Computer-Aided Diagnosis of Label-Free 3-D Optical Coherence Microscopy Images of Human Cervical Tissue

Yutao Ma, Tao Xu, Sharon Xiaolei Huang, Xiaofang Wang, Canyu Li, Jason Jerwick, Yuan Ning, Xianxu Zeng, Baojin Wang, Yihong Wang, Zhan Zhang, Xiaoan Zhang, Chao Zhou

Research output: Contribution to journalArticle

Abstract

Objective: Ultrahigh-resolution optical coherence microscopy (OCM) has recently demonstrated its potential for accurate diagnosis of human cervical diseases. One major challenge for clinical adoption, however, is the steep learning curve clinicians need to overcome to interpret OCM images. Developing an intelligent technique for computer-aided diagnosis (CADx) to accurately interpret OCM images will facilitate clinical adoption of the technology and improve patient care. Methods: 497 high-resolution 3-D OCM volumes (600 cross-sectional images each) were collected from 159 ex vivo specimens of 92 female patients. OCM image features were extracted using a convolutional neural network (CNN) model, concatenated with patient information (e.g., age, HPV results), and classified using a support vector machine classifier. Ten-fold cross-validations were utilized to test the performance of the CADx method in a five-class classification task and a binary classification task. Results: An 88.3<formula><tex>$\pm$</tex></formula>4.9% classification accuracy was achieved for five fine-grained classes of cervical tissue, namely normal, ectropion, low-grade and high-grade squamous intraepithelial lesions (LSIL and HSIL), and cancer. In the binary classification task (low-risk [normal, ectropion and LSIL] vs. high-risk [HSIL and cancer]), the CADx method achieved an area-under-the-curve (AUC) value of 0.959 with an 86.7<formula><tex>$\pm$</tex></formula>11.4% sensitivity and 93.5<formula><tex>$\pm$</tex></formula>3.8% specificity. Conclusion: The proposed deep-learning based CADx method outperformed four human experts. It was also able to identify morphological characteristics in OCM images that were consistent with histopathological interpretations. Significance: Label-free OCM imaging, combined with deep-learning based CADx methods, hold a great promise to be used in clinical settings for the effective screening and diagnosis of cervical diseases.

Original languageEnglish (US)
JournalIEEE Transactions on Biomedical Engineering
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Labels
Microscopic examination
Tissue
Computer aided diagnosis
Support vector machines
Screening
Classifiers
Neural networks
Imaging techniques

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering

Cite this

Ma, Yutao ; Xu, Tao ; Huang, Sharon Xiaolei ; Wang, Xiaofang ; Li, Canyu ; Jerwick, Jason ; Ning, Yuan ; Zeng, Xianxu ; Wang, Baojin ; Wang, Yihong ; Zhang, Zhan ; Zhang, Xiaoan ; Zhou, Chao. / Computer-Aided Diagnosis of Label-Free 3-D Optical Coherence Microscopy Images of Human Cervical Tissue. In: IEEE Transactions on Biomedical Engineering. 2018.
@article{7d8db9af68ab45a694a92d1fc9f8804d,
title = "Computer-Aided Diagnosis of Label-Free 3-D Optical Coherence Microscopy Images of Human Cervical Tissue",
abstract = "Objective: Ultrahigh-resolution optical coherence microscopy (OCM) has recently demonstrated its potential for accurate diagnosis of human cervical diseases. One major challenge for clinical adoption, however, is the steep learning curve clinicians need to overcome to interpret OCM images. Developing an intelligent technique for computer-aided diagnosis (CADx) to accurately interpret OCM images will facilitate clinical adoption of the technology and improve patient care. Methods: 497 high-resolution 3-D OCM volumes (600 cross-sectional images each) were collected from 159 ex vivo specimens of 92 female patients. OCM image features were extracted using a convolutional neural network (CNN) model, concatenated with patient information (e.g., age, HPV results), and classified using a support vector machine classifier. Ten-fold cross-validations were utilized to test the performance of the CADx method in a five-class classification task and a binary classification task. Results: An 88.3$\pm$4.9{\%} classification accuracy was achieved for five fine-grained classes of cervical tissue, namely normal, ectropion, low-grade and high-grade squamous intraepithelial lesions (LSIL and HSIL), and cancer. In the binary classification task (low-risk [normal, ectropion and LSIL] vs. high-risk [HSIL and cancer]), the CADx method achieved an area-under-the-curve (AUC) value of 0.959 with an 86.7$\pm$11.4{\%} sensitivity and 93.5$\pm$3.8{\%} specificity. Conclusion: The proposed deep-learning based CADx method outperformed four human experts. It was also able to identify morphological characteristics in OCM images that were consistent with histopathological interpretations. Significance: Label-free OCM imaging, combined with deep-learning based CADx methods, hold a great promise to be used in clinical settings for the effective screening and diagnosis of cervical diseases.",
author = "Yutao Ma and Tao Xu and Huang, {Sharon Xiaolei} and Xiaofang Wang and Canyu Li and Jason Jerwick and Yuan Ning and Xianxu Zeng and Baojin Wang and Yihong Wang and Zhan Zhang and Xiaoan Zhang and Chao Zhou",
year = "2018",
month = "1",
day = "1",
doi = "10.1109/TBME.2018.2890167",
language = "English (US)",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE Computer Society",

