- Email: [email protected]
Predictive value of diagnoses of endocervical glandular abnormalities in cervical smears
Pathology (2003) 35, pp. 198–203
ANATOMICAL PATHOLOGY
Predictive value of diagnoses of endocervical glandular abnormalities in cervical smears AMANDA SEGAL*, FELICITY A. FROST*, ALINA MIRANDA*, CHRISTINE FLETCHER{ GREGORY F. STERRETT*
AND
*Division of Tissue Pathology, The Western Australian Centre for Pathology and Medical Research, Nedlands; {Western Australian Cervical Cytology Registry, Perth, Western Australia
Summary Aims: To determine the positive predictive value (PPV) of cervical smear diagnoses of ‘definite’ and ‘possible’ endocervical adenocarcinoma in situ or invasive adenocarcinoma, and whether diagnostic accuracy can be improved. Methods: The study examined cervical smears reported as definite or possible high-grade glandular abnormality between 1992 and 1998. PPV was calculated by comparing smear diagnoses with the subsequent histopathology report. All available smears were reviewed without knowledge of follow-up results, and were reclassified by consensus. Results: Thirty-two smears were diagnosed as high-grade glandular lesions, with adequate biopsy follow-up in 31 cases (96.9%). A high-grade epithelial abnormality (HGEA) was detected in 29 cases (PPV, 93.5%), with a high-grade glandular lesion in 24 (PPV, 77.4%). Very few smears were reclassified on review. Seventy-three smears were initially diagnosed in the ‘inconclusive’ glandular or indeterminate cell-type category. There was adequate biopsy follow up for 54 cases (74.0%). On follow-up, 31 cases had a HGEA (PPV, 57.4%), with 14 cases having a high-grade glandular abnormality (PPV, 25.9%). In the review of ‘inconclusive’ smears, 12 were reclassified as squamous abnormalities and none of these had a glandular lesion on biopsy. Eight were reclassified as negative; seven contained endometrial stroma and the glandular cells in question were considered to be of lower uterine segment (LUS) origin. No significant lesion was present on follow-up of these cases. Conclusions: For clinicians using our laboratory, large loop excision of the transformation zone (LLETZ) or cone biopsy should follow a ‘definite’ cytological diagnosis of a highgrade endocervical glandular lesion. However, cone biopsy may not be the appropriate initial management in the ‘possible’ high-grade glandular group because of a significantly lower predictive value of the diagnosis. The slide review highlighted the importance of (1) caution in classifying sheets of abnormal cells as glandular, and (2) endometrial stroma as a marker of LUS material. Key words: Cervical smear, adenocarcinoma in situ, predictive value, inconclusive glandular. Abbreviations: AIS, adenocarcinoma in situ; HSIL, high-grade squamous intra-epithelial lesion; LSIL, low-grade squamous intra-epithelial lesion. Received 23 December 2002, revised 12 February, accepted 17 February 2003
INTRODUCTION In contrast to screening for squamous lesions of the cervix, cervical smears have not been shown to reduce the incidence of invasive adenocarcinoma.1,2 The reasons for this are not clear. It may partly reflect a sampling problem; however, the criteria and pitfalls in the diagnosis of precursors of adenocarcinoma have only been well defined in the last decade or so, and we may be yet to see results in this area. Difficulty in interpretation of cytological findings may be responsible for both over-diagnosis and underdiagnosis of glandular lesions. Under-calling may result from lack of awareness of certain patterns of adenocarcinoma in situ (AIS), for example, those resembling reactive endocervical cells and those with crowded small cells mimicking endometrial or high canal sampling.3,4 Overcalling of benign processes such as repair, tubo-endometrial metaplasia, cervical endometriosis and lower uterine segment sampling are also well-recognised problems.5–8 Several recent studies have examined the diagnosis of glandular lesions in cervical smears, from the point of view of sensitivity of cytological detection, accuracy of diagnosis and the most appropriate means of establishing a biopsy diagnosis.9–19 A cytological prediction of a possible or definite glandular lesion raises various management issues. As significant glandular lesions may be present in the setting of a normal colposcopic examination, either cone biopsy or large loop excision of the transition zone (LLETZ) biopsy is usually necessary for diagnosis. Knowledge of outcomes of cytological predictions of glandular lesions helps to guide the decision for a biopsy. The positive predictive value (PPV) of a diagnosis of AIS in screening laboratories can be very high, particularly in laboratories with an interest in this area. In a recent study, Roberts et al.,20 demonstrated a PPV of 95.7% for a definite diagnosis of AIS and 71.1% for an ‘inconclusive’ glandular report, enabling this laboratory to suggest that cone biopsy would be an appropriate next diagnostic step for patients in either of these cytological categories. However, there are few published papers addressing predictive value in this area of cervical cytology,21,22 and there is almost certainly wide variation between laboratories. It is important for laboratories to assess and monitor their predictive values and for clinicians to be aware of these data in order to determine appropriate management. With this in mind, the aims of this study were two-fold:
ISSN 0031–3025 printed/ISSN 1465–3931 online/03/030198–06 # 2003 Royal College of Pathologists of Australasia DOI: 10.1080/0031302031000123146
GLANDULAR ABNORMALITIES IN PAP SMEARS
1. To determine the PPV of cervical smears reported as ‘definite’ high-grade endocervical glandular abnormality or ‘possible’ high-grade glandular abnormality (‘inconclusive’) with or without associated squamous abnormality at PathCentre, a public sector screening laboratory which reports approximately 25 000 smears per year. 2. To determine whether any improvement in diagnostic accuracy can be achieved on review of smears originally diagnosed in these categories, particularly the ‘inconclusive’ glandular category.
MATERIALS AND METHODS Protocols for reporting glandular abnormalities in smears and biopsies All cervical smears were reported according to Australian National Health and Medical Research Council (NHMRC) categories and recommendations (Table 1). For biopsy reporting we did not use the diagnosis of glandular dysplasia. Although we acknowledge this may be controversial, we do not ourselves recognise a spectrum of glandular cell abnormalities of lesser degree than AIS.23–25 In interpreting histological reports from other laboratories, therefore, we have only accepted a definite diagnosis of AIS as confirmatory. Where a diagnosis of dysplasia was made (two cases), we obtained sections from these biopsies for review and used our review diagnosis in the results.
Case selection and histological follow-up All cervical smears reported in the high-grade glandular abnormality or ‘inconclusive’ glandular categories over a 7-year period between 1992 and 1998 were retrieved and reviewed by the authors. The study included smears reported as containing combined glandular and squamous abnormalities, and those where cell type was indeterminate, i.e., either squamous or glandular. Follow-up biopsy data from all pathology laboratories in Western Australia were obtained through the Western Australian Cervical Cytology Registry, and were based on codes and histology reports; histological sections of biopsies were not routinely reviewed. The follow-up included any available histology at the time of the review. The followup period ranged from 3 to 9 years, although relevant histological follow-up was usually within 6 months to 2 years of the index smear. PPV was assessed by comparing an initial smear diagnosis of a high-grade glandular lesion or an ‘inconclusive’ glandular report and the subsequent histopathology results. Cases without biopsy follow-up were reviewed, but these cases were not included in the calculation of PPV.
