[Application of optical coherence tomography in the evaluation of cervical lesions: a multicenter study].

Objective: To explore the value of optical coherence tomography (OCT) imaging system in evaluating cervical lesions in vivo. Methods: A total of 1 214 patients with cervical lesions were collected from January 2020 to December 2021 in the Third Affiliated Hospital of Zhengzhou University, Maternal and Chlid Heaith Hospital of Gushi County, Xinyang City, Henan Province, and Maternal and Chlid Heaith Hospital of Sui County, Shangqiu City, Henan Province. The age of the patients was (38.9±10.5) years (range: 16-77 years). All patients underwent in vivo cervical OCT examination and cervical biopsy pathology examination, and summarized the OCT image features of in vivo cervical lesions. Using the pathological diagnosis as the "gold standard", the accuracy, specificity, sensitivity, positive predictive value (PPV) and negative predictive value (NPV) of OCT image interpretation results were evaluated, as well as the consistency of OCT image diagnosis and pathological diagnosis. At the same time, the in vivo cervical OCT imaging system, as a newly developed screening tool, was compared with the traditional combined screening of human papillomavirus (HPV) and Thinprep cytologic test (TCT), to assess the screening effect. Results: By comparing the OCT images of the cervix in vivo with the corresponding HE images, the OCT image characteristics of the normal cervix and various types of cervical lesions in vivo were summarized. The accuracy, sensitivity, specificity, PPV and NPV of OCT image in the diagnosis of high-grade squamous intraepithelial lesion (HSIL) and above (HSIL+) were 93.4%, 88.5%, 95.0%, 85.0% and 96.2%, respectively. The accuracy, sensitivity, specificity, PPV and NPV of OCT for low-grade squamous intraepithelial lesion (LSIL) were 84.7%, 61.7%, 96.3%, 89.3% and 83.2%, respectively. The consistency between OCT image diagnosis and pathological diagnosis was strong (Kappa value was 0.701).The accuracy, sensitivity and specificity of OCT screening, HPV and TCT combined screening were 83.7% vs 64.9% (χ²=128.82, P<0.001), 77.8% vs 64.5% (χ²=39.01, P<0.001), 91.8% vs 65.4% (χ²=98.12, P<0.001), respectively. The differences were statistically significant. Conclusions: OCT imaging system has high sensitivity and specificity in the evaluation of cervical lesions in vivo, and has the characteristics of non-invasive, real-time and high efficiency. OCT examination is expected to become an effective method for the diagnosis of cervical lesions and cervical cancer screening.

The added value of digital imaging to reflex cytology for triage of high-risk human papillomavirus positive self-sampled material in cervical cancer screening: A prospective cohort study.

OBJECTIVE: Cytology performed directly on hrHPV-positive self-samples (reflex cytology) is feasible and for women with abnormal cytology, an additional cytology test at the general practitioner could be omitted. The aim of this study is to assess the added value of digital imaging (ThinPrep® Imaging System) on the clinical utility of reflex cytology by reducing screening error. DESIGN: A secondary analysis of a prospective cohort study. SETTING: One of five Dutch screening laboratories. POPULATION: Women tested hrHPV-positive on self-samples between December 2018 and August 2019. METHODS: Self-samples were used for reflex cytology with and without digital imaging. The follow-up data (cytological and histological results within 1 year of follow-up) were obtained through the Dutch Pathology Registry (PALGA). MAIN OUTCOME MEASURES: Test performance of the reflex cytology was determined by comparing it with physician-collected follow-up results. RESULTS: The sensitivity for detecting abnormal cells by reflex cytology on self-samples increased significantly from 26.3% (42/160; 95% confidence interval [CI] 19.6-33.8) without digital imaging to 35.4% (56/158; 95% CI 28-43.4) with digital imaging (P < 0.05) without compromising specificity. Importantly, 41.7% of women with ≥CIN2 (35/84) and 45.6% with ≥CIN3 (26/57) were detected by reflex cytology with digital imaging on hrHPV-positive self-samples. CONCLUSION: Digital imaging is of added value to reflex cytology on hrHPV-positive self-samples with a 9% increase in sensitivity. If reflex cytology on self-samples analysed with digital imaging had been implemented in the screening programme, 35.4% of the hrHPV-positive women with abnormal cytology on additional physician-collected samples could have been referred directly for colposcopy.

