Screening for anal cancer precursors

Anal cancer is an increasing problem in HIV-positive men who have sex with men (MSM), with incidence rates exceeding those of cervical cancer before the introduction of cervical screening programs [1,2]. Both cervical and anal cancer are preceded by a precursor lesion, respectively, cervical and anal intraepithelial neoplasia (AIN). AIN is categorized in low grade (AIN 1; LGAIN) and high grade (AIN 2–3; HGAIN) [3]. AIN is highly prevalent in HIV-positive MSM and AIN screening is the subject of discussion [1]. Unfortunately, screening with anal cytology underestimates the presence of HGAIN [4]. Therefore, screening with high-resolution anoscopy (HRA) appears to be the preferred screening method. HRA combines proctoscopy with camera-assisted magnification and is in many respects comparable to cervical colposcopy, as described previously [5,6]. Currently, the lack of experienced high-resolution anoscopists is a major obstacle for the expansion of AIN screening programs. The quality of HRA depends on the experience of the anoscopist in recognizing abnormal mucosa and adequately targeting biopsies. Since the early 1990s, HRA is practised at the University of California, San Francisco (UCSF). This group reports that HRA anoscopists need extensive training and that they have a long learning curve [6]. However, to date no studies have been conducted to quantify this learning curve. In this report, we analyzed the HRA learning curve of the first author. Between 5 March 2008 and 1 December 2010, one physician (OR) screened 388 HIV-positive MSM for AIN, of whom 383 were included in the current study (five HRAs were of insufficient quality). Prior to inclusion, OR was trained four afternoons by an experienced HRA anoscopist (A. Kreuter, St Joseph Hospital, Bochum, Germany). Patients were invited at the HIV outpatient clinic to participate and all participants were screened by HRA, with biopsies of suspect lesions. Patients with AIN were treated in a study setting, as reported elsewhere [7]. Patients without AIN were invited for a second screening after 1 year. To evaluate the learning curve, we analyzed prevalence trends in time. Therefore, we divided the 383 participants into six time-subsequent groups of 64 participants, and analyzed AIN and HGAIN prevalence for each group at first screening. Next, for the patients without AIN at the first screening, we did the same for each group at the second screening. At baseline (first screening), 208 of 383 (54%) had biopsy-proven AIN, of whom 94 (25%) had LGAIN and 114 (30%) HGAIN. Both the AIN and HGAIN prevalence increased over time, from 42% and 31%, respectively, in the first group to 67% and 41% in the last group (Fig. 1).Fig. 1: (High grade) anal intraepithelial neoplasia prevalence with increasing experience.(High grade) anal intraepithelial neoplasia [(HG)AIN] prevalence of six time-subsequent groups of patients screened by high-resolution anoscopy (HRA) by a single anoscopist. Patients without AIN were invited for a second HRA screening after 1 year.The 175 patients who did not have AIN at baseline were invited for the second screening after 1 year. Thirty three patients were lost to follow-up [19 patients could not be reached, one patient had died (not study related), and 13 refused a second HRA], leading to 142 evaluable patients for the second screening. Thirty seven of 142 (26%) had AIN, of whom 24 (17%) LGAIN and 13 (9%) HGAIN, with a gradual decrease in AIN and HGAIN prevalence over time, from 44% and 9% to 7% and 0%, respectively (Fig. 1). The increasing/decreasing trends over time appear to level off after the third patient group, that is, after approximately 200 patients. So, for both AIN and HGAIN prevalence, an obvious increasing trend in time is seen at the baseline screening. At the second screening, a decreasing trend is seen (despite a peak in the fifth group). We do not think that the increasing AIN prevalence at baseline is caused by a changing study population. All patients were HIV-positive MSM, selected in the same way over the whole study period. We think that the increasing prevalence is caused by the HRA learning curve of OR. This is also reflected by the decreasing AIN prevalence at the second screening of patients who were ‘AIN-negative’ at baseline. Probably, a part of these AIN cases were missed at the first screening. In addition, the prevalence numbers at baseline in the last group (67% AIN, 41% HGAIN) are among the highest prevalence numbers ever reported in HIV-positive MSM [1]. We, therefore, think that, over time, we reached a high quality of performance of HRA. Although our analysis only involves one anoscopist, our data support and now quantify the previously reported experience that HRA practitioners show a long learning curve [6]. HRA is a complicated procedure, which should not be underestimated. It requires extensive training and experience. Given the lack of other qualitative screening methods, training programs need to be expanded to facilitate further introduction of AIN screening programs. Acknowledgements Conflicts of interest There are no conflicts of interest.