Testing for HPV as an Objective Measure for Quality Assurance in Gynecologic Cytology

Transcription

1 67 Testing for HPV as an Objective Measure for Quality Assurance in Gynecologic Cytology Positive Rates in Equivocal and Abnormal Specimens and Comparison With the ASCUS to SIL Ratio Vincent Ko, MD Shabin Nanji, MD Rosemary H. Tambouret, MD David C. Wilbur, MD Department of Pathology, Massachusetts General Hospital, Boston, Massachusetts. This study was presented at the 2006 United States and Canadian Academy of Pathology Annual Meeting, February 11 17, Atlanta, Georgia. David C. Wilbur and Rosemary Tambouret receive grant support from and are on the speakers bureau of Tripath. Approved Protocol #2005-P /1 by the Institutional Review Board of the Massachusetts General Hospital. Address for reprints: Vincent Ko, MD, Department of Pathology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114; Fax: (617) ; Received October 16, 2006; accepted November 1, BACKGROUND. Inappropriate use of the category of atypical squamous cells of undetermined significance (ASCUS) can result in overtreatment or undertreatment of patients, which may decrease the cost effectiveness of screening. Quality assurance tools, such as the ASCUS to squamous intraepithelial lesion ratio (ASCUS:SIL) and case review, are imperfect. High-risk HPV (hrhpv) testing is an objective test for a known viral carcinogen, and hrhpv may be more useful in monitoring the quality of ASCUS interpretations. METHODS. hrhpv rates for cytologic diagnoses and patient age groups were calculated for a 2-year period. All hrhpv results for ASCUS and SIL over a 17-month period were analyzed by patient age group, over time, and by individual cytopathologist to compare hrhpv rates with the corresponding ASCUS:SIL. RESULTS. The hrhpv positive rate for SIL was >90%, and it was 32.6% for ASCUS. Stratification by patient age showed that approximately 50% of patients younger than 30 years and older than 70 years of age were hrhpv positive, whereas other patients had a lower rate ranging from 14% to 34%. The overall ASCUS:SIL was 1.42, and the overall hrhpv positive rate was 39.9%. Over time and by individual cytopathologist, the hrhpv rate performed similarly to the ASCUS:SIL. The analysis by patient age showed a high statistical correlation (R 2 ¼ ) between the 2 methods. CONCLUSIONS. Despite differences between these techniques, the hrhpv rate closely recapitulates the ASCUS:SIL. When used together, the 2 methods can complement each other. The desirable hrhpv-positive range appears to be 40% to 50%; however, this may vary based on the patient population. The hrhpv rate is as quick and cost effective as determining the ASCUS:SIL. Cancer (Cancer Cytopathol) 2007;111: Ó 2007 American Cancer Society. KEYWORDS: human Papillomavirus, HPV, ASCUS, ASCUS:SIL, quality assurance. In cervical cytology screening, atypical squamous cells of undetermined significance (ASCUS) is an equivocal category assigned to specimens with morphologic changes suggestive of cervical intraepithelial neoplasia but which may also represent non-neoplastic conditions of various causes. 1,2 ASCUS is the most common abnormal interpretation in cervical cytology, but a significant lesion is usually not found on follow-up. An interpretation of ASCUS affects the clinical management of patients and because of its high prevalence, if incorrectly used, can result in overtreatment or undertreatment of substantial numbers of patients. 3 Therefore, quality assurance for ASCUS interpretations is important to limit inap- ª 2007 American Cancer Society DOI /cncr Published online 28 February 2007 in Wiley InterScience (

