Cervicography screening for cervical cancer among 8460 women in a high-risk population

RGB Channel Superposition Algorithm with Acetowhite Mask Images in a Cervical Cancer Classification Deep Learning Model

Cervical cancer is one of the main causes of death from cancer in women. However, it can be treated successfully at an early stage. This study aims to propose an image processing algorithm based on acetowhite, which is an important criterion for diagnosing cervical cancer, to increase the accuracy of the deep learning classification model. Then, we mainly compared the performance of the model, the original image without image processing, a mask image made with acetowhite as the region of interest, and an image using the proposed algorithm. In conclusion, the deep learning classification model based on images with the proposed algorithm achieved an accuracy of 81.31%, which is approximately 9% higher than the model with original images and approximately 4% higher than the model with acetowhite mask images. Our study suggests that the proposed algorithm based on acetowhite could have a better performance than other image processing algorithms for classifying stages of cervical images.

Noninvasive Point-of-Care Nanobiosensing of Cervical Cancer as an Auxiliary to Pap-Smear Test

A potential cancer antigen (Ag), protein-phosphatase-1-gamma-2 (PP1γ2), with a restricted expression in testis and sperms has been identified as a biomarker specific to cervical cancer (CaCx). Detection of this cancer biomarker antigen (NCB-Ag) in human urine opens up the possibility of noninvasive detection of CaCx to supplement the dreaded and invasive Pap-smear test. A colorimetric response of an assembly of gold nanoparticles (Au NPs) has been employed for the quantitative, noninvasive, and point-of-care-testing of CaCx in the urine. In order to fabricate the immunosensor, Au NPs of sizes ∼5-20 nm have been chemically modified with a linker, 3,3'-di-thio-di-propionic-acid-di(n-hydroxy-succinimide-ester) (DTSP) to attach the antibody (Ab) specific to the NCB-Ag. Interestingly, the addition of Ag to the composite of Ab-DTSP-Au NPs leads to a significant hypsochromic shift due to a localized surface plasmon resonance phenomenon, which originates from the specific epitope-paratope interaction between the NCB-Ag and Ab-DTSP-Au NPs. The variations in the absorbance and wavelength shift during such attachments of different concentrations of NCB-Ag on the Ab-DTSP-Au NPs composite have been employed as a calibration to identify NCB-Ag in human urine. An in-house prototype has been assembled by integrating a light-emitting diode of a narrow range wavelength in one side of a cuvette in which the reaction has been performed while a sensitive photodetector to the other side to transduce the transmitted signal associated with the loading of NCB-Ag in the Ab-DTSP-Au NPs composite. The proposed immunosensing platform has been tested against other standard proteins to ensure noninterference alongside proving the proof-for-specificity of the NCB detection.

Network Visualization and Pyramidal Feature Comparison for Ablative Treatability Classification Using Digitized Cervix Images

Uterine cervical cancer is a leading cause of women’s mortality worldwide. Cervical tissue ablation is an effective surgical excision of high grade lesions that are determined to be precancerous. Our prior work on the Automated Visual Examination (AVE) method demonstrated a highly effective technique to analyze digital images of the cervix for identifying precancer. Next step would be to determine if she is treatable using ablation. However, not all women are eligible for the therapy due to cervical characteristics. We present a machine learning algorithm that uses a deep learning object detection architecture to determine if a cervix is eligible for ablative treatment based on visual characteristics presented in the image. The algorithm builds on the well-known RetinaNet architecture to derive a simpler and novel architecture in which the last convolutional layer is constructed by upsampling and concatenating specific RetinaNet pretrained layers, followed by an output module consisting of a Global Average Pooling (GAP) layer and a fully connected layer. To explain the recommendation of the deep learning algorithm and determine if it is consistent with lesion presentation on the cervical anatomy, we visualize classification results using two techniques: our (i) Class-selective Relevance Map (CRM), which has been reported earlier, and (ii) Class Activation Map (CAM). The class prediction heatmaps are evaluated by a gynecologic oncologist with more than 20 years of experience. Based on our observation and the expert’s opinion, the customized architecture not only outperforms the baseline RetinaNet network in treatability classification, but also provides insights about the features and regions considered significant by the network toward explaining reasons for treatment recommendation. Furthermore, by investigating the heatmaps on Gaussian-blurred images that serve as surrogates for out-of-focus cervical pictures we demonstrate the effect of image quality degradation on cervical treatability classification and underscoring the need for using images with good visual quality.

An Observational Study of Deep Learning and Automated Evaluation of Cervical Images for Cancer Screening

BACKGROUND

Human papillomavirus vaccination and cervical screening are lacking in most lower resource settings, where approximately 80% of more than 500 000 cancer cases occur annually. Visual inspection of the cervix following acetic acid application is practical but not reproducible or accurate. The objective of this study was to develop a "deep learning"-based visual evaluation algorithm that automatically recognizes cervical precancer/cancer.

METHODS

A population-based longitudinal cohort of 9406 women ages 18-94 years in Guanacaste, Costa Rica was followed for 7 years (1993-2000), incorporating multiple cervical screening methods and histopathologic confirmation of precancers. Tumor registry linkage identified cancers up to 18 years. Archived, digitized cervical images from screening, taken with a fixed-focus camera ("cervicography"), were used for training/validation of the deep learning-based algorithm. The resultant image prediction score (0-1) could be categorized to balance sensitivity and specificity for detection of precancer/cancer. All statistical tests were two-sided.

