The incidence of anal cancer in the general population has been steadily increasing since the 1970s, with current rates in men and women of 1.2 and 1.8 per 100,000, respectively.1 Although this is an important trend, it should be appreciated that the rates of anal cancer in specific demographic groups are of greater concern. Specifically, men who have sex with men (MSM) and both HIV-positive MSM and HIV-positive women have rates of 35/100,000, 131/100,000, and 30/100,000, respectively.2,3 Broadly available interventions to address these cancer rates are urgently required.
Anal squamous cell cancer is causally related to persistent infection with high-risk human papillomavirus (HPV) in a manner analogous to that of cervical squamous cell cancer. Other similarities between these 2 cancers include the evolution of increasing grades of dysplastic change in tandem with the increasing presence of chromosomal abnormalities (3q), before malignant transformation.4,5 This progression from initial HPV infection to low-grade (L) and then high-grade squamous intraepithelial lesions (HSILs) and eventually cancer takes years in most cases. There is therefore an opportunity to intervene to abrogate progression to cancer, as has been effectively demonstrated in the cervix after the introduction of cervical cytology screening programs and colposcopy with ablative treatment of cervical HSIL.6 Thus, any intervention to reduce anal cancer is predicated on the identification of anal HSIL.
Similar to the cervix, exfoliative cytology has been used as a screening method to diagnose anal SIL, and in the most frequently followed clinical management algorithm, abnormal anal cytology is routinely followed by high-resolution anoscopy (HRA). High-resolution anoscopy is a similar procedure to cervical colposcopy and involves the application of 3% to 5% acetic acid to the anal canal to induce an acetowhite change in HPV-infected epithelium and facilitate detection of blood vessel abnormalities. These abnormalities such as “mosaicism” and “punctation” are similar to those seen in the cervix and are indicative of anal HSIL, which can then be biopsied to both confirm the diagnosis and exclude microinvasive disease before ablative treatment is undertaken.7 Internal anal SIL is most frequently seen at the anorectal transition zone that may be biopsied without the use of local anesthetic because of an absence of the somatic sensory nerves that innervate the more distal anal canal.
After a diagnosis of anal HSIL, there are a number of ablative treatment options available that include cold scalpel excision, laser ablation, electrofulguration, infrared coagulation (IRC), and topical methods such as 5 fluorouracil, 80% trichloroacetic acid (TCA), and the immune modifier imiquimod.8–11 These treatment methods vary greatly in their efficacy in the small number of published reports available. The interventions also vary in their need for specialized equipment such as an IRC, hyfrecator, or an operating room and anesthetic support, all of which impact the cost. Procedural cost is an important factor to consider because in most studies of the natural history and treatment of HSIL, there are high rates of both recurrent and new HSILs that will require treatment.12 This necessitates repeated ablation procedures over what will currently be a lifetime in the case of HIV-positive individuals, given their risk profile and the limited impact of combined antiretroviral therapy to alter the natural history of dysplastic change.13,14 Furthermore, the more invasive operating room procedures may necessitate time off work and place the patient at risk for procedural adverse effects such as bleeding and infection, which may lead to scarring, altered anal sensation, and, rarely, stenosis or anal incontinence. These adverse outcomes are particularly germane in a population that uses the anal canal for sexual pleasure.
The University of Pittsburgh Anal Dysplasia Clinic (UPADC) was established in 2008 at the Pittsburgh AIDS Center for Treatment. The UPADC instituted a treatment program for all patients with anal HSIL to initially receive treatment with topical TCA based on a previous efficacy study,10 low billing cost relative to other modalities, and low side effect profile based on the author’s (R.D.C.) clinical experience. The method of TCA application in this study varied from the previous report of TCA efficacy to treat internal anal HSIL, as is described below.
Approval for this retrospective chart review study was obtained from the University of Pittsburgh Institutional Review Board. The UPADC received referrals of HIV-positive male and female Pittsburgh AIDS Center for Treatment patients older than 18 years with abnormal anal cytology. These patients were seen by 1 of 2 certified HRA providers (R.D.C. and J.R.B.). Initial assessment included HRA and biopsy of any lesion with characteristics (acetowhite change, mosaicism, punctation, absent Lugol iodine uptake) consistent with HSIL. Images of the lesion(s) were captured using Second Opinion (Second Opinion Telemedicine Solutions, Inc, Torrance, CA) image capture software that facilitated reidentification of the index lesion during follow-up. Patients with HSIL were treated and then, similar to all other patients with abnormal anal cytology, followed up every 3 to 6 months with HRA ± anal biopsy.