2 68 CANCER (CANCER CYTOPATHOLOGY) April 25, 2007 / Volume 111 / Number 2 propriate use, to monitor consistency of cytologic interpretation, and, therefore, to maintain overall cost effectiveness of current cervical cancer screening guidelines. Quality assurance guidelines recommend that the ASCUS rate and the squamous intraepithelial lesion (SIL) rate in the laboratory be monitored by comparing them to historical averages for the laboratory or to published data. 1,4,5 The usefulness of this method depends on the assumption that variations of ASCUS and SIL rates in stable patient populations should be relatively small over time. However, differences in patient populations may be found between laboratories and may shift over time, factors that will alter what is considered an appropriate ASCUS or SIL rate. The ASCUS to SIL ratio (ASCUS:SIL) attempts to normalize these figures by taking the prevalence of SIL, or the overall risk status of a population into consideration. Thus, the ASCUS:SIL is 1 of few commonly used tools for implementing ASCUS quality assurance. Unfortunately, the ASCUS: SIL does not reflect diagnostic accuracy, because cytologists can overinterpret or underinterpret either or both ASCUS and SIL while still maintaining a desirable ratio. By using this quality assurance tool, individuals who fall out of established ASCUS rate ranges determined by the laboratory may require recalibration and possible remediation of interpretation criteria. Another standard quality assurance tool is the review of cases by another cytologist. However, ASCUS is by definition an equivocal category, so it is not surprising that ASCUS interpretations by referees may also demonstrate poor reproducibility. 6 Therefore, case review may not be an effective method for monitoring accuracy. Unlike the ASCUS:SIL, which as a number has no biologic relevance of its own, a test for high-risk human Papillomavirus (hrhpv) can provide an objective assessment of ASCUS interpretations. Because hrhpv is known to be a causative agent in virtually all cervical cancers, performing a test for hrhpv DNA in patient samples is very useful because it measures the presence or absence of an analyte that is central to pathogenesis of cervical neoplasia. 7,8 hrhpv testing has been clinically validated as a useful method to appropriately triage ASCUS cases, with established performance statistics such as sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for detection of important high-grade lesions. On the basis of data from the ASCUS/LSIL Triage Study (ALTS) and other studies, Hybrid Capture II (Digene, Gaithersburg, Md) is an assay that has an approximately 99% NPV for the presence of CINIIþ, which means that patients with hrhpv-negative ASCUS specimens have a very low chance of harboring a high-grade dysplasia or carcinoma. Thus, current guidelines recommend that patients with hrhpv-negative ASCUS specimens can be safely rescreened annually. Conversely, patients with hrhpv-positive ASCUS specimens should undergo immediate colposcopy because such specimens have approximately a 15% to 20% chance of harboring a high-grade lesion. 3,9 Morphologic abnormalities in the presence of hrhpv in a specimen are likely due to HPV viral cytopathic effect when compared with hrhpv-negative cases. When all ASCUS cases are tested for hrhpv (ie, reflex testing), one would anticipate that overuse of ASCUS would lead to a lower hrhpv-positive ASCUS rate if cases with reactive changes are overcalled, whereas undercalling low-grade squamous intraepithelial lesions (LSIL) would lead to a higher hrhpv-positive ASCUS rate. To test this hypothesis, we evaluated 3 aspects of cervical cytology and hrhpv testing at our hospital. First, the baseline hrhpv rate for known abnormal cases and for different age groups was examined to determine whether the test performed as expected compared with current literature. Second, the hrhpv rate was compared with the ASCUS:SIL in our laboratory to determine the relation between these methods and to better determine the applicability of the hrhpv rate as a quality assurance tool. Finally, after we demonstrated that the hrhpv rate is a useful and reproducible method, the average historical