A retrospective study of focused ultrasound versus cryotherapy in treatment of cervical squamous intraepithelial lesions.

PURPOSE: To compare the efficacy and safety of focused ultrasound (FUS) therapy and cryotherapy for cervical squamous intraepithelial lesion (SIL). METHODS: In this retrospective study, data pertaining to women treated for cervical SIL with FUS therapy or cryotherapy at the Second Affiliated Hospital of Chongqing Medical University between 21 April 2018 and 31 August 2020 were obtained. The patients were followed up after 3-6 and 6-12 months. The proportions of women with no evidence of disease, recurrent disease, clearance of the human papillomavirus (HPV) and adverse effects or complications were determined. RESULTS: Of the 250 women with complete data who were included in the study, 144 and 106 received FUS therapy and cryotherapy, respectively. Overall, FUS therapy was observed to be more effective than cryotherapy (91.7 vs. 79.2%, p = 0.005). Statistically significant differences were noted in the treatment efficacy for patients with low-grade SIL (LSIL) (92.3 vs. 80.2%, p = 0.011). However, there were no significant differences in the treatment efficacy for patients with high-grade SIL (HSIL) (88.9 vs. 75.0%, p = 0.390). The recurrence rates in patients with LSIL treated with FUS therapy or cryotherapy showed no significant differences at the 6-12-month follow-up (1.0 vs. 6.0%, p = 0.163). Furthermore, there was no recurrence in patients with HSIL, either in the FUS or cryotherapy group. FUS therapy and cryotherapy resulted in similar HPV clearance at the 3-6-month follow-up (77.1 vs. 64.8%, p = 0.057). No statistically significant differences were observed in the complication rates between the two groups (3.5 vs. 1.9%, p = 0.717). CONCLUSION: The results of this study suggest that FUS therapy is superior to cryotherapy in the treatment of cervical LSIL.

An Improved Image Classification Method for Cervical Precancerous Lesions Based on ShuffleNet.

With the rapid development of deep learning, automatic lesion detection is used widely in clinical screening. To solve the problem that existing deep learning-based cervical precancerous lesion detection algorithms cannot meet high classification accuracy and fast running speed at the same time, a ShuffleNet-based cervical precancerous lesion classification method is proposed. By adding channel attention to the ShuffleNet, the network performance is improved. In this study, the image dataset is classified into five categories: normal, cervical cancer, LSIL (CIN1), HSIL (CIN2/CIN3), and cervical neoplasm. The colposcopy images are expanded to solve the problems of the lack of colposcopy images and the uneven distribution of images from each category. For the test dataset, the accuracy of the proposed CNN models is 81.23% and 81.38%. Our classifier achieved an AUC score of 0.99. The experimental results show that the colposcopy image classification network based on artificial intelligence has good performance in classification accuracy and model size, and it has high clinical applicability.

Comparison of accuracy and reproducibility of colposcopic impression based on a single image versus a two-minute time series of colposcopic images.

