scholarly journals Automatic model for cervical cancer screening based on convolutional neural network: a retrospective, multicohort, multicenter study

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Xiangyu Tan ◽  
Kexin Li ◽  
Jiucheng Zhang ◽  
Wenzhe Wang ◽  
Bian Wu ◽  
...  

Abstract Background The incidence rates of cervical cancer in developing countries have been steeply increasing while the medical resources for prevention, detection, and treatment are still quite limited. Computer-based deep learning methods can achieve high-accuracy fast cancer screening. Such methods can lead to early diagnosis, effective treatment, and hopefully successful prevention of cervical cancer. In this work, we seek to construct a robust deep convolutional neural network (DCNN) model that can assist pathologists in screening cervical cancer. Methods ThinPrep cytologic test (TCT) images diagnosed by pathologists from many collaborating hospitals in different regions were collected. The images were divided into a training dataset (13,775 images), validation dataset (2301 images), and test dataset (408,030 images from 290 scanned copies) for training and effect evaluation of a faster region convolutional neural network (Faster R-CNN) system. Results The sensitivity and specificity of the proposed cervical cancer screening system was 99.4 and 34.8%, respectively, with an area under the curve (AUC) of 0.67. The model could also distinguish between negative and positive cells. The sensitivity values of the atypical squamous cells of undetermined significance (ASCUS), the low-grade squamous intraepithelial lesion (LSIL), and the high-grade squamous intraepithelial lesions (HSIL) were 89.3, 71.5, and 73.9%, respectively. This system could quickly classify the images and generate a test report in about 3 minutes. Hence, the system can reduce the burden on the pathologists and saves them valuable time to analyze more complex cases. Conclusions In our study, a CNN-based TCT cervical-cancer screening model was established through a retrospective study of multicenter TCT images. This model shows improved speed and accuracy for cervical cancer screening, and helps overcome the shortage of medical resources required for cervical cancer screening.

2020 ◽  
Vol 10 (1) ◽  
pp. 1639-1644
Author(s):  
Indrani Krishnappa ◽  
Kalyani R. ◽  
Raja Parthiban ◽  
Abhishek Agrawal

Background: Pap smear examination has been universally used as an effective screening tool for early detection of cervical carcinoma. The aim of this study was to assess the utility of Cervical Acid Phosphatase staining as an adjunct to routine Pap smear testing to improvethe sensitivity and specificity of routine Pap smear examination for cervical cancer detection. Materials and Methods: Cervical smears were taken from patients attending the gynecology department and a few cervical cancer screening programmes. One set of slides were alcohol fixed and stained with rapid pap stain and another set of slides were fixed in a special fixative and stained with Cervical Acid Phosphatase -Pap stain. The nuclear features of these Cervical Acid Phosphatase stained dysplastic cells was studied on Pap stain to diagnose cervical intraepithelial lesion/ malignancy. Results: Out of 489 cases included in the study 6 cases were diagnosed with intraepithelial lesion/ malignancy. On Cervical Acid Phosphatase -Pap stain 2 of the cases diagnosed as inflammatory smears on pap stain showed Cervical Acid Phosphatase positivity and thus were re evaluated. Mild nuclear atypia was observed in the Cervical Acid Phosphatase positive cells and these cases were diagnosed as Low grade squamous intraepithelial lesion and later biopsy proven to be Cervical intraepithelial Neoplasia I. Therefore Cervical Acid Phosphatase -Pap test was 100% sensitive and specific for cervical cancer detection. Conclusions: With 100% sensitivity Cervical Acid Phosphatase -Pap test satisfies the criteria of an efficient screening test.


Biomédica ◽  
2019 ◽  
Vol 39 (1) ◽  
pp. 65-74
Author(s):  
Óscar Gamboa ◽  
Mauricio González ◽  
Jairo Bonilla ◽  
Joaquín Luna ◽  
Raul Murillo ◽  
...  