data obtained in our laboratory was compared with published data to determine what may be a desirable hrhpv-positive range to use as a benchmark. MATERIALS AND METHODS To determine the baseline hrhpv rate, 2 years of high-risk HPV testing with the Hybrid Capture II (HCII) assay with the SurePath (Tripath, Burlington, NC, is now part of BD [Becton, Dickinson and Company]) collection method were examined. We obtained hrhpv-positive rates by cytologic interpretation and age group. A subset analysis of ASCUS cases was performed to compare the hrhpv rate with the ASCUS:SIL. A custom-search protocol was designed and performed on the PowerPath (IMPAC Medical Systems, an Elekta company, Mountain View, Calif) laboratory information system to query for all cases of ASCUS that also had an associated hrhpv test. In addition, all SILþ cases (low-grade intraepithelial squamous lesion [LSIL], high-grade squamous intraepithelial lesion [HSIL], and carcinoma) having a concurrent

3 Use of HPV Testing for QA in ASCUS/Ko et al. 69 hrhpv test over a 17-month period were queried for comparison. These data were analyzed by patient age group, over time, and by the 11 cytopathologists who sign cases in our laboratory. To directly compare the ASCUS:SIL (a number between 0 and infinity) to the hrhpv rate (a percentage between 0 and 100), the standard ASCUS to SIL ratio (ASCUS:SIL) was converted to a modified ASCUS:SIL (Number of ASCUS7[Number of ASCUS þ Number of SILþ]100). This conversion to the modified ASCUS:SIL allows a more effective sideby-side analysis of the hrhpv rate and the ASCUS:SIL methods. The hrhpv-negative rate (Number of hrhpv negative7[number of hrhpv negative þ Number of hrhpv positive]100) was used in the tables and graphs because a high hrhpv-negative rate corresponds to a high ASCUS to SIL ratio. For both situations, a very high TABLE 1 hrhpv Rate by Cytologic Diagnosis Cytologic diagnosis hrhpv positive rate (%) ASCUS 146/448 (32.6) LSIL 94/105 (89.5) HSIL 22/23 (96.0) hrhpv indicates high-risk human papilloma virus; ASCUS, atypical squamous cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesion; HSIL, high-grade squamous intraepithelial lesion. TABLE 2 hrhpv Rate by Age Group Age, y hrhpv Positive Rate (%) /123 (47) /770 (50) /601 (34) /339 (22) /213 (14) /93 (15) 70þ 12/23 (52) rate probably reflects a tendency to overcall normal to ASCUS, whereas a very low hrhpv-negative rate indicates a tendency to undercall SIL, and a low ASCUS:SIL may correspond to an interpretation of cases with borderline morphologic changes as SIL. To determine an optimal hrhpv-positive range, hrhpv rates of the individual cytopathologists were examined and compared with the overall laboratory hrhpv rate and to their respective ASCUS to SIL ratios. The hrhpv-positive rate of ASCUS from the ALTS trial, and the rate from published studies that examined the relation between the ASCUS to SIL ratio and hrhpv-positive rates, were taken into consideration. RESULTS During the initial study, the hrhpv-positive rate by cytologic diagnosis showed a high-positive rate for LSIL (94 of 105; 89.5%) and HSIL (22 of 23; 96.0%), whereas ASCUS had a 32.6% (146 of 448) positive rate (Table 1). The hrhpv-positive rate stratified by age group for all cytologic diagnoses shows that nearly half of all patients with an age younger than 30 years and older than 70 years were hrhpv positive (47% to 52%). Patients between ages 30 years and 69 years had a lower hrhpv-positive rate that ranged from 14% to 34% (Table 2). In the subset analysis of ASCUS cases, the overall ASCUS to SIL ratio was 1.42 (n ¼ 5441 ASCUS þ SIL cases), and the overall hrhpv-positive rate was 39.9% (n ¼ 2422 ASCUS cases with corresponding hrhpv test). Over time, the hrhpv-negative rate closely followed the modified ASCUS:SIL in value and slope over a 17-month period (Table 3, Fig. 1). The mean hrhpv-negative rate for the 6 time periods was 59.1%, with a standard