Detection of cervical intraepithelial neoplasia by using optical coherence tomography in combination with microscopy

Characteristics and utility of high-resolution optical coherence microscopy images of endocervical canal lesions

OBJECTIVES

To investigate optical coherence microscopy (OCM) imaging features and the application value of these high-resolution images for identifying endocervical canal lesions (ECLs), which is a clinical dilemma in cervical cancer screening programs.

METHODS

In total, 520 OCM images were obtained by scanning the cervical canal lesions with an ultra-high-resolution OCM system (204 specimens from 73 patients). The OCM morphologic characteristics of ECLs were observed and summarized, and then 3 researchers performed a diagnostic test of OCM images of cervical canal lesions. The accuracy, sensitivity, specificity, positive predictive value, negative predictive value, 95% confidence interval of each parameter, and interinvestigator agreement (κ) were calculated.

RESULTS

Normal endocervix, cysts, squamous metaplasia, high-grade squamous intraepithelial lesions involving glands, and invasive carcinoma had distinct OCM characteristics, which correlated well with corresponding H&E histologic sections. The accuracy, sensitivity, and specificity of the 3 researchers were 90.6%, 89.3% (95% CI, 86.5%-91.7%) and 91.6% (95% CI, 89.2%-93.5%), respectively. The positive predictive value was 90.1% (95% CI, 87.3%-92.4%), and the negative predictive value was 90.9% (95% CI, 88.5%-92.9%), with almost perfect agreement (κ = 0.874).

CONCLUSIONS

The application of the OCM system in cervical canal lesions is feasible and could help improve detection of occult ECLs in cervical cancer screening programs. This study lays the foundation for further research on OCM in cervical canal lesions in vivo, which also has a potential impact on projecting pathologic evaluation beyond what is currently possible, perhaps globally.

Cross-Attention Based Multi-Resolution Feature Fusion Model for Self-Supervised Cervical OCT Image Classification

Cervical cancer seriously endangers the health of the female reproductive system and even risks women's life in severe cases. Optical coherence tomography (OCT) is a non-invasive, real-time, high-resolution imaging technology for cervical tissues. However, since the interpretation of cervical OCT images is a knowledge-intensive, time-consuming task, it is tough to acquire a large number of high-quality labeled images quickly, which is a big challenge for supervised learning. In this study, we introduce the vision Transformer (ViT) architecture, which has recently achieved impressive results in natural image analysis, into the classification task of cervical OCT images. Our work aims to develop a computer-aided diagnosis (CADx) approach based on a self-supervised ViT-based model to classify cervical OCT images effectively. We leverage masked autoencoders (MAE) to perform self-supervised pre-training on cervical OCT images, so the proposed classification model has a better transfer learning ability. In the fine-tuning process, the ViT-based classification model extracts multi-scale features from OCT images of different resolutions and fuses them with the cross-attention module. The ten-fold cross-validation results on an OCT image dataset from a multi-center clinical study of 733 patients in China indicate that our model achieved an AUC value of 0.9963 ± 0.0069 with a 95.89 ± 3.30% sensitivity and 98.23 ± 1.36 % specificity, outperforming some state-of-the-art classification models based on Transformers and convolutional neural networks (CNNs) in the binary classification task of detecting high-risk cervical diseases, including high-grade squamous intraepithelial lesion (HSIL) and cervical cancer. Furthermore, our model with the cross-shaped voting strategy achieved a sensitivity of 92.06% and specificity of 95.56% on an external validation dataset containing 288 three-dimensional (3D) OCT volumes from 118 Chinese patients in a different new hospital. This result met or exceeded the average of four medical experts who have used OCT for over one year. In addition to promising classification performance, our model has a remarkable ability to detect and visualize local lesions using the attention map of the standard ViT model, providing good interpretability for gynecologists to locate and diagnose possible cervical diseases.

Characterization of cervical tissue using Mueller matrix polarimetry

The cervix is composed of layers of squamous epithelium and connective tissue. The main component of the cervical connective tissue is collagen, which has specific orientations in different parts of the cervix and provides mechanical strength. Cervical pathologies such as cervical intraepithelial neoplasia (CIN), cancer, pregnancy, and spontaneous preterm birth (sPTB) allow for structural remodeling of both squamous epithelium and connective tissue. Mueller matrix (MM) polarimetry is an optical imaging technique that uses polarized light to characterize the morphologic changes in pathological cervix. In this study, advances in MM polarimetry in characterizing cervical tissue and associated pathologies were reviewed. In particular, the basic structure of the MM polarimeter is described. The interaction of polarized light with cervical tissue in terms of polarimetric parameters such as depolarization and birefringence is discussed. The assessment of cervical pathologies including CIN, cancer, pregnancy, and sPTB with MM polarimetry and the underlying reasons that produce the contrast in optical imaging are outlined. The clinical implementation of MM polarimetry, especially the Müller polarimetry colposcope, is also discussed. Finally, the challenges for MM polarimetry in cervical clinics are also speculated.