OBJECTIVE: Colposcopy is an important part of cervical screening/management programs. Colposcopic appearance is often classified, for teaching and telemedicine, based on static images that do not reveal the dynamics of acetowhitening. We compared the accuracy and reproducibility of colposcopic impression based on a single image at one minute after application of acetic acid versus a time-series of 17 sequential images over two minutes. METHODS: Approximately 5000 colposcopic examinations conducted with the DYSIS colposcopic system were divided into 10 random sets, each assigned to a separate expert colposcopist. Colposcopists first classified single two-dimensional images at one minute and then a time-series of 17 sequential images as 'normal,' 'indeterminate,' 'high grade,' or 'cancer'. Ratings were compared to histologic diagnoses. Additionally, 5 colposcopists reviewed a subset of 200 single images and 200 time series to estimate intra- and inter-rater reliability. RESULTS: Of 4640 patients with adequate images, only 24.4% were correctly categorized by single image visual assessment (11% of 64 cancers; 31% of 605 CIN3; 22.4% of 558 CIN2; 23.9% of 3412 < CIN2). Individual colposcopist accuracy was low; Youden indices (sensitivity plus specificity minus one) ranged from 0.07 to 0.24. Use of the time-series increased the proportion of images classified as normal, regardless of histology. Intra-rater reliability was substantial (weighted kappa = 0.64); inter-rater reliability was fair ( weighted kappa = 0.26). CONCLUSION: Substantial variation exists in visual assessment of colposcopic images, even when a 17-image time series showing the two-minute process of acetowhitening is presented. We are currently evaluating whether deep-learning image evaluation can assist classification.

Multi-task network for automated analysis of high-resolution endomicroscopy images to detect cervical precancer and cancer.

Cervical cancer is a public health emergency in low- and middle-income countries where resource limitations hamper standard-of-care prevention strategies. The high-resolution endomicroscope (HRME) is a low-cost, point-of-care device with which care providers can image the nuclear morphology of cervical lesions. Here, we propose a deep learning framework to diagnose cervical intraepithelial neoplasia grade 2 or more severe from HRME images. The proposed multi-task convolutional neural network uses nuclear segmentation to learn a diagnostically relevant representation. Nuclear segmentation was trained via proxy labels to circumvent the need for expensive, manually annotated nuclear masks. A dataset of images from over 1600 patients was used to train, validate, and test our algorithm; data from 20% of patients were reserved for testing. An external evaluation set with images from 508 patients was used to further validate our findings. The proposed method consistently outperformed other state-of-the art architectures achieving a test per patient area under the receiver operating characteristic curve (AUC-ROC) of 0.87. Performance was comparable to expert colposcopy with a test sensitivity and specificity of 0.94 (p = 0.3) and 0.58 (p = 1.0), respectively. Patients with recurrent human papillomavirus (HPV) infections are at a higher risk of developing cervical cancer. Thus, we sought to incorporate HPV DNA test results as a feature to inform prediction. We found that incorporating patient HPV status improved test specificity to 0.71 at a sensitivity of 0.94.

Association of lesion area measured by colposcopy and cervical neoplasia.

Studies have demonstrated that the size of lesion in colposcopic imaging can be associated with the grade of CIN. We evaluated 36 patients and at the time of colposcopy, the images were recorded and analysed for lesion area measurement. A ROC curve was used to obtain the area under the curve and to determine the best cut-off values between area lesion (pixels2) and biopsy result. Fisher's exact test was performed (p < .05). Half of the sample had a cervical biopsy showing HPV or LSIL, and 18 (50%)a biopsy showing HSIL or invasive cancer. HSIL and invasive cancer were associated with a lesion area greater than 30,337.03 pixels2 (cut off) with p = .04. Thus the area of the colposcopic lesion is related to the severity of that; so small lesions can be more conservatively followed.IMPACT STATEMENT:What is already known on this subject? Studies have proposed that the size of lesion in colposcopic imaging can be associated with the grade of CIN, and the size of CIN lesions may be a factor in determining the risk of progression.What do the results of this study add? This is the first study in the literature that uses the measurement of the lesion area in pixels2 in comparison with the severity of the lesion, which provides greater accuracy of the lesion area than the mere measurement of its diameter.What are the implications of these findings for clinical practice and/or further research? The size of the lesion should be considered in the management of cervical intraepithelial lesions. This approach also leads to lower cost and is less invasive. Small lesions will have the best prognosis and would be treated in the way more conservative, bringing to the patients more comfort and less complications with the treatment.

Optimizing ethyl cellulose-ethanol delivery towards enabling ablation of cervical dysplasia.