deviation of 6.2%. The mean modified ASCUS:SIL was 58.6%, with a standard deviation of 3.2%. Of note, the difference between TABLE 3 hrhpv Negative Rate Versus Modified ASCUS:SIL Ratio Versus Standard ASCUS:SIL Ratio by Date Date hrhpv negative rate (%) Modified ASCUS:SIL ratio (%) Standard ASCUS:SIL ratio 2004-Q3 177/312 (56.7) 533/843 (63.2) 533:310 (1.72) 2004-Q4 135/274 (49.3) 441/782 (56.4) 441:341 (1.29) 2005-Q1 259/428 (60.5) 540/978 (55.2) 540:438 (1.23) 2005-Q2 275/483 (56.9) 621/1068 (58.1) 621:447 (1.39) 2005-Q3 406/609 (66.7) 697/1127 (61.8) 697:430 (1.62) 2005-Q4 203/316 (64.2) 365/643 (56.8) 365:278 (1.31) Overall 1455/2422 (60.1) 3197/5441 (58.8) 3197:2244 (1.42) Mean (59.1) (58.6) 1.43 Std Deviation (6.2) (3.2) 0.20

4 70 CANCER (CANCER CYTOPATHOLOGY) April 25, 2007 / Volume 111 / Number 2 the hrhpv-negative rate and the modified ASCUS:SIL fell within 7.4%. The analysis by patient age group also closely followed the ASCUS:SIL in value and slope over all age groups for the same 17-month period (Table 4, Fig. 2). Linear regression analyses on the hrhpv-negative rate and the modified ASCUS:SIL showed high statistical correlation (R 2 ¼ ) (Fig. 3). In addition, although all patients in this subset had ASCUS, the hrhpv-positive rate was much higher in patients aged younger than 30 years (>56%) compared with patients aged older than 30 years (19% to 39%). Of note, the difference between the hrhpv-negative rate and the modified ASCUS:SIL fell within 8.5%. Analyses performed by individual cytopathologist showed variability similar to the ASCUS:SIL for most individuals (Table 5, Fig. 4). The mean hrhpv-negative rate for the 11 cytopathologists was 59.8%, with a standard deviation of 6.5%. The mean modified ASCUS:SIL was 57.1%, with a standard deviation of FIGURE 1. hrhpv-negative rate versus modified ASCUS:SIL by date. Blue represents hrhpv-negative rate; red represents modified ASCUS:SIL. 9.3%. Three pathologists (2, 5, and 9) had modified ASCUS:SIL that differed from the hrhpv-negative rate by >10%, whereas the other 8 pathologists had a difference of <10% (Table 5). Pathologists 2 and 5 had low ASCUS to SIL ratios (1.20 and 0.96), but they also had relatively low hrhpv-positive rates (33.1% and 32.8%). Pathologist 9 had the lowest ASCUS to SIL ratio of the group (0.74) and had an hrhpv-positive rate of 40.7%. DISCUSSION The ASCUS:SIL is a commonly used ratio in quality assurance for determining an appropriate rate of ASCUS interpretation. The range of an appropriate ASCUS:SIL is often determined by historical data, but this number does not reflect the underlying biology of cervical neoplasia, nor does it directly measure accuracy of either ASCUS or SIL interpretations. hrhpv testing is clinically used as a tool to manage women with ASCUS; hrhpv-positive tests lead to treatment similar to SIL, and hrhpv-negative tests are followed with repeat cytologic samples. 3 Because hrhpv testing detects a virus that represents a necessary factor in underlying pathogenesis of cervical cancer, it is more objective than a tool that is based solely on average group performance. The baseline data on known clinical cases demonstrate that when analyzing hrhpv-positive rates by cytologic interpretation, SIL cases had a high rate (90%) as expected, and this finding is similar to published data. 9 The hrhpv rate by patient age for all cytologic categories showed a bimodal distribution of high rates, with approximately a 50% positive rate for patients younger than 30 years and older than 70 years of age, and a rate of 14% to 34% for ages in between, which is also similar to published data. 10 The subset analysis of ASCUS cases shows that the hrhpv rate is as useful and reliable as the ASCUS to SIL ratio. The ASCUS:SIL for individual pathologists, as well as the overall laboratory (1.46), fell TABLE 4 hrhpv Negative Rate Versus Modified ASCUS:SIL Ratio Versus Standard ASCUS:SIL Ratio by Patient Age Age hrhpv negative rate (%) Modified ASCUS:SIL ratio (%) Standard ASCUS:SIL ratio /100 (44.0) 132/257 (51.4) 132:125 (1.06) /795 (42.9) 1085/2181 (49.7) 1085:1096 (0.99) /681 (61.2) 891/1462 (60.9) 891:571 (1.56) /468 (77.1) 566/825 (68.6) 566:259 (2.19) /237 (75.9) 327/453 (72.2) 327:126 (2.60) /94 (78.7) 135/182 (74.2) 135:47 (2.87) 70þ 38/47 (80.9) 61/81 (75.3) 61:20 (3.05) Overall 1455/2422 (60.1) 3197/5441 (58.8) 3197:2244 (1.42)

