The ASCUS : SIL ratio and the reference laboratory pathologist

The atypical squamous cells of undetermined significance (ASCUS) : squamous intraepithelial lesion (SIL) ratio was proposed to monitor laboratory use of the ASCUS diagnosis. This study addresses problems associated with comparing pathologists by this means. An intuitive example showed the ASCUS : SIL ratio depends on the prevalence of smears from patients who actually have SIL. In this study of 2000 cervical smears, each of five pathologists made 400 diagnoses. Differences among proportions of SIL diagnoses were statistically significant; differences among proportions of ASCUS diagnoses were not. Had an ASCUS : SIL ratio upper limit of 3.0 been used, two pathologists would have been misidentified as having high ASCUS diagnosis rates. Unlike the situation for laboratories, potential variability in SIL prevalence requires caution in the use of this ratio in assessing pathologists. An alternative measure that is independent of prevalence, the ASCUS : SIL odds ratio, is posited.