Abstract
Purpose:
Comparison of anal pre-cancer screening strategies in men who have sex with men (MSM).
Methods:
MSM in the Multicenter AIDS Cohort Study underwent repeated anal cytology (aCyt), oncogenic human papillomavirus (oncHPV) testing. A subset received High-Resolution Anoscopy (HRA). We evaluated three screening strategies for their ability to predict anal histological High-Grade Squamous Intraepithelial lesion (HSIL): single aCyt, sequential aCyt, and oncHPV co-testing. Multivariable logistic regression models evaluated risk of HSIL among participants undergoing HRA within 5 years of screening. Sensitivity and specificity were estimated among participants with HRA, and results corrected for verification bias using weighted generalized estimating equations.
Results:
There were 1426 MSM with aCyt screening (48% people with HIV [PWH]) and 428 that underwent HRA. Median age was 57 years, 14% of PWH had CD4< 350 cells/mm3. HSIL probability was higher in MSM with one (39%, p<0.01) or two abnormal aCyt results (46%, p<0.01), vs. those with normal aCyt (23–24%). Among men with abnormal aCyt, men with oncHPV+ had significantly higher risk than those who were oncHPV- (47% vs. 16%, p<0.01).
Specificity was modest with single aCyt+ (50%) but increased with sequential aCyt+ (79%) or oncHPV+ (67%). Sensitivity was high with single oncHPV+ (88%), moderate with single aCyt+ (66%) and oncHPV+ co-testing (61%), and low with sequential aCyt+ (39%). After correcting for potential verification bias, specificity increased and sensitivity decreased, but inferences were similar.
Conclusion:
None of the screening strategies evaluated had both sufficient specificity and sensitivity to warrant routine widespread use.
Keywords: anal cancer, HPV, HIV, longitudinal cohort, MSM, screening
INTRODUCTION
With widespread use of antiretroviral therapy (ART), people with HIV (PWH) now can experience a lifespan similar to HIV-uninfected adults. 1,2 However, higher incidence of comorbidities, including non-AIDS defining cancers, has been observed accompanying an increased lifespan among PWH. 3–5 In particular, PWH have higher burden of non-AIDS defining cancers that are associated with long term chronic viral coinfections (i.e., hepatitis B virus, hepatitis C virus, and human papillomavirus [HPV]), than similar individuals in the general population. 4,6 Anal cancer is mainly caused by HPV infection, 7 and incidence is nearly 40 times greater among men who have sex with men (MSM) with HIV than among men in the general U.S. population. 4,8 Oncogenic HPV (oncHPV) infections are associated with 3- to 4-fold increased risk of anal cancer after less than 10 years of follow-up. 9
Anal cancer incidence and mortality have increased significantly over the past 20 years. 10 In the U.S. the mortality rate from anal cancer is 0.23 per 100,000 overall, 10 and is much higher (16 per 10,000) among those living with HIV. 11 Early anal cancer diagnosis using effective screening strategies may increase survival and improve quality of life. 9 Anal cancer is typically preceded by a high-grade Squamous Intraepithelial Lesions (HSIL), a precancerous growth 12 that is common among PWH. 13 Large randomized clinical trials (i.e. the Anal Cancer/HSIL Outcomes Research [ANCHOR] study) showed that risk of anal cancer was significantly lower with treatment for anal HSIL than with active monitoring of progression among PWH with HSIL. 14–16 These findings highlight the need for early detection of anal pre-cancer/HSIL patients among high risk populations (i.e., PWH), but presently optimal guidelines for anal pre-cancer screening remain unclear. Clinical experts recommend that providers inquire about anal symptoms for PWH≥35 years old and some providers perform visual and digital anorectal examinations (DARE) annually, 17,18 referring patients with abnormalities detected by DARE for diagnostic follow-up using High-Resolution Anoscopy (HRA) and biopsy. 19 While DARE and visual inspection are considered to be useful strategies, they are insufficient because they usually only detect masses at advanced stages. 20 Strategies for earlier diagnosis and treatment are needed among high-risk groups.
HRA with biopsy is currently considered the gold standard for anal cancer and pre-cancer detection. However, the procedure requires specialized training, is costly, and is not widely available. 21,22 Whether routine HRA should be considered among select high-risk groups, or whether screening strategies have utility in determining whom to refer for more specialized HRA testing, are areas currently being explored. 23,24 Existing screening strategies used for cervical cancer have been studied for anal cancer, including anal cytology [aCyt] and HPV DNA from an anal swab, with some studies suggesting potential benefit 23,24 and others not. 25–27
The prevalence of oncHPV infection is high among MSM. 28,29 Thus, oncHPV testing alone may be highly sensitive, but has poor specificity for anal cancer screening among MSM. 27,30,31 An earlier Multicenter AIDS Cohort Study (MACS) looking at anal cancer screening via aCyt in 2011 suggested acceptance of screening by MSM is high (86%), but only 23% of MSM reported ever having undergone aCyt screening at that time. 32 Data on anal cancer and pre-cancer screening and test performance among MSM is still limited. 33 The current study sought to compare the test performance, as measured by the ability to predict anal pre-cancer risk, of three screening strategies (single aCyt, oncHPV co-testing, and sequential aCyt) using data from a long-standing community cohort of MSM with and without HIV.
METHODS
Study Population
The current study used a prospective cohort study design. Data were collected prospectively from MSM with and without HIV in the MACS (now called the MACS/WIHS Combined Cohort Study [MWCCS]). 34 Descriptions of the study design, enrollment, and data collection were previously reported. 35,36 Briefly, data were gathered at four study sites including Baltimore/Washington DC, Chicago, Pittsburgh/Columbus, and Los Angeles. Participants attended semi-annual study visits which included standardized interviews, laboratory tests, and physical examinations. 37 All sites’ IRBs reviewed and approved this study protocol. All participants provided written informed consent.