RESULTS

Automated visual evaluation of enrollment cervigrams identified cumulative precancer/cancer cases with greater accuracy (area under the curve [AUC] = 0.91, 95% confidence interval [CI] = 0.89 to 0.93) than original cervigram interpretation (AUC = 0.69, 95% CI = 0.63 to 0.74; P < .001) or conventional cytology (AUC = 0.71, 95% CI = 0.65 to 0.77; P < .001). A single visual screening round restricted to women at the prime screening ages of 25-49 years could identify 127 (55.7%) of 228 precancers (cervical intraepithelial neoplasia 2/cervical intraepithelial neoplasia 3/adenocarcinoma in situ [AIS]) diagnosed cumulatively in the entire adult population (ages 18-94 years) while referring 11.0% for management.

CONCLUSIONS

The results support consideration of automated visual evaluation of cervical images from contemporary digital cameras. If achieved, this might permit dissemination of effective point-of-care cervical screening.

Searching for an ideal cervical cancer screening model to reduce false-negative errors in a country with high prevalence of cervical cancer

The purpose of this study was to develop an ideal cervical cancer screening model to reduce false-negative errors in Korea where there is a high prevalence of cervical cancer. We conducted a cross-sectional study including 33,531 women who underwent routine cervical cancer screening in Korea. Colposcopic examinations were performed after abnormal results on their screening tests. Diagnostic capacities including sensitivity, specificity, and false-negative rate of each screening scenario were analysed at the CIN1 or worse (CIN1+) threshold with colposcopic biopsy results considered the gold standard. A total of 4117 women had valid results for Papanicolaou (Pap) cytology, human papilloma virus (HPV) tests, cervicography, and colposcopically directed biopsy were included in this study. The disease prevalence of CIN1+ was 38.1%. Pap-alone resulted in the highest false-negative rate of 46.9%, followed by HPV-alone at 25.1%, cervicography-alone at 18.7%, Pap/HPV-combined at 15.0%, Pap/cervicography-combined at 6.9% and Pap/HPV/cervicography-combined at 2.9% in a sample of 1570 women with CIN1+ lesions. Therefore, cervicography demonstrated excellent performance for the detection of CIN or cervical cancer and markedly reduced false-negative errors when used in combination with Pap cytology and HPV tests.IMPACT STATEMENTWhat is already known on this subject? False-negative rate of Pap smears is as high as approximately 40-50%. Limitations of the Papanicolaou (Pap) test have led to the development of new screening programmes for cervical cancer, such as combination screenings with human papillomavirus (HPV) tests or cervicography.What do the results of this study add? Pap-alone resulted in the highest false-negative rate of 46.9%, followed by HPV-alone at 25.1%, cervicography-alone at 18.7%, Pap/HPV-combined at 15.0%, Pap/cervicography-combined at 6.9% and Pap/HPV/cervicography-combined at 2.9% in a sample of 1570 women with CIN1+ lesions.What are the implications of these findings for clinical practice and/or further research? Cervicography demonstrated excellent performance for the detection of CIN or cervical cancer and markedly reduced false negative errors when used in combination with Pap cytology and HPV tests.

Biomarkers in Cervical Cancer

Cervical cancer, a potentially preventable disease, remains the second most common malignancy in women worldwide. Human papillomavirus (HPV) is the single most important etiological agent in cervical cancer, contributing to neoplastic progression through the action of viral oncoproteins, mainly E6 and E7. Cervical screening programs using Pap smear testing have dramatically improved cervical cancer incidence and reduced deaths, but cervical cancer still remains a global health burden. The biomarker discovery for accurate detection and diagnosis of cervical carcinoma and its malignant precursors (collectively referred to as high-grade cervical disease) represents one of the current challenges in clinical medicine and cytopathology.

The performance of tele-cervicography for detection of preinvasive and invasive disease of the uterine cervix as an adjunctive test to Pap smears

Aim of the study

To evaluate the diagnostic capacity of tele-cervicography for the detection of cervical neoplasia as an adjunctive test with Papanicolaou (Pap) smears.

Material and methods

Pap smear and tele-cervicography were performed on each subject. Histologic results were obtained for all patients.

Results

Of the 863 females who had a tele-cervigram, 252 (29.2%) had a positive result. Of the 60 histologically confirmed cases of high-grade squamous intraepithelial lesions (HSILs), 56 (93.3%) were detected by tele-cervicography, including 16 (26.7%) with a positive grade of 1 and 40 (66.7%) with a positive grade of 2. With the positive threshold of tele-cervicography set as any positive grade (P0 to P3), the overall sensitivity was 94.0% (95% CI: 88.0–97.3%), the specificity was 80.9% (95% CI: 80.0–81.5%), and the positive likelihood ratio was 4.94 (95% CI: 4.23–5.77) for the detection of HSILs or cancer. The combination of tele-cervicography with Pap smear testing for the detection of HSILs or cancer resulted in an increase in sensitivity from 84.6% (Pap only: cutoff = atypical squamous cells of undetermined significance or more severe) to 98.3% (Pap plus tele-cervicography: cutoff = P0 or more severe).