At the treatment visit, each HSIL area was identified as previously described. Individual lesion(s) that did not extend individually beyond 25% or collectively beyond 50% of the anal circumference was treated with TCA. This restriction was based on the theoretical possibility of inducing scarring if extensive areas were treated with TCA. Five Q-tip cotton heads were dipped in TCA and applied sequentially one after the other to the lesion and a small collar of surrounding normal anal tissue. Care was taken not to overload the Q-tip with TCA so as to allow a controlled application to the anal epithelium without dripping. Patients were informed that no symptoms beyond what is expected from HRA with biopsy (discomfort and bleeding with bowel movements for up to 7 days) should be expected and, if they were concerned with any emergent symptom, to call the covering physician (R.D.C.).
Follow-up was arranged for between 3 and 6 months after the initial treatment (the time of follow-up varied because of the introduction of practice rules by the University of Pittsburgh Medical Center that only allowed HRA screening to be performed every 6 months). At review, the entire anal epithelium was examined by HRA as previously described, and the index lesion site was identified with reference to previously captured images and chart note descriptions.
The UPADC database was searched using the following criteria: male or female HIV-positive patients with biopsy-proven HSIL who had TCA treatment only and had at least 1 follow-up HRA visit at least 3 months after the initial ablation procedure between July 1, 2009, and June 30, 2012. High-grade squamous intraepithelial lesion was defined as the presence of anal intraepithelial neoplasia (AIN) 2 or AIN 3 on biopsy reports. Only patients who were treated with TCA alone during the reporting period were included in the analysis; others with concurrent use of imiquimod or any other modality for the treatment of HSIL such as IRC were excluded. Lesion resolution was defined as the presence of LSIL or normal epithelium on anal biopsy or by the absence of visual markers of HSIL, as defined above, at the site of the index lesion. Data addressing recurrence or evolution of new HSILs during follow-up was censored in July 1, 2013
The data set included the presence of multiple lesions in the same patient. As a result, characteristic factors were divided into 2 levels (patient-level and lesion-level). Factors at the patient-level included basic demographic factors (age and smoking status), HIV-related information (HIV treatment status, CD4 T-cell count, and plasma HIV-1 viral load [VL] at diagnosis), and number of HSIL lesions at diagnosis. Factors at the lesion level included the histology type at diagnosis, number of TCA treatments given, and location of the lesion. A summary of the basic statistics of all patient level factors (e.g., age, smoking status, and HIV surrogate markers) was calculated among all patients in the study. Sample mean and standard deviation (SD) was reported for continuous variables. Sample median was also reported for data with a highly skewed distribution. For categorical variables, the percentage of each category was reported. Moreover, selected factors were compared between 2 outcome groups (lesion resolved and not resolved) using the generalized estimating equation (GEE) model, a common method to deal with correlated data. For continuous factors (age, VL, CD4) and categorical factors with 2 groups (smoking), we fit the univariate GEE model with the lesion resolution status as the covariate and each factor as the response variable. However, for categorical factors with more than 2 groups (age group, histology type, location, VL group, CD4 group, and VL * CD4 group), we fit the univariate GEE model with the lesion resolution as a response variable with each factor as covariate (age group, VL group, CD4 group, and VL * CD4 group). We report the P value for corresponding covariates in each model. In addition, we compared the patient characteristics between patients with all index HSIL/s resolved (defined as complete resolution) and patients with less than all the HSILs resolved (defined as partial resolution) using corresponding statistical methods.
The Kaplan-Meier (K-M) method was used to estimate the survival rate of HSIL resolution among all lesions at different time points. Because patients may have had more than 1 lesion, lesions within the same patient might be correlated. The variance formula of Ying and Wei15 was used to adjust for the dependence in the data. Kaplan-Meier curves by selected factors were used to compare the HSIL resolution among groups with different characteristics such as smoking status, CD4 T-cell count, and VL groups or histology types. We also studied the shortest time to both any and index/adjacent HSIL recurrence of all resolved HSILs in patients with a follow-up visit after index lesion resolution before July 1, 2013. We used 2 definitions of recurrence: (1) index/adjacent site recurrence and (2) different location recurrence.
All data analysis was performed using SAS software (version 9.3; Cary, NC) and R 3.0.1.
