Abstract
Background
Anal high-grade squamous intraepithelial lesions (HSILs) ablation may reduce the incidence of invasive cancer, but few data exist on treatment efficacy and natural regression without treatment.
Methods
An open-label, randomized, multisite clinical trial of human immunodeficiency virus (HIV)–infected adults aged ≥27 years with 1–3 biopsy-proven anal HSILs (index HSILs) without prior history of HSIL treatment with infrared coagulation (IRC). Participants were randomized 1:1 to HSIL ablation with IRC (treatment) or no treatment (active monitoring [AM]). Participants were followed every 3 months with high-resolution anoscopy. Treatment participants underwent anal biopsies of suspected new or recurrent HSILs. The AM participants underwent biopsies only at month 12. The primary end point was complete clearance of index HSIL at month 12.
Results
We randomized 120 participants. Complete index HSIL clearance occurred more frequently in the treatment group than in the AM (62% vs 30%; risk difference, 32%; 95% confidence interval [CI], 13%–48%; P < .001). Complete or partial clearance (clearance of ≥1 index HSIL) occurred more commonly in the treatment group (82% vs 47%; risk difference, 35%; 95% CI, 16%–50%; P < .001). Having a single index lesion, compared with having 2–3 lesions, was significantly associated with complete clearance (relative risk, 1.96; 95% CI, 1.22–3.10). The most common adverse events related to treatment were mild or moderate anal pain and bleeding. No serious adverse events were deemed related to treatment or study participation.
Conclusion
IRC ablation of anal HSILs results in more clearance of HSILs than observation alone.
Keywords: HSIL, anal cancer, human papillomavirus, high-grade dysplasia, ablation
Ablation of anal canal high-grade squamous intraepithelial lesions (HSILs) is more likely to result in a complete or partial resolution of HSILs than monitoring alone. Spontaneous HSIL regression occurred, and a 1-year delay in treatment did not affect ultimate outcomes.
Anal cancer is preceded by high-grade squamous intraepithelial lesions (HSILs), and treatment of HSILs identified by high-resolution anoscopy (HRA)–directed biopsy is advocated to prevent the development of invasive anal cancer [1]. The mainstay of anal HSIL treatment is targeted ablation using infrared coagulation (IRC) or hyfrecation (electrocautery) [1–7]. IRC delivers short pulses of a narrow beam of visible and infrared light transmitted down the rigid quartz glass of the light guide, causing thermal coagulation necrosis [8]. Depth of destruction is controlled with length of pulse. Healing occurs with minimal scarring [9].
An initial retrospective review of IRC ablation of anal HSILs in 68 human immunodeficiency virus (HIV)–infected men followed up for as long as 3.5 years showed that of 165 individual lesions, 72% did not recur during follow-up after a single treatment [2]. However, approximately 65% of treated patients had HSILs diagnosed again within a median of 203 days, primarily reflecting development of new HSILs in untreated areas (metachronous). HSIL recurrence after ablation is higher in HIV-infected than in HIV-uninfected individuals [3].
HSILs may regress without treatment. An observational study of anal HSILs in HIV-infected and HIV-uninfected men who have sex with men found a spontaneous regression rate of anal HSILs to low-grade squamous intraepithelial lesions (LSILs) or normal findings of 23.5 per 100 person-years (95% confidence interval [CI], 15.73–35.02) [10]. Unfortunately, we cannot predict which lesions will persist, regress, or progress. To our knowledge, there are no randomized clinical trial data establishing the relative efficacy of IRC HSIL ablation as compared with no treatment.
METHODS
General Study Design
We conducted a randomized, open-label trial in HIV-infected participants comparing the 1-year clearance of biopsy-proven intra-anal HSILs treated with IRC with regression of HSILs absent treatment at 7 US AIDS Malignancy Consortium (AMC) trial sites (ClinicalTrials.gov No. NCT01164722). Participants were recruited between 2011 and 2014. All clinicians performing HRA and IRC were certified according to standards developed by the AMC HPV Working Group [4].
Study participants were HIV-infected adults aged ≥27 years with 1–3 anal canal HSILs (index lesions). Each index lesion was ≤15 mm in greatest diameter. Table 1 details other inclusion and exclusion criteria. The institutional review board for each participating site approved the study, and all participants signed informed consent.
Table 1.
