Room for improvement: initial experience with anal cytology: observations from the college of American pathologists interlaboratory comparison program in nongynecologic cytology.
Anal cytology was included in the 2001 Bethesda System conference and atlas of cervicovaginal cytology because of the morphologic similarity between human papillomavirus-associated squamous lesions in these sites. (5) In the decade since the conference, more cytologists have been seeing these specimens in their practice, particularly as knowledge of increasing anal cancer rates and awareness of the potential of anal cancer screening in at-risk individuals have become more widespread. Anal cytology was first introduced into the educational slide sets of the College of American Pathologists' (CAP) Interlaboratory Comparison Program in Non-Gynecologic Cytology (NGC) in 2005. Because clinical experience with anal cytology is relatively recent and is concentrated in a few laboratories, few cytologists have experience with it. Familiarity with this specimen type may lead to enhanced diagnostic accuracy. Here, we report the performance of anal cytology during its initial years in the CAP NGC program.
MATERIALS AND METHODS
The CAP NGC program is an educational glass-slide program mailed quarterly. In 2011, approximately 2000 laboratories participated in the program and approximately 1400 kits had anal cytology slides included in the set. Conventional smears, ThinPrep (Hologic Inc, Boxborough, Massachusetts) preparations, and SurePath (BD, Franklin Lakes, New Jersey) preparations of anal cytology are included in the slide sets; however, the vast majority of anal cytology challenges in the program are ThinPrep preparations. Prior to acceptance of a slide into the CAP NGC program, 2 members of the Cytopathology Resource Committee must review it and agree on the slide's target diagnosis.
Anal cytology is categorized using the same interpretive categories as the 2001 Bethesda System for cervical cytology, modified for this specimen type. (5) For the purposes of the CAP NGC program, anal cytology with interpretations of low-grade squamous intraepithelial lesion (SIL) or higher is categorized under the general rubric of "suspicious or positive for malignancy." Benign interpretations, including negative for SIL or malignancy, reactive changes, and organisms are classified in the general category of "negative for malignancy." Concordance with the general category is classified as a participant response of positive or suspicious for a "Positive" case, or a participant response of negative for a "Negative" slide challenge. Concordance with the target diagnosis indicates the participant's specific response matched the target diagnosis for the slide challenge. Specific target diagnoses include negative for intraepithelial lesion or malignancy, low-grade SIL (LSIL), high-grade SIL (HSIL), squamous cell carcinoma (SCC), and organisms such as Ameba, Candida, and herpes virus. Questionable participant responses (for example, pituitary adenoma) that are not included on the NGC site-specific diagnostic list for anal cytology have been classified as" Invalid." Concordance with the target diagnosis indicates the participant's specific response matched the target diagnosis for the slide.
Using participant responses (pathologist, cytotechnologist, and laboratory) from 2006 to 2011, this study describes the performance of anal cytology in the CAP NGC educational slide program during a 6-year time span. Data from 2005 were excluded because of issues with the CAP NGC kit instructions. Individual laboratories vary in how the "laboratory response" is generated; it is not proscribed by the educational slide program. Some laboratories use the majority response of the participating pathologists and cytotechnologists in the laboratory, others use the pathologist's response, and others use the supervisor's response.
Two nonlinear mixed models were used to analyze the concordance rates for the target diagnosis and the general category. Four main factors were included in the analysis: (1) general category or target diagnoses, (2) program year from 2006 to 2011, (3) participant type (pathologist, cytotechnologist, or overall laboratory), and (4) preparation type (liquid based or conventional). The interaction terms between these factors were also included. Both models included a repeated-measures component to model the slide factor correlation structure, which controls for the slide-specific performance.
Statistical method: a significance level of .05 was used for these analyses.
This analysis includes 37 670 general category responses and 37 325 specific responses based on 708 anal cytology slides that were evaluated in the CAP NGC program during a 6-year period between 2006 and 2011. Pathologists responded to 18 960 challenges, cytotechnologist responses totaled 11 306, and overall laboratory responses totaled 7404. Overall, there was a 91.0% concordance with the "Negative" general category (N = 5238) and a 90.4% concordance with the "Positive" general category (N = 32 432).