The first column in Table 1 shows the basic demographic and HIV-related factors of the study population. All patients were male and 68 (94.4%) were white, with a median age of 48 years (range, 25–70 years; mean, 49 years). There were 98 HSILs in 72 patients, including 21 with 2 or 3 lesions. Among the 72 patients, 51 (70.8%) were current nonsmokers. Eighty-eight (89.8%) of 98 lesions had histology showing AIN 2. The site of histological diagnosis within the anal canal was evenly distributed between all quadrants. The median and mean CD4 T-cell counts at HSIL diagnosis were 555 and 573 cells/μL, respectively. Twenty-nine (40.3%) of the patients has CD4 T-cell counts less than or equal to 500 cells/μL. The plasma VL at HSIL diagnosis is right-skewed distributed among all patients, with median and mean of 50 and 17457 copies/mL, respectively. Fifty-seven (79.2%) of the patients had a VL less than 50 copies/mL. Combining the CD4 T-cell count with the VL group, approximately 50% of the patients had a CD4 T-cell count more than 500 cells/μL and VL less than 50 copies/mL.
A total of 98 HSILs from 72 patients (51 with 1 lesion, 16 with 2 lesions, and 5 with 3 lesions) were treated, and 77 (78.6%) lesions resolved to normal epithelium or LSIL during follow-up. Forty-eight (49.0%) and 27 (27.6%) of lesions resolved with 1 and 2 TCA treatments, respectively, whereas 1 lesion (1%) each resolved with 3 and 4 TCA treatments. Twenty-one (21.4%) lesions remained without evidence of resolution after TCA treatments (12 after 2, 6 after 3, 1 after 4, and 2 lesions that progressed from AIN 2 to AIN 3 after 1 and 3 treatments). These patients were offered an alternative treatment. Per patient, 52 (73.6%) of 72 were treated to resolution with TCA. Of the 20 patients with unresolved HSIL, 9 had multiple lesions.
The second and third columns in Table 1 show factors of interest (both demographic and clinical) stratified by HSIL resolution among all lesions. There are no apparent differences observed between the resolved and unresolved groups based on demographic parameters, except for race. Results from the GEE model indicated that the distribution of the patient and lesion-level factors in patients with resolved lesions are not significantly different from those of patients with unresolved lesions (age in years, age groups, smoking status, histology type, anal canal location, CD4 T-cell count, CD4 T-cell groups, VL, VL group, and CD4 T-cell and VL groups, respectively). There are only 4 African American patients in the study, all with resolved lesions, and so no statistical test was performed for the race parameter.
Table 2 shows the comparison between patients with complete resolution and those with partial resolution. For the 60 patients who had resolved lesions, 52 (86.7%) had complete resolution, whereas 8 (13.3%) had partial resolution. The average age was 56 ± 9 years in the complete resolution group and 49 ± 8 years in the partial resolution group (P = 0.009). Partial and complete resolution groups had a greater percentage of patients in the 56- to 70- and 41- to 55-year age range, respectively. The median CD4 T-cell count in the complete resolution group was 555 cells/μL, which was significantly less than that of 720 cells/μL in the partial resolution group (P = 0.031). No other factors were found to be significantly different between the 2 groups.
The K-M estimated survival probability (free of HSIL resolution) across time is shown in Figure 1. The estimated 3-, 6-, 9-, and 12-month survival (resolution) is 0.98 (95% confidence interval [CI], 0.95–1.00), 0.66 (95% CI, 0.55–0.76), 0.29 (95% CI, 0.17–0.40), and 0.19 (95% CI, 0.10–0.28), respectively. The median time to HSIL resolution was 196 days, indicating that 50% of the lesions resolved by day 196 after the first TCA treatment.
Survival curves were also assessed by age (25–40, 41–55, and 55–70 years), current smoking (yes or no), anal transition zone location (quadrants 1–4), CD4 cell count (≥500 or <500 cells/μL) and VL (≤50 or >50 copies/mL) at HSIL diagnosis. No apparent difference was observed in outcome for each variable by group specific survival curves (data not shown).
Patients with lesions treated to resolution with TCA at initial data censure were followed up per local standard of care, and follow-up data were censored in July 1, 2013. Among the 60 patients with at least 1 resolved HSIL, there are 53 patients with 67 resolved lesions with information on recurrence before July 1, 2013. Among the 67 resolved HSILs, 34 (50.7%) were recurrent at the index, adjacent or different location. Twelve lesions were recurrent at the index/adjacent sites (8 at the index and 4 at an adjacent site). In addition, 22 new HSILs were diagnosed, with some index recurrences at both index/adjacent and other sites.