}

Computer-Aided Diagnosis of Label-Free 3-D Optical Coherence Microscopy Images of Human Cervical Tissue. / Ma, Yutao; Xu, Tao; Huang, Sharon Xiaolei; Wang, Xiaofang; Li, Canyu; Jerwick, Jason; Ning, Yuan; Zeng, Xianxu; Wang, Baojin; Wang, Yihong; Zhang, Zhan; Zhang, Xiaoan; Zhou, Chao.

In: IEEE Transactions on Biomedical Engineering, 01.01.2018.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Computer-Aided Diagnosis of Label-Free 3-D Optical Coherence Microscopy Images of Human Cervical Tissue

AU - Ma, Yutao

AU - Xu, Tao

AU - Huang, Sharon Xiaolei

AU - Wang, Xiaofang

AU - Li, Canyu

AU - Jerwick, Jason

AU - Ning, Yuan

AU - Zeng, Xianxu

AU - Wang, Baojin

AU - Wang, Yihong

AU - Zhang, Zhan

AU - Zhang, Xiaoan

AU - Zhou, Chao

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Objective: Ultrahigh-resolution optical coherence microscopy (OCM) has recently demonstrated its potential for accurate diagnosis of human cervical diseases. One major challenge for clinical adoption, however, is the steep learning curve clinicians need to overcome to interpret OCM images. Developing an intelligent technique for computer-aided diagnosis (CADx) to accurately interpret OCM images will facilitate clinical adoption of the technology and improve patient care. Methods: 497 high-resolution 3-D OCM volumes (600 cross-sectional images each) were collected from 159 ex vivo specimens of 92 female patients. OCM image features were extracted using a convolutional neural network (CNN) model, concatenated with patient information (e.g., age, HPV results), and classified using a support vector machine classifier. Ten-fold cross-validations were utilized to test the performance of the CADx method in a five-class classification task and a binary classification task. Results: An 88.3$\pm$4.9% classification accuracy was achieved for five fine-grained classes of cervical tissue, namely normal, ectropion, low-grade and high-grade squamous intraepithelial lesions (LSIL and HSIL), and cancer. In the binary classification task (low-risk [normal, ectropion and LSIL] vs. high-risk [HSIL and cancer]), the CADx method achieved an area-under-the-curve (AUC) value of 0.959 with an 86.7$\pm$11.4% sensitivity and 93.5$\pm$3.8% specificity. Conclusion: The proposed deep-learning based CADx method outperformed four human experts. It was also able to identify morphological characteristics in OCM images that were consistent with histopathological interpretations. Significance: Label-free OCM imaging, combined with deep-learning based CADx methods, hold a great promise to be used in clinical settings for the effective screening and diagnosis of cervical diseases.

AB - Objective: Ultrahigh-resolution optical coherence microscopy (OCM) has recently demonstrated its potential for accurate diagnosis of human cervical diseases. One major challenge for clinical adoption, however, is the steep learning curve clinicians need to overcome to interpret OCM images. Developing an intelligent technique for computer-aided diagnosis (CADx) to accurately interpret OCM images will facilitate clinical adoption of the technology and improve patient care. Methods: 497 high-resolution 3-D OCM volumes (600 cross-sectional images each) were collected from 159 ex vivo specimens of 92 female patients. OCM image features were extracted using a convolutional neural network (CNN) model, concatenated with patient information (e.g., age, HPV results), and classified using a support vector machine classifier. Ten-fold cross-validations were utilized to test the performance of the CADx method in a five-class classification task and a binary classification task. Results: An 88.3$\pm$4.9% classification accuracy was achieved for five fine-grained classes of cervical tissue, namely normal, ectropion, low-grade and high-grade squamous intraepithelial lesions (LSIL and HSIL), and cancer. In the binary classification task (low-risk [normal, ectropion and LSIL] vs. high-risk [HSIL and cancer]), the CADx method achieved an area-under-the-curve (AUC) value of 0.959 with an 86.7$\pm$11.4% sensitivity and 93.5$\pm$3.8% specificity. Conclusion: The proposed deep-learning based CADx method outperformed four human experts. It was also able to identify morphological characteristics in OCM images that were consistent with histopathological interpretations. Significance: Label-free OCM imaging, combined with deep-learning based CADx methods, hold a great promise to be used in clinical settings for the effective screening and diagnosis of cervical diseases.

UR - http://www.scopus.com/inward/record.url?scp=85059619299&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85059619299&partnerID=8YFLogxK

U2 - 10.1109/TBME.2018.2890167

DO - 10.1109/TBME.2018.2890167

M3 - Article

C2 - 30605087

AN - SCOPUS:85059619299

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

ER -