Development of a multimodal mobile colposcope for real-time cervical cancer detection

Cervical cancer remains a leading cause of cancer death among women in low-and middle-income countries. Globally, cervical cancer prevention programs are hampered by a lack of resources, infrastructure, and personnel. We describe a multimodal mobile colposcope (MMC) designed to diagnose precancerous cervical lesions at the point-of-care without the need for biopsy. The MMC integrates two complementary imaging systems: 1) a commercially available colposcope and 2) a high speed, high-resolution, fiber-optic microendoscope (HRME). Combining these two image modalities allows, for the first time, the ability to locate suspicious cervical lesions using widefield imaging and then to obtain co-registered high-resolution images across an entire lesion. The MMC overcomes limitations of high-resolution imaging alone; widefield imaging can be used to guide the placement of the high-resolution imaging probe at clinically suspicious regions and co-registered, mosaicked high-resolution images effectively increase the field of view of high-resolution imaging. Representative data collected from patients referred for colposcopy at Barretos Cancer Hospital in Brazil, including 22,800 high resolution images and 9,900 colposcope images, illustrate the ability of the MMC to identify abnormal cervical regions, image suspicious areas with subcellular resolution, and distinguish between high-grade and low-grade dysplasia.

Cervical optical coherence tomography image classification based on contrastive self-supervised texture learning

BACKGROUND

Cervical cancer seriously affects the health of the female reproductive system. Optical coherence tomography (OCT) emerged as a non-invasive, high-resolution imaging technology for cervical disease detection. However, OCT image annotation is knowledge-intensive and time-consuming, which impedes the training process of deep-learning-based classification models.

PURPOSE

This study aims to develop a computer-aided diagnosis (CADx) approach to classifying in-vivo cervical OCT images based on self-supervised learning.

METHODS

In addition to high-level semantic features extracted by a convolutional neural network (CNN), the proposed CADx approach designs a contrastive texture learning (CTL) strategy to leverage unlabeled cervical OCT images' texture features. We conducted ten-fold cross-validation on the OCT image dataset from a multi-center clinical study on 733 patients from China.

RESULTS

In a binary classification task for detecting high-risk diseases, including high-grade squamous intraepithelial lesion and cervical cancer, our method achieved an area-under-the-curve value of 0.9798 ± 0.0157 with a sensitivity of 91.17 ± 4.99% and a specificity of 93.96 ± 4.72% for OCT image patches; also, it outperformed two out of four medical experts on the test set. Furthermore, our method achieved a 91.53% sensitivity and 97.37% specificity on an external validation dataset containing 287 3D OCT volumes from 118 Chinese patients in a new hospital using a cross-shaped threshold voting strategy.

CONCLUSIONS

The proposed contrastive-learning-based CADx method outperformed the end-to-end CNN models and provided better interpretability based on texture features, which holds great potential to be used in the clinical protocol of "see-and-treat." This article is protected by copyright. All rights reserved.

Study on the application and imaging characteristics of optical coherence tomography in vulva lesions

In this study, a prospective study was conducted by using optical coherence tomography (OCT) in the in vivo detection of vulvar diseases. The clinical efficacy of the OCT we investigated in the detection of vulvar diseases, and the characteristics of the OCT images were defined. Overall, this study recruited 63 patients undergoing the colposcopy for vulvar lesions in three Chinese hospitals from December 20th, 2018 and September 24th, 2019. The colposcopy and the OCT examination were performed successively, and the OCT images were compared with the relevant tissue sections to characterize different lesions. The OCT diagnoses where categorized into 7 types, including normal and inflammatory vulva, condyloma acuminata, papilloma, lichen sclerosus, atrophic sclerosing lichen, fibrous epithelial polyp as well as cysts. The structural characteristics of the vulva tissue can be clearly observed in the OCT image, which are consistent with the characteristics of the tissue section. Compared with the pathological results, the sensitivity, specificity and accuracy of the OCT examination reached 83.82% (95% confidence interval, CI 72.5%–91.3%), 57.89% (95% CI 34.0%–78.9%) and 78.16%, respectively. The OCT is found with the advantages of being noninvasive, real-time and sensitive and with high resolution. It is of high significance to screening vulva diseases, and it is expected as one of the methods to clinically diagnose vulva diseases.