By 2008, response concordance rates for the general category improved to present levels, ranging from a low of 58.4% concordance in 2006 to 96.1% concordance in 2008 (Table 1). Between 2006 and 2011, there was no significant change in the concordance of the specific participant responses for the target diagnosis, with concordance rates for the specific responses varying between 68.4% and 72.0%.
Cytotechnologists more accurately categorized challenges as "Positive" or "Negative" than did individual pathologists or the overall laboratory responses (P < .001). However, the laboratory's specific response more often matched the target diagnosis (72.0% concordance) than did either individual pathologist (70.6% concordance) or cytotechnologist (68.6% concordance) responses (P = .03; Table 2).
The concordance rates for the target diagnoses varied significantly (P < .001; Table 3). Participants most frequently correctly classified LSIL (85.0% concordance rate), with no significant difference between pathologists, cytotechnologists, or laboratories. Most anal cytology challenges with the target diagnosis of negative for intraepithelial lesion or malignancy and herpes simplex virus infection were also correctly classified, with 78.8% and 80.2% concordance rates, respectively. In fact, the best performance of any of the slide challenges was by cytotechnologists on slide challenges with herpes simplex virus infection, at a concordance rate of 90.7%. This rate was significantly better than pathologists' specific responses for slide challenges with the target diagnosis of herpes simplex virus, which was at 72.4% concordance.
Amebic organisms were infrequently recognized, with the lowest overall concordance rate for the target diagnosis at only 41.5%; pathologists and laboratory responses matched the target diagnosis significantly more often than cytotechnologists (P < .001; Figure 1, A). The poorest performance of any of the slide challenges was for cytotechnologist responses on slide challenges with Ameba having a concordance rate of 26.0%. Negative for intraepithelial lesion or malignancy, as well as other infectious organisms (for example, Trichomonas species and fungus), made up most of the other specific responses for slide challenges with the target diagnosis of Ameba. Only 5.1% of the slide challenges with Ameba were classified as LSIL or worse (data not shown).
The concordance rates for the slide challenges with target diagnoses of HSIL and SCC fared only slightly better than the Ameba challenges, at 57.1% and 56.2%, respectively. Here, there were no significant differences in performance between pathologists' and cytotechnologists' specific responses; however, the overall laboratory-specific response was significantly better (P = .01) than that of cytotechnologists for the target diagnosis of HSIL (Figure 1, B).
During the 6-year period of experience with anal cytology as part of the NGC program, there was significant improvement in the concordance rates for participants specific responses to the target diagnosis of LSIL but not HSIL (P < .001; Table 4). There are too few data to test for improvement over time for the target diagnosis of SCC.
Although the participants' specific responses matched the target diagnosis for the HSIL and SCC slide challenges less than 60.0% of the time, participants categorized most slides as abnormal, with more than 95.0% diagnosed as LSIL or greater (Table 5). The HSIL challenges were both under-called as LSIL (24.9%) and overcalled as SCC (14.3%). The SCC challenges were frequently interpreted as HSIL (40.4%).
There was no significant difference by slide preparation Type--either liquid-based or conventional--for the concordance rates for the target diagnosis or the general category (Table 6).
The inclusion of anal cytology in the CAP NGC educational slide program allows more cytologists the opportunity to view these specimens, even if not part of their routine practice. Although there was a significant increase in participants correctly classifying slide challenges as either "Positive" or "Negative" between 2006 and 2011, most improvement occurred within the first 2 years after anal cytology was introduced into the CAP NGC educational program. Temporally, this improvement occurred following clarification of the kit instructions after the introduction of these specimens into the program in 2005. Although properly considered nongynecologic specimens, which require pathologist review, anal cytology uses Bethesda System terminology in a fashion that is more akin to the Papanicolaou test for cervical cancer screening. However, issues with the CAP SCORES computer system preclude the use of the gynecologic cytology template for anal cytology in its educational program. This led to initial confusion as to the appropriate categorization of anal SIL as negative or as suspicious/positive for malignancy. After 2005, the kit instructions were clarified to label SIL with the general categorization of "Suspicious or Positive" for malignancy.