Conclusions

The sensitivity of tele-cervicography was higher than that of cytology for the detection of cervical neoplasia, and combining the two tests increased the sensitivity. Tele-cervicography can be considered a useful complementary tool to cytology.

Trends in cervical cancer screening in developing countries

Developing countries suffer the highest burden of cervical cancers but have the lowest resources. Effective cervical cytology screening programme, along with a network of diagnostic and therapeutic colposcopy centres, like developed countries, is almost impossible to be reproduced in developing countries. Visual inspection methods [e.g., Visual inspection with Lugol’s iodine (VILI) and Visual Inspection with Acetic Acid (VIA)] which are cheaper, require less expertise and have the advantage of possible treatment in one setting have been shown to be effective alternatives. The sensitivity to detect CIN2+, by VIA and VILI, have been shown to be 80% and 91% respectively, with a specificity rate of 92% and 85% respectively. Screening by human papillomavirus (HPV) testing has high sensitivity (96.4%) but low specificity (94.1%) to detect CIN2+, when compared to Pap Smear (sensitivity, 55.4% and specificity, 96.8%). A single lifetime HPV testing in a large unscreened population has been shown to significantly reduce cervical cancer incidence and mortality when compared to cervical cytology, VIA or no screening. HPV testing of self-collected vaginal specimens also helps to overcome religious and socio-cultural barriers towards pelvic examination amongst women in developing countries. Current HPV testing methods are expensive, skill/infrastructure demanding and takes time to produce results. A cheaper HPV test, called careHPV™, which is able to provide results within 2.5 h and requires minimal skill/infrastructure to operate, was designed for use in developing countries. One stop screen and treat facilities using VIA or rapid HPV testing, and cryotherapy, can overcome non-compliance to follow-up which is a major issue in developing countries. Cure rates of 81.4% for CIN1, 71.4% for CIN2 and 68.0% for CIN3 at 6 mo after treatment have been reported. Incorporating telemedicine with cervicography of VIA or VILI or even telecolposcopy, has great potential in cervical cancer screening, especially in countries with vast geographical areas.

Correlation between the digital cervicography and pathological diagnosis performed at private clinics in Korea

OBJECTIVE

To evaluate the correlation tendency between abnormal findings of digital cervicography and cervical pathology at private clinics in Korea.

METHODS

Abnormal finding of digital cervicography performed at private clinics in Korea between January 1, 2010 and May 31, 2012 were analysed retrospectively. The patient's age, abnormal findings of digital cervicography, cervical cytology, human papillomaviru (HPV) test and cervical pathology were investigated and the rate of agreement between abnormal finding of digital cervicography and cervical pathology results was calculated. Abnormal findings of digital cervicography were divided into 4 categories: atypical, compatible with CIN1, compatible with CIN2/3 and compatible with cancer.

RESULTS

The study group was composed of 1547 women with a mean (range) age of 37.4 (14-91 years). The agreement rate between abnormal findings of digital cervicography and cervical pathology was 52.0% in "compatible with CIN1", 78.9% in "compatible with CIN2/3", and 90.2% in "compatible with cancer".

CONCLUSIONS

Abnormal findings of digital cervicography were highly concordant with cervical intraepithelial neoplasia (CIN) and cancer examined at outpatient clinics in Korea. Therefore, abnormal interpretations of digital cervicography can be used as an excellent auxiliary technique with cervical cytology for CIN and cancer.

Long-term risk of recurrent cervical human papillomavirus infection and precancer and cancer following excisional treatment

Risk of recurrent CIN2+ (including cervical intraepithelial neoplasia grade 2 [CIN2], CIN3, carcinoma and in situ, adenocarcinoma in situ or cancer) remains elevated for years following treatment. The role of long-term post-treatment human papillomavirus (HPV) presence on subsequent risk of CIN2+ was evaluated in the 10,049-women Guanacaste cohort. Six hundred eighty-one women were referred to colposcopy because of high-grade cytology, positive cervicography and/or suspicion of cancer based on visual assessment; 486 were judged to require treatment. After excluding women with <12 months of follow-up (N = 88), prior cancer or hysterectomy (N = 37) or other reasons (N = 14), 347 were included in the analysis. Infections were categorized as persistent if present at both pre- and post-treatment visits and new if detected only post-treatment. Median time between the treatment and post-treatment visits was 6.7 years (interquartile range 3.8-7.8). At the post-treatment visit, 8 (2.4%), 2 (0.6%) and 8 (2.4%) of the 347 treated women had persistent HPV16, HPV18 or other carcinogenic HPV, respectively. Two (0.8%), 3 (1.0%) and 13 (4.0%) had new HPV16, HPV18 and other carcinogenic HPV, respectively. Six CIN2+ cases were identified at the post-treatment visit, all with persistent infections (three HPV16, one HPV18 and two other carcinogenic HPV). No recurrent disease was observed among women with new HPV infections during the follow-up period. Thus, persistence of HPV infection a median of six years after treatment was uncommon but, when present, posed a substantial risk of subsequent CIN2+. Serial follow-up data from other studies would further strengthen these conclusions.

Longitudinal study of human papillomavirus persistence and cervical intraepithelial neoplasia grade 2/3: critical role of duration of infection

BACKGROUND

The natural history of human papillomavirus (HPV) infections in older women is critical for preventive strategies, including vaccination and screening intervals, but is poorly understood. In a 7-year population-based cohort study in Guanacaste, Costa Rica, we examined whether women's age and the duration of carcinogenic HPV infections influenced subsequent persistence of infection and risk of cervical intraepithelial neoplasia grade 2 (CIN 2) or worse disease.