199
Slide review Detailed review of all available smears was undertaken to establish whether we agreed with the original diagnosis and cell type assigned, and whether we could identify any systematic errors in reporting. For the ‘inconclusive’ glandular smears, an attempt was made to refine the diagnosis into negative, low-grade or high-grade categories, and into squamous or glandular abnormality. The review was undertaken without knowledge of biopsy outcomes.
RESULTS Predictive value and review of smears diagnosed as high-grade endocervical glandular abnormality The results of the predictive value component of the study are summarised in Table 2. There were 32 smears diagnosed as either invasive adenocarcinoma, AIS alone or AIS plus high-grade squamous intra-epithelial lesion (HSIL) during the 7-year study period, with adequate biopsy follow-up in 30 cases (93.75%). One patient refused biopsy; however, this patient had a diagnosis of AIS extending to margins on cone biopsy 6 years previously and a repeat smear in this patient was also diagnosed as AIS; this case was considered to have positive follow-up for the purposes of calculation of PPV. A high-grade epithelial abnormality (HGEA) was detected on biopsy in 28 of the cases (PPV, 93.5%). A high-grade glandular abnormality was found in 24 cases (PPV, 77.4%); two were endometrial lesions and the remainder endocervical lesions. Twenty-seven slides were available for review. In 19 cases there was agreement with the original diagnosis and all 19 were confirmed to have high-grade glandular lesions on biopsy. Of four cases downgraded to ‘inconclusive’ possible high-grade glandular, one proved to be invasive adenocarcinoma, one AIS, one high-grade squamous abnormality and one negative on biopsy follow-up. Two cases on review were considered to be high-grade squamous lesions and this was confirmed on biopsy, with no glandular abnormality identified. One case was downgraded to ‘inconclusive’ of indeterminate cell type and had negative biopsy follow-up. One case was downgraded to ‘inconclusive’ squamous and had a high-grade squamous lesion with cervical endometriosis on follow-up histology. In total, there were three cases in which a high-grade lesion was not found on biopsy. In one of these cases the smear
TABLE 1 Summary of Australian National Health and Medical Research Council (NHMRC) terminology for reporting gynaecological cytology, and recommended management. Category High-grade epithelial abnormality (HGEA) Invasive carcinoma, squamous, glandular Cervical intraepithelial neoplasia (CIN) Grades 2,3 (HSIL) Adenocarcinoma in situ (AIS) Inconclusive: Possible HGEA Squamous, glandular Low-grade epithelial abnormality (LGEA) CIN, Grade 1 HPV effect NSMC squamous, endocervical glandular Negative for dysplasia and malignancy Unsatisfactory for evaluation Abbreviations: HPV, human papilloma virus; NSMC, non-specific minor changes.
Recommended management
Gynaecological referral Colposcopy Gynaecological referral Colposcopy Colposcopy Repeat smear 6–12 months Repeat smear 6–12 months Repeat smear in 2 years Repeat smear in 3 months
200
Pathology (2003), 35, June
SEGAL et al.
TABLE 2
Histological follow-up in cases diagnosed as ‘definite’ high-grade endocervical glandular abnormality. Histological diagnosis Invasive endocervical adenocarcinoma
Cytological diagnosis Invasive adenocarcinoma AIS AISzHSIL Total
5 1 6
AIS
AISzHSIL
HSIL
Endometrial carcinoma/ atypical hyperplasia
2 5 4 11
1 1 3 5
1
2
3 4
Negative
Unsatisfactory biopsy
No biopsy
2
1
1
2
1
1
2
Total 11 11 10 32
Abbreviations: AIS, adenocarcinoma in situ; HSIL, high-grade squamous intra-epithelial lesion.
glandular/squamous, two endometrial and one endometrial/ squamous). Of these 21 cases, ten had high-grade lesions on biopsy (9 HSIL and 1 HSILzAIS). Twelve cases were reclassified as either low-grade epithelial abnormality or negative, and two of these cases proved to be high-grade glandular lesions on biopsy follow-up. In total, eight ‘inconclusive’ cases were reclassified as negative and five had biopsy follow-up. In seven of these cases, stroma was present on the smear, indicating lower uterine sampling (LUS); no significant lesion was identified on follow-up (smear and/or biopsy) in these seven cases (Fig. 1). Twelve cases (23.1%) were reclassified as squamous (either HSIL or ‘inconclusive’ squamous categories). Of these reclassified cases, 66.7% had HSIL, 25% had LSIL and one case (8.3%) had negative biopsy follow-up; none of these cases had a glandular abnormality on biopsy follow-up (Fig. 2).
was considered to be ‘inconclusive’ rather than definite high-grade on review. A second case which had also been considered ‘inconclusive’ on review had human papillomavirus (HPV) changes on biopsy; however, only punch biopsies had been performed and follow-up was therefore considered unsatisfactory. A third case had negative LLETZ biopsy follow-up; however, the smear was not available for review. Predictive value and review of smears diagnosed as ‘inconclusive’: possible high grade endocervical glandular abnormality The results of the predictive value component of the study are summarised in Table 3. There were 73 smears diagnosed in the ‘inconclusive’ glandular category. Subgroups of the ‘inconclusive’ category are shown in the table. Overall, 31 cases (42.5%) demonstrated a HGEA on biopsy; however, adequate biopsy follow-up was only available for 54 cases (74.0%). Using only cases with adequate followup, the PPV for smears in the ‘inconclusive’ glandular category was 57.4%. Only 14 cases (25.9%) demonstrated a high-grade glandular abnormality, the remainder being high-grade squamous lesions. Nevertheless, in the ‘inconclusive’ cases classified as endocervical cell type on initial reporting, 13 of the 22 cases (59.1%) with an adequate biopsy had a high-grade lesion on biopsy and 10 of the 13 cases (76.9%) with high-grade lesions on biopsy proved to have a glandular component. Of the 54 ‘inconclusive’ glandular cases with adequate biopsy follow-up, there were 52 smears available for review. Nineteen smears (36.5%) were reclassified as HGEA (eight AIS, eight HSIL and three AISzHSIL) and, of these, 14 (73.7%) had a biopsy-proven high-grade lesion (five AIS, seven HSIL and two AISzHSIL). Twenty-one smears were considered to remain in the ‘inconclusive’ category following review (four squamous, four glandular, ten
DISCUSSION There were various limitations to the study including a proportion of cases where the follow-up was in the form of punch biopsy or endometrial curettings rather than cone or LLETZ biopsy; for the purposes of calculating predictive values, these cases were considered to have ‘unsatisfactory’ follow-up. There were also cases in which hysterectomy had been performed with no lesion detected. These cases may not have had an adequate examination of the cervix if the preceding smear diagnosis was not known; however, these cases were considered to have adequate (negative) follow-up for the purposes of the study. Only biopsies reported as showing endocervical ‘dysplasia’ (two cases) were reviewed; one was confirmed as AIS and the other was considered to represent minor reactive changes in endocervical cells. Otherwise, histological diagnoses were accepted on the basis of reports only.
