Cervical cytology has helped to reduce cervical cancer mortality in the developed world,1,2 but the lack of trained personnel and limited laboratory and patient-recall infrastructure has hindered implementation of cytology-based screening in much of the developing world. Visual inspection with acetic acid (VIA) is a low-cost alternative that can be performed by nonphysician health providers and has become a popular screening option in resource-constrained countries.3
HIV-infected women in developing countries are a high-risk group for cervical cancer, particularly with longer life spans on affordable antiretroviral therapy, but generally have little or no access to quality cervical cancer screening services.4,5 Cytology and VIA-based screening have been compared in several studies,6,7 but few have focused on HIV-infected women,8–10 and none of the studies in HIV-infected women have evaluated VIA enhanced by digital cervicography (DC). DC is an adjunct to VIA and involves digital photography of the cervix, using a commercial brand camera, to allow for magnified visualization of surface morphology, while also facilitating telemedicine support, patient and provider education, and quality assurance of screening.11
Zambia has a particularly high burden of cervical cancer, with the second highest incidence and highest mortality rates in the world.1,2 The Cervical Cancer Prevention Program in Zambia (CCPPZ), a public-sector initiative, offers nurse-led services with DC with same-day cryotherapy for eligible precancerous lesions, or referral for loop electrosurgical excision procedure (LEEP) treatment for cryotherapy-ineligible lesions.12,13 Surgical, radiation, and chemotherapy services for the management of invasive cervical cancer are offered through Zambian Ministry of Health facilities.
To assess the clinical performance of DC, a resource-appropriate screening technology, as well as cytology in HIV-infected women, we enrolled HIV-infected women in Zambia and calculated the clinical performance of each screening test to detect cervical lesions on histopathology.
Participants were enrolled between January 2008 and December 2011 from Matero public health clinic in Lusaka. After counseling by a nurse provider, HIV-infected women were invited to participate in the study if they were nonpregnant by self-report, aged between 20 and 45 years, and deemed healthy enough to undergo a pelvic examination (assessed by the nurse enrolling for the study and defined as patients who were not bedridden or physically incapacitated and were mobile enough to undergo a pelvic examination without discomfort). Informed consent was obtained from all participants, and a nurse-administered questionnaire was used to collect sociodemographic data.
Trained, experienced nurses performed the study procedures, starting with the collection of cervical specimens for thin layer cytology using a cytobrush (for endocervical sampling) and an Ayre spatula (for ectocervical sampling). Both the cytobrush and spatula were rinsed in PreservCyt vials (Cytyc Corporation, Marlborough, MA) and stored at room temperature locally for <4 weeks before batched-shipping to a US-based laboratory for processing, analysis, and interpretation by a certified senior cytotechnologist according to the revised (2001) Bethesda classification system. All abnormal slides and 10% of normals were subsequently reviewed by a board-certified senior cytopathologist.
Immediately after the collection for cytology, the nurse conducted VIA enhanced by DC, performed by washing the cervix with 5% acetic acid, waiting for 2–3 minutes, and evaluating acetowhite lesions by real-time digital imaging of the cervix.11 To capture the DC images, the study nurse used a 7–8 megapixel digital camera with 10× optical zoom and a built-in flash. The image was reviewed in real-time, and the results of the DC were recorded as being positive or negative. Next, the nurse performed DC-directed cervical punch biopsies with a 2 × 4 mm tip Tischler biopsy forceps. A biopsy was taken from the lesion that appeared to have the most advanced degree of neoplasia, and from a normal appearing area of the cervical transformation zone. If the cervix had no abnormal area, only a normal area biopsy was taken of the transformation zone; conversely, if the cervix had no normal area, only an abnormal area biopsy was taken. Biopsy specimens were immediately placed in 10% formalin and sent to the pathology department of the University Teaching Hospital in Lusaka, Zambia for review by a UK-trained, board-certified Zambian senior pathologist. A combined histopathology variable was created to represent the most severe diagnosis from the normal and abnormal areas for each woman.
Patients with cervical intraepithelial neoplasia (CIN) grade 2 or 3 on biopsy underwent therapeutic LEEP. Women with evidence of invasive cancer on biopsy were immediately referred to the University Teaching Hospital in Lusaka for further management.