METHODS

At enrollment, of the 9466 participants eligible for pelvic examination, 9175 were screened for cervical neoplasia using multiple methods; those with CIN 2 or worse disease were censored and treated. Participants at low risk of CIN 2 or worse (n = 6029) were rescreened at 5-7 years (passively followed), whereas higher-risk participants (n = 2115) and subsets of low-risk women (n = 540) and initially sexually inactive women (n = 410) were rescreened annually or semiannually (actively followed) for up to 7 years. HPV testing was done using a polymerase chain reaction-based method. We determined, by four age groups (18-25, 26-33, 34-41, and > or =42 years), the proportion of prevalent infections (found at baseline) and newly detected infections (first found during follow-up) that persisted at successive 1-year time points and calculated absolute risks of CIN 2 and CIN grade 3 (CIN 3) or worse during follow-up. P values are two-sided.

RESULTS

Regardless of the woman's age, newly detected infections were associated with very low absolute risks of persistence, CIN 2, or worse disease. For newly detected infections, the rate of progression to CIN 2+ (or CIN 3+), after 3 years of follow-up, was not higher for women aged 34 years and older than for younger women. Moreover, rates of newly detected infections declined sharply with age (in the actively followed group, at ages 18-25, 26-33, 34-41, and > or =42 years, rates were 35.9%, 30.6%, 18.1%, and 13.5%, respectively; P < .001). Among prevalent infections, persistent infections among older women (> or =42 years) was higher than that among younger age groups or new infections at any age (P < .01 for comparison of eight groups). Most (66 of 85) CIN 2 or worse detected during follow-up was associated with prevalent infections. Only a small subset (25 of 1128) of prevalent infections persisted throughout follow-up without apparent CIN 2 or worse.

CONCLUSIONS

The rate of new infections declines with age, and new infections typically do not progress to CIN 2 or worse disease in older women; thus, overall potential benefit of prophylactic vaccination or frequent HPV screening to prevent or detect new carcinogenic HPV infections at older ages is low.

Treatability by cryotherapy in a screen-and-treat strategy

OBJECTIVES

We estimated the percentage of women infected with human papillomavirus (HPV+) who cannot be immediately treated with cryotherapy.

MATERIALS AND METHODS

In a 10,000-woman Costa Rican cohort, we analyzed the 559 HPV+ women aged 25 to 55 years and estimated the proportion for whom immediate cryotherapy was not indicated (i.e., invasive cancer, large precancerous lesions, or benign abnormalities that risk failure such as large ectopy, squamocolumnar junction not visualized, polyps, ulcers, or distorted or atrophied cervix). To determine whether cryotherapy at time of baseline HPV screening would effectively treat HPV+ women, 2 expert gynecologists independently judged entire clinical histories (5-7 years of cytology, histology, and HPV tests) and a full longitudinal series of digitized cervical images.

RESULTS

Reviewers judged 144 (25.8%) of 559 HPV+ women as not treatable by immediate cryotherapy. Among 72 women with cervical intraepithelial neoplasia grade 3 who would benefit most from a screening program, 35 (48.6%) were not treatable. In particular, 29 women (40.3%) were determined not treatable for reasons most likely associated with cryotherapy's inadequacy (lesion was large, suspected cancerous or in the endocervical canal or fornix).

CONCLUSIONS

"Screen-and-treat" programs in low-resource settings will soon use a rapid HPV test to screen older women once or twice in their lifetime, identifying women at higher risk for precancer. Our findings suggest that cryotherapy might not effectively treat many precancers, and other safe, low-technology treatment options could be required, in a scenario where all HPV+ women in this targeted group would receive cryotherapy at the same visit.

New approaches to cervical cancer screening in Latin America and the Caribbean

Cervical cancer remains an important public health problem in the Latin America and Caribbean region (LAC), with an expected significant increase in disease burden in the next decades as a result of population ageing. Prophylactic human papillomavirus (HPV) vaccine is currently unaffordable in LAC countries. However, even if vaccination was implemented, an additional two decades will be required to observe its impact on HPV related disease and cancer. With some exceptions, cytology-based screening programs have been largely ineffective to control the problem in the region, and there is a need for new approaches to the organization of screening and for use of newly developed techniques. Several research groups in LAC have conducted research on new screening methods, some of which are summarized in this paper. A recommendation to reorganize screening programs is presented considering visual inspection for very low resource areas, improvement of cytology where it is operating successfully and HPV DNA testing followed by visual inspection with acetic acid (VIA) or cytology as soon as this method becomes technically and economically sustainable. This could be facilitated by the incorporation of new, low-cost HPV DNA testing methods and the use of self-collected vaginal specimens for selected groups of the population. An important requisite for screening based on HPV testing will be the quality assurance of the laboratory and the technique by validation and certification measures.

The Natural History of Human Papillomavirus Infection and Cervical Intraepithelial Neoplasia Among Young Women in the Guanacaste Cohort Shortly After Initiation of Sexual Life

OBJECTIVE

Cross-sectional analyses of our 10,000-woman, population-based Guanacaste cohort suggest a lag of > or =10 years between the peak of human papillomavirus (HPV) infection and the later peak of cervical intraepithelial neoplasia grade 3 (CIN 3). We wanted to explore early HPV natural history and CIN 3 prospectively.