Inclusion and Exclusion Criteria
| Inclusion Criteria | Exclusion Criteria |
|---|---|
| • HIV-infected adults aged ≥27 y with 1–3 anal canal HSILs (index lesions) | • History of anal cancer |
| • Prior treatment of anal HSILs with IRC at any time or treatment of anal HSILs by any means other than IRC within 2 months before enrollment | |
| • Each index lesion was ≤15 mm in greatest diametera | |
| • Within 90 d of enrollment: | |
| CD4 cell count ≥200 /µL | • Concurrent perianal HSIL, perianal condyloma, or vulvar HSIL or condyloma requiring treatment |
| Absolute neutrophil count >750 /µL | |
| Hemoglobin ≥9.0 g/dL | |
| • LSIL or condyloma covering >25% of the anal canal circumference | |
| Platelet count ≥75/µL | |
| INR and aPTT ≤ ULN (as defined by each site’s local laboratory) | • Ongoing use of anticoagulant therapy other than aspirin or nonsteroidal anti-inflammatory medications |
| • Treatment for an acute infection or other serious medical illness within 14 d before study entry; cancer requiring systemic therapy | |
| • Concurrent immunomodulatory therapy | |
| • History of HPV vaccination (added after protocol implementation) |
Abbreviations: aPTT, activated partial thromboplastin time; HIV, human immunodeficiency virus; HPV, human papillomavirus; HSIL, high-grade squamous intraepithelial lesion; INR, international normalized ratio; IRC, infrared coagulation; LSIL, low-grade squamous intraepithelial lesion; ULN, upper limit of normal.
aThe initial protocol specified a maximum lesion greatest diameter of 10 mm, but this was increased to 15 mm after enrollment began.
At enrollment HRA determined the presence of ≥1 previously identified, biopsy-proven index HSILs. Study participants were randomized 1:1 to HSIL ablation with IRC or active monitoring (AM) using a random permuted block randomization scheme stratified by AMC site that was not shared with investigators. The sample size was not large enough to account for multiple stratification factors. After randomization, only participants randomized to treatment had all index lesions treated with IRC.
Treatment Arm
IRC was performed under HRA guidance after infiltration with local anesthetic, as described elsewhere [2]. Briefly, the IRC tip was placed in direct contact with the lesion and treated with 1.5-second pulses of infrared light sequentially applied to the entire lesion. Necrotic tissue was debrided with sharp and blunt dissection to the level of the submucosal vessels. This was repeated until all index HSILs were completely removed.
Treatment arm (TA) participants were seen every 3 months for digital anorectal examination, anal cytology, and HRA with biopsy of areas with suspected HSILs. After ablation, HSILs were characterized in 2 ways: at the site of the original index lesion (recurrence) or at a new site (metachronous). Biopsy-proven recurrent or metachronous HSILs were treated with IRC within 4 weeks of diagnosis. Participants were followed up for 24 months after baseline treatment. At 12 and 24 months, TA participants underwent biopsy at index lesion sites (even if no lesion was seen) and sites with lesions suspected of being new HSILs. All treated participants completed a symptom diary for 10 days after treatment.
AM Arm
Study participants in the AM arm were examined by means of digital anorectal examination, anal cytology, and HRA every 3 months. Anal biopsy specimens were obtained only for lesions suspected of possible progression to invasive cancer, to prevent possible treatment effect from multiple biopsies. At 12 months after enrollment, AM arm participants underwent biopsy at index lesion sites (even if no lesion was seen) and sites with lesions suspected of being new HSILs. Study participants in the AM arm with HSILs at 12 months could cross over to IRC treatment for an additional 12 months, undergoing HRA every 3 months with ablation of metachronous or recurrent HSILs.
Anal biopsy specimens were interpreted at local site laboratories and classified as no evidence of intraepithelial lesion or cancer, LSIL (anal intraepithelial neoplasia grade 1), HSIL (anal intraepithelial neoplasia grade 2 or 3), or invasive cancer. A single pathologist (T. D.) reviewed biopsy specimens centrally. Discordant interpretations were reviewed and adjudicated with a second central pathologist (A. v. Z.). The central pathology interpretation was used for analysis. If a biopsy specimen was not available for central review (for 4 follow-up biopsy specimens through 12 months), the local histology result was used.
The primary end point was complete index lesion clearance (CILC) at month 12. In the TA, CILC was defined as no biopsy-proven HSIL at the site of all index lesions from the time of the initial treatment through month 12. In the AM arm, CILC was defined as regression of all index HSILs to LSILs or normal findings at month 12. Partial index lesion clearance (PILC) was defined as resolution of ≥1 (but not all) index lesions at month 12. CILC through month 24 was also evaluated from those continuing in the study in the TA, and for those in the AM arm who crossed over to IRC treatment at 12 months.
Statistical Analysis
The primary study population was the full analysis set based on intention to treat, including all participants who were randomized and completed the baseline visit. All participants not attending the month 12 visit or participants without ≥1 biopsy sample at month 12 were considered failures for the primary end point of 12-month CILC. Month 24 responses were based on evaluable participants.