5 Use of HPV Testing for QA in ASCUS/Ko et al. 71 FIGURE 2. hrhpv-negative rate versus modified ASCUS:SIL by patient age. Blue represents hrhpv-negative rate; red represents modified ASCUS:SIL. within established guidelines. 11 The hrhpv-positive rate for all ASCUS cases in our population was 40%. Interestingly, the hrhpv-positive rate for ASCUS had a similar age distribution to the analysis for all cytologic diagnoses, indicating that hrhpv prevalence follows a well-defined age association. Individual cytopathologists had overall hrhpv-positive rates that ranged from 32.8% to 53.4%. Over time, the whole laboratory hrhpv-negative rate trended well with the whole laboratory modified ASCUS:SIL, and the overall hrhpv-positive rate ranged from 33.3% to 50.7% by calendar quarter. The standard deviation of the hrhpv rate over time and by cytopathologist was 6.2 and 6.5, respectively, which fell within the modified ASCUS:SIL corresponding standard deviations of 3.2 and 9.3; the hrhpv rate was less variable than the modified ASCUS:SIL by cytopathologist but was more variable when analyzed over time, indicating that the variability of the 2 methods is similar. The hrhpv-negative rate also strongly correlated (R 2 ¼ ) with the modified ASCUS:SIL over different age groups (Fig. 3). In addition, the volume of cases examined did not appear to correlate with the hrhpv rate or the ASCUS:SIL (Table 5). This close match of the hrhpv rate with the standard quality assurance tool lends additional evidence to support its use as a quality assurance tool for laboratory and individual cytologist performance. A recent study by Chhieng et al did not find an association between the HPV rate and the ASC:SIL for their laboratory. 12 In their study, the overall HPVpositive rate was 34% with an overall ASC:SIL ratio of 2.0. The 5 pathologists studied had an HPV-positive rate that ranged from 31% to 38% and an ASC:SIL that ranged from 1.2 to 4.0. The pathologists with the FIGURE 3. hrhpv-negative rate versus modified ASCUS:SIL by patient age. highest (4.0) and lowest (1.2) ASC to SIL ratios had essentially the same HPV-positive rates (31% and 32%, respectively). In comparison, our study converted the ASCUS:SIL to a modified ASCUS:SIL (ASCUS 7 [ASCUS þ SIL]), and we compared it with the hrhpv-negative rate by patient age, over time, and by individual pathologist. This conversion enabled a more effective comparison to hrhpv-rate data and a more clear-cut demonstration of correlation and trends. Although this conversion is not intended to replace the standard way of calculating the ASCUS to SIL ratio, the technique can be useful for laboratories that choose to graphically compare the 2 methods. If our study examined data solely by individual pathologist, we may have also concluded that the hrhpv rate does not correlate very well with the ASCUS:SIL. However, by analyzing the data over time, and especially by patient age, we found strong evidence that the hrhpv rate and ASCUS to SIL ratio do have a correlation. Having shown that the hrhpv rate is as useful as the ASCUS:SIL, we took a final step in the study to determine a desirable benchmark hrhpv rate for quality assurance purposes. If ASCUS is defined as a perfectly equivocal diagnosis that falls midway between normal and SIL, then with a perfect hrhpv test, 50% of ASCUS cases should be hrhpv positive and 50% should be hrhpv negative. Individuals who have a high rate of hrhpv-positive ASCUS cases compared with the benchmark are most likely undercalling SIL, and, conversely, individuals who have a low rate of hrhpv-positive ASCUS cases are most

