In 1988, The Bethesda System (TBS) introduced the phrase atypical

Transcription

1 CANCER CYTOPATHOLOGY 93 Should the Cytologic Diagnosis of Atypical Squamous Cells of Undetermined Significance Be Qualified? An Assessment Including Comparison between Conventional and Liquid-Based Technologies Rose C. Anton, M.D. Ibrahim Ramzy, M.D. Mary R. Schwartz, M.D. Pamela Younes Subhendu Chakraborty Dina R. Mody, M.D. Department of Pathology, Baylor College of Medicine, Houston, Texas. BACKGROUND. The diagnosis of atypical squamous cells of undetermined significance (ASCUS) remains an enigma for the treating physician, because it encompasses both benign, reactive, as well as preneoplastic and possibly neoplastic conditions. To address this problem, The Bethesda System recommends qualifying the ASCUS diagnosis. This study analyzes the difference in the follow-up results between the various groups of patients with an ASCUS diagnosis as outlined by the Criteria Committee: favor premalignant (ASCUS-P), favor reactive (ASCUS-R), and unqualified (ASCUS-C). The outcome, based on follow-up biopsies and/or cytologies, for both the conventional Papanicolaou smear (CS) and liquid-based (LB) methodologies is compared. METHODS. The CS and LB biopsies and/or cytologies included 590 patients and 137 patients, respectively, who had an initial diagnosis of ASCUS. The final outcome after subsequent biopsy and cytology within a 1-year period for each methodology was tabulated. Furthermore, the addition of qualifiers for each diagnosis was tabulated for both cytology and biopsy follow-up and compared between CS and LB technologies. RESULTS. For CS, 29 patients (8.6%) were found to have squamous intraepithelial lesions (SIL) on subsequent cytologic smears, and 176 patients (63.7%) had SIL on biopsy follow-up. For LB, these numbers were 17 patients (19.1%) and 56 patients (65.1%), respectively. Regardless of the qualification used in the initial CS biopsy or cytology, over 90% of the subsequent smears resulted in a benign diagnosis. Biopsy outcomes after CS diagnoses of ASCUS-P, ASCUS-R, and ASCUS-C showed SIL in 80%, 53%, and 66% of patients, respectively. For LB diagnoses, subsequent smears detected SIL in 33% of patients with a diagnosis of ASCUS-P, 22% of patients with a diagnosis of ASCUS-R, and 12% of patients with a diagnosis of ASCUS-C. Biopsy outcomes after LB diagnoses of ASCUS-P, ASCUS-R, and ASCUS-C showed SIL in 68%, 64%, and 66% of patients, respectively. CONCLUSIONS. There appears to be no statistically valid benefit from a clinical point of view in qualifying an ASCUS interpretation by either CS or LB, except for CS evaluation of patients with a diagnosis of ASCUS-P. Cancer (Cancer Cytopathol) 2001;93: American Cancer Society. Address for reprints: Rose C. Anton, M.D., The Methodist Hospital, MS-205, 6565 Fannin St., Houston, Texas ; Fax: ; ranton@bcm.tmc.edu. Received May 1, 2000; revision received November 30, 2000; accepted December 6, KEYWORDS: cervical cytology, smears, cervix, carcinoma, Papanicolaou, screening, dysplasia. In 1988, The Bethesda System (TBS) introduced the phrase atypical squamous cells of undetermined significance (ASCUS) as the preferred terminology for the arbitrary and often overused term atypia American Cancer Society

2 94 CANCER (CANCER CYTOPATHOLOGY) April 25, 2001 / Volume 93 / Number 2 In this new classification, ASCUS is defined as squamous cell abnormalities that are not directly attributable to reactive cellular changes but that fall short of the criteria diagnostic for squamous intraepithelial lesion (SIL). Because the term ASCUS does not represent a specific diagnostic pathologic entity but, rather, a spectrum of cellular changes ranging from benign to preinvasive lesions, this entity often creates a dilemma for the treating physician. The clinical and biologic significance of ASCUS on cytologic smears remains unknown due in large part to the broad range ( %) in reported prevalence rates of SIL in women whose Papanicolaou (Pap) smears were reported as ASCUS The majority of these diagnoses, however, represent low grade lesions, with very few women having invasive disease, and only 15 30% of all SIL representing high grade lesions. 