TABLE 3 Histological follow-up in cases diagnosed as ‘possible’ high-grade epithelial lesion of endocervical glandular type (‘inconclusive’ glandular category). Histological diagnosis Invasive adenocarcinoma/ Ovarian Unsatisfactory No adenosquamous carcinoma AIS HSILzAIS HSIL carcinoma LSIL Neg biopsy biopsy Total
Cytological diagnosis Inconclusive Inconclusive Inconclusive Inconclusive Total
endocervical endocervical/squamous endocervical/endometrial endocervical/endometrial/squamous
2 1
3 1 1
5
3
5
5
3 11 2 1 17
1 1
5 8
4 4 2
2 2 2
6 6 1
13
10
6
13
Abbreviations: AIS, adenocarcinoma in situ; HSIL, high-grade squamous intra-epithelial lesion; LSIL, low-grade squamous intra-epithelial lesion.
29 34 8 2 73
GLANDULAR ABNORMALITIES IN PAP SMEARS
201
Fig. 1 Various appearances of stromal tissue in cytological smears indicating lower uterine segment or upper endocervical canal sampling. In our own experience, this material has often been overlooked in routine screening. (a,b) The material can easily be identified as endometrial. (c) Distorted stromal tissue; it is difficult to be certain whether such material is of endocervical or lower uterine segment origin (Papanicolaou stain, original magnification, a,b: 6250; c: 6400).
Fig. 2 Mimics of glandular differentiation. Cell aggregates from cases originally diagnosed as having a possible high-grade glandular component, but on review considered to be squamous. Histological follow-up: high-grade squamous lesion only. (a,b) Smooth edges to aggregates, but no well-defined layer of apical cytoplasm or columnar shapes. Fragile cytoplasm imparting an endocervical appearance, but multi-layered and disorganised sheet indicating a squamous origin. (c) Small rounded aggregate, but not a tight three-dimensional group. (Papanicolaou stain, original magnification, a: 6250; b,c: 6400).
The predictive value for a definite smear diagnosis of a glandular lesion was considered satisfactory. For all cases, prediction of a high-grade glandular abnormality on smear was followed by a biopsy proven high-grade epithelial abnormality (squamous or glandular) in 87.5% of cases, and in 75% of cases a high-grade glandular lesion was present. If only cases with adequate biopsy follow-up are included, these figures improve to 93.5 and 77.4%, respectively. These results suggest that LLETZ or cone biopsy should follow a definite cytological diagnosis of a high-grade glandular lesion, even if a lesion is not seen at colposcopy. For the ‘inconclusive’ glandular category, the overall yield of HGEA was 42.5%, with 19.2% being high-grade glandular lesions; however, when only cases with adequate
follow-up are considered, these figures improve to 57.4 and 21.7%, respectively. For our laboratory, colposcopy should still be the next diagnostic procedure for cases in the ‘inconclusive’ glandular category, given the presence of a high proportion of squamous lesions found in these cases. However, if no lesion is seen on colposcopy, close followup is still necessary; either diagnostic LLETZ biopsy or a repeat smear with a liquid-based sample for HPV DNA assessment may be the management of choice for this group.26 We would consider the latter approach more valuable than a random punch biopsy. In addition, a liquidbased sample can be used to produce a cell block; we believe such preparations can be a useful adjunct for morphological assessment and immunohistochemical studies (Fig. 3).27 There are few published studies regarding the predictive
202
SEGAL et al.
Pathology (2003), 35, June
Fig. 3 (a–c) Adenocarcinoma in situ of cervix. Cell block material produced from liquid based samples. Palisaded and pseudostratified strips with hyperchromatic, enlarged nuclei and a high nuclear:cytoplasmic ratio. Adjacent normal endocervical epithelium (H&E, original magnification, 6400).
value of glandular abnormalities in cervical smears20–22 and, in some instances, comparison with other studies is difficult due to differences in terminology. For example, Mathers et al.22 accept a biopsy diagnosis of low-grade glandular dysplasia as positive follow-up. While our results do not achieve the impressive figures reported by Roberts et al.20 or Laverty et al.,21 these two studies come from a laboratory with a special interest in cervical glandular lesions and the authors state: ‘we recognise that our long-term interest in the prediction of glandular lesions has undoubtedly contributed to our predictive accuracy’.20 We hope that the review processes described in this study will improve our expertise in this difficult area of cervical cytology. This view is supported by the results of the second part of the study involving review of ‘inconclusive’ glandular cases. The review highlighted the importance of recognising endometrial stroma as a marker of LUS material (Fig. 1). The majority of cases that had initially been included in the ‘inconclusive’ glandular category, but which on review were reclassified as negative, contained stromal tissue indicating high canal sampling. Seven cases were reclassified as negative based largely on the presence of stroma, and all had negative follow-up (five with negative biopsy follow-up and two with negative smear follow-up). In some instances, stromal tissue had not been highlighted at screening and may have been overlooked by the reporting pathologist. Increased awareness of stromal tissue as an indicator of LUS material can help in interpreting glandular epithelium which may have otherwise worrying features (e.g., crowding, nuclear hyperchromasia, granular chromatin). Obviously, the presence of stroma cannot be used to downgrade epithelial cells that are clearly neoplastic, as high sampling may be present concurrently in a smear that also contains AIS or a squamous lesion. The other potential pitfall in this area is false-negative diagnosis in cases of ‘endometrioid’ AIS, where cells are misinterpreted as being of endometrial origin when they actually represent small-celled AIS; in this situation, stroma should not be a feature.3,4 Improvement may also be possible in the correct
classification of smears as squamous rather than glandular or possible glandular type. In 12 (23.1%) of the ‘inconclusive’ glandular smears, there were aggregates of cells that had been misinterpreted as showing ‘glandular’ features. These aggregates were syncytial in appearance, with rather pale fragile cytoplasm and sometimes with an ‘edge’ to the cell aggregate. (Fig. 2) A few of these cases may represent endocervical gland involvement by HSIL;28,29 however, on closer analysis of the cell groups, features of definite glandular differentiation (columnar cells, palisading, rosettes/acinar arrangements, gland openings) were invariably absent. These cases were reclassified as squamous (either HSIL or ‘inconclusive’ squamous categories). Of these reclassified cases, 66.7% had HSIL and 25% had LSIL on biopsy; none of these cases had a glandular abnormality on biopsy follow-up. Improvement in accuracy of cell typing is important, as there are management implications: most squamous lesions are colposcopically detectable, while an ‘inconclusive’ glandular report may eventually necessitate a LLETZ or cone biopsy to prove or exclude AIS. Other outcomes of the review process included reclassification of more than one-third of ‘inconclusive’ glandular cases as HGEA with a PPV of 73.7%, and the downgrading of two cases which subsequently proved to be AIS on biopsy. This latter finding highlights some of the inherent difficulties in such review processes, as previously reported in a study by Raab et al.30 One of the downgraded cases of proven AIS contained endocervical cells showing features interpreted on review as florid reactive changes, a recognised area of difficulty.6,7 The other case contained only minimal material, interpreted as showing minor non-specific changes in endocervical cells by the reviewing panel. In conclusion, there is almost certainly wide variation between laboratories in the predictive value of diagnoses of glandular abnormalities of the cervix. It is important for laboratories to assess and monitor their predictive values at regular intervals and for clinicians to be aware of these data in order to ensure acceptable patient management and outcome; this applies to both glandular and squamous
GLANDULAR ABNORMALITIES IN PAP SMEARS
diagnoses. Australian policies for management of screendetected glandular abnormalities of the cervix will require careful consideration of data from multiple laboratories. While this study has highlighted two main areas where improvement in cytological diagnosis of ‘inconclusive’ glandular lesions may be possible, HPV DNA typing (and/ or other ancillary studies) will also almost certainly play an increasing role in determining future management in this area.26,31 In our view, the ability to perform HPV studies, as well as other potential uses of a liquid-based sample, e.g., production of a cell block, render a liquidbased cytological sample of greater value than a random punch biopsy, and perhaps endocervical curettings, in the follow-up of ‘inconclusive’ glandular smears. Address for correspondence: Dr G. Sterrett, The Western Australian Centre for Pathology and Medical Research, Locked Bag 2009, Nedlands, WA 6909, Australia. E-mail: [email protected]
14.