Anal cytology with LSIL was the best performer among slide challenges, having overall concordance rates of 85.0% for the target diagnosis. This compares favorably to the 84.8% rate for validated LSIL slides in the CAP PAP program for cervicovaginal cytology. (6) Among the squamous abnormalities, the correct diagnosis of LSIL also improved during the 6 years.
Disappointingly, with the exception of LSIL, there has been no significant improvement in correctly matching the specific response to the target diagnosis over time. Overall concordance rates hover around 70.0% during the 6-year time period reported here. Indeed, HSIL and SCC were correctly classified less than 60.0% of the time, with no significant improvement noted during the 6 years. Taken together, however, squamous abnormalities were recognized in more than 95.0% of challenges.
How does the correct classification of the degree of squamous abnormality have an impact on the clinical management of patients with abnormal anal cytology? That depends. In an ideal world with unlimited resources, all high-risk patients with atypical squamous cells or with more significant abnormalities would be triaged to high-resolution anoscopy. (4) Although the morphologic features of human papillomavirus-associated squamous lesions in the anus are similar to those in cervix, the extent and distribution of abnormal anal cytology results in this high-risk population are quite different from abnormal cytology results seen in cervical screening populations. The rate of histologic detection of high-grade squamous lesions is much higher in patients with atypical squamous cells of undetermined significance on anal cytology than in patients with atypical squamous cells of undetermined significance on a cervical Papanicolaou test. (7,8)
As anal cytology becomes more accepted as a routine screening test, however, accurate classification of the degree of SIL is increasingly important to expedite the evaluation of those with anal precancer because of the very limited availability of those with expertise in high-resolution anoscopy. Clearly, there is room for improvement. We need to get better at evaluating these specimens.
Identification of organisms, such as herpes simplex virus and Candida, in anal cytology mirrors that of gynecologic cytology. However, other organisms both pathologic and commensal can be seen on anal cytology. These include those rarely encountered in gynecologic cytology, such as Ameba, pinworm eggs, and other helminths. Recognition of amebic trophozoites and cysts was very poor on the CAP NGC challenges. Proper recognition of Ameba on anal cytology is essential for prompt appropriate clinical correlation with patient symptoms and additional studies, as warranted.
The level of performance on anal cytology in the CAP NGC program, especially with regard to specific interpretations, indicates that there is a need for continued education about anal cytology. Increased exposure evaluating a specific specimen type should theoretically improve performance, although little improvement was noted. More experience is obviously needed. At most, one anal cytologic slide was included per educational slide set. This may not be enough exposure to this sample type for individuals to become proficient. Perhaps concentrated training with multiple anal cytology samples viewed contemporaneously would improve performance. Continued educational efforts are needed.
Caption: Figure 1. A, Amebic cysts on anal cytology. Note their small size relative to the intermediate squamous cell and their internal structure. B, High-grade squamous intraepithelial lesion on anal cytology. The morphologic features are similar to those on gynecologic cytology (ThinPrep, Papanicolaou stains, original magnifications X600 [A]; original magnification X600 [B]).
(1.) Palefsky JM, Holly EA, Hogeboom CJ, Berry JM, Jay N, Darragh TM. Anal cytology as a screening tool for anal squamous intraepithelial lesions. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;14(5):415-422.
(2.) HPV and men-fact sheet. Centers for Disease Control and Prevention Web site. http://www.cdc.gov/std/hpv/STDFact-HPV-and-men.htm. Accessed July 18, 2012.
(3.) Anal cancer. American Cancer Society Web site. http://www.cancer.org/ Cancer/AnalCancer/DetailedGuide/index.htm. Accessed July 18, 2012.