Sample size was based on the assumption that CILC would be achieved in 50% of study participants in the TA and 20% in the AM arm by the primary end point. With 60 participants in each arm, we had 91.3% power to detect this difference between arms at a 1-sided α level of .025.
For the primary analysis, the percentages of participants with CILC were compared according to arm across sites, using a stratified Mantel-Haenszel test at a 1-sided α level of .025. Lesion-level responses were compared according to arm, with adjustment for clinical site in a generalized estimating equations binary model for relative risk (RR) that accounted for clustering of lesions within participants. Using RR regression, we explored predictors of CILC after adjusting for extent of disease, HIV treatment status, and other participant characteristics. We used Kaplan-Meier methods to analyze time to HSIL detection (in index or metachronous lesions) after IRC treatment.
RESULTS
Cohort Characteristics
Between May 2011 and June 2014, a total of 176 participants were screened with 121 randomized. One participant withdrew from observation immediately before any procedures were performed and was excluded from the analysis (Figure 1). The most common reason for screen failure was no HSIL (42%). Ultimately, 120 participants were enrolled, divided evenly between TA and AM arm. Seventy-one participants (59%) were enrolled from a single site (S.E.G. site). Table 2 describes participant characteristics. Most were white, non-Hispanic men receiving antiretroviral therapy (ART) (90%), with good viral suppression and immune reconstitution. Although most characteristics were similar between arms, significantly more Hispanic participants were randomized to treatment, and significantly more with >1 HSIL index lesion were randomized to the AM arm. At 12 and 24 months, cytology showed Atypical Squamous Cells-cannot exclude HSIL (ASC-H) or HSIL in 7 and 5 participants, respectively, and histology showed HSIL in all participants.
Figure 1.
Participant disposition flow diagram. The modified intention-to-treat (ITT) analysis included all participants who attended the baseline visit. Both participants categorized as not evaluable had 2 index lesions at enrollment, and at 12 months, 1 biopsy site was benign and 1 was not evaluable; for both participants, the result was considered a failure for the primary outcome. Abbreviations: HSILs, high-grade squamous intraepithelial lesions; IRC, infrared coagulation.
Table 2.
Participant Demographics and Disease Characteristics
| Characteristic | IRC Treatment (n = 60) | AM Arm (n = 60) | P Valuea |
|---|---|---|---|
| Male sex, No. (%) | 54 (90) | 58 (97) | .14 |
| Race/ethnicity, No. (%) | |||
| Non-Hispanic white | 36 (60) | 42 (70) | .05 |
| Non-Hispanic black | 7 (12) | 7 (12) | |
| Hispanicb | 16 (27) | 6 (10) | |
| Other | 1 (2) | 5 (8) | |
| Age, mean (range), y | 49.0 (25–78) | 50.5 (27–67) | .37 |
| ART, No. (%) | 54 (90) | 54 (90) | >.99 |
| Undetectable HIV viral load, No. (%) | 35 (59) | 42 (71) | .18 |
| Detectable viral load, median (range), copies/mL | 412 (77–1830) | 300 (82–1948) | .56 |
| CD4 cell count, median (IQR), cells/µL | |||
| Baseline | 669 (501–816) | 620 (514–736) | .34 |
| Nadir | 271 (106–395) | 300 (153–438) | .38 |
| History of AIDS diagnosis, No. (%) | 20 (33) | 19 (32) | .85 |
| History of HSIL, No. (%) | 29 (48) | 29 (48) | >.99 |
| Prior HSIL treatment, No. (%)c | 21/29 (72) | 24/29 (83) | .34 |
| Baseline No. of lesions, No. %d | |||
| 1 | 41 (68) | 32 (53) | .03 |
| 2 | 17 (28) | 19 (32) | |
| 3 | 2 (3) | 9 (15) |
Abbreviations: AM, active monitoring; ART, antiretroviral therapy; HIV, human immunodeficiency; HSIL, high-grade squamous intraepithelial lesion; IQR, interquartile range; IRC, infrared coagulation.
a P values determined with χ2 test or Wilcoxon rank sum test for categorical or continuous data, and with stratified (by site) Mantel-Haenszel test for number of lesions.
bAlthough the 4 categories of race/ethnicity did not significantly differ overall by arm, the difference between Hispanic and non-Hispanic participants was statistically significant (P = .02).
cOf those with a history of high-grade dysplasia.
dIdentified by local pathologist.
