| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Departments of Gynecologic Oncology, and Obstetrics and Gynecology, University of California at Irvine, Irvine, California; Clinical Pathology and Obstetrics and Gynecology, Keck School of Medicine, University of Southern California, Los Angeles, California; Department of Obstetrics and Gynecology, Southern California Permanente Medical Group, Anaheim, California; Department of Pathology, University of Utah School of Medicine, Salt Lake City, Utah; Department of Women’s Health, Arrowhead Regional Medical Center, Colton, California; Institute for Clinical and Experimental Pathology, Associated Regional and University Pathologists Inc., Salt Lake City, Utah; Department of Medicine, University of California at Los Angeles, Los Angeles, California; and The Trylon Corporation, Torrance, California.
Address reprint requests to: Stewart A. Lonky, MD, The Trylon Corporation, 970 West 190th Street, Suite 850, Torrance, CA 90502; E-mail: slonky{at}tryloncorp.com.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: Before large loop excision of the transformation zone, women with cervical intraepithelial neoplasia underwent a transepithelial brush biopsy of a portion of a colposcopically identified lesion, followed by a punch biopsy of the remaining portion. Brush biopsy samples were processed using liquid-based cytology and cell block techniques. Diagnoses were made using a consensus of three pathologists. Brush biopsy samples without basal cells were considered inadequate. The histological diagnosis was compared with the brush biopsy and punch biopsy samples. Patient-reported pain and physician-reported bleeding for punch and brush biopsies were compared.
RESULTS: Fifty-two women were enrolled in the study; 47 successfully completed the study protocol. Eight brush biopsy specimens were inadequate. Thirty-nine women showed abnormal pathology (human papillomavirus/cervical intraepithelial neoplasia I or worse) on large loop excision of the transformation zone, and 32 women had high-grade (or worse) lesions. The punch biopsy correlated with high-grade disease in 53.1% of these women. The brush biopsy result correlated with high-grade disease in 79.3% of these women using a cell block technique and 76.7% using liquid cytology. There was significantly less pain (P < .001) and significantly less bleeding (P < .001) with the brush biopsy.
CONCLUSION: When an adequate sample is collected, spiral brush biopsy is as good as a standard punch biopsy for detecting cervical pathology, with substantially less pain and bleeding. User training and guidelines for sampling are needed to assure that an adequate sample is collected.
The traditional evaluation and management of an abnormal Papanicolaou smear or a visually abnormal cervix includes colposcopy and a directed punch biopsy of any suspicious lesion.1,2 Although trainees in colposcopy are instructed to biopsy lesions at the area of worst appearance, multiple studies have demonstrated less than perfect correlation of punch biopsy specimens with subsequent large loop excision of the transformation zone.3–5 Whether this poor correlation is due to operator error ("geographic miss") or whether it is secondary to equipment shortfalls, such as the biopsy specimen being limited to the size of the bite of the biopsy jaws, it can result in incorrect diagnoses, with under-treatment as a result.
Alternatives to a colposcopy-directed punch biopsy are few. Multiple biopsies, using either a four-quadrant approach or multiple biopsies of abnormally appearing areas, yield correlations to subsequent large loop excision of the transformation zone procedures that are no better than those of directed punch biopsies.6,7 Although some investigators have suggested that the use of primary large loop excision of the transformation zone is both diagnostic and therapeutic,8–10 the complication rate of bleeding and the finding of significant numbers of specimens with no pathology have caused others to caution against the use of primary large loop excision of the transformation zone in a see and treat approach.7,11 As a result, most clinicians require a degree of certainty regarding cervical pathology before treatment with loop excision.