STUDY DESIGN

As part of the Guanacaste cohort, we followed 206 initially virginal women aged 18 to 26 semiannually for a median of 3.6 years after initiation of sexual life.

RESULTS

A total of 53.4% of women tested positive during the study for > or =1 HPV type. Very few infections persisted for >1 to 2 years. Three women had histologically confirmed CIN 3, of which 2 showed persistent HPV 16. The other had serologic evidence of HPV 31.

CONCLUSIONS

HPV infection occurs frequently and clears rapidly in most young women initiating sexual intercourse. Persistent HPV 16 can cause early CIN 3. The peak age for CIN 3 will decline with the increased screening intensity and sensitivity typical of longitudinal studies.

Cervical Cancer: The Indian Perspective

Cancer of the uterine cervix is the commonest gynaecologic cancer in India, with most women presenting with disease extending beyond the cervix. The majority of women belong to the lower socioeconomic status, are rural, aged between 35 and 64 years and highly noncompliant for complete treatment and follow-up. Opportunistic screening with cytology, colposcopy and test for Human Papilloma Virus and appropriate treatment are available on payment at urban private medical centres but are not available at urban and rural government health centres that are accessed by women of the lower socioeconomic status. The Government's investment in health is 0.9% of the GDP. Thus cytology screening as a government health measure is not feasible. The 'social vaccine' of health empowerment along with visual inspection and appropriate referral by the rural and urban health personnel (Department of Health and Family Welfare); with an additional input of health awareness and motivation by Anganwadi Workers (Department of Women and Child Development), elected women representatives in the Panchayats (Department of Rural Development and Panchayati Raj) and non-governmental development agencies could be a collaborative effort towards "downstaging" cervical cancer. This could lay the foundation for the introduction of cytology screening when resources are available.

The use of electrical impedance spectroscopy in the detection of cervical intraepithelial neoplasia

The objective of this study was to assess the performance of cervical impedance spectroscopy in the detection of cervical intraepithelial neoplasia (CIN) using the new MKIII impedance probe. A prospective observational study recruited women referred to colposcopy with an abnormal Papanicolaou smear. A pencil probe incorporating four gold electrodes was used to measure electrical impedance spectra from cervical epithelium. Colposcopy examinations, including probe positioning, were video recorded to allow for correlation between results obtained from colposcopic impression, histopathologic examination of colposcopic punch biopsies, and impedance measurements. Cervical impedance-derived parameters R, S, R/S, C, and Fc were assessed to see if significant difference in values obtained in CIN and normal epithelium existed. The performance of the probe in identifying women with CIN was also assessed. One hundred seventy-six women were recruited and 1168 points analyzed. Parameters R, S, and Fc showed significant separation of CIN or squamous intraepithelial lesion (SIL) from squamous, mature metaplastic, and columnar epithelium. Sensitivities of 74% and specificity of 53% can be achieved in identifying CIN 2/3 (High-grade SIL) in screened women. We conclude that cervical impedance spectrometry provides a potentially promising real-time screening tool for CIN with similar sensitivity and specificity to currently used screening tests. Further research is ongoing to develop the probe for potential clinical use.

Lack of agreement between cervicography and cytology and the effect of human papillomavirus infection and viral load

OBJECTIVE

Cohort studies of the natural history of human papillomavirus (HPV) infection and cervical squamous intraepithelial lesions with repeated screening allow the comparison of different macroscopic and microscopic diagnostic methods.

MATERIALS AND METHODS

Concurrent visual inspection using cervicography and conventional Pap cytology tests were performed during multiple visits in a cohort of women attending a maternal and child health clinic in São Paulo, Brazil. HPV infection status at the same visits was determined by polymerase chain reaction followed by typing with specific oligonucleotide probing and viral load quantification. Information on reproductive health and hygiene habits was also collected at each visit.

RESULTS

Overall agreement between cervicography and cytology was low (kappa = 0.046), which increased only slightly when high oncogenic-risk HPV types (kappa = 0.120) or high viral burden (>100 copies/cell) (kappa = 0.170) was present. Analysis of reproductive health and hygiene habits revealed somewhat different risk factors for cervical lesions detected by these tests. However, presence of oncogenic HPV DNA (odds ratio = 36.0, 95% CI = 16.6-77.8) and high viral burden (odds ratio = 67.34; 95% CI = 27.1-167.0) were strongly associated with lesions detected by cytology but not by cervicography.

CONCLUSIONS

Although changes in the cervix (because of age, gravidity, or hormonal effects) may influence the performance of morphology-based screening tests, the lack of agreement and the different degrees of association with HPV infection measures indicate that a visual inspection method such as cervicography may detect different cervical abnormalities relative to cytology.

Preparing digitized cervigrams for colposcopy research and education: determination of optimal resolution and compression parameters

OBJECTIVE

Visual assessment of digitized cervigrams through the Internet needs to be optimized. The National Cancer Institute and National Library of Medicine are involved in a large effort to improve colposcopic assessment and, in preparation, are conducting methodologic research.