Response to Treatment and Spontaneous Regression
At 12 months, 52 of 60 (participants 87%) in the TA and 57 of 60 (95%) in the AM arm returned and had evaluable biopsy specimens (Figure 1). During the 12-month follow-up, 32 TA participants (53%) underwent a single ablation, and 18 (30%) and 9 (15%) had 2 and 3 treatments, respectively. In no participants did the lesions progress to cancer. Table 3 presents the index lesion response rate at 12 months. Overall, 37 (62%) TA participants had CILC (no index lesion recurrence), compared with 18 (30%) in the AM arm, for an absolute difference of 32% (95% CI, 13–48%; P < .001). Forty-nine participants (82%) in the TA and 28 (47%) in the AM arm had CILC or PILC, for an absolute difference of 35% (95% CI 16–50%; P < .001). In subsequent central pathological review, 4 participants in the TA and 5 in the AM arm did not have HSILs confirmed at baseline; when these participants were excluded, the absolute difference, (33%; 95% CI, 13%–50%) was unchanged. Overall study results did not change significantly when only the local versus central pathology results were used (absolute difference, 34% vs 32%, respectively).
Table 3.
12-Month Response at Participant and Lesion Levelsa
| Response | Participants or Lesions, % [95% CI] (No./Total No.) | Risk Difference (95% CI), % | P Valueb | |
|---|---|---|---|---|
| IRC Treatment | AM Arm | |||
| Overall CILC rate (primary end point) | 62 [48–74] (37/60) | 30 [19–43] (18/60) | 32 (13–48) | <.001 |
| Overall CILC/PILC rate | 82 [70–90] (49/60) | 47 [33–60] (28/60) | 35 (16–50) | <.001 |
| Reason for failure | ||||
| HSIL at index biopsy | 25 [15–38] (15/60) | 65 [52–77] (39/60) | −40 (−56 to −22) | <.001 |
| Withdrawal (without prior HSIL) | 8 [3–18] (5/60) | 5 [1–14] (3/60) | … | |
| Not evaluable at pathology or biopsy refused | 5 [1–14] (3/60) | 0 [0–6] (0/60) | … | |
| Free of HSIL (index or metachronous) at 12-mo visitc | 71 [56–83] (36/51) | 28 [17–42] (16/57) | 43 (22–59) | <.001 |
| Incident metachronous lesionsd | 47 [33–61] (25/53) | 21 [11–34] (12/57) | 26 (6–43) | .004 |
| Index lesion–level clearance | 63 [52–74] (51/81 lesions) | 42 [30–54] (41/97 lesions) | … | .001 |
Abbreviations: AM, active monitoring; CI, confidence interval; CILC, complete index lesion clearance; HSIL, high-grade squamous intraepithelial lesion; IRC, infrared coagulation; PILC, partial index lesion clearance.
aThere was no index lesion HSIL recurrence after the initial ablation in IRC arm or regression of index lesion to normal or low-grade squamous intraepithelial lesions absent treatment in the AM arm. All results are at the participant level except for Index lesion level clearance (bottom row).
b P values determined with on stratified (by site) Mantel-Haenszel test or generalized estimating equation relative risk model with adjustment for site (lesion-level analysis).
cCross-sectional estimate at 1 point in time in returning evaluable participants.
eCumulative estimate in returning evaluable participants.
When response or regression rate was examined on the index lesion level, TA participants had significantly superior outcome (63% vs 42%, respectively; P = .001). (Table 3) Metachronous lesions were identified through 12 months in 25 (47%) of TA participants and at 12 months in 12 (21%) of the AM-arm participants (P = .004). The disease-free rate, defined as absence of index or metachronous HSIL at 12 months, was 71% in TA participants and 28% in AM arm participants, for an absolute difference of 43% (95% CI, 22%–59%; P < .001).
Risk Factors for CILC (Absence of Index Lesion)
In multivariable analysis (Table 4), the only factors significantly associated with CILC after adjustment for ART were being in the TA (RR, 2.22; 95% CI, 1.49–3.29), undergoing treatment at 1 trial site (with S. E. G.) (RR, 1.47; 95% CI, 1.03–2.09) and having a single index HSIL (RR, 1.96; 95% CI, 1.22–3.10). There was no significant difference in spontaneous regression rates across trial sites (data not shown).
Table 4.
