The Era of “Digital Colposcopy” Will Be Here Soon

Abstract

Although the basic technique has remained unchanged, the practice of colposcopy has evolved tremendously over the past half century. The sensitivity of colposcopy for the detection of high-grade cervical intraepithelial neoplasia was estimated to be 85% in a meta-analysis of studies from 1960 to 1996.1 In the 1980s, colposcopic grading became popular as a way of applying scientific method and critical analysis to a technique that had relied primarily on pattern recognition and expertise developed through experience.2 In the 1980s, colposcopy began to be taught universally in obstetrics and gynecology residency training programs. In the 1990s, the American Society for Colposcopy and Cervical Pathology, under the leadership of Tom Sedlacek, began teaching colposcopy to nongynecologists, including family practice physicians, nurse practitioners, physician assistants, and others. As more community practitioners began preforming colposcopy, colposcopy became decentralized; and the large colposcopy clinics that served as the training grounds for the previous generation of colposcopists began to disappear. This was highlighted in a 2001 survey of residency program directors that reported the limited experience available in colposcopy at some residency training programs.3 Cervicography is a technique developed by Adolph Stafl that uses a proprietary camera to take high-resolution colpophotographs that could be transmitted to expert colposcopists for interpretation.4 Several authors studied the use of cervicography as an intermediate triage test for ASC-US Papanicolaou tests and as a supplement to cytology screening, but it was soon superseded by human papillomavirus testing, which was a much more sensitive and specific tool. Cervicography became an excellent and easy-to-use tool for colpophotography and was used in the ASCUS/LSIL Triage Study for Cervical Cancer and reused in subsequent studies. By the mid-2000s, everything had changed. Newer studies reported that the sensitivity of colposcopy for the detection of high-grade lesions was only 54% to 57%5,6; colposcopic grading systems were shown to be poorly reproducible, even among experts,7,8 and colposcopists with experience and expertise did not identify high-grade lesions as well as less experienced colposcopists who just took more biopsies or even random biopsies.6,9,10 Tom Cox, then president of the American Society for Colposcopy and Cervical Pathology, wrote an editorial trying to explain these findings and suggested that the cause of low colposcopic sensitivity was small high-grade lesions being identified by ever more sensixtive screening tests. He said that these small lesions, of unknown premalignant potential, pushed the limits of detection by colposcopy and predicted that "eventually, new procedures or tests that are better equipped to triage at-risk women identified by increasingly sensitive tests may replace the (traditional) colposcope."11 Well, those days are rapidly approaching. Researchers and developers are using technology, miniaturization, computing power, and "big data" to develop tools that seek to improve the detection of significant high-grade lesions compared with traditional colposcopy. Among the approaches are the use of high-definition imaging and monitors; the use of computer algorithms that measure the development, fading, and intensity of the acetowhitening effect; or through computer analysis of a combination of optical and multispectral signals to identify potential biopsy sites. Other groups have sought to make the colposcope cheaper and more portable and to use cloud technology to bring colposcopy within the reach of low-resource settings. Two papers in this issue of the Journal report on the application of colposcopic technology within low-resource settings. Low-resource settings lack the health infrastructure and trained personnel to read cytology or perform colposcopy, and they do not have the funding to afford human papillomavirus testing. Wit