4,11 14 Although a Pap smear interpretation of high grade SIL or a macroscopically obvious cervical lesion are accepted indications for colposcopy and biopsy, the ideal management of patients with lesser lesions remains controversial. Despite the relatively low but clinically significant risk for a woman to be diagnosed with a high grade SIL after an initial diagnosis of ASCUS on Pap smear, many advocate immediate referral to colposcopy to avoid the possibility of failing to diagnose a high grade lesion. Other experts say that the safest and most cost-effective management of these patients is the performance of serial Pap smears with the addition of colposcopy only if a significant lesion becomes evident on the subsequent smears. To assist the treating physician in selecting the most appropriate treatment plan, TBS recommends that the cytologist further qualify the ASCUS diagnosis on the basis of whether a reactive process or a preneoplastic process is favored. Although proper communication with the treating physician is ideal in conveying a level of concern regarding an ASCUS smear, it is not always feasible. This is especially true with the advent of managed care, in which smears may be interpreted in laboratories at some distance from the patient. In addition, the diagnosis of ASCUS is inherently subjective and susceptible to intraobserver and interlaboratory variation. 4,12,14 It appears that, in today s medical practice, ascertaining the true optimal, cost-effective patient management after an ASCUS interpretation is innately unattainable. In recent years, liquid-based (LB), thin-layer technologies have shown promise in reducing the false negative rate of the conventional Pap smear (CS). A similar reduction in the ASCUS rate has been suggested. These LB preparations may help decrease the often wastebasket use of ASCUS by improved sensitivity: increasing cell number and, thus, curtailing the possibility of sampling error; reducing obscuring material (such as blood, mucus, and inflammation); and limiting preparation artifacts. To the best of our knowledge, no published report has examined the results and follow-up data after an initial ASCUS diagnosis by both LB and CS technologies and the diagnostic significance of the addition of qualifiers to subcategorize the ASCUS diagnosis. Therefore, this study was designed to evaluate the differences in patient outcome on follow-up cytology or biopsy of the three categories of ASCUS, as outlined by the Bethesda System Criteria Committee: reactive (ASCUS-R), premalignant (ASCUS-P), and unqualified (ASCUS-C). In addition, ASCUS follow-up for LB and CS are compared to determine whether the use of qualifiers is more applicable for either of the methodologies. MATERIALS AND METHODS Eight hundred sixty-five patients had a diagnosis of ASCUS with follow-up biopsies or cytologies within a 1-year period at one of the laboratories in our institution (Baylor College of Medicine, Houston, TX), which provides cytology and histology services to a single, large group of gynecologists in private practice. These included 704 patients with CS and 161 patients with LB preparations. Patients who had follow-up cytologies consisting only of ASCUS or AGUS were excluded, leaving 590 patients and 137 patients respectively for the study. The cytologic criteria used to determine ASCUS were those outlined by TBS. At the time of the original sign-out, the cytopathologist indicated whether an ASCUS-R, ASCUS-C, or ASCUS-P process was favored. The cytologic and histologic outcomes obtained after the initial ASCUS diagnosis were tabulated under the following categories: within normal limits (WNL), benign cellular changes (BCC), low grade SIL (LSIL), or high grade SIL (HSIL). The numbers of patients with benign (WNL or BCC) or SIL (LSIL or HSIL) diagnoses on cytologic and histologic follow-up for each ASCUS qualification were recorded for both groups. Statistical Analysis A Fisher exact chi-square test was performed to determine the significance of cytology and/or biopsy outcome data after an initial ASCUS diagnosis by either CS or LB technologies (Stata Inc., College Station, TX). The use of qualifiers was then assessed to determine whether subcategorizing an ASCUS diagnosis by either CS or LB technologies proved significant. P values 0.05 were considered statistically significant.