MATERIALS AND METHODS

We selected 50 cervigrams with diagnoses ranging from normal to cervical intraepithelial neoplasia 3 or invasive cancer. Those pictures were scanned at 5 resolution levels from 1,550 to 4,000 dots per inch (dpi) and were presented to 4 expert colposcopists to assess image quality. After the ideal resolution level was determined, pictures were compressed at 7 compression ratios from 20:1 to 80:1 to determine the optimal level of compression that permitted full assessment of key visual details.

RESULTS

There were no statistically significant differences between the 3,000 and 4,000 dpi pictures. At 2,000 dpi resolution, only one colposcopist found a slightly statistically significant difference (p = 0.02) compared with the gold standard. There was a clear loss of quality of the pictures at 1,660 dpi. At compression ratio 60:1, 3 of 4 evaluators found statistically significant differences when comparing against the gold standard.

CONCLUSIONS

Our results suggest that 2,000 dpi is the optimal level for digitizing cervigrams, and the optimal compression ratio is 50:1 using a novel wavelet-based technology. At these parameters, pictures have no significant differences with the gold standard.

Systems analysis of real-world obstacles to successful cervical cancer prevention in developing countries

Papanicolaou screening is feasible anywhere that screening for cervical cancer, the leading cause of cancer-related death among women in developing countries, is appropriate. After documenting that the Vietnam War had contributed to the problem of cervical cancer in Vietnam, we participated in a grass roots effort to establish a nationwide cervical cancer prevention program in that country and performed root cause analyses of program deficiencies. We found that real-world obstacles to successful cervical cancer prevention in developing countries involve people far more than technology and that such obstacles can be appropriately managed through a systems approach focused on programmatic quality rather than through ideological commitments to technology. A focus on quality satisfies public health goals, whereas a focus on technology is compatible with market forces.

Digitized cervical images: problems, solutions, and potential medical impact

OBJECTIVE

To demonstrate compression, illumination enhancement, registration, segmentation, automated classification and steganography using digitized cervical images.

MATERIALS AND METHODS

The Hybrid Multi-Scale Vector Quantization algorithm developed at Texas Technological University and other automated systems were used to improve digitized cervical images.

RESULTS

We demonstrated high levels of image compression, illumination enhancement, registration, automated segmentation and classification and steganography of digitized cervical images.

CONCLUSIONS

Digitized cervical images can be altered to facilitate research of cervical neoplasia.

Cervicography as an adjunctive test to visual inspection with acetic acid in cervical cancer detection screening

OBJECTIVE

Validate Digital Cervicography as adjunctive cervical cancer screening test to VIA and Cytology.

METHOD

Women (1292) were submitted to both cytological and VIA tests to obtain a sample of 301 positive VIA cases (23.3%) which have had the Digital Cervicography (DC) taken and evaluated. Just cases considered positive by DC (149/301 cases--49.5%) and/or positive Pap smear tests were referred to colposcopy and biopsy whenever indicated.

RESULTS

Cervical smear was positive in 5.4%, including LGSIL (4%), HGSIL (1%) and one case of invasive cancer (0.4%), and 1.6% of ASCUS or AGUS). DC identified 81 positive cases (74 LGSIL; 5 HGSIL; 2 cancers). The sensitivity of cytology was 14.8%, and the DC was 100%. Corresponding specificities were 95.4% and 69.1%, respectively. The NPV and PPV for DC were 100% and 54.4%.

CONCLUSION

DC increases sensitivity, specificity and positive predictive value of VIA, becoming a valid adjunctive screening test for cervical cancer in low resource settings.

Combining human papillomavirus testing or cervicography with cytology to detect cervical neoplasia

CONTEXT

Cervicography and oncogenic human papillomavirus (HPV) testing have been proposed for improving the accuracy of cervical cancer screening.

OBJECTIVE

To examine whether cervicography and HPV testing can improve beyond chance the detection of cervical intraepithelial neoplasia (CIN) 2 or 3 in women with atypical cells of undetermined significance or low-grade squamous intraepithelial lesions on cytology.

DESIGN

Cross-sectional analysis. Oncogenic HPV testing by Hybrid Capture II assay or cervicography combined with cytology was compared with the reference standard of colposcopy with directed biopsy.

SETTING

Community family practices.

PARTICIPANTS

Three hundred four women with low-grade cytologic abnormality.

MAIN OUTCOME MEASURES

The gain in accuracy for detecting histologic CIN 2 or 3 or carcinoma. Because an adjunct test may improve sensitivity by chance alone, the sensitivity or specificity if the second test performed randomly was estimated.

RESULTS

Cervical intraepithelial neoplasia 2 or 3 was found in 11.8% (36/304) of the women and invasive squamous cell carcinoma in 0.3% (1/304). The sensitivity of cytology for detecting CIN 2 or 3 was 73.0% and increased by 21.6% to 94.6% with the addition of a cervigram showing a low-grade lesion or higher or a positive HPV test result. These gains were reduced to 8.1% and 10.8% above the sensitivities expected if the additional tests performed randomly. The corresponding specificities decreased from 49.1% to 32.2% and 33.0%. There was insufficient power to determine whether observed sensitivities were statistically significantly higher than the expected sensitivities.

CONCLUSION

Adjunctive HPV testing or cervicography may provide similar gains in sensitivity, but they can appear misleadingly large if chance increases are not taken into account.