In low-income countries, up to 80% of women diagnosed with cervical dysplasia do not return for follow-up care, primarily due to treatment being inaccessible. Here, we describe development of a low-cost, portable treatment suitable for such settings. It is based on injection of ethyl cellulose (EC)-ethanol to ablate the transformation zone around the os, the site most impacted by dysplasia. EC is a polymer that sequesters the ethanol within a prescribed volume when injected into tissue, and this is modulated by the injected volume and delivery parameters (needle gauge, bevel orientation, insertion rate, depth, and infusion rate). Salient injection-based delivery parameters were varied in excised swine cervices. The resulting injection distribution volume was imaged with a wide-field fluorescence imaging device or computed tomography. A 27G needle and insertion rate of 10 mm/s achieved the desired insertion depth in tissue. Orienting the needle bevel towards the outer edge of the cervix and keeping infusion volumes ≤ 500 µL minimized leakage into off-target tissue. These results guided development of a custom hand-held injector, which was used to locate and ablate the upper quadrant of a swine cervix in vivo with no adverse events or changes in host temperature or heart rate. After 24 h, a distinct region of necrosis was detected that covered a majority (> 75%) of the upper quadrant of the cervix, indicating four injections could effectively cover the full cervix. The work here informs follow up large animal in vivo studies, e.g. in swine, to further assess safety and efficacy of EC-ethanol ablation in the cervix.

Cervical lesion assessment using real-time microendoscopy image analysis in Brazil: The CLARA study.

We conducted a prospective evaluation of the diagnostic performance of high-resolution microendoscopy (HRME) to detect cervical intraepithelial neoplasia (CIN) in women with abnormal screening tests. Study participants underwent colposcopy, HRME and cervical biopsy. The prospective diagnostic performance of HRME using an automated morphologic image analysis algorithm was compared to that of colposcopy using histopathologic detection of CIN as the gold standard. To assess the potential to further improve performance of HRME image analysis, we also conducted a retrospective analysis assessing performance of a multi-task convolutional neural network to segment and classify HRME images. One thousand four hundred eighty-six subjects completed the study; 435 (29%) subjects had CIN Grade 2 or more severe (CIN2+) diagnosis. HRME with morphologic image analysis for detection of CIN Grade 3 or more severe diagnoses (CIN3+) was similarly sensitive (95.6% vs 96.2%, P = .81) and specific (56.6% vs 58.7%, P = .18) as colposcopy. HRME with morphologic image analysis for detection of CIN2+ was slightly less sensitive (91.7% vs 95.6%, P < .01) and specific (59.7% vs 63.4%, P = .02) than colposcopy. Images from 870 subjects were used to train a multi-task convolutional neural network-based algorithm and images from the remaining 616 were used to validate its performance. There were no significant differences in the sensitivity and specificity of HRME with neural network analysis vs colposcopy for detection of CIN2+ or CIN3+. Using a neural network-based algorithm, HRME has comparable sensitivity and specificity to colposcopy for detection of CIN2+. HRME could provide a low-cost, point-of-care alternative to colposcopy and biopsy in the prevention of cervical cancer.

GRAND: A large-scale dataset and benchmark for cervical intraepithelial Neoplasia grading with fine-grained lesion description.