15.
16.
17.
18.
19.
References 1. Vizcaino AP, Moreno V, Bosch FX, et al. International trends in incidence of cervical cancer: I. Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer 1998; 75: 536–45. 2. Vizcaino AP, Moreno V, Bosch FX, et al. International trends in incidence of cervical cancer: II. Squamous cell carcinoma. Int J Cancer 2000; 86: 429–35. 3. Lee KR. Adenocarcinoma in-situ with a small cell (endometrioid) pattern in cervical smears. Cancer Cytopathol 1999; 87: 254–8. 4. Lee KR, Genest DR, Minter LJ, Granter SR, Cibas ES. Adenocarcinoma in situ in cervical smears with a small cell (endometrioid) pattern: distinction from cells directly sampled from the upper endocervical canal or lower segment of the endometrium. Am J Clin Pathol 1999; 111: 567–8. 5. Hanau CA, Begley N, Bibbo M. Cervical endometriosis: a potential pitfall in the evaluation of glandular cells in cervical smears. Diagn Cytopathol 1997; 16: 127–80. 6. Lundeen SJ, Horwitz CA, Larson CJ, Stanley MW. Abnormal cervicovaginal smears due to endometriosis: a continuing problem. Diagn Cytopathol 2002; 26: 35–40. 7. Mulvany NJ, Surtees V. Cervical/vaginal endometriosis with atypia: a cytohistopathologic study. Diagn Cytopathol 1999; 21: 188–93. 8. de Peralta-Venturino MN, Purslow MJ, Kini SR. Endometrial cells of the ‘‘lower uterine segment’’ (LUS) in cervical smears obtained by endocervical brushings: a source of potential diagnostic pitfall. Diagn Cytopathol 1995; 12: 263–8. 9. Krane JF, Granter SR, Trask CE, Hogan CL, Lee KR. Papanicolaou smear sensitivity for the detection of adenocarcinoma of the cervix: a study of 49 cases. Cancer 2001; 93: 8–15. 10. Schoolland M, Segal A, Allpress S, Miranda A, Frost F, Sterrett G. Adenocarcinoma in situ of the cervix. Cancer 2002; 96: 330–7. 11. Raab SS, Snider TE, Potts SA, et al. Atypical glandular cells of undetermined significance. Diagnostic accuracy and interobserver variability using select cytologic criteria. Am J Clin Pathol 1997; 107: 299–307. 12. Burja IT, Thompson SK, Sawyer WL Jr, et al. Atypical glandular cells of undetermined significance on cervical smears. A study with cytohistologic correlation. Acta Cytol 1999; 43: 351–6. 13. DiTomasso JP, Ramzy I, Mody DR. Glandular lesions of the cervix.
20.
21.
22. 23.
24.
25. 26.
27.
28.
29.
30.
31.
203
Validity of cytologic criteria used to differentiate reactive changes, glandular intraepithelial lesions and adenocarcinoma. Acta Cytol 1996; 40: 1127–35. Kim TJ, Kim HS, Park CT, et al. Clinical evaluation of follow-up methods and results of atypical glandular cells of undetermined significance (AGUS) detected on cervicovaginal Pap smears. Gynecol Oncol 1999; 732: 292–8. Manetta A, Keefe K, Lin F, et al. Atypical glandular cells of undetermined significance in cervical cytologic findings. Am J Obstet Gynecol 1999; 180: 883–8. Korn AP, Judson PL, Zaloudek CJ. Importance of atypical glandular cells of uncertain significance in cervical cytologic smears. J Reprod Med 1998; 43: 774–8. Chhieng DC, Elgert PA, Cangiarella JF, Cohen JM. Clinical significance of atypical glandular cells of undetermined significance. A follow-up study from an academic medical center. Acta Cytol 2000; 44: 557–66. Hecht JL, Sheets EE, Lee KR. Atypical glandular cells of undetermined significance in conventional cervical/vaginal smears and thinlayer preparations. Cancer 2002; 96: 1–4. Geier CS, Wilson M, Creasman W. Clinical evaluation of atypical glandular cells of undetermined significance. Am J Obstet Gynecol 2001; 184: 64–9. Roberts JM, Thurloe JK, Bowditch RC, Laverty CR. Subdividing atypical glandular cells of undetermined significance according to the Australian modified Bethesda, system: analysis of outcomes. Cancer 2000; 90: 87–95. Laverty CR, Farnsworth A, Thurloe J, Bowditch R. The reliability of a cytological prediction of cervical adenocarcinoma in situ. Aust NZ Obstet Gynaecol 1988; 4: 307–12. Mathers ME, Johnson J, Wadehra V. How predictive is a cervical smear suggesting glandular neoplasia? Cytopathology 2002; 13: 83–91. Lee KR, Sun D, Crum CP. Endocervical intraepithelial glandular atypia (dysplasia): a histopathologic, human papillomavirus, and MIB-1 analysis of 25 cases. Hum Pathol 2000; 31: 656–64. Zaino RJ. Symposium Part I: adenocarcinoma in situ, glandular dysplasia, and early invasive adenocarcinoma of the uterine cervix. Int J Gynecol Pathol 2002; 21: 314–26. Zaino RJ. Glandular lesions of the uterine cervix. Mod Pathol 2000; 13: 261–74. Ronnett BM, Manos MM, Ransley JE, et al. Atypical glandular cells of undetermined significance (AGUS): Cytopathologic features, histopathologic results, and human papillomavirus DNA detection. Hum Pathol 1999; 30: 816–25. Keyhani-Rofagha S, Vesey-Shecket M. Diagnostic value, feasibility, and validity of preparing cell blocks from fluid-based gynaecologic cytology specimens. Cancer Cytopathol 2002; 96: 204–9. Selvaggi SM. Cytologic features of squamous cell carcinoma in situ involving endocervical glands in endocervical cytobrush specimens. Acta Cytol 1994; 38: 687–92. Levine PH, Waisman J, Mittal K. Significance of the cytologic diagnosis of endocervical glandular involvement in high grade squamous intraepithelial lesions. Diagn Cytopathol 2002; 26: 217–21. Raab SS, Geisinger KR, Silverman JF, Thomas PA, Stanley MW. Interobserver variability of a Papanicolaou smear diagnosis of atypical glandular cells of undetermined significance. Am J Clin Pathol 1998; 110: 653–9. Riethdorf L, Riethdorf S, Lee KR, Cviko A, Loning T, Crum CP. Human papillomaviruses, expression of p16, and early endocervical glandular neoplasia. Hum Pathol 2002; 33: 899–904.
