neural network (CNN) was implemented using an 11-layer architecture consisting of serial 3 × 3 convolution kernels. Four tissue types were classified, including adipose, stroma, ductal carcinoma in situ, and invasive ductal carcinoma. RESULTS:The binary classification of cancer versus noncancer with the proposed CNN achieved 94% accuracy, 96% sensitivity, and 92% specificity. The mean five-fold validation F1 score was highest for invasive ductal carcinoma (mean standard deviation, 0.89 ± 0.09) and adipose (0.79 ± 0.17), followed by stroma (0.74 ± 0.18), and ductal carcinoma in situ (0.65 ± 0.15). CONCLUSION:It is feasible to use CNN based algorithm to accurately distinguish cancerous regions in OCT images. This fully automated method can overcome limitations of manual interpretation including interobserver variability and speed of interpretation and may enable real-time intraoperative margin assessment.
Automatic Detection of Early Gastrointestinal Cancer Lesions Based on Optimal Feature Extraction from Gastroscopic Images

D. Y. Liu,T. Gan,N. N. Rao,G. G. Xu,B. Zeng,H. L. Li

DOI: https://doi.org/10.1166/jmihi.2015.1390

2015-01-01

Journal of Medical Imaging and Health Informatics

Abstract:The conventional upper gastrointestinal endoscopy, the most commonly used method of diagnosing early gastrointestinal cancers, mainly depends on the experience of digestive endoscopists to detect the suspicious lesion. This study developed a computer-aided automatic method to detect lesion regions of early gastrointestinal cancer in gastroscopic images. First, we developed a technique to eliminate the useless regions from a gastroscopic image for enhancing the accuracy of suspicious lesion detection. Then Canny-dilation algorithm was applied to segment the image into smaller regions according to different image features. Next, we combined multiple kinds of features (including color histogram, color moment and rotation invariant local binary patterns) to describe each partitioned region. Therefore the optimal features can be selected according to the respective classification accuracy of each feature between lesions and normal regions. Finally, we detected the early gastrointestinal cancer lesions based on the optimal features in a region and used the clinical data of 407 gastroscopic images to testify the validation of the proposed method. The experimental results indicated the proposed method has the sensitivity, specificity, accuracy and G-mean of 0.93, 0.83, 0.85 and 0.88 in stomach and those of 0.83, 0.88, 0.87 and 0.86 in esophagus, and has low computation complexity. The performance of the proposed method is better than co-occurrence matrix algorithm and color wavelet covariance method treating the same data set. This pilot study shows that our established method has high sensitivity and specificity in the detection of the early gastric cancers and can provide worthy guidance to the further clinical trials.
An SVM Based Approach for the Analysis of Mammography Images

X. Gan,L. Kapsokahvas,A. Skahotis,K. Steinoefel,S. Tangaro

DOI: https://doi.org/10.1063/1.2790139

2007-01-01

AIP Conference Proceedings

Abstract:Mammography is among the most popular imaging techniques used in the diagnosis of breast cancer. Nevertheless distinguishing between healthy and ill images is hard even for an experienced radiologist, because a single image usually includes several regions of interest (ROIs). The hardness of this classification problem along with the substantial amount of data, gathered from patients' medical history, motivates the use of a machine learning approach as part of a CAD (Computer Aided Detection) tool, aiming to assist radiologists in the characterization of mammography images. Specifically, our approach involves: i) the ROI extraction, ii) the Feature Vector extraction, iii) the Support Vector Machine (SVM) classification of ROIs and iv) the characterization of the whole image. We evaluate the performance of our approach in terms of the SVM's training and testing error and in terms of ROI specificity - sensitivity. The results show a relation between the number of features used and the SVM's performance.

SVM Based Automatic Classification of Human Stomach Cancer with Optical Coherence Tomography Images

Classification of Human Stomach Cancer Using Morphological Feature Analysis from Optical Coherence Tomography Images

Classification of in Vivo Autofluorescence Spectra Using Support Vector Machines

Classification of Hepatic Tissues from CT Images Based on Texture Features and Multiclass Support Vector Machines

Classification of Oral Salivary Gland Tumors Based on Texture Features in Optical Coherence Tomography Images

Classification of Gastric Cancerous Tissues by a Residual Network Based on Optical Coherence Tomography Images

Visualization and Tissue Classification of Human Breast Cancer Images Using Ultrahigh-Resolution Oct

Identifying gastric cancer by optical coherence tomography based on an optimized depth-resolved attenuation estimation

Gastric and Colon Cancer Imaging with Swept Source Optical Coherence Tomography

Identification of Oral Squamous Cell Carcinoma in Optical Coherence Tomography Images Based on Texture Features

Boundary Region of Stomach Mucinous Carcinoma with Swept Source Optical Coherence Tomography

Deep-Learning-Based Automated Identification and Visualization of Oral Cancer in Optical Coherence Tomography Images

Identification of Oral Precancerous and Cancerous Tissue by Swept Source Optical Coherence Tomography.

Human colorectal cancer tissue assessment using optical coherence tomography catheter and deep learning

Convolutional neural network (CNN) classification of breast cancer in optical coherence tomography (OCT) images

Application of Hyperosmotic Agent to Determine Gastric Cancer with Optical Coherence Tomography Ex Vivo in Mice.

Method for Optical Coherence Tomography Image Classification Using Local Features and Earth Mover's Distance.

Machine-Learning Classification of Port Wine Stain with Quantitative Features of Optical Coherence Tomography Image

Fully Automated Postlumpectomy Breast Margin Assessment Utilizing Convolutional Neural Network Based Optical Coherence Tomography Image Classification Method

Automatic Detection of Early Gastrointestinal Cancer Lesions Based on Optimal Feature Extraction from Gastroscopic Images

An SVM Based Approach for the Analysis of Mammography Images