6 72 CANCER (CANCER CYTOPATHOLOGY) April 25, 2007 / Volume 111 / Number 2 TABLE 5 hrhpv Negative Rate Versus Modified ASCUS:SIL Ratio Versus Standard ASCUS:SIL Ratio by Cytopathologist hrhpv neg. rate (%) Modified ASCUS:SIL (%) Standard ASCUS: SIL ratio (%) 1 345/534 (64.6) 696/1059 (65.7) 696:363 (1.92) 2 107/160 (66.9) 220/403 (54.6) 220:183 (1.20) 3 181/325 (55.7) 444/685 (64.8) 444:241 (1.84) 4 66/99 (66.7) 144/214 (67.3) 144:70 (2.06) 5 123/183 (67.2) 232/474 (48.9) 232:242 (0.96) 6 63/101 (62.4) 124/176 (70.5) 124:52 (2.38) 7 199/348 (57.2) 482/870 (55.4) 482:388 (1.24) 8 31/58 (53.4) 78/173 (45.1) 78:95 (0.82) 9 35/59 (59.3) 78/183 (42.6) 78:105 (0.74) /424 (57.5) 515/863 (59.7) 515:348 (1.48) 11 61/131 (46.6) 184/341 (54.0) 184:157 (1.17) Overall 1455/2422 (60.1) 3197/5441 (58.8) 3197:2244 (1.42) Mean 59.8% 57.1% 1.44 Std Deviation 6.5% 9.3% 0.54 FIGURE 4. hrhpv-negative rate versus modified ASCUS:SIL by cytopathologist. likely overcalling negative cases. In real practice, cytologists appear to show a tendency to overcall normal specimens to ASCUS more than any other misclassification, which will depress the hrhpv-positive rate below the ideal. Some experts believe a perfectly equivocal ASCUS category in conjunction with the Hybrid Capture II assay should result in an hrhpv-positive rate of 55%, representing 50% equivocality plus a 5% baseline of hrhpv-positive cases in the normal population (MH Stoler, unpublished data). The ALTS trial found that ASCUS cases had a 50.6% hrhpvpositive rate when cases were adjudicated by an expert panel who used rigorous morphologic criteria, with a range of 31.0% to 59.7% by individual casereferral centers. 13 A College of American Pathologists (CAP) Q-PROBES 2005 article reported on 3302 Papaniculaou tests from 68 institutions that had an interpretation of ASCUS or atypical squamous cells cannot exclude HSIL (ASC-H) and an HPV test. The overall HPV-positive rate was 43.74% with an overall ASCUS to SIL ratio of Data from our laboratory are very similar to the CAP Q-PROBES 2005 study and to the overall data from the ALTS trial. Given published data as well as data generated in this study, an hrhpv-positive rate from 40% to 50% appears to be the practical ideal, and large deviations from this rate may be used to track a tendency to overcall or undercall the ASCUS category. Comparisons of the hrhpv rate and the ASCUS: SIL may provide additional insight into individual performance. For example, by using an admittedly arbitrary cutoff of >10%, 3 cytopathologists (2, 5, and 9) had modified ASCUS:SIL that differed by this amount from the mean hrhpv-negative rate (Table 5). Interestingly, Pathologists 2 and 5 also appeared to have ideal ASCUS to SIL ratios (1.20 and 0.96), but they had relatively low hrhpv-positive rates (33.1% and 32.8%), which fall below our proposed desirable hrhpv rate. Pathologist 9 had the lowest ASCUS:SIL of the group (0.74) but had an hrhpv-positive rate of 40.7%, which is on the borderline of a desirable hrhpv rate. The volume of cases examined does not correlate with these findings. These apparent discrepancies between the hrhpv-negative rate and modified ASCUS:SIL suggest that in comparison to the overall laboratory, these pathologists may be overcalling both ASCUS and SIL, which would lower their hrhpv-positive rate while maintaining a desirable ASCUS to SIL ratio. This demonstrates the utility of the hrhpv rate, which can identify potential situations in which individuals may be overcalling or undercalling ASCUS and SIL while continuing to maintain a satisfactory ASCUS:SIL. In addition, the hrhpv-positive rate may be falsely increased to a