Cervical cancer causes the fourth most cancer-related deaths of women worldwide. Early detection of cervical intraepithelial neoplasia (CIN) can significantly increase the survival rate of patients. World Health Organization (WHO) divided the CIN into three grades (CIN1, CIN2 and CIN3). In clinical practice, different CIN grades require different treatments. Although existing studies proposed computer aided diagnosis (CAD) systems for cervical cancer diagnosis, most of them are fail to perform accurate separation between CIN1 and CIN2/3, due to the similar appearances under colposcopy. To boost the accuracy of CAD systems, we construct a colposcopic image dataset for GRAding cervical intraepithelial Neoplasia with fine-grained lesion Description (GRAND). The dataset consists of colposcopic images collected from 8,604 patients along with the pathological reports. Additionally, we invite the experienced colposcopist to annotate two main clues, which are usually adopted for clinical diagnosis of CIN grade, i.e., texture of acetowhite epithelium (TAE) and appearance of blood vessel (ABV). A multi-rater model using the annotated clues is benchmarked for our dataset. The proposed framework contains several sub-networks (raters) to exploit the fine-grained lesion features TAE and ABV, respectively, by contrastive learning and a backbone network to extract the global information from colposcopic images. A comprehensive experiment is conducted on our GRAND dataset. The experimental results demonstrate the benefit of using additional lesion descriptions (TAE and ABV), which increases the CIN grading accuracy by over 10%. Furthermore, we conduct a human-machine confrontation to evaluate the potential of the proposed benchmark framework for clinical applications. Particularly, three colposcopists on different professional levels (intern, in-service and professional) are invited to compete with our benchmark framework by investigating a same extra test set-our framework achieves a comparable CIN grading accuracy to that of a professional colposcopist.

Endocervical microglandular hyperplasia: Colposcopic aspects, physiopathology and differential diagnosis.

Endocervical microglandular hyperplasia (MGH) is a reactive type of glandular lesion that may be confused with endocervical adenocarcinoma from the macroscopic and the colposcopic findings, as well as from a histological. Differential diagnosis is very important. Here, we report a case of a 21 years-old women with a challenging differential diagnosis in the colposcopy and a MGH as histological finding.

Raman spectroscopic study of cervical precancerous lesions and cervical cancer.

Early detection of cervical lesions, accurate diagnosis of cervical lesions, and timely and effective therapy can effectively avoid the occurrence of cervical cancer or improve the survival rate of patients. In this paper, the spectra of tissue sections of cervical inflammation (n = 60), CIN (cervical intraepithelial neoplasia) I (n = 30), CIN II (n = 30), CIN III (n = 30), cervical squamous cell carcinoma (n = 30), and cervical adenocarcinoma (n = 30) were collected by a confocal Raman micro-spectrometer (LabRAM HR Evolution, Horiba France SAS, Villeneuve d'Ascq, France). The Raman spectra of six kinds of cervical tissues were analyzed, the dominant Raman peaks of different kinds of tissues were summarized, and the differences in chemical composition between the six tissue samples were compared. An independent sample t test (p ≤ 0.05) was used to analyze the difference of average relative intensity of Raman spectra of six types of cervical tissues. The difference of relative intensity of Raman spectra of six kinds of tissues can reflect the difference of biochemical components in six kinds of tissues and the characteristic of biochemical components in different kinds of tissues. The classification models of cervical inflammation, CIN I, CIN II, CIN III, cervical squamous cell carcinoma, and cervical adenocarcinoma were established by using a support vector machine (SVM) algorithm. Six types of cervical tissues were classified and identified with an overall diagnostic accuracy of 85.7%. This study laid a foundation for the application of Raman spectroscopy in the clinical diagnosis of cervical precancerous lesions and cervical cancer.

Ultrasonic Elastography Combined with Human Papilloma Virus Detection Based on Intelligent Denoising Algorithm in Diagnosis of Cervical Intraepithelial Neoplasia.

The aim of this research was to study the application of ultrasonic elastography combined with human papilloma virus (HPV) detection based on bilateral filter intelligent denoising algorithm in the diagnosis of cervical intraepithelial neoplasia (CIN) and provide a theoretical basis for clinical diagnosis and treatment of CIN. In this study, 100 patients with cervical lesions were selected as research objects and randomly divided into control group and experimental group, with 50 cases in each group. Patients in control group and experimental group were diagnosed by ultrasonic elastography combined with HPV detection. The experimental group used the optimized image map of bilateral filter intelligent denoising algorithm for denoising and optimization, while the control group did not use optimization, and the differences between them were analyzed and compared. The diagnostic effects of the two groups were compared. As a result, the three accuracy rates of the experimental group were 95%, 95%, and 98%, respectively; the three sensitivity rates were 96%, 92%, and 94%, respectively; and the three specificity rates were 99%, 97%, and 98%, respectively. In the control group, the three accuracy rates were 84%, 86%, and 84%, respectively; the three sensitivity rates were 88%, 84%, and 86%, respectively; and the three specificity rates were 81%, 83%, and 88%, respectively. The accuracy, sensitivity, and specificity of experiment group were significantly higher than those of control group, and the difference was statistically significant (P < 0.05). In summary, the bilateral filter intelligent denoising algorithm has a good denoising effect on the ultrasonic elastography. The ultrasonic image processed by the algorithm combined with HPV detection has a better diagnosis of CIN.