Advances in Imaging Modalities, Artificial Intelligence, and Single Cell Biomarker Analysis, and Their Applications in Cytopathology

Several advances in recent decades in digital imaging, artificial intelligence, and multiplex modalities have improved our ability to automatically analyze and interpret imaging data. Imaging technologies such as optical coherence tomography, optical projection tomography, and quantitative phase microscopy allow analysis of tissues and cells in 3-dimensions and with subcellular granularity. Improvements in computer vision and machine learning have made algorithms more successful in automatically identifying important features to diagnose disease. Many new automated multiplex modalities such as antibody barcoding with cleavable DNA (ABCD), single cell analysis for tumor phenotyping (SCANT), fast analytical screening technique fine needle aspiration (FAST-FNA), and portable fluorescence-based image cytometry analyzer (CytoPAN) are under investigation. These have shown great promise in their ability to automatically analyze several biomarkers concurrently with high sensitivity, even in paucicellular samples, lending themselves well as tools in FNA. Not yet widely adopted for clinical use, many have successfully been applied to human samples. Once clinically validated, some of these technologies are poised to change the routine practice of cytopathology.

Near-Histologic Resolution Images of Cervical Dysplasia Obtained With Gabor Domain Optical Coherence Microscopy

OBJECTIVE

Histopathology is the criterion standard for evaluating cervical squamous intraepithelial neoplasia (dysplasia). In this pilot feasibility study, we examined whether a novel 3-dimensional imaging device using Gabor-domain optical coherence microscopy (GDOCM) could distinguish features of cervical dysplasia comparable with histopathology.

METHODS

A prospective observational pilot study enrolled a small sample of women undergoing loop electrosurgical excision procedure for cervical squamous intraepithelial neoplasia. Fresh ex vivo specimens were imaged with the GDOCM device. Digital images were reviewed by a pathologist who was blinded to the histopathology results. Histopathologic features were then compared with the digital observations.

RESULTS

Standard histologic features of cervical squamous epithelium and of squamous intraepithelial neoplasia could be observed in GDOCM images. Cervical epithelium, stroma, basement membrane, and squamous papilla could all be identified. Human papillomavirus effects, such as vacuolization and cellular density, were also observed.

CONCLUSIONS

A GDOCM imaging system has the potential to obtain histologic resolution images of the cervix in the evaluation of squamous intraepithelial neoplasia. This pilot study allowed for optimizing the imaging system and paved the way for a future diagnostic accuracy study. The development of this technology could streamline the evaluation of patients at risk for cervical neoplasia.

Microscope integrated optical coherence tomography of a cerebral arachnoid cyst: A new technique to increase intraoperative security

PURPOSE

This technical note illustrates microscope integrated optical coherence tomography (iOCT) as an imaging technique to delineate concealed micro anatomical structures not displayable by conventional intraoperative imaging methods in the context of a cerebral arachnoid cyst.

METHODS

iOCT was used for the first time to scan a cerebral arachnoid cyst in vivo. Scanning sites were defined at the outer membrane of the arachnoid cyst, the inner membrane at the temporal cortex as well as at the fenestration site to the basal cisterns - a point out of reach and resolution for conventional intraoperative imaging methods like e. g. ultrasound or neuroendoscopy.

RESULTS

iOCT was feasible during microsurgical fenestration of an arachnoid cyst. A clear delineation of the arachnoid cyst membrane was possible. The differentiation of the arachnoid cyst membrane and underlying arachnoid barrier cell membrane was possible. Trans cystic scanning at the temporal cortex could delineate the content of the subarachnoid space like subarachnoid blood vessels, trabecular sytem and vessel wall morphology of a M4 middle cerebral artery branch. Scanning of the inner membrane of the arachnoid cyst at site of fenestration to the basal cisterns excluded underlying micro anatomical structures.

CONCLUSION

This case demonstrates that iOCT achieved to delineate concealed micro anatomical structures which are occult to conventional intraoperative imaging methods. Further studies are necessary to value iOCT as a tool to improve intraoperative security.

Ten Years of Gabor-Domain Optical Coherence Microscopy

Gabor-domain optical coherence microscopy (GDOCM) is a high-definition imaging technique leveraging principles of low-coherence interferometry, liquid lens technology, high-speed imaging, and precision scanning. GDOCM achieves isotropic 2 μm resolution in 3D, effectively breaking the cellular resolution limit of optical coherence tomography (OCT). In the ten years since its introduction, GDOCM has been used for cellular imaging in 3D in a number of clinical applications, including dermatology, oncology and ophthalmology, as well as to characterize materials in industrial applications. Future developments will enhance the structural imaging capability of GDOCM by adding functional modalities, such as fluorescence and elastography, by estimating thicknesses on the nano-scale, and by incorporating machine learning techniques.