Between 2010 and 2015, all active MACS participants were offered the possibility of participating in the Anal Health Sub-Study (AHS). Among them,1550 MACS participants enrolled in the AHS and underwent up to four annual aCyt and HPV test visits. 37 Inclusion criteria for each of the screening test groups in the primary analysis, as well as the risk verification bias analysis (see statistical analysis section), are detailed in Figure 1. Specifically, participants were included in the primary analysis if they: (1) had history of sex with men; (2) underwent HRA within 5 years after aCyt (or for co-testing group within 5 years after aCyt and oncHPV test); and (3) were not diagnosed with anal cancer before the first aCyt. Individuals with aCyt and HPV testing who never underwent HRA were evaluated as a reference group in the analysis correcting for potential verification bias.
Figure 1.
Study population, with inclusion and exclusion criteria.
AIN2+ = Anal intraepithelial neoplasia grade 2+; HPV = Human papillomavirus; HRA = High-resolution anoscopy; MSM = Men who have sex with men.
At every semiannual MACS visit, whole blood was collected and processed for testing and cryopreservation, and self- and examiner-administered sexual, behavior, and health surveys were administered. 35,36 HIV tests were performed using enzyme-linked immunosorbent assay and confirmed by Western blot; testing was discontinued after infection was detected. CD4 T-lymphocyte counts were measured by flow cytometry using standardized methods. 38
Anal Cytology (aCyt) Testing
aCyt was measured annually among MSM with HIV and biennially among MSM without HIV. The results were classified as: No Intraepithelial Lesions (NIL); Atypical Squamous Cells of Undetermined Significance (ASCUS); Atypical Squamous Cells but cannot rule out HSIL (ASC-H); Low-grade Squamous Intraepithelial Lesions (LSIL/Anal Intraepithelial Neoplasia [AIN]1); HSIL/AIN2,3; and anal squamous cell cancer. Any level of abnormal findings from aCyt were classified together as abnormal for this analysis.39
Ascertainment of HSIL (Anal Pre-cancer)
High-resolution anoscopy (HRA) was performed on 236 (48%) of 488 men with abnormal aCyt. Given the potential for verification bias in observational studies if gold standard tests are only conducted among individuals with abnormal findings based on clinical decisions 40,41we also performed HRA on 192 (20%) of 938 men with normal aCyt as part of two funded studies: U01-AI-35043 supplement (PI D’Souza G) and R01CA169508 (PI Wiley D). This included a random subset of 50 men with normal aCyt who were sent for HRA as part of U01-AI-35043 supplement (PI D’Souza G). And men participating in R01CA169508 (PI Wiley D) who all had HRA as part of that study regardless of aCyt result (an unbiased sample). Results from the group of men with normal aCyt will serve as a comparison, as well as provide an opportunity to correct for verification bias using statistical tool (details in statistical approach). Among individuals with abnormal HRA, biopsy was performed on the tissues to further differentiate the severity of disease. Results were classified into low grade/mild dysplasia, moderate grade/moderate dysplasia, severe dysplasia, cancer in Situ, and anal cancer. Moderate grade/moderate dysplasia and severe dysplasia were defined as HSIL in the current study. The majority (78%) of the HRA tests were conducted in two (Pittsburgh and Los Angeles) of the four study sites, due to better availability of HRA providers at those sites.
Oncogenic HPV (oncHPV) Testing
oncHPV were measured at the same visit as the aCyt test (>95% of cases) or within 18 months (remaining cases) by a single laboratory (Tricore Reference Laboratories, Albuquerque, NM), using Linear Array HPV genotyping (Roche Molecular Diagnostics, Pleasanton, CA). oncHPV were defined as DNA detection of any of the following thirteen HPV genotypes: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68.
Statistical Analysis
We evaluated the test performance of three screening strategies to predict biopsy confirmed HSIL: (1) sequential aCyt was defined as two consecutive aCyt of no more than 30 months apart, with the last aCyt within study period prior to their HRA; (2) single aCyt was defined as the aCyt that either aligned with the second aCyt in the sequential strategy (for men included in the sequential strategy), or was the closest aCyt prior to their HRA (for men not included in the sequential strategy); and (3) co-testing was defined as aCyt and oncHPV tested at the same time (expanded to include testing within 18 months of each other when not available at the same visit [n = 18 men]), both prior to their HRA.
Baseline characteristics of participants were stratified by HIV serostatus and shown for all men screened by aCyt (used for verification bias correction analyses, detailed methods below), and the subset who also had HRA performed (primary analyses). Differences by HIV serostatus were evaluated through chi-squared or Fisher’s exact test for categorical variables and Kruskal-Wallis equality-of-populations rank test for continuous variables. Analysis of the screening strategies were conducted within the subset of participants who had HRA testing.
Multivariable logistic regression models estimated the odds of HSIL within 5 years of each of the three screening strategies. Models were adjusted for the effects of age, race and ethnicity, HIV infection, number of male sexual intercourse partners since last visit (0–1, >1–10, >10), education (any college experience, yes vs. no), smoking status (never, former, current), and study site. A sub-analysis evaluated the effects of CD4 T-lymphocyte count among PWH. Confounding factors included in the statistical models were selected based on review of the literature 18,32,42 and statistical significance (p<0.05) in preliminary univariate models. Test performance of the screening strategies, including sensitivity and specificity, were calculated using standard formulae.