Elastography of the uterine cervix in gynecology: normal appearance, cervical intraepithelial neoplasia and cancer. A systematic review.

AIMS: To revise the current literature about the usefulness of elastography in cervical cancer (CC) and cervical intraepithelial neoplasia (CIN), from methods and technical limitations, to diagnosis, staging and the ability of predicting the response to oncologic treatment. METHODS: An electronic database search was performed (PubMed, EMBASE, Web of Science) with the data range from January 2000 until May 2020. All studies, fully-available in English, assessing elastography of the uterine cervix in CC and CIN were selected. Studies were reviewed and discussed according to the elastographic technique and to the purpose of the research. RESULTS: Twenty-three articles were found: 11 articles regarding strain elastography, 4 articles assessing shear wave elastography and 8 papers with matter-related information. Elastography was used in the study of normal variants of the uterine cervix as well as: the positive diagnosis of CC and CIN, clinical staging and the prediction of therapeutic response in CC. Comparison of the elastographic techniques was also performed. CONCLUSIONS: Elastography has multiple applications in the gynecological pathology of the cervix. The methods used to assess the cervix are diverse, and none have become universally accepted. With regard to CC and CIN, elastography is still an ongoing research field.

Surgical management of cervical cancer in a resource-limited setting: One year of data from the National Cancer Institute, Sri Lanka.

OBJECTIVE: To evaluate the surgical management of cervical cancer without the use of preoperative pelvic imaging in a resource-limited setting. METHODS: A retrospective study was carried out using clinical records and the ongoing electronic database at the Gynaecological Oncology Unit, National Cancer Institute (Apeksha Hospital), Maharagama, Sri Lanka. Details regarding the radical hysterectomies carried out from January 1, 2019, to December 31, 2019, were retrospectively studied. RESULTS: Out of nearly 700 patients with cervical cancer admitted during the year 2019, 57 surgically managed radical hysterectomies were included. Of these, seven cases were ineligible and excluded and 50 cases of radical hysterectomies were included for analysis. Mean age was 53.6 ± 9.5 years and median parity was 3 (range 2-4). Of the cases, 94% were found to have no parametrial involvement showing the success of clinical examination in assessing local tumor spread. Overall, 11 (22.0%) were upstaged due to lymph node metastasis that was statistically significant. CONCLUSION: Preoperative clinical staging is a practical method in selecting surgically treatable cervical cancer in low- and middle-income countries (LMICs). Combining clinical assessment with comparatively more readily available computed tomography scans could be helpful in triaging patients for treatment of cervical cancer in LMICs.

MSCI: A multistate dataset for colposcopy image classification of cervical cancer screening.