Clinical and pathology data were entered into a Microsoft Access (Microsoft Corporation, Redmond, WA) database and cleaned using Microsoft Excel and SAS version 9.2 (SAS Institute Inc., Cary, NC). SAS and Open Epi (www.openepi.com) were used to calculate the point estimates and 95% confidence intervals (CIs) of sensitivity, specificity, positive predictive value, and negative predictive value of DC and cytology. DC results were dichotomized as positive and negative, whereas cytology results were dichotomized at 3 clinically-relevant cutoffs: atypical squamous cells of undetermined significance or worse (ASC-US+), low-grade squamous intraepithelial lesions or worse (LSIL+), or high-grade squamous intraepithelial lesions or worse (HSIL+).
Ethical approval for this study was obtained from the University of Zambia Biomedical Research Ethics Committee and the University of Alabama at Birmingham Institutional Review Board (affiliation of CIDRZ at the time of this study).
We enrolled 303 women into the study; all women were screened by both cytology and DC, and had histopathology results from a punch biopsy. The median age was 32 years, 10.6% had completed high school, and 61.8% were married (Table 1). A total of 86.4% were antiretroviral experienced, and 56.5% had a baseline CD4+ count <200 cells per cubic millimeter.
Half of all women (50.5%) screened positive by DC, and nearly half (45.5%) of all women had HSIL+ (Table 1). Using the most severe histopathologic diagnosis from the individual biopsy results for each woman, 63.7% of women had CIN1 or worse (CIN1+), 20.1% had CIN2+, and 10.9% had CIN3+ lesions (Table 1).
The sensitivity of DC for identifying CIN2+ was 84% (95% CI: 72 to 91), and the specificity was 58% (95% CI: 52 to 64) (Table 2). The sensitivity estimates of cytology for identifying CIN2+ were as follows: HSIL+, 61% (95% CI: 48 to 72); LSIL+, 90% (95% CI: 80 to 95); and ASC-US+, 100% (95% CI: 94 to 100). The specificity estimates of cytology for identifying CIN2+ were HSIL+, 58% (95% CI: 52 to 64); LSIL+, 35% (95% CI: 29 to 41); and ASC-US+, 13% (95% CI: 10 to 18). The positive predictive values were low (23% to 33%) for both tests, whereas the negative predictive values were correspondingly high (86 to 100). A similar pattern of results was observed at the CIN3+ diagnostic threshold on histopathology (Table 2).
We have demonstrated that among HIV-infected women in Zambia, the point estimates for sensitivity of DC to detect CIN2+ and CIN3+ lesions were higher than those of cytology at the HSIL+ cutoff. Although previous studies have reported that VIA has higher sensitivity than cytology for both HIV-uninfected women6,7 and HIV-infected women,8–10 our study is the first to provide estimates of the clinical performance of DC.
The sensitivity point estimate of DC for CIN2+ that we report (84%) is slightly higher than 3 previous studies of HIV-infected women who reported 65%–80% sensitivity for unaided VIA.8–10 The specificity point estimate of DC for CIN2+ that we report (58%) lies near the lower end of the range (51%–83%) reported for unaided VIA in these studies.8–10 The specificity point estimate of cytology for CIN2+ that we report (58%) is slightly lower but comparable with that of Mabeya et al9 (66%), while both are substantially lower than that reported by Sahasrabuddhe et al8 (83%). Our lower specificity of cytology could be because our histopathology gold standard was based solely on punch biopsy specimens. Punch biopsies are small, and in women who screen DC positive, the punch biopsies could lead to underascertainment of the true amount of cervical disease if the lesion is not adequately sampled in the (relatively small) punch biopsy specimen. Our histopathology specimens, and that of Mabeya et al, were from punch biopsy alone, whereas those of Sahasrabuddhe et al were based on real-time colposcopically guided cervical punch biopsies, endocervical curettage, and LEEP, which result in a more extensive sampling of at-risk areas on the cervix.
Strengths of our study include the number of women enrolled, leading to relatively precise estimates of test performance characteristics. In addition, all women had a punch biopsy taken, and while biopsy placement was guided by DC impression, biopsies were also obtained from normal appearing areas of the cervix. Thus, histopathology was obtained regardless of DC or cytology test results, and we have minimized (if not eliminated) any verification bias that can result from only performing histopathology on screen-positive women.