3 Should ASCUS Be Qualified?/Anton et al. 95 FIGURE 1. The results of cytologic evaluations after an initial diagnosis of atypical squamous cells of undetermined significance from conventional smears and liquid-based preparations. WNL: within normal limits; BCC: benign cellular changes; LSIL: low grade squamous intraepithelial lesion; HSIL: high grade squamous intraepithelial lesion. FIGURE 2. The results of biopsies after an initial diagnosis of atypical squamous cells of undetermined significance from conventional smears and liquid-based preparations. WNL: within normal limits; BCC: benign cellular changes; LSIL: low grade squamous intraepithelial lesion; HSIL: high grade squamous intraepithelial lesion. Finally, positive and negative predictive values for an ASCUS qualification by CS or LB were generated. RESULTS CS Three hundred thirty-six patients had follow-up cytologic smears obtained within 1 12 months after the initial ASCUS diagnosis, with a mean of 8.3 months. After the initial diagnosis of ASCUS, the results in 52 patients (15.5%) were WNL, the results in 255 patients (75.9%) showed BCC, the results in 25 patients (7.4%) showed LSIL, and the results in 4 patients (1.2%) showed HSIL. Hence, 307 patients (91.4%) had follow-up cytologic results that were benign, and 29 patients (8.6%) has results that showed SIL. These results are demonstrated in Figure 1. The overall results for 276 patients who had follow-up biopsies within a 1-year period are summarized in Figure 2. The follow-up biopsies were obtained within 1 12 months after the initial ASCUS diagnosis, with a mean of 2.6 months. After the initial interpretation of ASCUS, the follow-up biopsies for 100 patients (36.3%) were nonneoplastic. Eleven patients (4.0%) had biopsies that demonstrated no pathologic alteration, and 89 patients (32.3%) had biopsies that showed BCC. One hundred seventy-six patients (63.7%) demonstrated SIL on a follow-up biopsy, of which 166 (60.1%) were LSIL, and 10 (3.6%) were HSIL. Seventy-five patients had a qualified ASCUS-P diagnosis on a conventional smear. Seventy of these patients (93%) had a benign diagnosis on a follow-up cytologic smear, with only 5 patients (7%) demonstrating SIL. Of the 172 patients with an initial qualified ASCUS-R diagnosis, 162 patients (94%) demonstrated WNL or BCC on a follow-up smear, and 10 patients FIGURE 3. The results from patients who had benign cytologic evaluations after the initial qualification of a diagnosis of atypical squamous cells of undetermined significance (ASCUS) from conventional smears and liquid-based preparations. ASCUS-R: reactive ASCUS; ASCUS-P: premalignant ASCUS; ASCUS-C: unqualified ASCUS. (6%) had SIL. Among the 93 ASCUS patients who could not be subcategorized further, 87 patients (94%) eventually showed WNL or BCC, and 6 patients (6%) demonstrated SIL. These results are summarized in Figures 3 and 4. Figures 5 and 6 summarize the outcome data for 250 patients who had a biopsy after their initial ASCUS diagnosis with qualification. For the 99 patients who were subcategorized with ASCUS-P, most (79 patients, 80%) were found to have SIL on the subsequent biopsy. The 89 patients who were favored to have a reactive process (ASCUS-R) showed results that were more divergent: 42 patients (47%) demonstrated a benign diagnosis on follow-up, whereas 47 patients (53%) showed SIL. Most of the 62 patients whose smears could not be subcategorized further showed SIL on follow-up (41 patients; 66%). LB Eighty-nine patients had follow-up cytologic smears obtained within 1 12 months after the initial ASCUS

4 96 CANCER (CANCER CYTOPATHOLOGY) April 25, 2001 / Volume 93 / Number 2 FIGURE 4. The results from patients who had squamous intraepithelial lesions on cytologic evaluation after the initial qualification of a diagnosis of atypical squamous cells of undetermined significance (ASCUS) with conventional smears and liquid-based preparations. ASCUS-R: reactive ASCUS; ASCUS-P: premalignant ASCUS; ASCUS-C: unqualified ASCUS. FIGURE 6. The results from patients who had squamous intraepithelial lesions on biopsy evaluation after the initial qualification of a diagnosis of atypical squamous cells of undetermined significance (ASCUS) with conventional smears and liquid-based preparations. ASCUS-R: reactive ASCUS; ASCUS-P: premalignant ASCUS; ASCUS-C: unqualified ASCUS. FIGURE 5. The results from patients with benign biopsies after the initial