Introduction: Direct visual inspection for cervical cancer screening remains controversial, whereas colposcopy-biopsy is considered the gold standard for diagnosis of preneoplastic cervical lesions.Objectives: To determine the rates of cervical intraepithelial neoplasia grade 2 or more and of false positives for colposcopy and direct visual inspection.Materials and methods: Women aged 25-59 underwent direct visual inspection with acetic acid (VIA), Lugol’s iodine (VIA-VILI), and colposcopy. Punch biopsies were obtained for all positive tests. Using histology as the gold standard, detection and false positive rates were compared for VIA, VIA-VILI, and colposcopy (two thresholds). Sensitivity and false positive ratios with the corresponding 95% confidence intervals were estimated.Results: We included 5,011 women in the analysis and we obtained 602 biopsies. Positivity rates for colposcopy high-grade and low-grade diagnosis were 1.6% and 10.8%. Positivity rates for VIA and VIA-VILI were 7.4% and 9.9%. VIA showed a significantly lower detection rate than colposcopy with low-grade diagnosis as the threshold (SR=0.72; 95% CI 0.57-0.91), and significantly lower false positive rate (FPR=0.70; 95% CI 0.65-0.76). No differences between VIA-VILI and colposcopy low-grade threshold were observed. VIA and VIA-VILI showed significantly higher detection and false positive rates than colposcopy high-grade threshold. Sensitivity rates for visual inspection decreased with age and false positive rates increased. For all age groups, false positive rates for VIA and VIA-VILI were significantly higher than colposcopy.Conclusions: Detection rates for VIA-VILI similar to colposcopy low-grade threshold represent a chance to reduce cervical cancer mortality through see-and-treat approaches among women with limited access to health care. Lower detection rates suggest reviewing high-grade colposcopy findings as the threshold for biopsy in certain settings.


2020 ◽  
Author(s):  
Yuwei Sun ◽  
Hideya Ochiai ◽  
Hiroshi Esaki

Abstract This article illustrates a method of visualizing network traffic in LAN based on the Hilbert Curve structure and the array exchange and projection, with nine types of protocols’ communication frequency information as the discriminators, the results of which we call them feature maps of network events. Several known scan cases are simulated in LANs and network traffic is collected for generating feature maps under each case. In order to solve this multi-label classification task, we adopt and train a deep convolutional neural network (DCNN), in two different network environments with feature maps as the input data, and different scan cases as the labels. We separate datasets with a ratio of 4:1 into the training dataset and the validation dataset. Then, based on the micro scores and the macro scores of the validation, we evaluate performance of the scheme, achieving macro-F-measure scores of 0.982 and 0.975, and micro-F-measure scores of 0.976 and 0.965 separately in these two LANs.


2020 ◽  
Vol 7 (2) ◽  
pp. 239-246
Author(s):  
J. S. Misra ◽  
A. N. Srivastava ◽  
Shivani Singh

Carcinoma cervix is a major health problem faced by Indian women and situation is very alarming in rural population because of illiteracy and poverty. There is urgent need of creating awareness regarding the risk factors of cervical cancer such as early marriage and multiparity in the rural women and the importance of early detection of the cervical cancer. With aim for providing single life time Pap smear cover to the rural women, cervical cancer screening program was initiated in May 2013 by Era's Lucknow, Medical College and Hospital, Lucknow, through organizing camps in the villages of Lucknow ,west. Till December 2019, a total of 183 camps have been organized and 5209 women attended these camps (28.5%). Pap smear examination could be done in only 2912 women (55.9%). The incidence of squamous intraepithelial lesions of cervix (SIL) was found to be very high in these women (17.1%) though majority of SIL were of low grade (LSIL). This may be related to the poor personal genital hygiene prevailing in rural women due to illiteracy. In the present review, the SIL incidence has been extensively analyzed in relation to different predisposing factors of cervical carcinogenesis and problems faced during camp organization and the benefits derived from these camps to rural women have been highlighted.


2019 ◽  
Vol 26 (3) ◽  
Author(s):  
K. Decker ◽  
N. Baines ◽  
C. Muzyka ◽  
M. Lee ◽  
M. H. Mayrand ◽  
...  

Background Colposcopy is a key part of cervical cancer control. As cervical cancer screening and prevention strategies evolve, monitoring colposcopy performance will become even more critical. In the present paper, we describe population-based colposcopy quality indicators that are recommended for ongoing measurement by cervical cancer screening programs in Canada.Methods The Pan-Canadian Cervical Cancer Screening Network established a multidisciplinary expert working group to identify population-based colposcopy quality indicators. A systematic literature review was conducted to ascertain existing population and program-level colposcopy quality indicators. A systems-level cervical cancer screening pathway describing each step from an abnormal screening test, to colposcopy, and back to screening was developed. Indicators from the literature were assigned a place on the pathway to ensure that all steps were measured. A prioritization matrix scoring system was used to score each indicator based on predetermined criteria. Proposed colposcopy quality indicators were shared with provincial and territorial screening programs and subsequently revised.Results The 10 population-based colposcopy quality indicators identified as priorities were colposcopy uptake, histologic investigation (biopsy) rate, colposcopy referral rate, failure to attend colposcopy, treatment frequency in women 18–24 years of age, re-treatment proportion, colposcopy exit-test proportion, histologic investigation (biopsy) frequency after low-grade Pap test results, length of colposcopy episode of care, and operating room treatment rate. Two descriptive indicators were also identified: colposcopist volume and number of colposcopists per capita.Summary High-quality colposcopy services are an essential component of provincial cervical cancer screening programs. The proposed quality and descriptive indicators will permit colposcopy outcomes to be compared between provinces and across Canada so as to identify opportunities for improving colposcopy services.