Considering the 53 patients with resolved lesions and at least 1 follow-up visit, the denominator 22 (41.5%) of 53 had a recurrence at the index site, another site or both. Eight (15.1%) of 53 had recurrent HSILs at the index site, 11 (22.6%) of 53 had recurrences that include both index and adjacent sites, and 17 (32.1%) of 53 developed new site HSIL/s.
For each resolved index HSIL with follow-up after resolution before July 1, 2013, we calculated the shortest time to the index/adjacent recurrence and the shortest time to all recurrence (index/adjacent and different). Lesions without recurrence in patients who had visit before July 1, 2013, are defined as a censored observation with the censored time calculated as the number of days between the time of HSIL resolution and the time of the last follow-up visit of the patient.
Figures 2 and 3 show the patient-adjusted K-M curve for the time to the shortest recurrence and the time to the shortest index/adjacent recurrence for all resolved lesions, respectively. The estimated median time to the shortest recurrence is 507 days. The estimated 1-year survival rate for all types of recurrence is 0.6082 (95% CI, 0.47–0.74). Because the smallest estimated survival rate for the index/adjacent recurrence is 0.7673, the median time to the shortest index/adjacent recurrence is nonestimable. However, the estimated median time will be greater than the last censored/event time, which is 937 days. The 1-year survival rate for the index/adjacent recurrence is 0.8917 (95% CI, 0.82–0.96).
To see if there are any prognostic factors related to the 2 types of the recurrence, we performed an univariate marginal Cox proportional hazard model for each selected factors (age at diagnosis, smoking status, histology type, VL group, CD4 group, number of TCA visits performed, and the number of lesions at diagnosis). None of the models are significant for both of the 2 types of the recurrence using the likelihood ratio test (data are not shown.)
The acceptability of TCA application was not formally assessed. However, no patient complained of pain or discomfort during TCA application, no calls were received by the on-call physician after the procedure, and no patients raised any unanticipated issues regarding the procedure at follow-up.
No participant progressed to invasive anal cancer.
In this retrospective study, we have demonstrated that application of TCA to internal anal HSIL results in resolution of approximately 80% of lesions, with 50% of all lesions showing resolution after 1 treatment visit. The application procedure was painless with no unexpected adverse events. Because of the natural history of anal SIL, its management requires a long-term commitment from both patient and provider. Ideally, the treatment offered for ablation of HSIL should have both high efficacy and minimal impact on the patient’s quality of life both during and after the procedure. Trichloroacetic acid application seems to be promising for this treatment indication where there is HRA capability with resolution rates similar to other treatments.9,16
The study population was composed predominantly of HIV-positive white MSM, and the most frequently diagnosed HSIL in this population was AIN 2. These lesions were detected equally in all quadrants of the anal transition zone, indicating visualization of the entire anal circumference during HRA. HIV infection per se is associated with an increase in anal cancer risk. The influence of surrogate markers in previous studies of anal cancer indicates that a lower CD4 T-cell count and higher VL are associated with increased risk of anal cancer.17,18 In our study population, a higher CD4 T-cell count and older age, but not HIV VL, were associated with partial when compared to complete lesion resolution with TCA treatment (see Tables 1 and 2).
Most published treatment studies of HSIL have been retrospective analyses of clinical practice.9,19,20 Surgical treatment most commonly uses cold scalpel excision with or without electrofulguration of the lesion(s). Recovery time from this procedure is lengthy, and postoperative pain is often difficult to effectively control.21 One early study was associated with HSIL resolution of only 21% at 12 months after treatment in a population of HIV-positive MSM. It should, however, be appreciated that this population had extensive disease and was considered for surgical intervention only after more conservative treatment had failed.19
Infrared coagulation involves applying a probe that directs 1.5-second bursts of infrared light to internal anal HSIL after instillation of local anesthetic. After the procedure, patients usually have some anal discomfort for several days and bleeding with bowel movement for up to 2 weeks. Per lesion, resolution rates of more than 60% have been achieved, with an increase in efficacy seen with sequential treatments of the same lesion.9,20
A double-blind randomized trial of HSIL treatment in HIV-positive MSM comparing intra-anal imiquimod with matched placebo showed sustained resolution in 61% of patients at 36 months in the imiquimod arm.16 More recently, a prospective comparison study between self-applied and clinician administered treatment was published. This study assessed intra-anal imiquimod, intra-anal topical fluorouracil, and electrofulguration in HIV-positive MSM and showed treatment response rates of 24% (95% CI, 15–37), 17% (95% CI, 8–304), and 39% (95% CI, 26–54), respectively, in a modified intention to treat analysis where resolution was defined as no AIN.12 The differing efficacy rates between these studies of imiquimod most likely reflect differing application procedures, dose of imiquimod used, grades of anal SIL treated, and time frame for assessment of resolution.