Predictors of 12-Month CILC at the Participant Level
| Predictor | Participants, No. | Univariate Analysis | Multivariable Modelb | ||
|---|---|---|---|---|---|
| CILC, No. (%) | P Valuea | RR (95% CI) | P Value | ||
| Study arm | |||||
| IRC treatment | 60 | 37 (62) | <.001 | 2.22 (1.49–3.29) | <.001 |
| AM arm | 60 | 18 (30) | 1 | ||
| Clinical site | |||||
| S.E.G. site | 71 | 39 (55) | .02 | 1.47 (1.03–2.09) | .03 |
| Other 5 sites | 49 | 16 (33) | 1 | ||
| Age, y | |||||
| 25–49 | 60 | 25 (42) | .36 | … | … |
| 50–78 | 60 | 30 (50) | … | ||
| Sex | |||||
| Male | 112 | 51 (46) | … | … | |
| Female | 8 | 4 (50) | >.99 | … | |
| Race | |||||
| White | 92 | 42 (46) | .94 | … | … |
| Nonwhite | 28 | 13 (46) | … | ||
| Ethnicity | |||||
| Hispanic or Latino | 22 | 14 (64) | .07 | … | … |
| Non-Hispanic or non-Latino | 97 | 41 (42) | … | ||
| Nadir CD4 cell count, cells/µL | |||||
| ≥200 | 78 | 35 (45) | .77 | … | … |
| <200 | 42 | 20 (48) | … | ||
| AIDS diagnosis | |||||
| Yes | 39 | 16 (41) | .46 | … | … |
| No | 81 | 39 (48) | … | ||
| History of HSIL | |||||
| Yes | 58 | 27 (47) | .88 | … | … |
| No | 62 | 28 (45) | … | ||
| Prior treatment for HSIL | |||||
| Yes | 45 | 23 (51) | .37 | … | … |
| No | 75 | 32 (42) | … | ||
| ART | |||||
| Yes | 108 | 54 (50) | .006 | 4.95 (0.77–31.60) | .09 |
| No | 12 | 1 (8) | 1 | ||
| Baseline CD4 cells/µL | |||||
| ≥500 | 93 | 41 (44) | .48 | … | … |
| <500 | 27 | 14 (52) | … | ||
| Baseline CD8 cells/µL | |||||
| ≥847 (median) | 57 | 26 (46) | .85 | … | … |
| <847 | 57 | 27 (47) | … | ||
| Detectable HIV viral load | |||||
| No | 77 | 36 (47) | .77 | … | … |
| Yes | 41 | 18 (44) | … | ||
| No. of index lesions | |||||
| 1 | 73 | 42 (58) | .001 | 1.96 (1.22–3.10) | .005 |
| 2–3 | 47 | 13 (28) | 1 | ||
| Sum of greatest diameters, mm | |||||
| 0 to <9 | 61 | 33 (54) | .06 | … | … |
| ≥9 | 59 | 22 (37) | … | ||
| Total | 120 | 55 (46) | … | ||
Abbreviations: AM, active monitoring; ART, antiretroviral therapy; CI, confidence interval; CILC, complete index lesion clearance; HIV, human immunodeficiency; HSIL, high-grade squamous intraepithelial lesion; IRC, infrared coagulation; RR, relative risk.
a P values determined with χ2 test.
bVariables significant at P < .10 in univariate analyses were considered in the multivariable generalized estimating equations binary model for RR, which adjusted for clinical site and accounted for clustering of lesions within participant; the final multivariable model included only variables that were significant at P < .10 after adjustment for other variables in the model. Ethnicity and sum of greatest diameters did not meet inclusion criterion after adjustment for treatment, site, ART, and number of index lesions.
24-Month End-point Results
Thirty-five participants in the AM arm underwent IRC ablation of persistent index and metachronous lesions at 12 months, and 30 (86%) were evaluable at 24 months. Twenty-five participants had a single treatment, 2 had 2 treatments, and 3 required 3 treatments. Of these, 25 participants (83%) had CILC at 24 months, and 41 of 51 treated index lesions (80%) did not recur. The disease-free rate was 80% (24 of 30) at study completion. None of these results significantly differed from what was seen in TA participants at 12 months (P > .05).
Among the 30 TA participants who returned at 24 months, 25 (83%) achieved CILC, and of the 51 HSILs treated, CILC was achieved in 41 (80%). Of 16 AM arm participants with no HSIL at 12 months, 11 returned at 24 months, with the index HSIL recurring in 4 of them by month 24.
Probability of HSIL-free Survival
Kaplan-Meier probability of HSIL recurrence is shown in Figure 2. It is estimated that with a single IRC treatment 63.5% (standard deviation, 6.4%) of participants will remain HSIL free at 6 months, a rate falling to 52.5% (6.7%) at 12 months and 42.1% (7.1%) by 24 months. The probabilities of HSIL-free survival in AM-arm participants who crossed over to treatment at 12 months were 87.5% (5.9%) at 6 months after treatment and 63.5% (9.6%) at 12 months. The Kaplan-Meier curves were not significantly different (log-rank test, P = .13).