Endoscopy ◽  
2019 ◽  
Vol 51 (12) ◽  
pp. 1121-1129 ◽  
Author(s):  
Bum-Joo Cho ◽  
Chang Seok Bang ◽  
Se Woo Park ◽  
Young Joo Yang ◽  
Seung In Seo ◽  
...  

Abstract Background Visual inspection, lesion detection, and differentiation between malignant and benign features are key aspects of an endoscopist’s role. The use of machine learning for the recognition and differentiation of images has been increasingly adopted in clinical practice. This study aimed to establish convolutional neural network (CNN) models to automatically classify gastric neoplasms based on endoscopic images. Methods Endoscopic white-light images of pathologically confirmed gastric lesions were collected and classified into five categories: advanced gastric cancer, early gastric cancer, high grade dysplasia, low grade dysplasia, and non-neoplasm. Three pretrained CNN models were fine-tuned using a training dataset. The classifying performance of the models was evaluated using a test dataset and a prospective validation dataset. Results A total of 5017 images were collected from 1269 patients, among which 812 images from 212 patients were used as the test dataset. An additional 200 images from 200 patients were collected and used for prospective validation. For the five-category classification, the weighted average accuracy of the Inception-Resnet-v2 model reached 84.6 %. The mean area under the curve (AUC) of the model for differentiating gastric cancer and neoplasm was 0.877 and 0.927, respectively. In prospective validation, the Inception-Resnet-v2 model showed lower performance compared with the endoscopist with the best performance (five-category accuracy 76.4 % vs. 87.6 %; cancer 76.0 % vs. 97.5 %; neoplasm 73.5 % vs. 96.5 %; P  < 0.001). However, there was no statistical difference between the Inception-Resnet-v2 model and the endoscopist with the worst performance in the differentiation of gastric cancer (accuracy 76.0 % vs. 82.0 %) and neoplasm (AUC 0.776 vs. 0.865). Conclusion The evaluated deep-learning models have the potential for clinical application in classifying gastric cancer or neoplasm on endoscopic white-light images.


2018 ◽  
Vol 4 (Supplement 2) ◽  
pp. 137s-137s
Author(s):  
C. Onwuka ◽  
C. Umezurike ◽  
C.-E. Hendricks ◽  
K. Eguzo

Background and context: Cervical cancer is the second most common cancer and a leading cause of mortality affecting Nigerian women. There is prevalent ignorance about the disease and reduced access to screening services in Nigeria due to a variety of reasons. Many religious leaders often view cancer as evil, thus perpetuating a culture of fear and denial. Religious events have been used to promote health awareness and access to preventive services in Nigeria, but not usually for cancer screening. A cancer survivor might use a religious event to provide a strong avenue for improving cancer awareness and access to screening. Aim: Report describes a successful collaborative integration of cervical cancer screening into a religious event led by a breast cancer survivor. Strategy/Tactics: The 2018 Annual Women's Gospel Forum at Susanna Homes was used by a cancer survivor to teach women about cervical cancer and early diagnosis. Information was circulated among the Church of Christ fellowship in southern Nigeria. Program/Policy process: A medical team (gynecologist and pathologist) provided free cancer screening to consenting women using Papanicolaou test VIA simultaneously, with Pap preceding VIA. All smears were handled according to standard protocol and the findings were classified using the Bethesda system for reporting Papanicolaou test. Participants were taught about cervical cancer, counseled on screening and follow-up care prior to consent. Data were analyzed using descriptive statistics. Outcomes: Seventy-nine women at the event consented to the screening. The average participant was aged 51 years, but 90.6% (58/64) were above 30 years. Most (92.4%; 73/79) women were negative for squamous intraepithelial lesions (SILs), and six women had a low grade squamous intraepithelial lesion (LSIL). The prevalence of abnormal cervical cytology in this outreach was 7.6% (6/79), and mostly affected women aged 61-60 years. Table 1 describes the Papanicolaou test results. The result of VIA was recorded as positive or negative. What was learned: Participants provided positive feedback regarding integrating cervical cancer screening into a religious event, as it improved access and addressed ignorance. Leadership by a cancer survivor helped demystify the disease and encouraged women to be screened. Having a pathologist on site improved slides quality and reporting turn-around time. Project shows that collaboration between clinicians and religious leaders can improve cancer prevention. More physicians should be trained to provide these services and collaborate with community/religious organizations to provide screening outside of the clinical setting.[Table: see text]