It is important to appreciate that the recurrence rate of HSIL after all treatment modalities is high, resulting in the need for repeated treatments of the index lesion(s) in addition to any other HSIL(s) that may evolve over time. In the current study, less than half of our patients had developed either recurrent or new HSILs. Twenty percent of patients had recurrent HSIL at the index or adjacent location after initial resolution, whereas approximately 32% developed a new site HSIL, with some patients developed recurrences at both the index and a new site.
It has been a broadly held opinion that intra-anal HSIL only infrequently regresses to lower grades of dysplasia or normal epithelium, and so when it is identified, the lesion should be treated to prevent progression to cancer. However, in a recent retrospective analysis from Australia of clinical practice in Sydney, investigators reported that almost one quarter of HSIL diagnosed by anal cytology or biopsy spontaneously regressed, mostly to AIN 1.22 Although this finding needs to be more fully assessed in prospective studies, if HSIL does indeed regress at this rate, it raises the question of how to identify lesions that will progress and allow clinicians to focus ablative treatment accordingly.
All studies of anal dysplasia are contingent upon the accurate diagnosis of SIL grade by pathologists. However, the interobserver correlation between pathologists to diagnose HSIL on biopsy specimens is described only as “moderate.” In one study, the κ value for agreement between pathologists to distinguish HSIL from non-HSIL biopsy specimens ranged from 0.76 to 0.94.23 Thus, is should also be appreciated that the potential exists for both undertreatment and overtreatment of HSIL.
This study has limitations that relate mainly to its retrospective nature and a demographic that is mostly limited to white HIV-positive MSM with controlled HIV infection. It is possible that the anoscopists had varying abilities to recognize HSIL and index lesions may not have been recognized at the treatment visit. We consider this risk to be minimal because both anoscopists are credentialed in HRA and use image capture software for relocation of lesions in addition to using written descriptions of the index lesion in the patient chart. The follow-up time between visits varied because of a change in hospital policy, and it is possible that lesions may have resolved and then recurred within 6 months. Based on the author’s clinical experience, this is less likely than lesions appearing to be inadequately resolved at 3 months and fully resolved at 6 months. Because the clinical practice of anal dysplasia screening and management depends on identification of visual markers, it was felt appropriate that resolution be determined by absence of these markers without the need for biopsy. However, it is possible that HSIL may have been present in visually normal or low-grade appearing epithelium. Finally, because there was no formal acceptability scoring for the TCA procedure, we have assumed acceptability based on the absence of concerns verbalized by patients at the time of the procedure and also during follow-up.
Our experience with TCA has demonstrated that this intervention has efficacy that is similar to other more invasive treatments and may be used in settings that have HRA capacity. The high recurrence rates of HSIL after treatment, possibility of spontaneous regression of HSIL, and the potential to overtreat LSIL that is read as HSIL make interventions that are only minimally invasive with no lasting anal sequelae more attractive to both patient and provider.
Although the potential limiting aspects of this study should be duly considered, the findings suggest that a prospective study of TCA compared with other treatment options that include collection of patient acceptability data as well as cost-effectiveness will be important as the field of anal dysplasia evolves.
1. Human papillomavirus–associated cancers—United States, 2004–2008. MMWR Morb Mortal Wkly Rep 2012; 61: 258–261.
2. Daling JR, Weiss NS, Klopfenstein LL, et al. Correlates of homosexual behavior and the incidence of anal cancer. JAMA 1982; 247: 1988–1990.
3. Silverberg MJ, Lau B, Justice AC, et al. Risk of anal cancer in HIV-infected and HIV-uninfected individuals in North America. Clin Infect Dis 2012; 54: 1026–1034.
4. Palefsky JM. Human papillomavirus–related disease in men: Not just a women’s issue. J Adolesc Health 2010; 46 (4 suppl): S12–S19.