ANATOMICAL PATHOLOGY
Predictive value of diagnoses of endocervical glandular abnormalities in cervical smears AMANDA SEGAL*, FELICITY A. FROST*, ALINA MIRANDA*, CHRISTINE FLETCHER{ GREGORY F. STERRETT*
AND
*Division of Tissue Pathology, The Western Australian Centre for Pathology and Medical Research, Nedlands; {Western Australian Cervical Cytology Registry, Perth, Western Australia
Summary Aims: To determine the positive predictive value (PPV) of cervical smear diagnoses of ‘definite’ and ‘possible’ endocervical adenocarcinoma in situ or invasive adenocarcinoma, and whether diagnostic accuracy can be improved. Methods: The study examined cervical smears reported as definite or possible high-grade glandular abnormality between 1992 and 1998. PPV was calculated by comparing smear diagnoses with the subsequent histopathology report. All available smears were reviewed without knowledge of follow-up results, and were reclassified by consensus. Results: Thirty-two smears were diagnosed as high-grade glandular lesions, with adequate biopsy follow-up in 31 cases (96.9%). A high-grade epithelial abnormality (HGEA) was detected in 29 cases (PPV, 93.5%), with a high-grade glandular lesion in 24 (PPV, 77.4%). Very few smears were reclassified on review. Seventy-three smears were initially diagnosed in the ‘inconclusive’ glandular or indeterminate cell-type category. There was adequate biopsy follow up for 54 cases (74.0%). On follow-up, 31 cases had a HGEA (PPV, 57.4%), with 14 cases having a high-grade glandular abnormality (PPV, 25.9%). In the review of ‘inconclusive’ smears, 12 were reclassified as squamous abnormalities and none of these had a glandular lesion on biopsy. Eight were reclassified as negative; seven contained endometrial stroma and the glandular cells in question were considered to be of lower uterine segment (LUS) origin. No significant lesion was present on follow-up of these cases. Conclusions: For clinicians using our laboratory, large loop excision of the transformation zone (LLETZ) or cone biopsy should follow a ‘definite’ cytological diagnosis of a highgrade endocervical glandular lesion. However, cone biopsy may not be the appropriate initial management in the ‘possible’ high-grade glandular group because of a significantly lower predictive value of the diagnosis. The slide review highlighted the importance of (1) caution in classifying sheets of abnormal cells as glandular, and (2) endometrial stroma as a marker of LUS material. Key words: Cervical smear, adenocarcinoma in situ, predictive value, inconclusive glandular. Abbreviations: AIS, adenocarcinoma in situ; HSIL, high-grade squamous intra-epithelial lesion; LSIL, low-grade squamous intra-epithelial lesion. Received 23 December 2002, revised 12 February, accepted 17 February 2003
INTRODUCTION In contrast to screening for squamous lesions of the cervix, cervical smears have not been shown to reduce the incidence of invasive adenocarcinoma.1,2 The reasons for this are not clear. It may partly reflect a sampling problem; however, the criteria and pitfalls in the diagnosis of precursors of adenocarcinoma have only been well defined in the last decade or so, and we may be yet to see results in this area. Difficulty in interpretation of cytological findings may be responsible for both over-diagnosis and underdiagnosis of glandular lesions. Under-calling may result from lack of awareness of certain patterns of adenocarcinoma in situ (AIS), for example, those resembling reactive endocervical cells and those with crowded small cells mimicking endometrial or high canal sampling.3,4 Overcalling of benign processes such as repair, tubo-endometrial metaplasia, cervical endometriosis and lower uterine segment sampling are also well-recognised problems.5–8 Several recent studies have examined the diagnosis of glandular lesions in cervical smears, from the point of view of sensitivity of cytological detection, accuracy of diagnosis and the most appropriate means of establishing a biopsy diagnosis.9–19 A cytological prediction of a possible or definite glandular lesion raises various management issues. As significant glandular lesions may be present in the setting of a normal colposcopic examination, either cone biopsy or large loop excision of the transition zone (LLETZ) biopsy is usually necessary for diagnosis. Knowledge of outcomes of cytological predictions of glandular lesions helps to guide the decision for a biopsy. The positive predictive value (PPV) of a diagnosis of AIS in screening laboratories can be very high, particularly in laboratories with an interest in this area. In a recent study, Roberts et al.,20 demonstrated a PPV of 95.7% for a definite diagnosis of AIS and 71.1% for an ‘inconclusive’ glandular report, enabling this laboratory to suggest that cone biopsy would be an appropriate next diagnostic step for patients in either of these cytological categories. However, there are few published papers addressing predictive value in this area of cervical cytology,21,22 and there is almost certainly wide variation between laboratories. It is important for laboratories to assess and monitor their predictive values and for clinicians to be aware of these data in order to determine appropriate management. With this in mind, the aims of this study were two-fold:
ISSN 0031–3025 printed/ISSN 1465–3931 online/03/030198–06 # 2003 Royal College of Pathologists of Australasia DOI: 10.1080/0031302031000123146
GLANDULAR ABNORMALITIES IN PAP SMEARS
1. To determine the PPV of cervical smears reported as ‘definite’ high-grade endocervical glandular abnormality or ‘possible’ high-grade glandular abnormality (‘inconclusive’) with or without associated squamous abnormality at PathCentre, a public sector screening laboratory which reports approximately 25 000 smears per year. 2. To determine whether any improvement in diagnostic accuracy can be achieved on review of smears originally diagnosed in these categories, particularly the ‘inconclusive’ glandular category.
MATERIALS AND METHODS Protocols for reporting glandular abnormalities in smears and biopsies All cervical smears were reported according to Australian National Health and Medical Research Council (NHMRC) categories and recommendations (Table 1). For biopsy reporting we did not use the diagnosis of glandular dysplasia. Although we acknowledge this may be controversial, we do not ourselves recognise a spectrum of glandular cell abnormalities of lesser degree than AIS.23–25 In interpreting histological reports from other laboratories, therefore, we have only accepted a definite diagnosis of AIS as confirmatory. Where a diagnosis of dysplasia was made (two cases), we obtained sections from these biopsies for review and used our review diagnosis in the results.
Case selection and histological follow-up All cervical smears reported in the high-grade glandular abnormality or ‘inconclusive’ glandular categories over a 7-year period between 1992 and 1998 were retrieved and reviewed by the authors. The study included smears reported as containing combined glandular and squamous abnormalities, and those where cell type was indeterminate, i.e., either squamous or glandular. Follow-up biopsy data from all pathology laboratories in Western Australia were obtained through the Western Australian Cervical Cytology Registry, and were based on codes and histology reports; histological sections of biopsies were not routinely reviewed. The follow-up included any available histology at the time of the review. The followup period ranged from 3 to 9 years, although relevant histological follow-up was usually within 6 months to 2 years of the index smear. PPV was assessed by comparing an initial smear diagnosis of a high-grade glandular lesion or an ‘inconclusive’ glandular report and the subsequent histopathology results. Cases without biopsy follow-up were reviewed, but these cases were not included in the calculation of PPV.