7 Use of HPV Testing for QA in ASCUS/Ko et al. 73 more ideal level if an individual has a tendency to undercall SIL; however, this would result in an increase in the ASCUS:SIL; so when used together, both methods can reflect accuracy to a better degree than either method alone. It is still important for an individual laboratory to determine an appropriate hrhpv rate for the population that it serves, given the much higher hrhpv-positive rate in patients aged younger than 30 years (>56%) compared with patients aged older than 30 years (19% to 39%) even when all patients have an interpretation of ASCUS. Clearly similar cytologic features in diverse populations may have substantial differences in predictive values for neoplasia; hence, the objectivity of hrhpv testing increases in relevance. In the current study, hrhpv tests were performed on approximately 75% of examined ASCUS specimens. As the use of reflex testing for all cases of ASC increases, this method clearly shows promise as an ongoing marker of the quality of cytologic interpretation. In summary, hrhpv testing is a more objective method than the ASCUS:SIL. hrhpv testing is highly sensitive in detecting the virus that is a necessary factor in the pathogenesis of cervical cancer. The ASCUS:SIL, on the other hand, is a useful number only when accurate interpretations are made of both atypical cells and squamous intraepithelial lesions by individual cytologists and the overall laboratory. Despite differences in each technique, the hrhpv rate closely recapitulates the ASCUS:SIL and may be more useful in analyzing not only global laboratory performance but also variation among individual pathologists. Calculation of the hrhpv rate is as quick and cost effective as finding the ASCUS to SIL ratio, and as hrhpv reflex testing for ASCUS cases increases, this tool can provide a substantial improvement in the specificity of all ambiguous cases. REFERENCES 1. Solomon D, Nayar R. The Bethesda System for Reporting Cervical Cytology. 2nd ed. New York: Springer-Verlag; Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA. 2002;287: Wright TC Jr, Schiffman M, Solomon D, et al. Interim guidance for the use of human papillomavirus DNA testing as an adjunct to cervical cytology for screening. Obstet Gynecol. 2004;103: Davey DD, Naryshkin S, Nielsen ML, Kline TS. Atypical squamous cells of undetermined significance: interlaboratory comparison and quality assurance monitors. Diagn Cytopathol. 1994;11: Jones BA, Davey DD. Quality management in gynecologic cytology using interlaboratory comparison. Arch Pathol Lab Med. 2000;124: Stoler MH, Schiffman M, Atypical Squamous Cells of Undetermined Significance-Low-grade Squamous Intraepithelial Lesion Triage Study (ALTS) Group. Interobserver reproducibility of cervical cytologic and histologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study. JAMA. 2001;285: Walboomers JMM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol. 1999;189: Comment in: J Pathol. 1999;189: zur Hausen H. Papillomaviruses causing cancer: evasion from host-cell control in early events in carcinogenesis. J Natl Cancer Inst. 2000;92: Stoler MH. Testing for human papillomavirus: data driven implications for cervical neoplasia management. Clin Lab Med. 2003;23: Eltoum IA, Chhieng DC, Roberson J, McMillon D, Partridge EE. Reflex human papilloma virus infection testing detects the same proportion of cervical intraepithelial neoplasia grade 2 3 in young versus elderly women. Cancer. 2005; 105: Cytopathology Checklist. Commission on Laboratory Accreditation, Laboratory Accreditation Program. College of American Pathologists. 10/06/ Chhieng DC, Chen J, Connolly K, Roberson J, Eltoum I. High-risk HPV DNA detection rate in patients with atypical squamous cells and its relationship to the atypical squamous cell: squamous intraepithelial lesion ratio. Acta Cytol. 2006;50: Solomon D, Schiffman M, Tarone R, ALTS Study group. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst. 2001;93: Tworek J, Raab S, Jones B, Clary K. Q-Probes HPV Testing: Data Analysis and Critique. College of American Pathologists, Northfield, IL; Jan