5. Gagne SE, Jensen R, Polvi A, et al. High-resolution analysis of genomic alterations and human papillomavirus integration in anal intraepithelial neoplasia. J Acquir Immune Defic Syndr 2005; 40: 182–189.
6. Cannistra SA, Niloff JM. Cancer of the uterine cervix. N Engl J Med 1996; 334: 1030–1038.
7. Chin-Hong PV, Palefsky JM. Natural history and clinical management of anal human papillomavirus disease in men and women infected with human immunodeficiency virus. Clin Infect Dis 2002; 35: 1127–1134.
8. Jay N, Berry JM, Hogeboom CJ, et al. Colposcopic appearance of anal squamous intraepithelial lesions: relationship to histopathology. Dis Colon Rectum 1997; 40: 919–928.
9. Cranston RD, Hirschowitz SL, Cortina G, et al. A retrospective clinical study of the treatment of high-grade anal dysplasia by infrared coagulation in a population of HIV-positive men who have sex with men. Int J STD AIDS 2008; 19: 118–120.
10. Singh JC, Kuohung V, Palefsky JM. Efficacy of trichloroacetic acid in the treatment of anal intraepithelial neoplasia in HIV-positive and HIV-negative men who have sex with men. J Acquir Immune Defic Syndr 2009; 52: 474–479.
11. Kreuter A, Potthoff A, Brockmeyer NH, et al. Imiquimod leads to a decrease of human papillomavirus DNA and to a sustained clearance of anal intraepithelial neoplasia in HIV-infected men. J Invest Dermatol 2008; 128: 2078–2083.
12. Richel O, de Vries HJ, van Noesel CJ, et al. Comparison of imiquimod, topical fluorouracil, and electrocautery for the treatment of anal intraepithelial neoplasia in HIV-positive men who have sex with men: An open-label, randomised controlled trial. Lancet Oncol 2013; 14: 346–353.
13. Palefsky JM, Holly EA, Ralston ML, et al. Effect of highly active antiretroviral therapy on the natural history of anal squamous intraepithelial lesions and anal human papillomavirus infection. J Acquir Immune Defic Syndr 2001; 28: 422–428.
14. Piketty C, Darragh TM, Heard I, et al. High prevalence of anal squamous intraepithelial lesions in HIV-positive men despite the use of highly active antiretroviral therapy. Sex Transm Dis 2004; 31: 96–99.
15. Ying Z, Wei LJ. The Kaplan-Meier estimate for dependent failure time observations. J Multivariate Anal 1994; 50: 17–29.
16. Fox PA, Nathan M, Francis N, et al. A double-blind, randomized controlled trial of the use of imiquimod cream for the treatment of anal canal high-grade anal intraepithelial neoplasia in HIV-positive MSM on HAART, with long-term follow-up data including the use of open-label imiquimod. AIDS 2010; 24: 2331–2335.
17. Chiao EY, Hartman CM, El-Serag HB, et al. The impact of HIV viral control on the incidence of HIV-associated anal cancer. J Acquir Immune Defic Syndr 2013; 63: 631–638.
18. Silverberg MJ, Chao C, Leyden WA, et al. HIV infection, immunodeficiency, viral replication, and the risk of cancer. Cancer Epidemiol Biomarkers Prev 2011; 20: 2551–2559.
19. Chang GJ, Berry JM, Jay N, et al. Surgical treatment of high-grade anal squamous intraepithelial lesions: A prospective study. Dis Colon Rectum 2002; 45: 453–458.
20. Goldstone SE, Hundert JS, Huyett JW. Infrared coagulator ablation of high-grade anal squamous intraepithelial lesions in HIV-negative males who have sex with males. Dis Colon Rectum 2007; 50: 565–575.
21. Goldie SJ, Kuntz KM, Weinstein MC, et al. Cost-effectiveness of screening for anal squamous intraepithelial lesions and anal cancer in human immunodeficiency virus–negative homosexual and bisexual men. Am J Med 2000; 108: 634–641.
22. Tong WY, Jin F, Mchugh LC, et al. Progression to and spontaneous regression of high-grade anal squamous intraepithelial lesions in HIV-infected and uninfected men. AIDS 2013; 27: 000.
© Copyright 2014 American Sexually Transmitted Diseases Association
23. Lytwyn A, Salit IE, Raboud J, et al. Interobserver agreement in the interpretation of anal intraepithelial neoplasia. Cancer 2005; 103: 1447–1456.