199
Slide review Detailed review of all available smears was undertaken to establish whether we agreed with the original diagnosis and cell type assigned, and whether we could identify any systematic errors in reporting. For the ‘inconclusive’ glandular smears, an attempt was made to refine the diagnosis into negative, low-grade or high-grade categories, and into squamous or glandular abnormality. The review was undertaken without knowledge of biopsy outcomes.
RESULTS Predictive value and review of smears diagnosed as high-grade endocervical glandular abnormality The results of the predictive value component of the study are summarised in Table 2. There were 32 smears diagnosed as either invasive adenocarcinoma, AIS alone or AIS plus high-grade squamous intra-epithelial lesion (HSIL) during the 7-year study period, with adequate biopsy follow-up in 30 cases (93.75%). One patient refused biopsy; however, this patient had a diagnosis of AIS extending to margins on cone biopsy 6 years previously and a repeat smear in this patient was also diagnosed as AIS; this case was considered to have positive follow-up for the purposes of calculation of PPV. A high-grade epithelial abnormality (HGEA) was detected on biopsy in 28 of the cases (PPV, 93.5%). A high-grade glandular abnormality was found in 24 cases (PPV, 77.4%); two were endometrial lesions and the remainder endocervical lesions. Twenty-seven slides were available for review. In 19 cases there was agreement with the original diagnosis and all 19 were confirmed to have high-grade glandular lesions on biopsy. Of four cases downgraded to ‘inconclusive’ possible high-grade glandular, one proved to be invasive adenocarcinoma, one AIS, one high-grade squamous abnormality and one negative on biopsy follow-up. Two cases on review were considered to be high-grade squamous lesions and this was confirmed on biopsy, with no glandular abnormality identified. One case was downgraded to ‘inconclusive’ of indeterminate cell type and had negative biopsy follow-up. One case was downgraded to ‘inconclusive’ squamous and had a high-grade squamous lesion with cervical endometriosis on follow-up histology. In total, there were three cases in which a high-grade lesion was not found on biopsy. In one of these cases the smear
TABLE 1 Summary of Australian National Health and Medical Research Council (NHMRC) terminology for reporting gynaecological cytology, and recommended management. Category High-grade epithelial abnormality (HGEA) Invasive carcinoma, squamous, glandular Cervical intraepithelial neoplasia (CIN) Grades 2,3 (HSIL) Adenocarcinoma in situ (AIS) Inconclusive: Possible HGEA Squamous, glandular Low-grade epithelial abnormality (LGEA) CIN, Grade 1 HPV effect NSMC squamous, endocervical glandular Negative for dysplasia and malignancy Unsatisfactory for evaluation Abbreviations: HPV, human papilloma virus; NSMC, non-specific minor changes.
Recommended management
Gynaecological referral Colposcopy Gynaecological referral Colposcopy Colposcopy Repeat smear 6–12 months Repeat smear 6–12 months Repeat smear in 2 years Repeat smear in 3 months
200
Pathology (2003), 35, June
SEGAL et al.
TABLE 2
Histological follow-up in cases diagnosed as ‘definite’ high-grade endocervical glandular abnormality. Histological diagnosis Invasive endocervical adenocarcinoma
Cytological diagnosis Invasive adenocarcinoma AIS AISzHSIL Total
5 1 6
AIS
AISzHSIL
HSIL
Endometrial carcinoma/ atypical hyperplasia
2 5 4 11
1 1 3 5
1
2
3 4
Negative
Unsatisfactory biopsy
No biopsy
2
1
1
2
1
1
2
Total 11 11 10 32
Abbreviations: AIS, adenocarcinoma in situ; HSIL, high-grade squamous intra-epithelial lesion.
glandular/squamous, two endometrial and one endometrial/ squamous). Of these 21 cases, ten had high-grade lesions on biopsy (9 HSIL and 1 HSILzAIS). Twelve cases were reclassified as either low-grade epithelial abnormality or negative, and two of these cases proved to be high-grade glandular lesions on biopsy follow-up. In total, eight ‘inconclusive’ cases were reclassified as negative and five had biopsy follow-up. In seven of these cases, stroma was present on the smear, indicating lower uterine sampling (LUS); no significant lesion was identified on follow-up (smear and/or biopsy) in these seven cases (Fig. 1). Twelve cases (23.1%) were reclassified as squamous (either HSIL or ‘inconclusive’ squamous categories). Of these reclassified cases, 66.7% had HSIL, 25% had LSIL and one case (8.3%) had negative biopsy follow-up; none of these cases had a glandular abnormality on biopsy follow-up (Fig. 2).
was considered to be ‘inconclusive’ rather than definite high-grade on review. A second case which had also been considered ‘inconclusive’ on review had human papillomavirus (HPV) changes on biopsy; however, only punch biopsies had been performed and follow-up was therefore considered unsatisfactory. A third case had negative LLETZ biopsy follow-up; however, the smear was not available for review. Predictive value and review of smears diagnosed as ‘inconclusive’: possible high grade endocervical glandular abnormality The results of the predictive value component of the study are summarised in Table 3. There were 73 smears diagnosed in the ‘inconclusive’ glandular category. Subgroups of the ‘inconclusive’ category are shown in the table. Overall, 31 cases (42.5%) demonstrated a HGEA on biopsy; however, adequate biopsy follow-up was only available for 54 cases (74.0%). Using only cases with adequate followup, the PPV for smears in the ‘inconclusive’ glandular category was 57.4%. Only 14 cases (25.9%) demonstrated a high-grade glandular abnormality, the remainder being high-grade squamous lesions. Nevertheless, in the ‘inconclusive’ cases classified as endocervical cell type on initial reporting, 13 of the 22 cases (59.1%) with an adequate biopsy had a high-grade lesion on biopsy and 10 of the 13 cases (76.9%) with high-grade lesions on biopsy proved to have a glandular component. Of the 54 ‘inconclusive’ glandular cases with adequate biopsy follow-up, there were 52 smears available for review. Nineteen smears (36.5%) were reclassified as HGEA (eight AIS, eight HSIL and three AISzHSIL) and, of these, 14 (73.7%) had a biopsy-proven high-grade lesion (five AIS, seven HSIL and two AISzHSIL). Twenty-one smears were considered to remain in the ‘inconclusive’ category following review (four squamous, four glandular, ten
DISCUSSION There were various limitations to the study including a proportion of cases where the follow-up was in the form of punch biopsy or endometrial curettings rather than cone or LLETZ biopsy; for the purposes of calculating predictive values, these cases were considered to have ‘unsatisfactory’ follow-up. There were also cases in which hysterectomy had been performed with no lesion detected. These cases may not have had an adequate examination of the cervix if the preceding smear diagnosis was not known; however, these cases were considered to have adequate (negative) follow-up for the purposes of the study. Only biopsies reported as showing endocervical ‘dysplasia’ (two cases) were reviewed; one was confirmed as AIS and the other was considered to represent minor reactive changes in endocervical cells. Otherwise, histological diagnoses were accepted on the basis of reports only.
TABLE 3 Histological follow-up in cases diagnosed as ‘possible’ high-grade epithelial lesion of endocervical glandular type (‘inconclusive’ glandular category). Histological diagnosis Invasive adenocarcinoma/ Ovarian Unsatisfactory No adenosquamous carcinoma AIS HSILzAIS HSIL carcinoma LSIL Neg biopsy biopsy Total
Cytological diagnosis Inconclusive Inconclusive Inconclusive Inconclusive Total
endocervical endocervical/squamous endocervical/endometrial endocervical/endometrial/squamous
2 1
3 1 1
5
3
5
5
3 11 2 1 17
1 1
5 8
4 4 2
2 2 2
6 6 1
13
10
6
13
Abbreviations: AIS, adenocarcinoma in situ; HSIL, high-grade squamous intra-epithelial lesion; LSIL, low-grade squamous intra-epithelial lesion.
29 34 8 2 73
GLANDULAR ABNORMALITIES IN PAP SMEARS
201
Fig. 1 Various appearances of stromal tissue in cytological smears indicating lower uterine segment or upper endocervical canal sampling. In our own experience, this material has often been overlooked in routine screening. (a,b) The material can easily be identified as endometrial. (c) Distorted stromal tissue; it is difficult to be certain whether such material is of endocervical or lower uterine segment origin (Papanicolaou stain, original magnification, a,b: 6250; c: 6400).
Fig. 2 Mimics of glandular differentiation. Cell aggregates from cases originally diagnosed as having a possible high-grade glandular component, but on review considered to be squamous. Histological follow-up: high-grade squamous lesion only. (a,b) Smooth edges to aggregates, but no well-defined layer of apical cytoplasm or columnar shapes. Fragile cytoplasm imparting an endocervical appearance, but multi-layered and disorganised sheet indicating a squamous origin. (c) Small rounded aggregate, but not a tight three-dimensional group. (Papanicolaou stain, original magnification, a: 6250; b,c: 6400).
The predictive value for a definite smear diagnosis of a glandular lesion was considered satisfactory. For all cases, prediction of a high-grade glandular abnormality on smear was followed by a biopsy proven high-grade epithelial abnormality (squamous or glandular) in 87.5% of cases, and in 75% of cases a high-grade glandular lesion was present. If only cases with adequate biopsy follow-up are included, these figures improve to 93.5 and 77.4%, respectively. These results suggest that LLETZ or cone biopsy should follow a definite cytological diagnosis of a high-grade glandular lesion, even if a lesion is not seen at colposcopy. For the ‘inconclusive’ glandular category, the overall yield of HGEA was 42.5%, with 19.2% being high-grade glandular lesions; however, when only cases with adequate
follow-up are considered, these figures improve to 57.4 and 21.7%, respectively. For our laboratory, colposcopy should still be the next diagnostic procedure for cases in the ‘inconclusive’ glandular category, given the presence of a high proportion of squamous lesions found in these cases. However, if no lesion is seen on colposcopy, close followup is still necessary; either diagnostic LLETZ biopsy or a repeat smear with a liquid-based sample for HPV DNA assessment may be the management of choice for this group.26 We would consider the latter approach more valuable than a random punch biopsy. In addition, a liquidbased sample can be used to produce a cell block; we believe such preparations can be a useful adjunct for morphological assessment and immunohistochemical studies (Fig. 3).27 There are few published studies regarding the predictive
202
SEGAL et al.
Pathology (2003), 35, June
Fig. 3 (a–c) Adenocarcinoma in situ of cervix. Cell block material produced from liquid based samples. Palisaded and pseudostratified strips with hyperchromatic, enlarged nuclei and a high nuclear:cytoplasmic ratio. Adjacent normal endocervical epithelium (H&E, original magnification, 6400).
value of glandular abnormalities in cervical smears20–22 and, in some instances, comparison with other studies is difficult due to differences in terminology. For example, Mathers et al.22 accept a biopsy diagnosis of low-grade glandular dysplasia as positive follow-up. While our results do not achieve the impressive figures reported by Roberts et al.20 or Laverty et al.,21 these two studies come from a laboratory with a special interest in cervical glandular lesions and the authors state: ‘we recognise that our long-term interest in the prediction of glandular lesions has undoubtedly contributed to our predictive accuracy’.20 We hope that the review processes described in this study will improve our expertise in this difficult area of cervical cytology. This view is supported by the results of the second part of the study involving review of ‘inconclusive’ glandular cases. The review highlighted the importance of recognising endometrial stroma as a marker of LUS material (Fig. 1). The majority of cases that had initially been included in the ‘inconclusive’ glandular category, but which on review were reclassified as negative, contained stromal tissue indicating high canal sampling. Seven cases were reclassified as negative based largely on the presence of stroma, and all had negative follow-up (five with negative biopsy follow-up and two with negative smear follow-up). In some instances, stromal tissue had not been highlighted at screening and may have been overlooked by the reporting pathologist. Increased awareness of stromal tissue as an indicator of LUS material can help in interpreting glandular epithelium which may have otherwise worrying features (e.g., crowding, nuclear hyperchromasia, granular chromatin). Obviously, the presence of stroma cannot be used to downgrade epithelial cells that are clearly neoplastic, as high sampling may be present concurrently in a smear that also contains AIS or a squamous lesion. The other potential pitfall in this area is false-negative diagnosis in cases of ‘endometrioid’ AIS, where cells are misinterpreted as being of endometrial origin when they actually represent small-celled AIS; in this situation, stroma should not be a feature.3,4 Improvement may also be possible in the correct
classification of smears as squamous rather than glandular or possible glandular type. In 12 (23.1%) of the ‘inconclusive’ glandular smears, there were aggregates of cells that had been misinterpreted as showing ‘glandular’ features. These aggregates were syncytial in appearance, with rather pale fragile cytoplasm and sometimes with an ‘edge’ to the cell aggregate. (Fig. 2) A few of these cases may represent endocervical gland involvement by HSIL;28,29 however, on closer analysis of the cell groups, features of definite glandular differentiation (columnar cells, palisading, rosettes/acinar arrangements, gland openings) were invariably absent. These cases were reclassified as squamous (either HSIL or ‘inconclusive’ squamous categories). Of these reclassified cases, 66.7% had HSIL and 25% had LSIL on biopsy; none of these cases had a glandular abnormality on biopsy follow-up. Improvement in accuracy of cell typing is important, as there are management implications: most squamous lesions are colposcopically detectable, while an ‘inconclusive’ glandular report may eventually necessitate a LLETZ or cone biopsy to prove or exclude AIS. Other outcomes of the review process included reclassification of more than one-third of ‘inconclusive’ glandular cases as HGEA with a PPV of 73.7%, and the downgrading of two cases which subsequently proved to be AIS on biopsy. This latter finding highlights some of the inherent difficulties in such review processes, as previously reported in a study by Raab et al.30 One of the downgraded cases of proven AIS contained endocervical cells showing features interpreted on review as florid reactive changes, a recognised area of difficulty.6,7 The other case contained only minimal material, interpreted as showing minor non-specific changes in endocervical cells by the reviewing panel. In conclusion, there is almost certainly wide variation between laboratories in the predictive value of diagnoses of glandular abnormalities of the cervix. It is important for laboratories to assess and monitor their predictive values at regular intervals and for clinicians to be aware of these data in order to ensure acceptable patient management and outcome; this applies to both glandular and squamous
GLANDULAR ABNORMALITIES IN PAP SMEARS
diagnoses. Australian policies for management of screendetected glandular abnormalities of the cervix will require careful consideration of data from multiple laboratories. While this study has highlighted two main areas where improvement in cytological diagnosis of ‘inconclusive’ glandular lesions may be possible, HPV DNA typing (and/ or other ancillary studies) will also almost certainly play an increasing role in determining future management in this area.26,31 In our view, the ability to perform HPV studies, as well as other potential uses of a liquid-based sample, e.g., production of a cell block, render a liquidbased cytological sample of greater value than a random punch biopsy, and perhaps endocervical curettings, in the follow-up of ‘inconclusive’ glandular smears. Address for correspondence: Dr G. Sterrett, The Western Australian Centre for Pathology and Medical Research, Locked Bag 2009, Nedlands, WA 6909, Australia. E-mail: [email protected]
14.
15.
16.
17.
18.
19.
References 1. Vizcaino AP, Moreno V, Bosch FX, et al. International trends in incidence of cervical cancer: I. Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer 1998; 75: 536–45. 2. Vizcaino AP, Moreno V, Bosch FX, et al. International trends in incidence of cervical cancer: II. Squamous cell carcinoma. Int J Cancer 2000; 86: 429–35. 3. Lee KR. Adenocarcinoma in-situ with a small cell (endometrioid) pattern in cervical smears. Cancer Cytopathol 1999; 87: 254–8. 4. Lee KR, Genest DR, Minter LJ, Granter SR, Cibas ES. Adenocarcinoma in situ in cervical smears with a small cell (endometrioid) pattern: distinction from cells directly sampled from the upper endocervical canal or lower segment of the endometrium. Am J Clin Pathol 1999; 111: 567–8. 5. Hanau CA, Begley N, Bibbo M. Cervical endometriosis: a potential pitfall in the evaluation of glandular cells in cervical smears. Diagn Cytopathol 1997; 16: 127–80. 6. Lundeen SJ, Horwitz CA, Larson CJ, Stanley MW. Abnormal cervicovaginal smears due to endometriosis: a continuing problem. Diagn Cytopathol 2002; 26: 35–40. 7. Mulvany NJ, Surtees V. Cervical/vaginal endometriosis with atypia: a cytohistopathologic study. Diagn Cytopathol 1999; 21: 188–93. 8. de Peralta-Venturino MN, Purslow MJ, Kini SR. Endometrial cells of the ‘‘lower uterine segment’’ (LUS) in cervical smears obtained by endocervical brushings: a source of potential diagnostic pitfall. Diagn Cytopathol 1995; 12: 263–8. 9. Krane JF, Granter SR, Trask CE, Hogan CL, Lee KR. Papanicolaou smear sensitivity for the detection of adenocarcinoma of the cervix: a study of 49 cases. Cancer 2001; 93: 8–15. 10. Schoolland M, Segal A, Allpress S, Miranda A, Frost F, Sterrett G. Adenocarcinoma in situ of the cervix. Cancer 2002; 96: 330–7. 11. Raab SS, Snider TE, Potts SA, et al. Atypical glandular cells of undetermined significance. Diagnostic accuracy and interobserver variability using select cytologic criteria. Am J Clin Pathol 1997; 107: 299–307. 12. Burja IT, Thompson SK, Sawyer WL Jr, et al. Atypical glandular cells of undetermined significance on cervical smears. A study with cytohistologic correlation. Acta Cytol 1999; 43: 351–6. 13. DiTomasso JP, Ramzy I, Mody DR. Glandular lesions of the cervix.
20.
21.
22. 23.
24.
25. 26.
27.
28.
29.
30.
31.
203
Validity of cytologic criteria used to differentiate reactive changes, glandular intraepithelial lesions and adenocarcinoma. Acta Cytol 1996; 40: 1127–35. Kim TJ, Kim HS, Park CT, et al. Clinical evaluation of follow-up methods and results of atypical glandular cells of undetermined significance (AGUS) detected on cervicovaginal Pap smears. Gynecol Oncol 1999; 732: 292–8. Manetta A, Keefe K, Lin F, et al. Atypical glandular cells of undetermined significance in cervical cytologic findings. Am J Obstet Gynecol 1999; 180: 883–8. Korn AP, Judson PL, Zaloudek CJ. Importance of atypical glandular cells of uncertain significance in cervical cytologic smears. J Reprod Med 1998; 43: 774–8. Chhieng DC, Elgert PA, Cangiarella JF, Cohen JM. Clinical significance of atypical glandular cells of undetermined significance. A follow-up study from an academic medical center. Acta Cytol 2000; 44: 557–66. Hecht JL, Sheets EE, Lee KR. Atypical glandular cells of undetermined significance in conventional cervical/vaginal smears and thinlayer preparations. Cancer 2002; 96: 1–4. Geier CS, Wilson M, Creasman W. Clinical evaluation of atypical glandular cells of undetermined significance. Am J Obstet Gynecol 2001; 184: 64–9. Roberts JM, Thurloe JK, Bowditch RC, Laverty CR. Subdividing atypical glandular cells of undetermined significance according to the Australian modified Bethesda, system: analysis of outcomes. Cancer 2000; 90: 87–95. Laverty CR, Farnsworth A, Thurloe J, Bowditch R. The reliability of a cytological prediction of cervical adenocarcinoma in situ. Aust NZ Obstet Gynaecol 1988; 4: 307–12. Mathers ME, Johnson J, Wadehra V. How predictive is a cervical smear suggesting glandular neoplasia? Cytopathology 2002; 13: 83–91. Lee KR, Sun D, Crum CP. Endocervical intraepithelial glandular atypia (dysplasia): a histopathologic, human papillomavirus, and MIB-1 analysis of 25 cases. Hum Pathol 2000; 31: 656–64. Zaino RJ. Symposium Part I: adenocarcinoma in situ, glandular dysplasia, and early invasive adenocarcinoma of the uterine cervix. Int J Gynecol Pathol 2002; 21: 314–26. Zaino RJ. Glandular lesions of the uterine cervix. Mod Pathol 2000; 13: 261–74. Ronnett BM, Manos MM, Ransley JE, et al. Atypical glandular cells of undetermined significance (AGUS): Cytopathologic features, histopathologic results, and human papillomavirus DNA detection. Hum Pathol 1999; 30: 816–25. Keyhani-Rofagha S, Vesey-Shecket M. Diagnostic value, feasibility, and validity of preparing cell blocks from fluid-based gynaecologic cytology specimens. Cancer Cytopathol 2002; 96: 204–9. Selvaggi SM. Cytologic features of squamous cell carcinoma in situ involving endocervical glands in endocervical cytobrush specimens. Acta Cytol 1994; 38: 687–92. Levine PH, Waisman J, Mittal K. Significance of the cytologic diagnosis of endocervical glandular involvement in high grade squamous intraepithelial lesions. Diagn Cytopathol 2002; 26: 217–21. Raab SS, Geisinger KR, Silverman JF, Thomas PA, Stanley MW. Interobserver variability of a Papanicolaou smear diagnosis of atypical glandular cells of undetermined significance. Am J Clin Pathol 1998; 110: 653–9. Riethdorf L, Riethdorf S, Lee KR, Cviko A, Loning T, Crum CP. Human papillomaviruses, expression of p16, and early endocervical glandular neoplasia. Hum Pathol 2002; 33: 899–904.