Figure 2.
Kaplan-Meir Probability of high-grade squamous intraepithelial lesion (HSIL) diagnosis after initial infrared coagulation (IRC) treatment for IRC arm and active monitoring (AM) arm after crossover to IRC treatment. Those in the AM arm were not treated for the first 12 months. Those with HSILs at month 12 could then be treated and followed up for recurrence for an additional 12 months.
Adverse Events
Through 12 months of follow-up, 11 serious adverse events occurred (grade 3–5) among 9 participants, with none deemed study related. (Table 5) Adverse events at least possibly related to study procedures were observed in 87% of TA participants and 7% of AM-arm participants, with anal pain (mild or moderate) in 48 (80%) and bleeding (mild or moderate) in 47 (78%) of TA participants. Six serious adverse events in 6 participants (3 in each arm) were identified during the second year, and none were related to study participation (Table 5).
Table 5.
Treatment-related Adverse Events and Non–treatment-related Serious Adverse Events
| Adverse Eventsa | Participants, No. (%) | |||
|---|---|---|---|---|
| IRC Treatment | AM Arm | |||
| Grade 1–2 | Grade 3–5 | Grade 1–2 | Grade 3–5 | |
| Treatment related | ||||
| Anal pain | 48 (80) | 0 | 3 (5) | 0 |
| Postoperative hemorrhage/anal hemorrhage | 47 (78) | 0 | 0 | 0 |
| Non–treatment related | ||||
| Anxiety | 0 | 0 | 0 | 11 (2) |
| Fracture | 0 | 0 | 0 | 1 (2) |
| Aortic valve disease | 0 | 0 | 1 (2) | 0 |
| Pericardial effusion | 0 | 1 (2) | 0 | 0 |
| Other cardiac disorder (coronary heart disease) | 0 | 1 (2) | 0 | 0 |
| Thromboembolic event | 0 | 1 (2) | 0 | 1 (2) |
| Retinal vascular disorders | 0 | 0 | 0 | 1 (2) |
| Other neoplasms (basal cell carcinoma) | 0 | 0 | 0 | 1 (2) |
| Other neoplasms (lymphoma) | 0 | 1 (2) | 0 | 0 |
| Other infections (Salmonella poisoning) | 0 | 1 (2) | 0 | 0 |
Abbreviations: AM, active monitoring; IRC, infrared coagulation.
aThe table lists treatment-related adverse events occurring in ≥5% of participants in either arm for grades 1–2 and in any participants for grades 3–5. When a participant had multiple occurrences of a specific adverse event, the most severe is shown. All serious adverse events are listed; none were treatment related. There were no grade 5 adverse events.
DISCUSSION
We performed the first randomized, prospective trial comparing treatment with IRC to AM in HIV-infected participants with limited anal canal HSILs. At 1 year after randomization, participants in the TA were significantly more likely to have clearance of their index lesions, more likely to have achieved PILC, and more likely to be free of HSILs (either index or metachronous) than those in the AM arm.
Our observed response rates are consistent with those reported in prior retrospective cohort studies of ablative therapy. In 2005, Goldstone et al [2] reported the first series of HSIL ablation in HIV-infected participants, demonstrating that within 1 year of their first ablation, HSILs recurred in almost two-thirds of patients. Moreover, they reported long-term results of HSIL ablation by any method (IRC, electrocautery, or laser) in 456 HIV-infected participants, and the Kaplan-Meier–adjusted probability of recurrent HSIL at 1 year was 53%, similar to our data [1]. In a prospective trial, Richel et al [11] reported a slightly lower recurrence rate of 43% at 48 weeks in 46 HIV-infected patients treated with cautery ablation, but participants could have LSILs or HSILs involving the canal or perianal skin. LSILs and perianal HSILs have lower recurrence rates than anal canal HSILs, which could account for their improved results [12, 13].
We report spontaneous HSIL regression in almost one-third of participants without treatment. HSIL regression was documented in a retrospective Australian study of 574 HIV-infected and HIV-uninfected participants, in whom about 25% demonstrated spontaneous regression during approximately 1 year of follow-up [10]. When only HIV-infected participants were considered, regression occurred in approximately 20%. Our data found a higher rate of regression, which could be related to our enrolling participants with relatively limited disease. Biopsy of lesions, especially when small, could also contribute to HSIL regression.
Compared with the AM arm, metachronous HSILs developed in more than twice as many TA participants, which could reflect the fact that AM arm participants had lesions biopsied only at 12 months, whereas in the TA, biopsies were performed whenever lesions were identified. Metachronous HSILs could have regressed before biopsy in AM arm participants. It is also possible that treatment could cause a host or viral response, with development of additional metachronous lesions. Irrespective of this difference, TA participants were significantly more likely year than those in the AM arm to be disease free at 1 year.