Comparison of pap smear, visual inspection with acetic acid, human papillomavirus DNA-PCR testing and cervicography

OBJECTIVE

To assess the test qualities of four screening methods to detect cervical intra-epithelial neoplasia in an urban African setting.

METHOD

Six hundred fifty-three women, attending a family planning clinic in Nairobi (Kenya), underwent four concurrent screening methods: pap smear, visual inspection with acetic acid (VIA), PCR for high risk human papillomavirus (HR HPV) and cervicography. The presence of cervical intra-epithelial neoplasia (CIN) was verified by colposcopy or biopsy.

RESULT

Sensitivity (for CIN2 or higher) and specificity (to exclude any CIN or cancer) were 83.3% (95% CI [73.6, 93.0]) and 94.6% (95% CI [92.6, 96.5]), respectively, for pap smear; 73.3% (95% CI [61.8, 84.9]) and 80.0% (95% CI [76.6, 83.4]) for VIA; 94.4% (95% CI [84.6, 98.8]) and 73.9% (95% CI [69.7, 78.2]) for HR HPV; and 72.3% (95% CI [59.1, 85.6]) and 93.2% (95% CI [90.8, 95.7]) for cervicography.

CONCLUSION

The pap smear had the highest specificity (94.6%) and HPV testing the highest sensitivity (94.4%). The visual methods, VIA and cervicography, were similar and showed an accuracy in between the former two tests.

Chapter 11: Future directions in cervical pathology

Technologic innovations, including the development of improved sampling devices, liquid-based collection systems, and computer-assisted screening, have revolutionized cervical cytology. Biologic discoveries in upcoming years promise to transform our understanding of the pathogenesis of cervical neoplasia, leading to another quantum leap in our approach to screening and prevention. This review summarizes the opportunities and challenges that recent and anticipated advances in pathology present for epidemiologic research.

Chapter 10: Cervical cancer screening using visualization techniques

There is a resurgence of interest in the use of visual techniques to identify cervical intraepithelial neoplasia (CIN). These visual techniques can be divided into two general categories. One is the simple visual screening method, such as direct visual inspection (DVI), during which the cervix is visualized with either the naked eye or a low-power magnification device after the application of a solution of 3% to 5% acetic acid that is used as a chemical contrast agent to highlight regions of CIN. DVI has been evaluated in a number of large clinical trials and is considered by some to be a possible alternative to cervical cytology for primary cervical cancer screening in low-resource settings. The advantages of DVI compared with cervical cytology for these settings are that it is inexpensive, it does not require a laboratory infrastructure, and it provides an immediate result, allowing the use of "screen and treat" protocols. The major disadvantage of DVI is that it is relatively nonspecific and that its sensitivity is low compared with testing for human papillomavirus. The other category of visual techniques includes devices that use electro-optical sensors and light of specific wavelengths produced by lasers or specialized light sources to identify and localize regions of CIN on the cervix. Although these "high-technology" devices are not yet in routine clinical use, several groups and companies have such devices in clinical trials.

Can cervicography be improved? An evaluation with arbitrated cervicography interpretations

OBJECTIVE

The purpose of this study was to estimate the optimal performance of cervicography. We compared an arbitrated cervigram classification with an arbitrated referent diagnosis of cervical neoplasia.

STUDY DESIGN

From an initial group of 8460 women, a stratified sample of cervigrams from 3645 women and histologic information from 414 women underwent arbitration. Interobserver agreement was assessed for cervicography and the referent diagnosis. Sensitivity, specificity, and predictive values were estimated for initial and arbitrated cervicography results, compared with the initial and arbitrated referent diagnoses.

RESULTS

For the detection of arbitrated high-grade lesions or cancer, arbitrated cervicography yielded an overall sensitivity of 63.9% and a specificity of 93.7%. Significantly higher sensitivity was associated with younger age and age-related visual characteristics.

CONCLUSION

Optimization of the cervigram classification improved performance over a single interpretation in this population but suggested the limits of static visual screening.

Effects of age and human papilloma viral load on colposcopy triage: data from the randomized Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesion Triage Study (ALTS)

BACKGROUND

Testing for oncogenic human papillomavirus (HPV) DNA at a 1.0-pg/mL threshold represents a promising approach for colposcopy triage of atypical squamous cells of undetermined significance (ASCUS), but not for low-grade squamous intraepithelial lesions (LSIL). Considering age or viral load could improve colposcopy triage.

METHODS

We determined the sensitivity for detecting Cervical Intraepithelial Neoplasia 3 (CIN3) and cancer and the percentage of referrals for colposcopy using HPV testing and repeat thin-layer cytopathology in 2198 women with ASCUS and in 848 women with LSIL enrolled in ALTS from November 1996 through December 1998. We analyzed results by age and at two thresholds for HPV load and repeat cytopathology.

RESULTS

For ASCUS, the overall sensitivity of HPV testing at 1.0 pg/mL was 96.1% (95% confidence interval [CI] = 92.8 to 99.5%) and varied minimally with age (range, 93.9% to 97.8%). HPV testing at this threshold would refer 31.2% (95% CI = 28.0% to 34.3%) of women aged 29 years or older as compared with more than 65% of younger women. Among women aged 29 years or older with ASCUS, referral for repeat cytopathology of ASCUS had a sensitivity of 90.9% (95% CI = 81.1% to 100.0%) and would refer 50.1% (95% CI = 46.7 to 53.5%). Among all ASCUS, HPV testing using a 10.0-pg/mL threshold decreased sensitivity to 91.5% and referrals to 41.7%. More than 63% of LSIL would have been referred using any strategy achieving 90% sensitivity.