Our clinical performance point estimates suggest that DC is as good as or better than cytology for identifying cervical lesions in our population of HIV-infected women, whereas the relatively lower specificity point estimate of DC (58%) likely leads to overtreatment and/or over-referral of women who, based on the CCPPZ clinical protocol, require excisional biopsy (LEEP) or diagnostic biopsy. The program scale-up in CCPPZ has used the advantage of the reasonably high sensitivity of DC,14 while overtreatment with cryotherapy is a lesser concern because this treatment modality has been shown to be a safe and acceptable treatment method.15 Nevertheless, the integration of other screening tests, such as point-of-care human papillomavirus DNA or E6 tests, either individually or in combination with DC may improve both the sensitivity and specificity of cervical cancer screening in HIV-infected women, and thus merits investigation.
The authors thank Suzanne Werneke from Hologic Corporation for her assistance in providing ThinPrep supplies and processing the ThinPrep specimens.
1. Ferlay J, Shin HR, Bray F, et al.. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer. 2010;127:2893–2917.
3. Sahasrabuddhe VV, Parham GP, Mwanahamuntu MH, et al.. Cervical cancer
prevention in low- and middle-income countries: feasible, affordable, essential. Cancer Prev Res (Phila). 2012;5:11–17.
4. Franceschi S, Jaffe H. Cervical cancer screening
of women living with HIV infection: a must in the era of antiretroviral therapy. Clin Infect Dis. 2007;45:510–513.
5. De Vuyst H, Lillo F, Broutet N, et al.. HIV, human papillomavirus, and cervical neoplasia and cancer in the era of highly active antiretroviral therapy. Eur J Cancer Prev. 2008;17:545–554.
6. Aggarwal P, Batra S, Gandhi G, et al.. Comparison of Papanicolaou test with visual detection tests in screening
for cervical cancer
and developing the optimal strategy for low resource settings. Int J Gynecol Cancer. 2010;20:862–868.
7. Sankaranarayanan R, Nessa A, Esmy PO, et al.. Visual inspection methods for cervical cancer
prevention. Best Pract Res Clin Obstet Gynaecol. 2012;26:221–232.
8. Sahasrabuddhe VV, Bhosale RA, Kavatkar AN, et al.. Comparison of visual inspection with acetic acid and cervical cytology
to detect high-grade cervical neoplasia among HIV-infected women in India. Int J Cancer. 2012;130:234–240.
9. Mabeya H, Khozaim K, Liu T, et al.. Comparison of conventional cervical cytology
versus visual inspection with acetic acid (VIA) among HIV-infected women in Western Kenya. J Low Genit Tract Dis. 2012;16:92–97.
10. Firnhaber C, Mayisela M, Mao L, et al.. Validation of cervical Cancer screening
methods in HIV positive women from Johannesburg South Africa. PLoS One. 2013;8:e53494.
11. Parham GP, Mwanahamuntu MH, Pfaendler KS, et al.. eC3–a modern telecommunications matrix for cervical cancer
prevention in Zambia. J Low Genit Tract Dis. 2010;14:167–173.
12. Mwanahamuntu MH, Sahasrabuddhe VV, Pfaendler KS, et al.. Implementation of “see-and-treat” cervical cancer
prevention services linked to HIV care in Zambia. AIDS. 2009;23:N1–N5.
13. Mwanahamuntu MH, Sahasrabuddhe VV, Kapambwe S, et al.. Advancing cervical cancer
prevention initiatives in resource-constrained settings: insights from the Cervical Cancer
Prevention Program in Zambia. PLoS Med. 2011;8:e1001032.
14. Mwanahamuntu MH, Sahasrabuddhe VV, Blevins M, et al.. Utilization of cervical Cancer screening
services and Trends in screening
Positivity rates in a “Screen-And-Treat” program Integrated with HIV/AIDS
care in Zambia. PLoS One. 2013;8:e74607.
15. Pfaendler KS, Mwanahamuntu MH, Sahasrabuddhe VV, et al.. Management of cryotherapy-ineligible women in a “screen-and-treat” cervical cancer
prevention program targeting HIV-infected women in Zambia: lessons from the field. Gynecol Oncol. 2008;110:402–407.