We identified multiple factors affecting response. As in other series, our data document that having a single lesion significantly improved response [1, 3]. There is a significant learning curve related to HRA, and more experience increases the ability to identify HSILs [14, 15]. Similarly, our data suggest that procedural experience may affect treatment outcome. One site (S. E. G.) had a significantly higher CILC rate than other sites. The investigator at this site (S. E. G.) has the longest experience with HSIL ablation, which could translate to improved outcome. Although receipt of ART significantly improved CILC in univariate analysis and had an almost 5-fold RR of achieving CILC in multivariable analysis, this effect was not statistically significant, perhaps because 90% of participants received ART.
We did not identify clinical factors predictive of HSIL regression without treatment. Future studies should evaluate possible predictors of regression that were not tested for such as human papillomavirus types or other biomarkers in addition to clinical characteristics.
No participants’ lesions progressed to anal cancer, which is encouraging but not unexpected. We did not enroll participants with large-volume disease. Moreover, participants were observed for only 1 year before being offered treatment. This short observation period in only 60 participants with limited disease was probably not sufficient to identify progression. An ongoing clinical trial that will include >5000 HIV-infected adults with anal HSILs randomized to treatment or AM will address this question (ANCHOR study; ClinicalTrials.gov No. NCT02135419).
Spontaneous regression occurred in AM arm participants, and waiting 1 year to treat those with limited disease did not reduce the CILC rate once participants were treated. It might be reasonable to adjust the treatment algorithms so that individuals with limited disease (and especially those with a single small HSIL) could be observed closely for a year to determine whether spontaneous regression occurs. Persistent HSILs could then be treated. Even if not all lesions regress, PILC resulting in fewer HSILs could facilitate ablative therapy with less recurrence and diminished morbidity. Even when complete regression occurs, individuals must continue follow-up. In approximately 15% of those with CILC achieved at 1 year, the index lesion recurred during the second year.
This study has several limitations. As previously mentioned, more participants with 2 or 3 index lesions were randomized to AM, which could affect results. It is possible that the absolute difference in HSILs is overestimated, given that at baseline the AM group had more HSILs than the treatment group, but in the multivariable model the RR of HSIL clearance was still significant. We enrolled participants with small lesions, which could overestimate response, because small lesions may be more likely than large lesions to resolve with treatment or regress without treatment. The investigators were all fairly experienced in treating HSILs and participants were mostly white, HIV-infected men with limited disease, receiving ART with satisfactory viral suppression and immune reconstitution, making results less generalizable to other providers and populations. The overall strength of this study rests on the fact that it is the first multicenter, prospective, randomized trial powered to determine whether or not HSIL ablation is superior to AM alone in clearing HSILs in HIV-infected individuals.
In conclusion, IRC ablation of anal canal HSILs is more likely to result in complete or partial resolution than AM alone in those with small and few lesions. Although there is treatment-related morbidity, treatment-related serious adverse events were not observed. Spontaneous HSIL regression can occur, and delaying definitive therapy for 1 year in those with limited disease did not affect ultimate outcomes. Future studies should identify more effective treatment options so that more patients can be effectively treated with a single intervention.
Notes
Author contributions. Trial design: S. E. G., S. Y. L., E. A. S., T. D., J. Y. L., N. J., J. M. B. L., R. D. C., R. M., D. A., J. M. P., and T. W.; collection of data: S. E. G., E. A. S., T. D., A. v. Z., N. J., J. M. B. L., R. D. C., R. M., D. A., J. M. P., and T. W.; analysis of data: S. Y. L. and J. Y. L.; interpretation of data: S. E. G., S. Y. L., E. A. S., J. Y. L., J. M. P., and T. W.; manuscript writing: S. E. G. S. Y. L., E. A. S., and T. W.; review of manuscript: S. E. G., S. Y. L., E. A. S., T. D., J. Y. L., A. v. Z., N. J., J. M. B. L., R. D. C., R. M., D. A., J. M. P., and T. W.; and figures and tables: S. Y. L. and J. Y. L.
Acknowledgments. We thank the National Cancer Institute (NCI) as well as the AMC for supporting this research. We thank the AMC Operations and Data Management Center at The Emmes Corporation and investigators for enrolling participants. We are indebted to all participants who helped us further the science of anal neoplasia.
Financial support. This work was supported by the AMC, an NCI-supported clinical trials group (grant UM1CA121947).