(4.) Park IU, Palefsky JM. Evaluation and management of anal intraepithelial neoplasia in HIV-negative and HIV-positive men who have sex with men. Curr Infect Dis Rep. 2010;12(2):126-133.
(5.) Darragh TM, Birdsong GG, Luff RD, Davey DD. Anal-rectal cytology. In: Solomon D, Nayar R, eds. The Bethesda System for Reporting Cervical Cytology. 2nd ed. New York, NY: Springer-Verlag; 2004:169-175.
(6.) Woodhouse SL, Stastny JF, Styer PE, Kennedy M, Praestgaard AH, Davey DD. Interobserver variability in subclassification of squamous intraepithelial lesions: results of the College of American Pathologists Interlaboratory Comparison Program in Cervicovaginal Cytology. Arch Pathol Lab Med. 1999; 123(11):1079-1084.
(7.) Bean SM, Chhieng DC. Anal-rectal cytology. Diagn Cytopathol. 2010;38(7): 538-546.
(8.) Zhao C, Domfeh AB, Austin RM. Histopathologic outcomes and clinical correlations for high-risk patients screened with anal cytology. Acta Cytol. 2012; 56(1):62-67.
Teresa M. Darragh, MD; Barbara Winkler, MD; Rhona J. Souers, MS; Rodolfo Laucirica, MD; Chengquan Zhao, MD; Ann T. Moriarty, MD; for the College of American Pathologists Cytopathology Committee
Accepted for publication January 1 6, 2013.
From the Departments of Pathology, and Obstetrics, Gynecology & Reproductive Sciences, University of California, San Francisco (Dr Darragh); the Department of Pathology, Mount Kisco Medical Group, Mount Kisco, New York (Dr Winkler);Biostatistics, College of American Pathologists, Northfield, Illinois (Ms Souers); the Department of Pathology and Immunology, Baylor College of Medicine, Houston, Texas (Dr Laucirica); the Department of Pathology, University of Pittsburgh Medical Center, Pennsylvania (Dr Zhao); and the Department of Pathology, AmeriPath, Indianapolis, Indiana (Dr Moriarty).
Research supplies for anal ThinPreps have been provided to Dr Darragh by Hologic Inc (Boxborough, Massachusetts). The authors have no other relevant financial interest in the products or companies described in this article.
Reprints: Teresa M. Darragh, MD, University of California, San Francisco/Mt. Zion Medical Center, Departments of Pathology and OB/Gyn, Box 1785, 1600 Divisadero St, Room B618, San Francisco, CA 94115 (e-mail: Teresa.firstname.lastname@example.org).
Table 1. Concordance Rates for the General Category and Target Diagnosis by Program Year No. of Concordance Program Year Responses Rate, % P Value General category .17 2006 1972 58.4 2007 4722 71.2 2008 5673 96.1 2009 6455 95.6 2010 9208 95.7 2011 9640 94.9 Target diagnosis .32 2006 1950 68.5 2007 4685 71.5 2008 5634 72.0 2009 6355 71.7 2010 9130 68.4 2011 9571 69.4 Table 2. Concordance Rates for the General Category and Target Diagnosis for General Category and Specific Response by Reader Participant Type No. of Concordance Participant Type Responses Rate, % P Value General category < .001 Pathologist 18 960 88.8 Cytotechnologist 11 306 93.4 Laboratory 7404 90.5 Target diagnosis .03 Pathologist 18 810 70.6 Cytotechnologist 11 225 68.6 Laboratory 7290 72.0 Table 3. Overall Concordance Rates for the Target Diagnosis and by Participant Type No. of Concordance Responses Rate, % P Value Target diagnosis < .001 LSIL 15 781 85.0 HSIL 16 025 57.1 SCC 329 56.2 NILM 3019 78.8 Amebic organisms 1696 41.5 HSV 475 80.2 Target diagnosis by < .001 participant type LSIL Path-Cyto P = .10 Pathologist 7844 84.4 Cytotechnologist 4838 85.8 Laboratory 3099 85.3 HSIL Lab-Cyto P = .01 Pathologist 8256 58.4 Cytotechnologist 4631 53.1 Laboratory 3138 59.6 SCC Path-Cyto P = .27 Pathologist 175 55.4 Cytotechnologist 93 60.2 Laboratory 61 52.5 NILM Path-Cyto P = .67 Pathologist 1490 77.9 Cytotechnologist 947 78.8 Laboratory 582 81.1 Amebic organisms Pathologist 813 49.6 Path-Cyto P < .001 Cytotechnologist 566 26.0 Lab-Cyto P < .001 Laboratory 317 48.3 HSV Path-Cyto P < .001 Pathologist 232 72.4 Cytotechnologist 150 90.7 Laboratory 93 82.8 Abbreviations: Cyto, cytotechnologist; HSIL, high-grade squamous intraepithelial lesion; HSV, herpes simplex virus; Lab, laboratory; LSIL, low-grade squamous intraepithelial lesion; NILM, negative for intraepithelial lesion or malignancy. Table 4. Concordance Rates for Target Diagnoses of Low-Grade Squamous Intraepithelial Lesion (LSIL), High-Grade Squamous Intraepithelial Lesion (HSIL), and Squamous Cell Carcinoma (SCC) by Year Target Diagnosis 2006 2007 LSIL concordance rate, 967 (79.8) 2094 (82.7) No. (%);n = 15 781 HSIL concordance rate, 841 (54.6) 1970 (56.8) No. (%); n = 16 025 SCC concordance rate, - - No. (%); n = 329 Year Target Diagnosis 2008 2009 LSIL concordance rate, 2701 (85.7) 2857 (84.4) No. (%);n = 15 781 HSIL concordance rate, 2097 (57.1) 2618 (59.2) No. (%); n = 16 025 SCC concordance rate, - - No. (%); n = 329 Target Diagnosis 2010 2011 LSIL concordance rate, 3433 (86.3) 3729 (86.5) No. (%); n = 15 781 HSIL concordance rate, 4130 (55.9) 4369 (57.7) No. (%); n = 16 025 SCC concordance rate, 97 (52.6) 232 (57.8) No. (%); n = 329 Target Diagnosis P Value LSIL concordance rate, < .001 No. (%); n = 15 781 HSIL concordance rate, .21 No. (%); n = 16 025 SCC concordance rate, Not tested No. (%); n = 329 Table 5. Percentage Concordance of Participants' Responses for Target Diagnosis for Low-Grade Squamous Intraepithelial Lesion (LSIL), High-Grade Squamous Intraepithelial Lesion (HSIL), and Squamous Cell Carcinoma (SCC) Participants' Response Target LSIL, HSIL, SCC, Total Abnormal Diagnosis % % % (LSIL+), % LSIL 85.0 8.5 1.9 95.4 HSIL 24.9 57.1 14.3 96.3 SCC 2.1 40.4 56.2 98.7 Table 6. Concordance Rates for the General Category and Target Diagnosis by Preparation Type Preparation No. of Concordance P Type Responses Rate, % Value General category .06 Liquid-based 37 096 90.4 Conventional 574 96.9 Target diagnosis .93 Liquid-based 36 758 70.4 Conventional 567 59.4
Please note: Illustration(s) are not available due to copyright restrictions.
|Printer friendly Cite/link Email Feedback|
|Author:||Darragh, Teresa M.; Winkler, Barbara; Souers, Rhona J.; Laucirica, Rodolfo; Zhao, Chengquan; Moriart|
|Publication:||Archives of Pathology & Laboratory Medicine|
|Date:||Nov 1, 2013|
|Previous Article:||Histopathologic diagnosis of pediatric neoplasms: a review of international consultations.|
|Next Article:||Should "indeterminate" diagnoses be used for thyroid fine-needle aspirates of nodules smaller than 1 cm?|