We evaluated the test performance of the three screening strategies stratified by HIV serostatus and overall. As HRA was not performed on all men, and a higher proportion of participants with abnormal aCyt underwent HRA compared to those men who had NIL in aCyt, verification bias could occur and overestimate the test sensitivity. 41 We performed additional analyses using a weighted generalized estimating equation approach to correct for verification bias, accounting for the sample without HRA testing. 43 This approach corrected the verification bias by using weights, obtained from the inverse probability of sampling in people with negative and positive screening results, in weighted generalized estimating equations. This analytic method has been validated in a cervical cancer screening study and showed robust performance. 43 All statistical analyses were conducted in Stata software version 14 (Statacorp, College Station, TX).
RESULTS
Characteristics of Study Participants
We included 1441 MSM with at least one aCyt in this study. There were 428 MSM who underwent HRA testing within 5 years of screening. Among those with HRA, HPV co-testing was available in 427 (99.8%), and 279 (65%) had HRA testing after sequential aCyt (Figure 1). Among the 428 screened men with HRA, the median age was 57 years, the majority were PWH (61%), White non-Latino (82%), and graduated from college (59%)(Table 1). Compared to MSM without HIV, MSM with HIV were younger (median 55 vs. 60 years), more likely to be Black non-Latino (16% vs. 4%), and current smokers (25% vs. 14%), all p<0.01. The majority of PWH (86%) had CD4 cell counts ≥350 cells/mm3 at the aCyt visit. Population characteristics were similar between participants with and without HRA testing (Table 1).
Table 1.
Characteristics of all men included in the study, and men who had high-resolution anoscopy (HRA) results, by HIV status.
| Characteristic | Men with anal cytology testing (population used to assess verification biasa) N (%) | Men with HRA N (%) | ||||||
|---|---|---|---|---|---|---|---|---|
| Overall (N = 1426) | Men with HIV (N = 683) | Men without HIV (N = 743) | p-value | Overall (N = 428) | Men with HIV (N = 261) | Men without HIV (N = 167) | p-value | |
| Age in years, median (IQR) | 56 (50–62) | 54 (49–60) | 58 (52–65) | < 0.01 | 57 (52–63) | 55 (51–60) | 60 (56–65) | < 0.01 |
| Race/ethnicity | < 0.01 | < 0.01 | ||||||
| White non-Latino | 1047 (73) | 440 (64) | 607 (82) | 350 (82) | 201 (77) | 149 (89) | ||
| Black non-Latino | 235 (16) | 161 (24) | 74 (10) | 48 (11) | 41 (16) | 7 (4) | ||
| Other | 144 (10) | 82 (12) | 62 (8) | 30 (7) | 19 (7) | 11 (7) | ||
| Anal cytology test site | < 0.01 | < 0.01 | ||||||
| Baltimore | 333 (23) | 148 (22) | 185 (25) | 49 (11) | 34 (13) | 15 (9) | ||
| Chicago | 306 (21) | 211 (31) | 95 (13) | 65 (15) | 57 (22) | 8 (5) | ||
| Pittsburgh | 350 (25) | 150 (22) | 200 (27) | 121 (28) | 80 (31) | 41 (25) | ||
| Los Angeles | 437 (31) | 174 (25) | 263 (35) | 193 (45) | 90 (34) | 103 (62) | ||
| Education | < 0.01 | 0.03 | ||||||
| ≤ High school | 183 (13) | 117 (17) | 66 (9) | 52 (12) | 34 (13) | 18 (11) | ||
| Some college | 384 (27) | 217 (32) | 167 (22) | 124 (29) | 86 (33) | 38 (23) | ||
| ≥ College graduate | 859 (60) | 349 (51) | 510 (69) | 252 (59) | 141 (54) | 111 (66) | ||
| Smoking status | < 0.01 | < 0.01 | ||||||
| Never smoker | 466 (34) | 214 (32) | 252 (36) | 140 (35) | 93 (37) | 47 (31) | ||
| Former smoker | 639 (47) | 274 (41) | 365 (52) | 179 (45) | 97 (38) | 82 (55) | ||
| Current smoker | 265 (19) | 177 (27) | 88 (12) | 83 (21) | 62 (25) | 21 (14) | ||
| Number of male sexual intercourse partners since last study visit | 0.06 | 0.25 | ||||||
| 0 | 418 (31) | 222 (34) | 196 (28) | 136 (34) | 84 (34) | 52 (35) | ||
| 1 | 330 (24) | 148 (22) | 182 (26) | 91 (23) | 53 (21) | 38 (26) | ||
| 2 – 10 | 455 (34) | 210 (32) | 245 (35) | 127 (32) | 79 (32) | 48 (32) | ||
| Greater than 10 | 151 (11) | 80 (12) | 71 (10) | 42 (11) | 32 (13) | 10 (7) | ||
| CD4 cell count < 350 cells/mm3 | - | 96 (14) | - | - | - | 37 (14) | - | - |
| Currently using ART | - | 626 (93) | - | - | - | 241 (93) | - | - |
| Cytology test resultsb,c | < 0.01 | < 0.01 | ||||||
| Normal/NIL | 938 (66) | 393 (58) | 545 (73) | 192 (45) | 108 (41) | 84 (50) | ||
| ASCUS | 298 (21) | 167 (24) | 131 (18) | 139 (32) | 88 (34) | 51 (31) | ||
| LSIL | 132 (9) | 95 (14) | 37 (5) | 60 (14) | 47 (18) | 13 (8) | ||
| HSIL/ASC-H | 58 (4) | 28 (4) | 30 (4) | 37 (9) | 18 (7) | 19 (11) | ||
Population of 1426 men used for the verification bias dataset for the main analytic group of MSM. This included 428 men with HRA in the main analysis as well as another 983 men without HRA and 15 men who had HRA more than five years after anal cytology so were excluded from the HRA analysis but included in the verification bias analysis.