Potential conflicts of interest. S. E. G. is a consultant, investigator, and speaker for Merck and a consultant and speaker for Medtronic, and he has received a research grant and travel support from Medtronic and personal fees from THD. S. Y. L., E. A. S., N. J., J. M. B. L., R, D. C., and T. D. have received grants from NCI, National Institutes of Health. E. A. S. had received equipment from Redfield and nonfinancial support from Antiva Biosciences. T. D. has received nonfinancial support from Hologic, fees from Antiva, and personal fees from TheVax, Roche, and BD. R. D. C. has received fees from UpToDate medical encyclopedia. J. M. P. has received grants from Merck, Agenovir, and Antiva Biosciences; nonfinancial support from Merck; personal fees from Agenovir and Antiva Biosciences; and stock options from Ubiome and Agenovir. T. W. has received grants from GlaxoSmithKline/ViiV Healthcare, Gilead Sciences, and Bristol Myers-Squibb and fees from GlaxoSmithKline/ViiV Healthcare. All other authors report no potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1. Goldstone SE, Johnstone AA, Moshier EL. Long-term outcome of ablation of anal high-grade squamous intraepithelial lesions: recurrence and incidence of cancer. Dis Colon Rectum 2014; 57:316–23. [DOI] [PubMed] [Google Scholar]
- 2. Goldstone SE, Kawalek AZ, Huyett JW. Infrared coagulator: a useful tool for treating anal squamous intraepithelial lesions. Dis Colon Rectum 2005; 48:1042–54. [DOI] [PubMed] [Google Scholar]
- 3. Goldstone RN, Goldstone AB, Russ J, Goldstone SE. Long-term follow-up of infrared coagulator ablation of anal high-grade dysplasia in men who have sex with men. Dis Colon Rectum 2011; 54:1284–92. [DOI] [PubMed] [Google Scholar]
- 4. Stier EA, Goldstone SE, Berry JM, et al. Infrared coagulator treatment of high-grade anal dysplasia in HIV-infected individuals: an AIDS Malignancy Consortium pilot study. J Acquir Immune Defic Syndr 2008; 47:56–61. [DOI] [PubMed] [Google Scholar]
- 5. Weis SE, Vecino I, Pogoda JM, Susa JS. Treatment of high-grade anal intraepithelial neoplasia with infrared coagulation in a primary care population of HIV-infected men and women. Dis Colon Rectum 2012; 55:1236–43. [DOI] [PubMed] [Google Scholar]
- 6. Sirera G, Videla S, Piñol M, et al. Long-term effectiveness of infrared coagulation for the treatment of anal intraepithelial neoplasia grades 2 and 3 in HIV-infected men and women. AIDS 2013; 27:951–959. [DOI] [PubMed] [Google Scholar]
- 7. Cranston RD, Hirschowitz SL, Cortina G, Moe AA. 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; 19:118–20. [DOI] [PubMed] [Google Scholar]
- 8. Colver G, Cherry G, Dawber R, Ryan T. Tattoo removal using infra-red coagulation. Br J Dermatol 1985; 112:481–485. [DOI] [PubMed] [Google Scholar]
- 9. Halasz C. Treatment of common warts using infrared coagulator. J Dermatol Surg Oncol 1994; 20:252–256. [DOI] [PubMed] [Google Scholar]
- 10. Tong WW, 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; 27:2233–43. [DOI] [PubMed] [Google Scholar]
- 11. Richel O, de Vries HJC, van Noesel CJM, Dijkgraaf MGW, Prins JM. 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; 14:346–53. [DOI] [PubMed] [Google Scholar]
- 12. Johnstone AA, Silvera R, Goldstone SE. Targeted ablation of perianal high-grade dysplasia in men who have sex with men: an alternative to mapping and wide local excision. Dis Colon Rectum 2015; 58:45–52. [DOI] [PubMed] [Google Scholar]
- 13. Silvera RJ, Smith CK, Swedish KA, Goldstone SE. Anal condyloma treatment and recurrence in HIV-negative men who have sex with men. Dis Colon Rectum 2014; 57:752–61. [DOI] [PubMed] [Google Scholar]
- 14. Swedish KA, Lee EQ, Goldstone SE. The changing picture of high-grade anal intraepithelial neoplasia in men who have sex with men: the effects of 10 years of experience performing high-resolution anoscopy. Dis Colon Rectum 2011; 54:1003–7. [DOI] [PubMed] [Google Scholar]
- 15. Hillman RJ, Gunathilake MP, Jin F, Tong W, Field A, Carr A. Ability to detect high-grade squamous anal intraepithelial lesions at high resolution anoscopy improves over time. Sex Health 2016; 13:177–81. [DOI] [PubMed] [Google Scholar]