CONCLUSION

For women with ASCUS, HPV testing was highly sensitive for detecting CIN3 and cancer with dramatically fewer referrals of older women. Neither a single HPV test nor repeat cytopathology provides useful triage for women with LSIL.

Cervicography for triage of women with mildly abnormal cervical cytology results

OBJECTIVE

To estimate the sensitivity and specificity of cervicography in detecting cervical cancer precursor lesions in women participating in the National Cancer Institute's multicenter atypical squamous cells of undetermined significance and low-grade squamous intraepithelial lesion triage study (ALTS).

STUDY DESIGN

Cervigrams were obtained from 3134 women with a referral Papanicolaou smear diagnosis of atypical squamous cells of undetermined significance or low-grade squamous intraepithelial lesion. Cervigram and cervical histology results were compared by using cervical intraepithelial neoplasia (CIN) 2 and CIN 3 disease end points.

RESULTS

Of 3134 women, 444 had histologic findings of more than or equal to CIN 2 and 222 had CIN 3. Cervicography interpretation by using a threshold of greater than or equal to atypical had a sensitivity, specificity, and positive and negative predictive values of 79.3%, 61.0%, 13.4%, and 97.5%, respectively, for detecting greater than or equal to CIN 3. Cervicography was more sensitive (80.8% vs 57.1%), but less specific (55.7% vs 81.8%), for detecting CIN 3 in women younger than 35 years compared with women 35 years or older, respectively.

CONCLUSIONS

Cervicography functioned moderately well at detecting CIN 2 or CIN 3 in women with atypical squamous cells of undetermined significance and low-grade squamous intraepithelial lesion Papanicolau smear results. Cervigram sensitivity was better for younger women.

An update on human papillomavirus infection and Papanicolaou smears in adolescents

Human papillomavirus infection is one of the most common sexually transmitted infections in young women, and may lead to clinical sequelae such as anogenital condylomata and cervical squamous cell carcinoma. Recent data on the biology and natural history of HPV infection in adolescents will have important implications for the development of adolescent-specific protocols for cervical cancer screening and for follow-up of abnormal cytology.

New technologies in cervical cancer precursor detection

The current literature reflects three routes toward improving cervical cancer screening. The first is to improve the test qualities of cytology-based screening. The use of liquid-based cytology and computerized analysis of Papanicolaou tests are examples of attempts at this approach. Secondly, through various combinations of parallel or sequential tests, either the sensitivity or the specificity of a given test could be improved depending on the tests chosen and the order in which they were performed (eg, Papanicolaou test followed by human papillomavirus [HPV] or vice versa). Several excellent studies have been published this year on the use of HPV DNA testing as a primary screening modality and as an adjunct to the triage of mildly abnormal cytologic findings. The recent literature also reflects increasing interest in visual inspection of the cervix and self-collected samples for HPV testing as an equally effective and viable alternative to cytology in low-resource settings. A third possibility is to make use of advances in digital and spectroscopic techniques. In these cost-conscious times, a significant number of articles address the cost-effectiveness of these technologies and the real value of cervical cancer screening. This article reviews the current literature concerning both the advent of new cervical cancer screening technologies and the rediscovery of old ones.

Population-based study of human papillomavirus infection and cervical neoplasia in rural Costa Rica

BACKGROUND

Human papillomavirus (HPV) is the main cause of cervical neoplasia. Because few population-based studies have investigated the prevalence of type-specific infection in relation to cervical disease, we studied a high-risk population, estimating the prevalence of HPV infection and the risk associated with various HPV types.

METHODS

We screened 9175 women in Guanacaste, Costa Rica, to obtain a referent standard final diagnosis, and tested 3024 women for more than 40 types of HPV with a polymerase chain reaction-based system.

RESULTS

Among women with normal cytology, HPV infections peaked first in women younger than 25 years, and they peaked again at age 55 years or older with predominantly non-cancer-associated types of HPV and uncharacterized HPV types. Low-grade squamous intraepithelial lesions (LSILs) (n = 189) decreased consistently with age. The prevalence of high-grade squamous intraepithelial lesions (HSILs) (n = 128) peaked first around age 30 years and again at age 65 years or older. Seventy-three percent of LSILs were HPV positive, with HPV16 being the predominant type (16% of positive subjects). HPV was found in 89% of HSILs and 88% of cancers, with HPV16 being strongly predominant (51% and 53% of positive subjects). Virtually all HSILs and cancers had cancer-associated HPV types, with high odds ratios (ORs) and attributable fractions around 80%. Risk for HPV16 was particularly high (OR for HSILs = 320, 95% confidence interval [CI] = 97-1000; OR for cancer = 710, 95% CI = 110-4500).

CONCLUSIONS

We confirm the early decline of HPV infection with age but note increased prevalence after menopause, which could be related to a second peak of HSILs, an observation that warrants further investigation. At least 80% of HPVs involved in cervical carcinogenesis in this population have been characterized. Polyvalent vaccines including the main cancer-associated HPV types may be able to prevent most cases of cervical disease in this region.