BACKGROUND: Cervical cancer is the second most common female cancer globally, and it is vital to detect cervical cancer with low cost at an early stage using automated screening methods of high accuracy, especially in areas with insufficient medical resources. Automatic detection of cervical intraepithelial neoplasia (CIN) can effectively prevent cervical cancer. OBJECTIVES: Due to the deficiency of standard and accessible colposcopy image datasets, we present a dataset containing 4753 colposcopy images acquired from 679 patients in three states (acetic acid reaction, green filter, and iodine test) for detection of cervical intraepithelial neoplasia. Based on this dataset, a new computer-aided method for cervical cancer screening was proposed. METHODS: We employed a wide range of methods to comprehensively evaluate our proposed dataset. Hand-crafted feature extraction methods and deep learning methods were used for the performance verification of the multistate colposcopy image (MSCI) dataset. Importantly, we propose a gated recurrent convolutional neural network (C-GCNN) for colposcopy image analysis that considers time series and combined multistate cervical images for CIN grading. RESULTS: The experimental results showed that the proposed C-GCNN model achieves the best classification performance in CIN grading compared with hand-crafted feature extraction methods and classic deep learning methods. The results showed an accuracy of 96.87 %, a sensitivity of 95.68 %, and a specificity of 98.72 %. CONCLUSION: A multistate colposcopy image dataset (MSCI) is proposed. A CIN grading model (C-GCNN) based on the MSCI dataset is established, which provides a potential method for automated cervical cancer screening.

In vivo optical cellular diagnosis for uterine cervical or vaginal intraepithelial neoplasia using flexible gastrointestinal endocytoscopy -a prospective pilot study.

BACKGROUUND: For patients with any kind of atypical squamous intraepithelial lesion of the uterine cervix or vagina, colposcopy and punch biopsy are common procedures for histological determination following cytology. However, colposcopy-guided biopsy does not provide a high level of diagnostic accuracy. The aim of this study was to determine the usefulness of optical biopsy in vivo using endocytoscopy compared with conventional procedures using colposcopy. METHODS: Between May 2018 and March 2019, patients who were scheduled for cervical conization or mapping biopsies of the vagina were prospectively enrolled. Endocytoscopy was performed by senior endoscopists prior to scheduled procedures, and endocytoscopic images and biopsy samples were taken from the most prominent site and surrounding area of the cervical or vaginal lesions. The collection process of images was randomized and anonymous, and three doctors separately evaluated the images according to the ECA classification. ECA 4 and 5 are indicative of endoscopic malignancy. The primary endpoint was diagnostic accuracy (benign or malignant: cervical intraepithelial neoplasia (CIN) 3 or vaginal intraepithelial neoplasia (VAIN) 3 or worse) of cell images at the most prominent site in each patient. RESULTS: A total of 28 consecutive patients were enrolled. Sensitivity, specificity, positive predictive value, negative predictive value and accuracy of endocytoscopic images were 95.0% (84.8-98.6%), 87.5% (61.9-96.5%), 95.0% (84.8-98.6%), 87.5% (61.9-96.5%) and 92.9% (78.2-98.0%), respectively. Inter-observer agreement among three reviewers was 0.78 (0.08-9.88, P < 0.01). On the other hand, the accuracy of colposcopy-guided biopsy was 74.1% (64.0-84.0%). CONCLUSIONS: Optical cell diagnosis of cervical or vaginal intraepithelial neoplasia using endocytoscopy provides a high level of diagnostic accuracy. TRIAL REGISTRATION: The study was registered with the UMIN database (ID: 000031712 ). UMIN000031712 . Registered 16 March 2017.

3D ultrasound as a surgical quality control of conization in patients with severe dysplasia: a prospective study.

PURPOSE: To compare the techniques for cone measurement with ultrasound to determine the size of the resected tissue and to evaluate parameters which may be relevant for stratifying women at risk who need surveillance when pregnant. METHODS: The present study included women with a pathological cervical biopsy. Cervical length and volume were determined by transvaginal ultrasound prior to conization. The pathologist measured the volume of the removed tissue by the fluid displacement technique and using a ruler. A repeat transvaginal ultrasound was performed during a follow-up visit. Factors affecting cone volume as well as the correlation between measurement techniques were analyzed. RESULTS: A total of 28 patients underwent cervical excision treatment. The mean cervical volumes measured sonographically before and after the operation were 17.72 ± 7.34 and 13.21 ± 5.43 cm3, respectively. The proportion of volume excised was 25.50 ± 17.43%. A significant correlation was found between the cone depth and the cone volume measured by the fluid displacement technique, and histopathologically and sonographically measured difference in cervical volume. The interobserver reliability coefficient was > 0.9. Analyzing influential parameters, only age affected the extent of cone volume and the correlation between the three measurement techniques. CONCLUSION: Commonly applied techniques of cervical and cone measurement are equivalent and interchangeable. Our ultrasound data show variety in the volume and length of the cervix, and in the proportion of the volume excised at conization. Ultrasound measurements may help the surgeon to estimate not only the dimension of the remaining cervix but also its function.