qualification of a diagnosis of atypical squamous cells of undetermined significance (ASCUS) with conventional smears and liquid-based preparations. ASCUS-R: reactive ASCUS; ASCUS-P: premalignant ASCUS; ASCUS-C: unqualified ASCUS. diagnosis, with a mean of 8.1 months. After the initial diagnosis of ASCUS, 72 patients (80.9%) had follow-up cytologies that were benign. Three patients (3.4%) had smears that showed WNL, and 69 (77.5%) demonstrated BCC. Seventeen patients (19.1%) showed SIL, of which 14 patients (15.7%) displayed LSIL, and 3 (3.4%) had HSIL. These results are summarized in Figure 1. The overall results for 86 patients who had follow-up biopsies within a 1-year period are summarized in Figure 2. The follow-up biopsies were obtained within 1 12 months after the initial ASCUS diagnosis, with a mean of 2.5 months. After the initial interpretation of ASCUS, the follow-up biopsies for 30 patients (34.9%) were benign. Twelve patients (14.0%) had biopsies that demonstrated no pathologic alteration, and 18 patients (20.9%) had biopsies that showed BCC. Fifty-six patients (65.1%) demonstrated SIL on a follow-up biopsy, of which 49 patients (57.0%) had LSIL, and 7 patients (8.1%) had HSIL. Eighteen patients had a qualified ASCUS-P diagnosis on a conventional smear. Twelve of these patients (67%) had a benign diagnosis on a follow-up cytologic smear. The remaining six patients (33%) demonstrated SIL on a subsequent smear. Of the 32 patients with an initial qualified ASCUS-R diagnosis, 25 (78%) demonstrated WNL or BCC on a follow-up smear. Seven patients (22%) had SIL on follow-up. Among the 40 patients with ASCUS who could not be subcategorized further, 35 patients (88%) eventually showed WNL or BCC, and 5 patients (12%) demonstrated SIL. These results are summarized in Figures 3 and 4. Figures 5 and 6 summarize the outcome data for 88 patients who had a colposcopy/biopsy after their initial ASCUS diagnosis with qualification. For the 31 patients who were subcategorized with ASCUS-P, most (21 patients; 68%) were found to have SIL on the subsequent biopsy, and the remaining 10 patients (32%) had a negative diagnosis. Sixteen of 25 patients (64%) who initially were believed to have a reactive process demonstrated SIL on histologic examination, whereas only 9 patients (36%) eventually were found to have a benign diagnosis. Most of the 32 patients who could not be subcategorized further showed SIL on follow-up (21 patients; 66%). Eleven patients (34%) demonstrated a benign process. Statistical Analysis An ASCUS diagnosis by LB technology was far more likely to result in an SIL diagnosis on cytology followup, with 17 of 89 patients who had an SIL outcome compared with 29 of 336 patients by CS (P ). The results remained statistically significant when cytology and biopsy follow-up data were combined, with

5 Should ASCUS Be Qualified?/Anton et al. 97 an SIL outcome in 73 of 175 patients by LB compared with 205 of 612 patients by CS (P ). However, when only the biopsy data were analyzed, the results were insignificant (56 of 86 patients by LB vs. 176 of 276 patients by CS; P ). The cytology, biopsy, and combined cytology and biopsy follow-up data subsequent to an initial ASCUS diagnosis with qualification on CS were then analyzed. The results for cytology follow-up were not significantly different for any ASCUS qualification: ASCUS-P versus ASCUS-R (P ), ASCUS-P versus ASCUS-C (P ), or ASCUS-R versus ASCUS-C (P ). The biopsy results proved highly significant for an ASCUS-P diagnosis compared with an ASCUS-R diagnosis (P ). However, it remained insignificant for an ASCUS-P diagnosis compared with an ASCUS-C diagnosis (P ) and an ASCUS-R compared with an ASCUS-C diagnosis (P ). When the cytology and biopsy follow-up results were combined, an ASCUS-P diagnosis compared with an ASCUS-R diagnosis was significant (P ), and an ASCUS-P diagnosis compared with an ASCUS-C diagnosis was significant (P ), whereas the ASCUS-R qualification compared with the ASCUS-C qualification approached significance (P ). The data for an ASCUS qualification by LB methodology indicate that these qualifications are even less helpful in suggesting SIL or a reactive lesion compared with the CS methodology. The cytology outcome values were insignificant for ASCUS-P compared with ASCUS-R (P ), ASCUS-P compared with ASCUS-C (P ), and ASCUS-R compared with ASCUS-C (P ). Likewise, the biopsy data were insignificant for all