Objective: To evaluate the value of high-resolution anoscopy (HRA) in the diagnosis of anal precancerous lesions. Methods: A retrospective cohort study was performed, which included 142 patients who underwent HRA in vulvar and anal clinic at Obstetrics and Gynecology Hospital of Fudan University from January 2014 to December 2019. With the perianal and anal canal biopsy pathology as the "gold standard", the diagnostic value of HRA and specific findings for anal precancerous lesions were evaluated and the clinical characteristics of patients with anal precancerous lesions were analyzed. Results: (1) Agreement between HRA and anal pathology were 76.6% (95/124, perianal) and 70.0% (84/120, anal canal), in which there was no significant difference (χ2=1.365, P=0.243). The strength of agreement with weighted Kappa statistic were 0.604 (perianal) and 0.455 (anal canal) respectively. HRA diagnosis were more overestimated [16.9% (21/124) in perianal and 25.0% (30/120) in anal canal] than underestimated [6.5% (8/124) in perianal and 5.0% (6/120) in anal canal]. The sensitivity and specificity of HRA in the diagnosis of anal precancerous lesions were all more than 60.0% (perianal lesions: 97.8% and 74.7%; anal canal lesions: 90.9% and 66.7%), and Youden's index were >0.5 (perianal lesions: 0.725; anal canal lesions: 0.576). The positive predictive value of HRA in anal canal lesions (50.8%) was lower than that in perianal lesions (68.8%). (2) The thin acetowhite epithelium was the most common finding in anal precancerous lesions, and its performance in anal canal lesions (the sensitivity, the specificity and the Youden's index were 78.8%, 62.4% and 0.412, respectively) were better than those in perianal lesions (the sensitivity, the specificity and the Youden's index were 28.9%, 96.9% and 0.258, respectively). The dense acetowhite epithelium and vascular patterns were only seen in anal canal lesions. Lugol's iodine was little useful for perianal lesions (the incidence of iodine negative was 0),but quite useful to evaluate the canal lesions (the positive predictive value was 83.3% and the negative predictive value was 82.3%). (3) The average age of patients with low-grade anal precancerous lesion was (37±10) years old, and high-grade anal precancerous lesion was (42±11) years old, and there was significant difference between them (P=0.034). Age curve showed that the peak age was 30-40 years old. Vulvar intraepithelial lesion was the risk factor of anal precancerous lesions (χ2=4.284, P<0.05). Conclusions: HRA patterns are reliable in the diagnosis of anal precancerous lesions, which is important for guiding biopsy. However, it is easy to be overestimated, especially for anal canal lesions. The acetowhite epithelium is the most common finding in anal precancerous lesions, but vascular patterns are only seen in anal canal lesions.
目的： 评价高分辨率肛门镜（HRA）检查对肛门癌前病变的诊断价值。 方法： 采用回顾性队列研究方法，选择2014年1月至2019年12月复旦大学附属妇产科医院行HRA检查的患者共142例，以肛门（包括肛周和肛管）活检病理诊断为“金标准”，探讨HRA检查对肛门癌前病变的诊断价值、HRA镜下具体征象及异常所见对肛门癌前病变的诊断效能，并分析肛门癌前病变患者的临床特征。 结果： （1）HRA诊断与病理诊断的完全符合率，肛周为76.6%（95/124）、肛管为70.0%（84/120），两者比较，差异无统计学意义（χ2=1.365，P=0.243）；HRA诊断与病理诊断的一致性检验Kappa值，肛周为0.604、肛管为0.455；肛周、肛管的过高诊断率分别为16.9%（21/124）、25.0%（30/120），均高于过低诊断率［分别为6.5%（8/124）、5.0%（6/120）］。HRA诊断肛门癌前病变［指低级别鳞状上皮内病变（LSIL）及以上病变］的敏感度和特异度均超过60.0%（肛周：分别为97.8%、74.7%；肛管：分别为90.9%、66.7%），约登指数均>0.5（肛周：0.725；肛管：0.576）；HRA诊断肛管病变的阳性预测值低于肛周病变（分别为50.8%、68.8%）。（2）HRA镜下异常所见，薄醋酸白上皮对肛管癌前病变的诊断效能（敏感度、特异度分别为78.8%、62.4%，约登指数为0.412）高于肛周癌前病变（敏感度、特异度分别为28.9%、96.9%，约登指数为0.258）。厚醋酸白上皮以及所有血管征象均仅见于肛管病变。碘染色对肛周病变的诊断效能低（碘染不着色的发生率为0），但对肛管病变的诊断效能高（阳性预测值、阴性预测值分别为83.3%和82.3%）。（3）肛门LSIL、高级别鳞状上皮内病变（HSIL）患者的年龄分别为（37±10）、（42±11）岁，两者比较，差异有统计学意义（P=0.034）；年龄曲线图显示，30~40岁为肛门癌前病变的好发年龄。伴发外阴病变患者的肛门癌前病变的发生率显著高于未伴发外阴病变者（分别为56.8%、35.1%；χ2=4.284，P<0.05）。 结论： HRA检查对肛门癌前病变的诊断效能稳定，镜下图像指导活检的意义较大；但易过高判别肛门癌前病变，尤其是对肛管病变。醋酸白上皮是HRA检查肛门最常见的镜下征象，但血管征象仅见于肛管病变。.