First in vivo visualization of the human subarachnoid space and brain cortex via optical coherence tomography

The present work explores optical coherence tomography (OCT) as a suitable in vivo neuroimaging modality of the subarachnoid space (SAS). Patients (n = 26) with frontolateral craniotomy were recruited. The temporal and frontal arachnoid mater and adjacent anatomical structures were scanned using microscope-integrated three-dimensional OCT, (iOCT). Analysis revealed a detailed depiction of the SAS (76.9%) with delineation of the internal microanatomical structures such as the arachnoid barrier cell membrane (ABCM; 96.2%), trabecular system (50.2%), internal blood vessels (96.2%), pia mater (26.9%) and the brain cortex (96.2%). Orthogonal distance measuring was possible. The SAS showed a mean depth of 570 µm frontotemporal. The ABCM showed a mean depth of 74 µm frontotemporal. These results indicate that OCT provides a dynamic, non-invasive tool for real-time imaging of the SAS and adjacent anatomical structures at micrometer spatial resolution. Further studies are necessary to evaluate the value of OCT during microsurgical procedures.

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

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 three-dimensional (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 and human papillomavirus (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 ± 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) versus high-risk (HSIL and cancer)], the CADx method achieved an area-under-the-curve value of 0.959 with an 86.7 ± 11.4% sensitivity and 93.5 ± 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, holds a great promise to be used in clinical settings for the effective screening and diagnosis of cervical diseases.

Ultrahigh-resolution optical coherence microscopy accurately classifies precancerous and cancerous human cervix free of labeling

Cervical cancer remains the fourth most common cause of cancer worldwide and the third leading cause of cancer deaths for women in developing countries. Traditional screening tools, such as human papillomavirus and Pap tests, cannot provide results in real-time and cannot localize suspicious regions. Colposcopy-directed biopsies are invasive in nature and only a few sites of the cervix may be chosen for investigation. A non-invasive, label-free and real-time imaging method with a resolution approaching that of histopathology is desirable for early detection of the disease.

Methods: Ultrahigh-resolution optical coherence microscopy (OCM) is an emerging imaging technique used to obtain 3-dimensional (3-D) “optical biopsies” of biological samples with cellular resolution. In this study, 497 3-D OCM datasets from 159 specimens were collected from 92 patients.

Results: Distinctive patterns for normal cervix, squamocolumnar junction, ectropion, low-grade and high-grade squamous intraepithelial lesions (LSIL and HSIL) and invasive cervical lesions were clearly observed from OCM images, which matched well with corresponding histological slides. OCM images demonstrated a sensitivity of 80% (95% confidence interval, CI, 72%-86%) and a specificity of 89% (95% CI, 84%-93%) for detecting high-risk lesions (HSIL and invasive lesions) when blindly tested by three investigators. A substantial inter-observer agreement was observed (κ=0.627), which showed high diagnostic consistency among three investigators.

Conclusion: These results laid the foundation for future non-invasive optical evaluation of cervical tissue in vivo, which could lead to a less invasive and more effective screening and “see-and-treat” strategy for the management of cervical cancer.

Optical coherence tomography in gynecology: a narrative review

Modern gynecologic practice requires noninvasive diagnostics techniques capable of detecting morphological and functional alterations in tissues of female reproductive organs. Optical coherence tomography (OCT) is a promising tool for providing imaging of biotissues with high resolution at depths up to 2 mm. Design of the customized probes provides wide opportunities for OCT use in gynecology. This paper contains a retrospective insight into the history of OCT employment in gynecology, an overview of the existing gynecologic OCT probes, including those for combination with other diagnostic modalities, and state-of-the-art application of OCT for diagnostics of tumor and nontumor pathologies of female genitalia. Perspectives of OCT both in diagnostics and treatment planning and monitoring in gynecology are overviewed.

Optical techniques for cervical neoplasia detection

This paper provides an overview of the current research in the field of optical techniques for cervical neoplasia detection and covers a wide range of the existing and emerging technologies. Using colposcopy, a visual inspection of the uterine cervix with a colposcope (a binocular microscope with 3- to 15-fold magnification), has proven to be an efficient approach for the detection of invasive cancer. Nevertheless, the development of a reliable and cost-effective technique for the identification of precancerous lesions, confined to the epithelium (cervical intraepithelial neoplasia) still remains a challenging problem. It is known that even at early stages the neoplastic transformations of cervical tissue induce complex changes and modify both structural and biochemical properties of tissues. The different methods, including spectroscopic (diffuse reflectance spectroscopy, induced fluorescence and autofluorescence spectroscopy, Raman spectroscopy) and imaging techniques (confocal microscopy, optical coherence tomography, Mueller matrix imaging polarimetry, photoacoustic imaging), probe different tissue properties that may serve as optical biomarkers for diagnosis. Both the advantages and drawbacks of these techniques for the diagnosis of cervical precancerous lesions are discussed and compared.