comparisons: ASCUS-P versus ASCUS-R (P ), ASCUS-P versus ASCUS-C (P ), and ASCUS-R versus ASCUS-C (P ). For the cytology and biopsy combined follow-up, ASCUS-P versus ASCUS-C was the only significant comparison (P ), whereas ASCUS-P versus ASCUS-R and ASCUS-R versus ASCUS-C remained insignificant (P and P , respectively). The positive predictive value of an ASCUS-P interpretation by CS was 6.7% for cytology follow-up, 79.8% for biopsy follow-up, and 48.3% for cytology and biopsy follow-up data combined. The corresponding positive predictive values by LB were 33.3%, 67.7%, and 55.1%, respectively. The negative predictive value of an ASCUS-R interpretation by CS was 94.2% for cytology follow-up, 53.9% for biopsy followup, and 78.2% for cytology and biopsy follow-up data combined. The corresponding negative predictive values by LB were 78.1%, 36.0%, and 59.6%, respectively. DISCUSSION Since its introduction over 50 years ago, the Pap smear has become a valuable screening tool for the detection of cervical carcinoma. More recently, newer methodologies, including LB smears and human papilloma virus testing, have been added in an attempt to increase the sensitivity of the procedure. Several studies have reported an increase in SIL detection with LB smears Comparisons of the ASCUS category between CS and LB have been less clear cut. Previously published studies have reported both increased and decreased ASCUS rates, further confusing the ASCUS 16,19 22 interpretation. Since the incorporation of LB preparations at our institution, the percentage of patients diagnosed with ASCUS decreased from an initial rate of 6.8% to 3.2%, with 70% of our volume consisting of LB preparations during the study period. Our study summarizes the outcome data (biopsy and cytology) after an initial ASCUS interpretation on CS and LB of a large private practice, low risk population. The data regarding subsequent cytologic smears demonstrate an enhanced SIL detection rate with the LB methodology (19% by LB compared with 9% by CS; P ). The reported improved quality and quantity of cells offered by this technique may play some role. However, this study is limited by the fact that the subsequent cytologic smears after the initial LB ASCUS interpretation consisted of both CS and LB smears. Therefore, no definite conclusion can be drawn between the two methodologies based on the cytology results alone. Similar to previous reports, our colposcopy with biopsy follow-up data were far more sensitive in the detection of SIL than subsequent smears for both CS (63.7% vs. 8.6%) and LB (65.1% vs. 19.1%). This is most likely because of the better visualization and direct sampling offered by colposcopy. However, patient selection bias by the treating physician cannot be excluded entirely. Significantly, however, our data recognize that ASCUS interpretations by LB and CS technologies yielded similar SIL detection rates on subsequent biopsies (65.1% vs. 63.7%, respectively; P ). These findings echo those of Carpenter and Davey 19 and indicate that an ASCUS interpretation by LB warrants management protocols similar to those for CS. Fortunately, the great majority of SIL on subsequent biopsy are low grade, with LB preparations having a slight edge in detecting HSIL (13% of SIL vs. 6% of HSIL, respectively). Overall, the above data parallel what has been reported previously in the literature, in which SIL has been demonstrated in 58 72% of women who had follow-up biopsies. 8,14,23 However, several other stud-

6 98 CANCER (CANCER CYTOPATHOLOGY) April 25, 2001 / Volume 93 / Number 2 ies reported a slightly lower SIL rate. Selvaggi and Haefner 11 noted SIL in 37% of subsequent tissue specimens. Forty-six percent of women in the study by Auger et al. 24 had SIL or invasive carcinoma on subsequent biopsy. Carpenter and Davey, 19 in a study comparing LB and CS smears, noted an SIL outcome in 21% and 24% of patients, respectively. Our data concerning ASCUS qualifications are unique. Regardless of the qualification used in the initial CS, over 90% of the subsequent smears resulted in a benign diagnosis. Ninety-three percent of women who were favored to harbor a preneoplastic lesion ultimately were diagnosed with a normal or reactive pattern. Similar results were noted for women whose initial ASCUS interpretation by LB resulted in a subsequent smear: two-thirds in the favor neoplastic category had a negative diagnosis on subsequent