2018 ◽  
Vol 29 (12) ◽  
pp. 1204-1214 ◽  
Author(s):  
Sheyla D Richards ◽  
Samantha Stonbraker ◽  
Mina Halpern ◽  
Silvia Amesty

Cervical cancer is the third leading cause of cancer-related death and the second most diagnosed cancer among women in developing countries. We determined the prevalence of abnormal Papanicolaou (Pap), high-risk HPV (hrHPV), and colposcopy among transactional female sex workers (FSWs) in La Romana, Dominican Republic. The results of 144 FSWs of ages 18–54 years who completed a demographic interview and Pap testing with hrHPV detection between June 2015 and April 2016 were analyzed. Women with abnormal results were referred for colposcopy. Risk factors for abnormal Pap were assessed through bivariate and multivariate analyses. Overall, 36.1% (52/144) of Paps were abnormal and 43.4% (62/143) had hrHPV. Of all women with hrHPV and/or abnormal Pap (68/144; 47.2%), 61 (89.7%) were referred and 16 (26.2) underwent colposcopy. HPV16 and/or 18/45 was detected in 33.3% (15/45) of low-grade Paps. Binge drinking, weekly (AOR 5.1, 95% CI: 1.8–14.5) or daily (AOR 4.9, 95% CI: 1.5–16.6), and age at first sexual relation (AOR 1.2, 95% CI: 1.0–1.5) were significantly associated ( p < 0.05) with abnormal Pap. Although almost half of participants had abnormal Pap or hrHPV, few underwent colposcopy. Improving access to cervical cancer screening and follow-up for FSWs is imperative.


Author(s):  
Luiz Zeferino ◽  
Joana Bastos ◽  
Diama Vale ◽  
Rita Zanine ◽  
Yara Melo ◽  
...  

AbstractEvidence-based clinical guidelines ensure best practice protocols are available in health care. There is a widespread use of human papillomavirus deoxyribonucleic acid (HPV-DNA) tests in Brazil, regardless of the lack of official guidelines. On behalf of the Brazilian Association for the Lower Genital Tract Pathology and Colposcopy (ABPTGIC, in the Portuguese acronym), a team of reviewers searched for published evidence and developed a set of recommendations for the use of HPV-DNA tests in cervical cancer screening in Brazil. The product of this process was debated and consensus was sought by the participants. One concern of the authors was the inclusion of these tests in the assessment of women with cytologic atypia and women treated for cervical intraepithelial neoplasia (CIN). Testing for HPV is recommended in an organized screening scenario to identify women with precursor lesions or asymptomatic cervical cancer older than 30 years of age, and it can be performed every 5 years. It also has value after the cytology showing atypical squamous cells of undetermined significance (ASC-US) or low-grade squamous intraepithelial lesions (LSILs) as a triage test for colposcopy, in the investigation of other cytological alterations when no abnormal findings are observed at colposcopy, seeking to exclude disease, or, further, after treatment of high-grade cervical intraepithelial neoplasia, to rule out residual disease.


2017 ◽  
Vol 2017 ◽  
pp. 1-5 ◽  
Author(s):  
Alexandra Aserlind ◽  
Karla Maguire ◽  
Lunthita Duthely ◽  
Stefan Wennin ◽  
JoNell Potter

Objective. Women living with HIV are at increased risk of human papillomavirus (HPV) infection, which can lead to cervical cancer. New guidelines recommend indefinite screening. The objective of this study is to describe cervical cancer screening practices and colposcopy results in a cohort of women living with HIV over age of 65 who were followed before the new guidelines. Comorbidities, sexually transmitted infections (STIs), and other risk factors were evaluated. Methods. We conducted a retrospective chart review on 75 women aged 65 or older living with HIV with at least one Pap smear. Results. The mean age of the cohort was 66.5 and at HIV diagnosis was 56. The majority of women were immunocompetent. 80% had serial Pap smears. Of these, 86% of 238 were negative or ASCUS. No women progressed to HSIL. 92% of colposcopies had negative or CIN I results. Three women were treated successfully for high-grade dysplasia. More than half of women had other STIs. 72% were screened for HPV; 50% were positive. Conclusion. The majority of women had negative and low-grade Pap smears. Questions remain regarding the utility of continued Pap screening and the added value of HPV testing in this unique population of older women living with HIV.


Sign in / Sign up

Export Citation Format

Share Document