Anal cytology results shown are from either: (1) the earliest cytology test (for men who did not have HRA), (2) the closest cytology before the first available HRA (among those with HRA who were not in the sequential cytology testing group), or (3) the second cytology selected in the sequential testing strategy (among those with HRA who were in the sequential cytology testing group).
The ASCUS cytology result category includes one participant with an “atypical cells present (may not be squamous)” result.
ART = Antiretroviral therapy; ASCUS = Atypical squamous cells of undetermined significance; HRA = High-resolution anoscopy; HSIL/ASC-H = High-grade squamous intraepithelial lesions/Atypical squamous cells – cannot rule out high-grade; IQR = Interquartile range; LSIL = Low-grade squamous intraepithelial lesions; MSM = Men who have sex with men; NIL = No intraepithelial lesions.
Predicted Probability of Biopsy Confirmed HSIL by Screening Strategy
The predicted probability of biopsy confirmed HSIL within 5 years was significantly higher in MSM with abnormal than normal cytology findings (39% vs. 24%, p<0.01, Table 2, Figure 2; adjusted odds ratio [aOR]: 2.0, 95% confidence interval [CI]: 1.3, 3.2). Risk was three-fold higher among MSM with two sequential positive aCyt tests than those with two negative aCyt tests (aOR: 3.1, 95% CI: 1.5, 6.4). HSIL risk was similar after two vs. one abnormal aCyt (46% vs. 39%) and after two vs. one negative aCyt test (23% vs. 24%). In the stratified estimates by HIV status, the predicted risk yields similar results between the two groups (supplementary Figure S1).
Table 2.
Comparison of anal pre-cancer (HSIL) risk within five years of screen test (with 95% confidence interval) by screening strategy among a high-risk group of men who have sex with men (MSM) living with HIV and without HIV, who had HRA.
| Testing Strategy | Adjusted Estimate of HSIL Probabilitya | ||
|---|---|---|---|
| Overall | Men with HIV | Men without HIV | |
| Single anal cytology | N = 396 | N = 246 | N = 148 |
| Normal [ref] | 24% (18–31%) | 26% (17–34%) | 23% (13–32%) |
| Abnormal | 39% (33–45%) | 42% (34–50%) | 33% (22–45%) |
| Anal cytology and oncHPV co-testing | N = 395 | N = 245 | N = 148 |
| Normal aCyt and oncHPV negative [ref] | 16% (7–26%) | 11% (−1–22%) | 21% (6–35%) |
| Abnormal aCyt and oncHPV negative | 16% (6–25%) | 19% (5–34%) | 10% (−3–23%) |
| Normal aCyt and oncHPV positive | 28% (20–36%) | 32% (21–42%) | 23% (11–36%) |
| Abnormal aCyt and oncHPV positive | 47% (39–55%) | 48% (39–57%) | 44% (29–59%) |
| Sequential anal cytologyb | N = 262 | N = 176 | N = 75 |
| Normal and Normal [ref] | 23% (14–32%) | 27% (15–39%) | 18% (4–31%) |
| Abnormal and Normal | 28% (15–41%) | 26% (10–43%) | 39% (17–61%) |
| Normal and Abnormal | 32% (20–43%) | 33% (18–48%) | 26% (7–46%) |
| Abnormal and Abnormal | 46% (34–57%) | 47% (34–60%) | 49% (24–73%) |
Boldface indicates difference in anal pre-cancer risk by screening strategy result p-value < 0.05.
Testing strategies included: Single anal cytology, Co-testing (anal cytology and anal oncHPV testing), and Sequential testing (first two sequential anal cytology tests taken within five years prior to an HRA).
All models controlled for age, race/ethnicity, study site, college education, current smoking status, and number of male sexual intercourse partners since last visit. The overall model also controlled for HIV status. The model of MSM with HIV also controlled for current CD4 cell count at time of screen test. Ten men were excluded from the MSM without HIV sequential testing adjusted model because they predicted anal pre-cancer risk perfectly in the limited sample size in their strata (these men were included in crude estimates where results were similar – see Supplementary Table S1).
Test characteristics for sequential testing describe risk in the five years after the second aCyt. The study population excludes 12 men who had anal pre-cancer between the first and second aCyt.
aCyt = Anal cytology test; HRA = High-resolution anoscopy; HSIL = High-grade squamous intraepithelial lesions; MSM = Men who have sex with men; oncHPV = Oncogenic human papillomavirus
Figure 2.
Predicted risk of anal pre-cancer (HSIL) by testing strategy results.
Models adjusted for age, race/ethnicity, study site, college education, current smoking status, number of male sexual intercourse partners since last visit, and HIV serostatus.
Vertical bars represent 95% confidence intervals. Horizontal dashed line is the overall prevalence of anal pre-cancer observed in the study population (32%), as assessed by HRA.
aCyt = Anal cytology test; HRA = High-resolution anoscopy; HSIL = High-grade squamous intraepithelial lesions; oncHPV = Oncogenic human papillomavirus.
HSIL risk was lowest among men with a negative oncHPV co-test (16%), regardless of whether aCyt co-test was normal or abnormal (Table 2). Probability of HSIL among MSM with dual positive co-test (oncHPV positive and abnormal aCyt) was 47%, five-fold higher than among men with dual negative results (aOR: 4.8, 95% CI: 2.2, 10.8). Crude and adjusted estimates of predicted probability of HSIL were similar in MSM with and without HIV (Table 2 and Supplementary Table S1).
Test Performance of Screening Strategies
Table 3 compares sensitivity and specificity of the three screening strategies among MSM overall, and when stratified by HIV serostatus. Specificity was low (crude estimate: 50%; after verification bias correction: 65%) when using a single abnormal aCyt result. Specificity increased (but remained sub-optimal) when using dual positive sequential aCyt testings (crude: 79%; after correction: 83%), or dual positive in oncHPV/aCyt co-testing (crude: 67%; after correction: 76%), as the threshold of positive screening results.
Table 3.
Comparison of test characteristics for anal pre-cancer (HSIL) by screening strategy, including: Baseline testing for only aCyt or only oncHPV, Co-testing for aCyt and oncHPV, and Sequential testing (two sequential anal cytology tests).
| HSIL within 5 years | |||||
|---|---|---|---|---|---|
|
| |||||
| Screening Test(s) Outcome Defining Positivity | Participants | Crude Estimation | Verification Bias Correctiona | ||
|
| |||||
| Sensitivity | Specificity | Sensitivity | Specificity | ||
|
| |||||
| Baseline aCyt | |||||
| aCyt positive | Overall | 66% | 50% | 45% | 65% |
| MSM without HIV | 60% | 54% | 37% | 67% | |
| MSM with HIV | 69% | 47% | 53% | 61% | |
|
| |||||
| Baseline oncHPV (regardless of aCyt) | |||||
| oncHPV positive | Overall | 88% | 36% | 78% | 46% |
| MSM without HIV | 82% | 43% | 70% | 54% | |
| MSM with HIV | 90% | 31% | 86% | 36% | |
|
| |||||
| Co-testing (aCyt and oncHPV) | |||||
| aCyt and/or oncHPV positive | Overall | 93% | 19% | 83% | 35% |
| MSM without HIV | 87% | 24% | 78% | 43% | |
| MSM with HIV | 97% | 16% | 91% | 27% | |
|
|
|||||
| Both aCyt and oncHPV positive | Overall | 61% | 67% | 44% | 76% |
| MSM without HIV | 56% | 74% | 33% | 78% | |
| MSM with HIV | 63% | 62% | 50% | 72% | |
|
| |||||
| Sequential aCyt testing | |||||
| Either aCyt positive (1st and/or 2nd) | Overall | 78% | 35% | 57% | 50% |
| MSM without HIV | 81% | 39% | 78% | 62% | |
| MSM with HIV | 77% | 34% | 56% | 45% | |
|
|
|||||
| Both aCyt positive | Overall | 39% | 79% | 27% | 83% |
| MSM without HIV | 31% | 87% | 14% | 88% | |
| MSM with HIV | 42% | 74% | 31% | 81% | |
Verification bias correction using weighted generalized estimating equation approach. Weighting was based on a model of the inverse probability of having an anal HRA test controlled for age, race/ethnicity, number of male sexual intercourse partners since last visit, college education, study site, smoking status, and HIV status (for overall analysis).
aCyt = Anal cytology test; HSIL = High-grade squamous intraepithelial lesions; MSM = men who have sex with men; oncHPV = Oncogenic human papillomavirus.
Sensitivity was moderate to low in those with a single abnormal aCyt (crude: 66%; after correction: 45%). In sequential aCyt testing, requiring both aCyts to be abnormal yielded low sensitivity (crude: 39%; after correction: 27%), but considering results as positive if either aCyt was abnormal had improved sensitivity (crude: 78%; after correction: 57%).
Single oncHPV positive yielded relatively high sensitivity (crude: 88%; after correction: 78%). When abnormal aCyt was combined with positive oncHPV, co-test sensitivity was only 61% in crude estimates and 44% in estimates after correction for verification bias. The highest sensitivity was observed when either an oncHPV positive co-test or abnormal aCyt were considered as positive (crude: 93%; after correction: 83%), although this was not a large increase from oncHPV testing alone. All test performance metrics were qualitatively similar among MSM with and without HIV (Table 3). Overall, after correcting for verification bias, specificity increased and sensitivity reduced in all strategies (Table 3), but inferences from findings remained the same. The change in sensitivity and specificity was largest in the sequential aCyt testing, with smaller changes observed for test performance related to oncHPV testing (co-testing strategy).
DISCUSSION
We used data from a large longitudinal cohort of MSM with and without HIV to evaluate multiple screening strategies to predict HRA-validated HSIL outcomes. Anal pre-cancer within 5 years of screening was predicted in almost half of MSM with dual positive aCyt and oncHPV co-tests or sequential abnormal aCyt tests, highlighting the high burden of HSIL among MSM and the possible utility of screening to identify those most at risk. However, specificity was moderate or low by many strategies, highlighting the many “false-positives” identified in screening, which is an important limitation.
Although this study population was composed of MSM who are at high risk of anal cancer, none of the testing strategies used had both high sensitivity and high specificity. Screening for HSIL by oncHPV alone had good sensitivity (78%) but moderate specificity (46%) after verification bias correction. Compared to single aCyt screening, sequential aCyt testing and oncHPV co-testing did not substantially improve test performance. For example, using oncHPV co-testing (dual positive definition) had similar sensitivity to aCyt alone (61% vs. 66% in crude estimates), and verification bias correction also suggested comparable sensitivity of these groups at 44% vs. 45%. Specificity for oncHPV co-testing vs. aCyt alone was also similar in the crude (67% vs. 50%) and verification bias (76% vs. 65%) analyses. oncHPV and aCyt co-testing where either oncHPV positivity or abnormal aCyt were considered positive yielded the highest sensitivity (83%), but specificity was very low (35%), suggesting more harm than benefit given the high number of false-positives that would be sent for HRA.
Consistent with previous studies, 25,44,45 a single abnormal aCyt had 66% sensitivity and 50% specificity of detecting biopsy confirmed HSIL. Verification bias correction suggested sensitivity of this method might be lower at 45% and specificity higher at 65%. This suggests that while aCyt might be considered for HSIL screening in select high-risk populations given the high risk of disease and few other screening options, test performance is not sufficient to clearly recommend.
Our observations were consistent with previous studies in which oncHPV alone or in combination with aCyt co-testing had high sensitivity but low specificity. 27,46 Our observations provide further evidence that the cumulative incidence of HSIL among aCyt and oncHPV positive MSM could be nearly 50% over a 5-year period. The low specificity of screening strategies explored here tempers enthusiasm for screening with aCyt or oncHPV since these strategies will indicate many people without high-grade disease may require follow-up with HRA. However, given that the pre-test probability for anal pre-cancer among MSM with abnormal aCyt or oncHPV positive tests is high, referral for HRA can lead to earlier detection and treatment in these men, providing some clear benefit for those with HSIL detected when strong referral paths from HRA are in place to ensure treatment. Other reports have promoted consideration of screening tools to triage who should undergo HRA. Cintia, et al. suggests positive oncHPV be used to triage patients with normal aCyt for HRA, 46 consistent with our finding that co-testing did not improve test performance over oncHPV testing alone.
The current study has several limitations and strengths. HRA was not obtained on all individuals in our cohort. Individuals with an abnormal aCyt result were more likely to obtain an HRA than individuals with a normal aCyt and therefore verification bias could have occurred. We addressed this issue by considering all participants with aCyt results, adding additional samples of MSM with negative screening results, and applying an established statistical method 43 of the weighted generalized estimating equation approach to address the potential verification bias. This statistical method has been shown to improve accuracy of sensitivity and specificity estimates in both clinical trials and simulation studies, 43,47,48 when verification bias is a potential issue. Strengths of our study include a centralized protocol for how to perform aCyt screening at study visits, centralized biomarker testing, and inclusion of a diverse group of MSM with and without HIV across several study sites around the U.S.
CONCLUSION
The current study provides epidemiological evidence that aCyt screening had moderate specificity and sensitivity for anal pre-cancer within 5 years among MSM with and without HIV. oncHPV alone or co-testing with aCyt screening (where positivity on either test was considered positive) provided high sensitivity but poor specificity for anal pre-cancer among MSM. No screening strategies explored demonstrated sufficient performance to warrant routine clinical recommendation. However, high rates of HSIL among MSM, especially PWH, may warrant consideration of provider-patient discussions as to whether undertake screening with oncHPV and/or aCyt. More research is clearly needed in this area to improve biomarker performance. It is also critical to reduce barriers to care and improve access to HRA in order for anal cancer screening efforts to be successful and equitable.
Supplementary Material
ACKNOWLEDGEMENT
The authors gratefully acknowledge the contributions of the study participants and dedication of the staff at the MWCCS sites.
Source of Funding:
U01-HL146202, R01CA169508, K01AI162247
MWCCS (Principal Investigators): Atlanta CRS (Ighovwerha Ofotokun, Anandi Sheth, and Gina Wingood), U01-HL146241; Baltimore CRS (Todd Brown and Joseph Margolick), U01-HL146201; Bronx CRS (Kathryn Anastos and Anjali Sharma), U01-HL146204; Brooklyn CRS (Deborah Gustafson and Tracey Wilson), U01-HL146202; Data Analysis and Coordination Center (Gypsyamber D’Souza, Stephen Gange and Elizabeth Golub), U01-HL146193; Chicago-Cook County CRS (Mardge Cohen and Audrey French), U01-HL146245; Chicago-Northwestern CRS (Steven Wolinsky), U01-HL146240; Northern California CRS (Bradley Aouizerat, Jennifer Price, and Phyllis Tien), U01-HL146242; Los Angeles CRS (Roger Detels and Matthew Mimiaga), U01-HL146333; Metropolitan Washington CRS (Seble Kassaye and Daniel Merenstein), U01-HL146205; Miami CRS (Maria Alcaide, Margaret Fischl, and Deborah Jones), U01-HL146203; Pittsburgh CRS (Jeremy Martinson and Charles Rinaldo), U01-HL146208; UAB-MS CRS (Mirjam-Colette Kempf, Jodie Dionne-Odom, and Deborah Konkle-Parker), U01-HL146192; UNC CRS (Adaora Adimora), U01-HL146194. The MWCCS is funded primarily by the National Heart, Lung, and Blood Institute (NHLBI), with additional co-funding from the Eunice Kennedy Shriver National Institute Of Child Health & Human Development (NICHD), National Institute On Aging (NIA), National Institute Of Dental & Craniofacial Research (NIDCR), National Institute Of Allergy And Infectious Diseases (NIAID), National Institute Of Neurological Disorders And Stroke (NINDS), National Institute Of Mental Health (NIMH), National Institute On Drug Abuse (NIDA), National Institute Of Nursing Research (NINR), National Cancer Institute (NCI), National Institute on Alcohol Abuse and Alcoholism (NIAAA), National Institute on Deafness and Other Communication Disorders (NIDCD), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institute on Minority Health and Health Disparities (NIMHD), and in coordination and alignment with the research priorities of the National Institutes of Health, Office of AIDS Research (OAR). MWCCS data collection is also supported by UL1-TR000004 (UCSF CTSA), UL1-TR003098 (JHU ICTR), UL1-TR001881 (UCLA CTSI), P30-AI-050409 (Atlanta CFAR), P30-AI-073961 (Miami CFAR), P30-AI-050410 (UNC CFAR), P30-AI-027767 (UAB CFAR), and P30-MH-116867 (Miami CHARM).
Footnotes
Conflicts of Interest: Hilary Hsu reports employment and stock ownership with Amgen, Inc. Frank Paella is consultant for Gilead sciences, Janssen pharmaceuticals and ViiV Healthcare. All other authors have no conflicts.
Disclaimer: The contents of this publication are solely the responsibility of the authors and do not represent the official views of the National Institutes of Health (NIH).
REFERENCES
- 1.Samji H, Cescon A, Hogg RS, et al. Closing the gap: increases in life expectancy among treated HIV-positive individuals in the United States and Canada. PloS One. 2013;8(12):e81355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Nakagawa F, Lodwick RK, Smith CJ, et al. Projected life expectancy of people with HIV according to timing of diagnosis. AIDS. 2012;26(3):335–343. [DOI] [PubMed] [Google Scholar]
- 3.Lerner AM, Eisinger RW, Fauci AS. Comorbidities in Persons With HIV: The Lingering Challenge. JAMA. 2020;323(1):19–20. [DOI] [PubMed] [Google Scholar]
- 4.Silverberg MJ, Lau B, Achenbach CJ, et al. Cumulative Incidence of Cancer Among Persons With HIV in North America: A Cohort Study. Annals of Internal Medicine. 2015;163(7):507–518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.De Martel C, Ferlay J, Franceschi S, et al. Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. The lancet oncology. 2012;13(6):607–615. [DOI] [PubMed] [Google Scholar]
- 6.Machalek DA, Poynten M, Jin F, et al. Anal human papillomavirus infection and associated neoplastic lesions in men who have sex with men: a systematic review and meta-analysis. The lancet oncology. 2012;13(5):487–500. [DOI] [PubMed] [Google Scholar]
- 7.Daling JR, Madeleine MM, Johnson LG, et al. Human papillomavirus, smoking, and sexual practices in the etiology of anal cancer. Cancer. 2004;101(2):270–280. [DOI] [PubMed] [Google Scholar]
- 8.Silverberg MJ, Chao C, Leyden WA, et al. HIV infection and the risk of cancers with and without a known infectious cause. AIDS (London, England). 2009;23(17):2337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Johnson LG, Madeleine MM, Newcomer LM, Schwartz SM, Daling JR. Anal cancer incidence and survival: the surveillance, epidemiology, and end results experience, 1973–2000. Cancer. 2004;101(2):281–288. [DOI] [PubMed] [Google Scholar]
- 10.Deshmukh AA, Suk R, Shiels MS, et al. Recent trends in squamous cell carcinoma of the anus incidence and mortality in the United States, 2001–2015. JNCI: Journal of the National Cancer Institute. 2020;112(8):829–838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Horner M-J, Shiels MS, Pfeiffer RM, Engels EA. Deaths Attributable to Cancer in the US Human Immunodeficiency Virus Population During 2001–2015. Clinical Infectious Diseases. 2021;72(9):e224–e231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Berry JM, Jay N, Cranston RD, et al. Progression of anal high‐grade squamous intraepithelial lesions to invasive anal cancer among HIV‐infected men who have sex with men. International Journal of Cancer. 2014;134(5):1147–1155. [DOI] [PubMed] [Google Scholar]
- 13.Clifford GM, Georges D, Shiels MS, et al. A meta‐analysis of anal cancer incidence by risk group: Toward a unified anal cancer risk scale. International Journal of Cancer. 2021;148(1):38–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Medicine USNo. Topical or Ablative Treatment in Preventing Anal Cancer in Patients With HIV and Anal High-Grade Squamous Intraepithelial Lesions (ANCHOR). https://clinicaltrials.gov/ct2/show/NCT02135419.
- 15.Consortium AM. The Anchor Study. In: Accessed; 2017. [Google Scholar]
- 16.Palefsky JM, Lee JY, Jay N, et al. Treatment of anal high-grade squamous intraepithelial lesions to prevent anal cancer. New England Journal of Medicine. 2022;386(24):2273–2282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hillman RJ, Berry-Lawhorn JM, Ong JJ, et al. International Anal Neoplasia Society guidelines for the practice of digital anal rectal examination. Journal of Lower Genital Tract Disease. 2019;23(2):138–146. [DOI] [PubMed] [Google Scholar]
- 18.Nyitray AG, D’Souza G, Stier EA, Clifford G, Chiao EY. The utility of digital anal rectal examinations in a public health screening program for anal cancer. Journal of Lower Genital Tract Disease. 2020;24(2):192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Brown G Screening for Anal Dysplasia and Cancer in Patients With HIV. 2020.
- 20.Leeds IL, Fang SH. Anal cancer and intraepithelial neoplasia screening: a review. World Journal of Gastrointestinal Surgery. 2016;8(1):41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hillman RJ, Cuming T, Darragh T, et al. 2016 IANS international guidelines for practice standards in the detection of anal cancer precursors. Journal of Lower Genital Tract Disease. 2016;20(4):283–291. [DOI] [PubMed] [Google Scholar]
- 22.Ghebre R, Berry-Lawhorn JM, D’Souza G. State of the Science: Screening, Surveillance, and Epidemiology of HPV-Related Malignancies. American Society of Clinical Oncology Educational Book. 2021;41:377–388. [DOI] [PubMed] [Google Scholar]
- 23.Lam JM, Hoch JS, Tinmouth J, Sano M, Raboud J, Salit IE. Cost-effectiveness of screening for anal precancers in HIV-positive men. AIDS. 2011;25(5):635–642. [DOI] [PubMed] [Google Scholar]
- 24.Deshmukh AA, Chiao EY, Cantor SB, et al. Management of precancerous anal intraepithelial lesions in human immunodeficiency virus–positive men who have sex with men: Clinical effectiveness and cost‐effectiveness. Cancer. 2017;123(23):4709–4719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Salit IE, Lytwyn A, Raboud J, et al. The role of cytology (Pap tests) and human papillomavirus testing in anal cancer screening. AIDS. 2010;24(9):1307–1313. [DOI] [PubMed] [Google Scholar]
- 26.Wang Y, Wang Y, Gaisa MM, et al. Negative predictive value of human papillomavirus testing: Implications for anal cancer screening in people living with HIV/AIDS. Journal of Oncology. 2020;2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Gaisa MM, Sigel KM, Deshmukh AA, et al. Comparing Anal Cancer Screening Algorithms Using Cytology and Human Papillomavirus DNA Testing in 3 High-Risk Populations. The Journal of infectious diseases. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Nyitray AG, Carvalho da Silva RJ, Baggio ML, et al. Six-month incidence, persistence, and factors associated with persistence of anal human papillomavirus in men: the HPV in men study. Journal of Infectious Diseases. 2011;204(11):1711–1722. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Patel P, Bush T, Kojic EM, et al. Prevalence, incidence, and clearance of anal high-risk human papillomavirus infection among HIV-infected men in the SUN study. The Journal of infectious diseases. 2018;217(6):953–963. [DOI] [PubMed] [Google Scholar]
- 30.Berry JM, Palefsky JM, Jay N, Cheng S-C, Darragh TM, Chin-Hong PV. Performance characteristics of anal cytology and human papillomavirus testing in patients with high-resolution anoscopy-guided biopsy of high-grade anal intraepithelial neoplasia. Diseases of the Colon and Rectum. 2009;52(2):239–247. [DOI] [PubMed] [Google Scholar]
- 31.Clarke MA, Wentzensen N. Strategies for screening and early detection of anal cancers: A narrative and systematic review and meta‐analysis of cytology, HPV testing, and other biomarkers. Cancer Cytopathology. 2018;126(7):447–460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.D’Souza G, Rajan SD, Bhatia R, et al. Uptake and predictors of anal cancer screening in men who have sex with men. American Journal of Public Health. 2013;103(9):e88–e95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Fuchs MA, Multani AG, Mayer KH, Keuroghlian AS. Anal Cancer Screening for HIV-Negative Men Who Have Sex with Men: Making Clinical Decisions with Limited Data LGBT health. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.MWCCS. MACS/WIHS Combined Cohort Study. https://statepi.jhsph.edu/mwccs/about-mwccs/.
- 35.Dudley J, Jin S, Hoover D, Metz S, Thackeray R, Chmiel J. The multicenter AIDS cohort study: retention after 9½ years. American Journal of Epidemiology. 1995;142(3):323–330. [DOI] [PubMed] [Google Scholar]
- 36.Schrack JA, Althoff KN, Jacobson LP, et al. Accelerated Longitudinal Gait Speed Decline in HIV-Infected Older Men. Journal of acquired immune deficiency syndromes (1999). 2015;70(4):370–376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Robbins HA, Wiley DJ, Ho K, et al. Patterns of repeated anal cytology results among HIV-positive and HIV-negative men who have sex with men. Papillomavirus Research. 2018;5:143–149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Detels R, Jacobson L, Margolick J, et al. The multicenter AIDS cohort study, 1983 to…. Public Health. 2012;126(3):196–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Darragh TM, Colgan TJ, Cox JT, et al. The lower anogenital squamous terminology standardization project for HPV-associated lesions: background and consensus recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. Archives of Pathology and Laboratory Medicine. 2012;136(10):1266–1297. [DOI] [PubMed] [Google Scholar]
- 40.Begg CB, Greenes RAJB. Assessment of diagnostic tests when disease verification is subject to selection bias. 1983:207–215. [PubMed] [Google Scholar]
- 41.O’Sullivan JW, Banerjee A, Heneghan C, Pluddemann A. Verification bias. BMJ evidence-based medicine. 2018. [DOI] [PubMed] [Google Scholar]
- 42.Keller MJ, Burk RD, Massad LS, et al. Cervical Precancer Risk in HIV-Infected Women Who Test Positive for Oncogenic Human Papillomavirus Despite a Normal Pap Test. Clinical Infectious Diseases. 2015;61(10):1573–1581. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Xue X, Kim MY, Castle PE, Strickler HDJJoce. A new method to address verification bias in studies of clinical screening tests: cervical cancer screening assays as an example. 2014;67(3):343–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Chiao EY, Giordano TP, Palefsky JM, Tyring S, Serag HE. Screening HIV-infected individuals for anal cancer precursor lesions: a systematic review. Clinical Infectious Diseases. 2006;43(2):223–233. [DOI] [PubMed] [Google Scholar]
- 45.Palefsky JM, Holly EA, Hogeboom CJ, Berry JM, Jay N, Darragh TM. Anal cytology as a screening tool for anal squamous intraepithelial lesions. JAIDS Journal of Acquired Immune Deficiency Syndromes. 1997;14(5):415–422. [DOI] [PubMed] [Google Scholar]
- 46.Kimura CM, Nahas CS, Silva-Filho EV, et al. High-risk human papillomavirus test in anal smears: can it optimize the screening for anal cancer? AIDS. 2021;35(5):737–745. [DOI] [PubMed] [Google Scholar]
- 47.Torres-Ibarra L, Cuzick J, Lorincz AT, et al. Comparison of HPV-16 and HPV-18 Genotyping and Cytological Testing as Triage Testing Within Human Papillomavirus–Based Screening in Mexico. JAMA network open. 2019;2(11):e1915781–e1915781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Umemneku Chikere CM, Wilson K, Graziadio S, Vale L, Allen AJ. Diagnostic test evaluation methodology: a systematic review of methods employed to evaluate diagnostic tests in the absence of gold standard–an update. PloS One. 2019;14(10):e0223832. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.