smear. Because the repeat smears obtained after the initial ASCUS diagnosis by LB consisted of both CS and LB smears, we cannot consider the cytology follow-up as our gold standard for this study. More relevant are the data obtained after subsequent biopsy, because issues concerning the final outcome procedure and sampling are similar. Eighty percent of women in the ASCUS favor preneoplastic category after CS were diagnosed with SIL on follow-up biopsy. Favor reactive was more equivocal, with an almost equal likelihood of indicating SIL (53%) or a negative finding (47%). The biopsy outcome for LB smears reflects the cytology data, with an equal probability of SIL regardless of the qualification. When comparing ASCUS qualifications for CS and LB technologies, the only value that proved statistically significant using the biopsy outcome data was the increased likelihood of finding SIL with an ASCUS-P interpretation compared with an ASCUS-R interpretation by CS (P ). In general, our results indicate little practical benefit from qualifying an ASCUS diagnosis. The uniformity of the data obtained by the LB technique may reflect a comfort level as cytologists adjust to a new technique. As cytotechnologists and pathologists gain more experience in LB interpretation, the outcome results for ASCUS qualifications may become more divergent. Our findings are not in accord with the majority of previous studies, the findings of which champion the value of qualification. Ettler et al. 25 found SIL in 73% of their favor dysplasia group and a negative diagnosis in 73% of the favor reactive group. The figures cited by Gonzalez et al. 26 demonstrate a very low SIL rate, regardless of further classification. Six percent of the favor reactive smears and 30% of the favor dysplasia smears resulted in SIL in the study by Genest et al. 27 Ten percent of women in the favor reactive group of Collins et al. 28 resulted in an SIL diagnosis. Kline and Davey 29 noted that only the favor dysplastic category was useful in their laboratories, with an SIL outcome rate of 30% for their favor reactive category. These findings were corroborated by Malik et al., 30 who, although they found that 78% of women in the favor dysplasia category had SIL, noted SIL in 48% of women in the favor reactive category. 30 It appears that the clinical utility of qualification varies among laboratories. When comparing the positive predictive values of an ASCUS-P interpretation by CS and LB technologies, we found that CS was associated more strongly with SIL on biopsy follow-up than LB (79.8% vs. 67.7%, respectively) due to the lower false positive rate seen with CS. Similarly, CS proved superior to LB when assessing the false negative rate of an ASCUS-R interpretation (53.9% vs. 36.0%, respectively). In other words, an ASCUS-P diagnosis qualified by CS was more likely to result in the detection of SIL on follow-up compared with an ASCUS-P diagnosis qualified by LB, and an ASCUS-R diagnosis qualified by CS was more likely to be benign on follow-up than an ASCUS-R diagnosis qualified by LB. One reason for our unique findings may be the interobserver variation and the lack of reproducibility of the criteria used to establish the favor reactive or preneoplastic categories, comparable to the hypothesis attempting to explain the wide variation in SIL rates after the initial ASCUS diagnosis. The use of cytologic and/or histologic correlations is a valid laboratory quality-assurance tool that may help standardize these rates. In summary, the current study demonstrates that a number of patients who are diagnosed with ASCUS in all categories by either CS or LB methodologies have ASCUS associated with a clinically significant number of SILs, warranting further patient evaluation. Although no definite conclusion concerning optimal patient management can be drawn, we feel that the benefit of adding qualifiers to the ASCUS interpretation will vary among laboratories but can play some role in assisting the treating physician. The ideal approach may be laboratory independent and may rest with individualizing each patient based on their history, clinical suspicion, and patient compliance. REFERENCES 1. National Cancer Institute Workshop. The 1988 Bethesda System for reporting cervical/vaginal cytologic diagnoses. JAMA 1989;262: Abu-Jawdeh GM, Trawinski G, Nang HH. Histocytological study of squamous atypia on pap smears. Mod Pathol 1994; 7:920 4.