The application of deep learning based diagnostic system to cervical squamous intraepithelial lesions recognition in colposcopy images.

Background Deep learning has presented considerable potential and is gaining more importance in computer assisted diagnosis. As the gold standard for pathologically diagnosing cervical intraepithelial lesions and invasive cervical cancer, colposcopy-guided biopsy faces challenges in improving accuracy and efficiency worldwide, especially in developing countries. To ease the heavy burden of cervical cancer screening, it is urgent to establish a scientific, accurate and efficient method for assisting diagnosis and biopsy. Methods The data were collected to establish three deep-learning-based models. For every case, one saline image, one acetic image, one iodine image and the corresponding clinical information, including age, the results of human papillomavirus testing and cytology, type of transformation zone, and pathologic diagnosis, were collected. The dataset was proportionally divided into three subsets including the training set, the test set and the validation set, at a ratio of 8:1:1. The validation set was used to evaluate model performance. After model establishment, an independent dataset of high-definition images was collected to further evaluate the model performance. In addition, the comparison of diagnostic accuracy between colposcopists and models weas performed. Results The sensitivity, specificity and accuracy of the classification model to differentiate negative cases from positive cases were 85.38%, 82.62% and 84.10% respectively, with an AUC of 0.93. The recall and DICE of the segmentation model to segment suspicious lesions in acetic images were 84.73% and 61.64%, with an average accuracy of 95.59%. Furthermore, 84.67% of high-grade lesions were detected by the acetic detection model. Compared to colposcopists, the diagnostic system performed better in ordinary colposcopy images but slightly unsatisfactory in high-definition images. Implications The deep learning-based diagnostic system could help assist colposcopy diagnosis and biopsy for HSILs.

Usefulness of real time elastography strain ratio in the assessment of cervical intraepithelial neoplasia and cervical cancer using a reference material.

AIMS: To assess the usefulness of real time elastography (RTE) strain ratio (SR) in diagnosing cervical cancer (CC) and cervical intraepithelial neoplasia (CIN), using a synthetic experimental device (ED) as reference material. MATERIAL AND METHODS: Seventy-nine participants were enrolled, divided in three groups: Group 1 - benign cervix (n=39); Group 2 - CIN (n=32); Group 3 - CC (n=8). Transvaginal RTE was performed, with SR determination, as the ratio between the ED and the cervical tissue. Mean SR values of the groups were compared; diagnostic performance was assessed by tracing the receiver operating characteristic (ROC) curve. Area under the curve (AUC) was analyzed. Cut-off values were established. Pathological results were considered as reference for data interpretation. RESULTS: SR means significantly differed in Group 1 as compared to Groups 2 and 3 (p=0.001). Excluding 2 aberrant values in Group 3, assigned to cases complicated by hemorrhagic necrosis, statistical difference was also noted between Groups 2 and 3 (p=0.02). For Groups 1 and 3, AUC was 0.966 with a 95%CI (0.914-1.000); the cut-off point of SR was 1.42, with a sensitivity of 100% and a specificity of 94.9%. AUC was 0.752 with a 95%CI (0.629-0.876) for Groups 1 and 2. For the cut-off value of 1.03, sensitivity and specificity were 75% and 74%, respectively. CONCLUSION: RTE SR, performed with a synthetic reference material, seems a reliable method for distinguishing between benign uterine cervix and malignancy, with promising results as a complementary investigation in diagnosing CIN. However, SR becomes inoperant in cases of cancer complicated with hemorrhagic necrosis.