Recently, interest has been kindled in using a less traumatic approach to obtaining diagnostic samples from squamous epithelium. A brush sampler for endocervical canal samples12 and the use of a cytobrush bent at an angle for sampling abnormally appearing cervical epithelium in pregnant women13 have been shown to be both safe and effective in providing diagnostic material. The introduction of OralCDx (OralScan Laboratories Inc., Suffern, NY) for sampling the squamous epithelium of the oral cavity has shown that it is possible to actually obtain a transepithelial tissue sample with a relatively nontraumatic approach.14 This device, using a spiral-shaped brush with bristles much stiffer than those of a cytobrush or endocervical brush, provides transepithelial samples that are rich in epithelial cells, including basal and parabasal cells, and can be studied using cytologic techniques. Because the cervix is covered with squamous epithelium similar to that of the oral cavity, we undertook the present investigation to evaluate the efficacy of this spiral-shaped brush device for cervical transepithelial sampling.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
).
|
At the beginning of the study, each subject was shown an example of the SpiraBrush device being investigated and provided a copy of the discomfort scale that they would be asked to use to report sampling-associated discomfort. The discomfort scale was available in English, Spanish, and Vietnamese. A urine pregnancy test was done and results documented on the study data form.
The clinical investigator performed the scheduled routine colposcopic evaluation of the cervix using the standard procedure for the study site. The investigator identified and documented on the study data form the most suspicious exocervical lesion. The SpiraBrush was placed on the lesion, covering at least 25% of the visualized abnormal area but no more than 75% of the area. This approximation was accomplished using unaided vision or low-power magnification through the colposcope. The tissue sampling was done by applying firm pressure with the SpiraBrush, keeping the flat surface of the brush in contact with the visualized lesion that was sampled. The brush was rotated one rotation clockwise and one rotation counterclockwise while firm pressure was maintained against the cervix. The back and forth rotation cycle of the brush was repeated for a minimum of three cycles, and the provider continued to rotate the brush, increasing the pressure applied, if necessary, until the head was abundantly covered with bloody, mucoid material. This protocol insured the fewest number of inadequate samples (samples containing no basal or parabasal cells). The SpiraBrush head was then snapped off at the scored mark and placed directly into a prelabeled vial of CytoRich (TriPath Imaging Inc., Burlington, NC), an alcohol-based cytologic preservative solution. The patient was then asked to score her discomfort associated with the sample collection using the following criteria: 0 = no discomfort; 1 = mild discomfort during the sampling procedure; 2 = discomfort after the sampling procedure, defined as sensation of "pain"; 3 = pain and/or discomfort associated with cramping after sampling; and 4 = pain and/or discomfort associated with cramping that persists for longer than 5 minutes.
After the scoring of discomfort, the clinical investigator documented post-SpiraBrush bleeding, using the following criteria: 0 = no bleeding, 1 = minimal bleeding controlled by dabbing with a swab (less than 1 mL), 2 = moderate bleeding lasting up to 1 minute (1–5 mL), and 3 = moderate to severe bleeding continuing longer than 1 minute (5 mL or more). The results of this bleeding assessment were recorded on the study data form.
A single punch biopsy of the same cervical lesion was next obtained, but only after complete hemostasis from the SpiraBrush sample had been obtained. The investigator was instructed to take efforts to have the punch biopsy sample area include both a portion of the remaining unsampled cervical lesion and part of the area that had been previously sampled with the SpiraBrush. At each of the study centers, the biopsy forceps used included the following: Tischler-Kevorkian punch biopsy (Cooper Surgical, Shelton, CT), Kevorkian-Pacific punch biopsy (Cooper Surgical), or Schubert punch biopsy device (Cooper Surgical). In all cases the "bite" of the biopsy forceps was between 8 x 3 mm and 9.5 x 3 mm.
The punch biopsy specimen was placed into a container of 10% formalin as a preservative in a prelabeled study container. After the punch biopsy was obtained, the subject was then asked to score her discomfort associated with the obtaining of the punch biopsy using the same discomfort scale (0–4). This information was documented on the study data form. The investigator documented post–punch biopsy bleeding, using the same criteria (0–3). The investigator was also required to indicate the initial size and characteristics of the cervical lesion being studied, the location of the SpiraBrush sampling, and the location of the punch biopsy sampling on a diagram of the cervix on the study data form.
After the recording of the post–punch biopsy data, the subject was told that the investigational portion of her visit had been concluded. The investigator proceeded with completing the scheduled large loop excision of the transformation zone procedure after administering local anesthesia. The specimens from the large loop excision of the transformation zone were then processed in the usual manner for each institution. The two investigational specimen samples (SpiraBrush biopsy sample and punch biopsy sample) were placed into a prepaid mailing box and forwarded to a study coordinator for further blinding, processing, and analysis.
Coded samples were sent to one of the participating laboratories (Associated Regional and University Pathologists Inc. Laboratory, Salt Lake City, Utah). All SpiraBrush specimens were processed in a uniform manner. A thin-layer cytology slide was prepared using the Prep technique (TriPath Imaging). The remaining sample was centrifuged, and the residual cell pellet was coagulated with plasma and thromboplastin and prepared as a standard paraffin cell block. The punch biopsy was prepared as a standard paraffin tissue block.
The cytologic SpiraBrush sample was interpreted using the cytologic criteria of the Bethesda System15 for reporting, whereas both the SpiraBrush cell block and punch biopsy tissue block were sectioned, stained with hematoxylineosin, and then studied by light microscopy. The histological data for both cell block and punch biopsy specimens were reported using both a modified Bethesda reporting system for tissue and the CIN I, II, III reporting system.16 A standard laboratory data form was used to report the results of all study specimens. Cytology and cell block slides were initially read at the Associated Regional and University Pathologists laboratory by one of two pathologists (JSB and CJM) who were blinded to any clinical data regarding these specimens. Punch biopsy specimens were initially read at the Associated Regional and University Pathologists laboratory by a study pathologist. All samples were then sent to the pathology laboratory at the Women and Children’s Hospital at the University of Southern California. The same set of slides was read by two additional pathologists, who were also blinded from clinical information as well as previous slide readings. The data forms were completed separately by each of the pathologists, and the data were available only to the study coordinator.
For each cytologic preparation and cell block, an assessment of the presence of basal and parabasal cells was made. If these cell types were not present in the specimen, the specimen was judged to be inadequate because the sample could not be considered to represent a true transepithelial sample of the cervical epithelium.
All specimens from the large loop excision of the transformation zone were prepared and read at the same institution where the subject was being treated. All readings were made according to the standard procedure for the institution, and diagnoses were reported using the modified Bethesda system of classification.
To provide sufficient power to detect differences between the brush sampling and punch biopsy techniques, we determined that, with 45 samples analyzed by both methods and using the McNemar test to analyze the data, a difference in differential responses of 0.2 could be detected (assuming the probability of a discordant pair of 0.25), with about 84% power and a one-sided (lack of inferiority) significance level of .05.
For the SpiraBrush samples, the individual cytology readings of the three pathologists and the cell block readings from the three reviewing pathologists were tabulated in terms of outcome. A diagnostic consensus was reported when two of three reviewing pathologists agreed on the final diagnosis. If there was diagnostic discordance among all three pathologists, then the worst pathologic diagnosis was used as the final outcome for final data analysis. Both thin-layer cytologic and cell block data are reported separately here.
All punch biopsy specimens were read by three reviewing pathologists who were blinded from the details regarding each specimen, and tabulated. A consensus reading was determined, just as was done for the SpiraBrush samples. Again, if there was discordance in the outcome of the three readings, then the worst pathology diagnosis was recorded in the consensus data analysis.
The data comparing the pain and discomfort for SpiraBrush and the conventional punch biopsy sampling were subjected to the Wilcoxon signed-rank test using a two-tailed analysis. The data for the quantification of bleeding (as judged by the clinician) for the SpiraBrush and for the conventional punch biopsy were also analyzed using the test for marginal homogeneity, also known as the Bowker test.17
The 2 x 2 tables comparing the diagnoses with SpiraBrush and punch biopsy were analyzed using the McNemar test (StatXact 5.0; Cytel Software Corp., Cambridge, MA). For all calculations, a P value of .05 or less was considered statistically significant.
A comparison of the sensitivity of the brush biopsy in this study with the sensitivity of punch biopsy as reported in previous studies was analyzed using 
2 statistics as well as the Fisher exact test. Weighted 
statistics based on the assumption of a linear score difference are reported for the comparison of the different methods.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Seven of the cell block preparations and eight of the cytologic preparations were deemed inadequate by the consensus analysis (no basal or parabasal cells identified). In five of these women, neither preparation yielded adequate samples. Of the two additional women with an inadequate cell block preparation, one had high-grade disease detected on large loop excision of the transformation zone and a high-grade consensus diagnosis from her cytologic preparation, whereas the other woman had a normal large loop excision of the transformation zone and a negative thin-layer cytology preparation. In each of these women the punch biopsy was negative.
Of the three women with inadequate thin-layer preparations only, one woman had high-grade disease on the large loop excision of the transformation zone specimen, the cell block preparation, and the punch biopsy. One of the remaining two women had a low-grade lesion on her large loop excision of the transformation zone specimen that was not identified on the corresponding punch biopsy or cell block, and the other woman had no pathology on any of the sampling and biopsy techniques.
The correlation between punch biopsy and the large loop excision of the transformation zone diagnosis from this study population is shown in Table 1. In 25 women there was diagnostic correlation between the large loop excision of the transformation zone interpretation and the punch biopsy consensus. Of the 32 women with a large loop excision of the transformation zone diagnosis of high-grade dysplasia or worse, 17 were correctly identified on the punch biopsy if either high- or low-grade dysplasia was used as the threshold for being considered abnormal. In 15 women with high-grade squamous intraepithelial lesions (SILs) in specimens from the large loop excision of the transformation zone, the punch biopsy was read as being free of pathology. Five cases of low-grade SILs were also read as normal on punch biopsy. The single case of microinvasive cancer seen on large loop excision of the transformation zone was diagnosed as CIN III on the corresponding punch biopsy. The weighted statistic for these data was 0.30.
|
shows the correlation between the diagnosis made from SpiraBrush samples and the diagnosis from samples from the large loop excision of the transformation zone. In the top half of Table 2 the data from the 40 women with adequate cell block preparations are presented. There was diagnostic correlation in 26 women. Of the 29 women with a large loop excision of the transformation zone diagnosis of high-grade SILs or worse, 26 were correctly identified on the cell block if either high- or low-grade dysplasia was used as the threshold for being considered abnormal. Three women with high-grade SILs on large loop excision of the transformation zone had no pathology detected on the cell block, and four women with low-grade SILs on their large loop excision of the transformation zone specimen had no abnormal cells identified.
|
shows the correlation between the diagnosis from the large loop excision of the transformation zone and thin-layer cytology preparations from the 39 women with adequate thin-layer cytology preparations. There was diagnostic correlation in 25 of these women. Of the 30 women with a diagnosis from large loop excision of the transformation zone of high-grade dysplasia or worse, 27 were correctly identified on the thin-layer cytology specimen if either high- or low-grade dysplasia was used as the threshold for being considered abnormal. Three women with high-grade SILs on large loop excision of the transformation zone had no pathology detected on thin-layer cytology, and four women with low-grade SILs on their large loop excision of the transformation zone specimen had no abnormal cells identified on their cytologic preparation. The weighted statistics for the top and bottom halves of Table 2 were 0.47 and 0.43, respectively.
Table 3 shows the ability of each of the biopsy techniques to indicate either high-grade cervical pathology or worse, or low-grade pathology or worse. When the large loop excision of the transformation zone diagnosis was high grade or worse, and the threshold for considering a biopsy to be abnormal was that it revealed high-grade disease, the punch biopsy correctly identified women in 17 of 32 cases (53.1%), whereas the SpiraBrush biopsy correctly identified disease in 79.3% and 76.7% of the cases (difference between SpiraBrush and punch biopsy P = .04 by the McNemar test). Lowering the threshold for a biopsy to be considered abnormal to low-grade disease or worse did not change this apparent discrepancy (difference between SpiraBrush and punch biopsy with low-grade SILs or worse threshold P = .004 by the McNemar test). When considering those women with a large loop excision of the transformation zone sample showing low-grade pathology or worse, there was no difference between the two SpiraBrush processing techniques, and the brush again appeared to have a higher sensitivity for finding disease than did the punch biopsy (difference between SpiraBrush and punch biopsy P = .004 by the McNemar test).
|
With regard to blood loss, the SpiraBrush procedure resulted in less bleeding and need for hemostasis. The median (± interquartile range) blood loss scores were 1 ± 0 for the SpiraBrush and 2 ± 1 for the punch biopsy (P < .001). In addition to this statistical calculation, an analysis of the data demonstrates that in 28 of the 47 women bleeding persisted for longer than 1 minute after the punch biopsy, requiring the examiner to use a cotton-tipped applicator repeatedly. For the SpiraBrush biopsy device, only two patients had bleeding that lasted more than 1 minute.
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We have demonstrated that sampling exocervical lesions with a spiral-shaped, stiff bristled brush provides a sufficient cellular and tissue sample to allow for the accurate identification of underlying cervical pathology. Considering women for whom the brush sample was adequate, the SpiraBrush instrument resulted in an exact correlation with high-grade dysplasia in 23 of 29 women (79.3%) using cell block processing or 23 of 30 women (76.7%) using thin-layer cytology (Table 2). The fact that the brush biopsy sample was split between cytology and cell block did not appear to influence the data. The methods appear to predict the underlying pathology equally well. In comparison, the corresponding punch biopsy showed an exact correlation in 17 of 32 women with high-grade dysplasia (53.1%) (Table 1) (P = .048 by 2). In addition, there was significantly less bleeding after SpiraBrush sampling than after punch biopsy, and patients experienced significantly less discomfort with the brush device.
These data would seem to indicate that the SpiraBrush sample correlates more closely with the diagnosis from the large loop excision of the transformation zone than does the punch biopsy. We caution against drawing such a conclusion from these data alone because the study design favored the brush device sampling by having SpiraBrush biopsies taken before a punch biopsy. Although one could argue the fact that the punch biopsy still should have been diagnostic because the subsequent large loop excision of the transformation zone did reveal pathology at the same location, we feel that the sampling order created a significant bias in favor of the brush biopsy device, negating the validity of any statements regarding enhanced sensitivity of the brush biopsy when compared with the punch biopsy. Although it is true that previous investigators have demonstrated a lack of agreement between a directed punch biopsy and a subsequent large loop excision of the transformation zone or cone biopsy, this is most often confined to a single-grade disagreement between the two procedures.7,24 The incidence of completely negative punch biopsies in the face of a large loop excision of the transformation zone or cone biopsy diagnoses of dysplasia or invasive cancer is more in the range of 5–25% of all dysplasias detected.3,4,25 Inadequate sampling by the punch biopsy device is most frequently cited as the most likely cause of this disagreement, and some authors have even suggested that large loop excision of the transformation zone should be performed on all suspicious lesions found at colposcopy.26,27
The nearly 50% false-negative punch biopsy rate in this study is, in our opinion, artifactually increased because of the order of sampling. In almost all cases, the biopsy brush resulted in a disturbance of the operator’s ability to identify the most abnormal area for sampling and, moreover, decreased the volume of dysplastic epithelial tissue remaining for punch biopsy sampling.
To eliminate this potential bias and evaluate the SpiraBrush device’s ability to correctly identify underlying pathology, a comparison was made between the results of the SpiraBrush sampling in the current study and the results from previous studies of colposcopically directed punch biopsies performed in women who were undergoing either loop excision or cone biopsy.3,7,24,25 These data are displayed in Table 4. In these studies, a punch biopsy of low-grade disease or worse was considered as positive, so the corresponding sensitivity for SpiraBrush from Table 3 was used in this analysis. A comparison of the five studies by use of the 2 statistic yields a P value of .14, and using the Fisher exact test to compare all of the punch biopsy studies versus Spirabrush sensitivity yields a P value of .60. These data indicate that, using the parameters mentioned above, there is no apparent difference between the sensitivity of SpiraBrush biopsy and that of a colposcopically directed punch biopsy.
|
It is also worth commenting on the fact that, unlike a directed punch biopsy, the spiral brush biopsy device did not require sampling of the "worst" area of pathology, but rather a placement that would cover at least half of the visualized abnormality. This characteristic of the brush device is likely a result of its larger volume of sampling of the epithelium than a directed punch biopsy. The punch biopsy devices used in the current study produced tissue specimens that ranged from 2 to 7 mm in length and from 2 to 4 mm in width. These dimensions are similar to those reported by Prendiville et al.30 For the 28 punch biopsy specimens from this study in which specific dimensions were given, the mean length of the tissue sample was 4.07 ± 1.46 mm and the mean width was 3.07 ± 0.86 mm. The volume of squamous epithelium available for study would be the product of these two dimensions and the depth of the squamous layer. Assuming a depth of 0.5 mm from surface to basement membrane, the volume of cells available for study would be 6.25 mm3. In contrast, the brush biopsy device has a diameter of 13 mm and a potential cellular volume of 132.67 mm3. This more than 20-fold increase in cellular sample likely contributes to the accuracy obtained with the SpiraBrush biopsy device in the absence of sample site selection.
Last, it is important to note that the brush biopsy device described in these studies, though capable of obtaining samples of epithelium to the depth of the basement membrane, suffers from the limitation of not being able to sample the cervical subepithelial stroma. This would limit the ability of this device to predictably diagnose invasive lesions, and would require an additional biopsy at the time of a subsequent colposcopy to verify this occurrence. Further studies of patients with more advanced disease would indicate exactly what role, if any, SpiraBrush data would play in management decisions related to women with invasive lesions.
|
Footnotes |
|---|
Financial Disclosure
This research was supported by a grant from The Trylon Corporation. Dr. Stewart A. Lonky is the Medical Director and Dr. Neal M. Lonky has served as a consultant to the Trylon Corporation; they have an equity interest in the company. No other contributor to this research project or manuscript has any financial relationship to the Trylon Corporation.
Received March 14, 2002. Received in revised form June 7, 2002. Accepted June 27, 2002.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Paraskevaidis E, Malamou-Mitsi V, Koliopoulos G, Pappa L, Lolis E, Georgiou I, et al. Expanded cytological referral criteria for colposcopy in cervical screening: Comparison with human papillomavirus testing. Gynecol Oncol 2001; 82:355–9.[Medline]
3. Buxton EJ, Luesley DM, Shafl MI, Rollason M. Colposcopically directed punch biopsy: A potentially misleading investigation. Br J Obstet Gynaecol 1991;98:1273–6.[Medline]
4. Matsuura Y, Toshinori K, Toki N, Sugihara K, Kashimura M. Early cervical neoplasia confirmed by conization. Diagnostic accuracy of cytology, colposcopy, and punch biopsy. Acta Cytol 1996;40:241–6.[Medline]
5. Massad LS, Halperin CJ, Bitterman P. Correlation between colposcopically directed biopsy and cervical loop excision. Gynecol Oncol 1996;60:400–3.[Medline]
6. Omigbodun AO, Thomas JO, Adewole IF, Ogunlusi ML. Triage of patients with abnormal cervical smears in the absence of colposcopy. Trop Geogr Med 1993;45:157–8.[Medline]
7. Howe DT, Vincenti AC. Is large loop excision of the transformation zone (LLETZ) more accurate than colposcopically directed punch biopsy in the diagnosis of cervical intraepithelial neoplasia? Br J Obstet Gynaecol 1991; 98:588–91.[Medline]
8. Chappette OA, Byrne DL, Raju KS, Nayagam M, Kenney A. Histologic differences between colposcopic directed biopsy and loop excision of the transformation zone (LETZ): A cause for concern. Gynecol Oncol 1991;43: 46–50.[Medline]
9. Darwish A, Gadallah H. One step management of cervical lesions. Int J Gynecol Obstet 1998;61:261–7.[Medline]
10. Herzog TJ, Williams S, Adler LM, Rader JS, Kubiniec RT, Camel HM, et al. Potential of cervical intraepithelial neoplasia. Gynecol Oncol 1995;57:286–93.[Medline]
11. Ferris DG, Hainer BL, Pfenninger JL, Zuber TJ. "See and treat" electrosurgical loop excision of the cervical transformation zone. J Fam Pract 1996;42:253–7.[Medline]
12. Klam S, Arseneau J, Mansour N, Franco E, Ferenczy A. Comparison of endocervical curettage and endocervical brushing. Obstet Gynecol 2000;96:90–4.
13. Lieberman RW, Henry MR, Laskin WB, Walenga J, Buckner S, O’Connor DM. Colposcopy in pregnancy: Directed brush cytology. Obstet Gynecol 1999;94: 198–203.
14. Sciubba JJ. Improving detection of precancerous and cancerous oral lesions. Computer assisted analysis of oral brush biopsy. J Am Dent Assoc 1999;130:1445–57.
15. Kurman RJ, Solomon D. Descriptive diagnosis: Definitions, criteria and explanatory notes in the Bethesda System for reporting cervical/vaginal cytologic diagnoses. New York: Springer-Verlag, 1994:9–44.
16. Sculley RF, Bontiglio TA, Kurman RV, Silverman SG, Wilkerson EJ. Histologic typing of female genital tract tumors. In: Sobin LH, ed. World Health Organization international network histological classification of tumors. 2nd ed. New York: Springer-Verlag, 1994:39–41.
17. Agresti A. Categorical data analysis. New York: John Wiley and Sons Inc., 1990:365–70.
18. Matsuda M, Horai T, Nakamura S, Nishio H, Sakuma T, Ikegami H, et al. Bronchial brushing and bronchial biopsy: Comparison of diagnostic accuracy and cell typing reliability in lung cancer. Thorax 1986;41:475–81.
19. Rosell A, Monso E, Lores L, Vila X, Llatjos M, Ruiz J, et al. Cytology of bronchial brush biopsy rinse fluid to improve diagnostic yield for lung cancer. Eur Respir J 1998;12:1415–8.[Abstract]
20. Farouk R, Edwards J, Thorne M, Young W, MacDonald AW, Monson JR, et al. Brush cytology for the diagnosis of rectal carcinoma. Br J Surg 1996;83:1456–8.[Medline]
21. Arroliga AC, Matthay RA. The role of bronchoscopy in lung cancer. Clin Chest Med 1993;14:87–98.[Medline]
22. Petrelli NJ, Letourneau R, Weber T, Nava ME, Rodriguez-Bigas M. Accuracy of biopsy and cytology for the preoperative diagnosis of colorectal adenocarcinoma. J Surg Oncol 1999;71:46–9.[Medline]
23. Tate KM, Strickland JL. A randomized controlled trial to evaluate the use of the endocervical brush after endocervical curettage. Obstet Gynecol 1997;90:715–7.[Abstract]
24. Bonardi R, Cecchini S, Grazzini G, Ciatto S. Loop electrosurgical excision procedure of the transformation zone and colposcopically directed punch biopsy in the diagnosis of cervical lesions. Obstet Gynecol 1992;80:1020–2.
25. Skehan M, Soutter WP, Lim K, Krausz T, Pryse-Davies J. Reliability of colposcopy and directed punch biopsy. Br J Obstet Gynaecol 1990;97:811–6.[Medline]
26. Megevand E, Van Wyk W, Knight B, Bloch B. Can cervical cancer be prevented by a see, screen, and treat program? A pilot study. Am J Obstet Gynecol 1996;174: 923–8.[Medline]
27. Das SS, Elias AH. Diagnosis and treatment of cervical intraepithelial neoplasia in a single visit. Aust N Z Obstet Gynaecol 1998;38:246–50.[Medline]
28. Hatch KD, Hacker NF. Intraepithelial disease of the cervix, vagina, and vulva. In: Berek JS, Adashi EY, Hillard PA, eds. Novak’s gynecology. 12th ed. Baltimore: Williams and Wilkins, 1996:447–86.
29. Litvinov SV, van Driel W, van Rhijn CM, Bakker HA, van Krieken H, Fleuren GJ, et al. Expression of Ep-CAM in cervical squamous epithelia correlates with increased proliferation and the disappearance of markers for terminal differentiation. Am J Pathol 1996;148:865–75.[Abstract]
30. Prendiville W, Davies R, Berry PJ. A low voltage diathermy loop for taking cervical biopsies: A qualitative comparison with punch biopsy forceps. Am J Obstet Gynecol 1986;93:773–6.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