7 Should ASCUS Be Qualified?/Anton et al Busseniers AE, Sidawy MD. Inflammatory atypia on cervical smears: a diagnostic dilemma for the gynecologist. J Reprod Med 1991;36: Davey DD, Naryshkin S, Nielsen ML, Kline TS. Atypical squamous cells of undetermined significance: interlaboratory comparison and quality assurance monitors. Diagn Cytopathol 1994;11: Davis GL, Hernandez E, Davis JL, Miyazawa K. Atypical squamous cells in Papanicolaou smears. Obstet Gynecol 1987;69: Jones DE, Creasman WT, Dombroski RA, Lentz SS, Waeltz JL. Evaluation of the atypical Pap smear. Am J Obstet Gynecol 1987;157: Noumoff JS. Atypia in cervical cytology as a risk factor for intraepithelial neoplasia. Am J Obstet Gynecol 1987;156: Sidawy MK, Tabbara SO. Reactive change and atypical squamous cells of undetermined significance in Papanicolaou smears: a cytohistologic correlation. Diagn Cytopathol 1993; 9: Slawson DC, Bennett JH, Herman JM. Follow-up Papanicolaou smear for cervical atypia: are we missing significant disease? Family Pract 1993;36: Soutter WP, Wisdom S, Brough AK, Minaghan JM. Should patients with mild atypia in a cervical smear be referred for colposcopy? Br J Obstet Gynaecol 1986;93: Selvaggi SM, Haefner HK. Reporting of atypical squamous cells of undetermined significance on cervical smears: is it significant? Diagn Cytopathol 1995;13: Solomon D, Frable WJ, Vooijs GP, Wilbur DC, Amma NS, Davey DD, et al. ASCUS and AGUS criteria: IAC Task Force summary. Acta Cytol 1998;42: Wilbur DC, Bonfiglio TA. Atypical squamous cells in cervical smears: resolving a controversy [editorial]. Diagn Cytopathol 1993;9: Williams ML, Rimm DL, Pedigo MA, Frable WJ. Atypical squamous cells of undetermined significance: correlative histologic and follow-up studies from an academic medical center. Diagn Cytopathol 1997;16: Austin RM, Ramzy I. Increased detection of epithelial cell abnormalities by liquid-based gynecologic cytology preparations. A review of accumulated data. Acta Cytol 1998;42: Dupree WB, Suprun HZ, Beckwith DG, Shane JJ, Lucente V. The promise and risk of new technology. The Lehigh Valley Hospital s experience with liquid-based cervical cytology. Cancer 1998;84: Linder J, Zahniser D. ThinPrep Papanicolaou testing to reduce false-negative cervical cytology. Arch Pathol Lab Med 1998;122: Roberts JM, Gurley AM, Thurloe JK, Bowditch R, Laverty CR. Evaluation of the ThinPrep Pap test as an adjunct to the conventional Pap smear. Med J Aust 1997;167: Carpenter AB, Davey DD. ThinPrep Pap test: performance and biopsy follow-up in a university hospital. Cancer 1999; 87: Bolick DR, Hellamn DJ. Laboratory implementation and efficacy assessment of the ThinPrep cervical cancer screening system. Acta Cytol 1998;42: Lee KR, Ashfaq R, Birdsong GG, Corkill ME, McIntosh KM, Inhorn SL. Comparison of conventional Papanicolaou smears and fluid-based, thin-layer system for cervical cancer screening. Obstet Gynecol 1997;90: Papillo JL, Zarka MA, St. John PL. Evaluation of the ThinPrep Pap test in clinical practice: a seven month, 16,314-case experience in northern Vermont. Acta Cytol 1998;42: Dvorak KA, Finnemore M, Maksem JA. Histology correlation with atypical squamous cells of undetermined significance (ASCUS) and low-grade squamous intraepithelial lesion (LSIL) cytology diagnoses: an argument to ensure ASCUS follow-up that is as aggressive as that for LSIL. Diagn Cytopathol 1999;21: Auger M, Charbonneau M, Arseneau J. Atypical squamous cells of undetermined significance: a cytohistologic study of 52 cases. Acta Cytol 1997;41: Ettler HC, Joseph MG, Downing PA, Suskin NG, Wright VC. Atypical squamous cells of undetermined significance: a cytohistological study in a colposcopy clinic. Diagn Cytopathol 1999;21: Gonzalez D, Hernandez E, Anderson L, Heller P, Atkinson BF. Clinical significance of cervical cytologic diagnosis of atypical squamous cells of undetermined significance: favoring a reactive process or low grade squamous intraepithelial lesion. J Reprod Med 1996;41: Genest DR, Dean B, Lee KR, Sheets E, Crum C, Cibas ES. Qualifying the cytologic diagnosis of atypical squamous cells of undetermined significance affects the predictive value of a squamous intraepithelial lesion on subsequent biopsy. Arch Pathol Lab Med 1998;122: Collins LC, Wang HH, Abu-Jawdeh MA. Qualifiers of atypical squamous cells of undetermined significance help in patient management. Mod Pathol 1996;9: Kline MJ, Davey DD. Atypical squamous cells of undetermined significance qualified: a follow-up study. Diagn Cytopathol 1996;14: Malik SH, Wilkinson EJ, Drew PA, Bennett BB, Hardt NS. Do qualifiers of ASCUS distinguish between low- and high-risk patients? Acta Cytol 1999;43: