The LAST Guidelines in Clinical Practice. Implementing Recommendations for p16 Use

Transcription

1 AJCP / Original Article The LAST Guidelines in Clinical Practice Implementing Recommendations for p16 Use Lani K. Clinton, MD, PhD, 1,2 Kyle Miyazaki, 1 Asia Ayabe, 1 James Davis, PhD, 2 Pamela Tauchi-Nishi, MD, 1,2 and David Shimizu, MD 1,2 From the 1 Hawaii Pathologists Laboratory, Queen s Medical Center, Honolulu, and 2 John A. Burns School of Medicine, University of Hawaii, Honolulu. Key Words: p16; LAST guidelines; Cervical pathology; Gynecologic pathology Am J Clin Pathol December 2015;144: ABSTRACT Objectives: To determine the impact of implementing p16 Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions (LAST) guidelines, we compared p16 use and follow-up data before and after implementation of the guidelines. Methods: We reviewed all cervical biopsy specimens diagnosed by two pathologists before and after implementation of the LAST guidelines and calculated the rate of and reason for p16 use across all biopsy specimens, high-grade squamous intraepithelial lesion (HSIL) detection, and follow-up. Results: In total, 1,829 and 1,623 cervical biopsy specimens were reviewed in periods A and B, respectively. Overall p16 use increased from 2.8% to 6.2% (P <.001). Recommendations 2 and 4 increased from 0.16% and 0% of all cervical biopsy specimens in period A to 1.4% and 1.9% in period B, respectively (). p16+ HSIL increased from 1.4% to 2.3% (P <.05). The positive predictive value of p16+ HSIL increased from 48% to 76% (P <.05). Conclusions: Implementation of the p16 LAST guidelines resulted in a significant increase in p16 use and a significant increase in the positive predictive value of p16+ HSIL. The human papillomavirus (HPV) associated squamous lesions of the lower anogenital tract (LAT) have long presented a challenge with respect to terminology and diagnosis. HPV-infected squamous epithelium responds biologically in a process that results in a transient low-grade lesion or a more stable precancerous lesion. Interestingly, it is nearly impossible to differentiate between intraepithelial lesions from different LAT sites, including the cervix, anus, or penis. 1,2 The Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions (LAST Project) 3 aimed to align terminology for HPV-associated squamous lesions of the LAT with current scientific knowledge by proposing a two-tiered system low-grade squamous intraepithelial lesion (LSIL) and high-grade squamous intraepithelial lesion (HSIL) across all LAT sites in both sexes. The LAST Project also suggested guidelines for the use of biomarkers, and immunohistochemical staining with p16 was selected as the preferred biomarker for use in cervical lesions. 3 Recommendation 1 is to use p16 to differentiate between HSIL and mimics such as atrophy, immature squamous metaplasia, or tangential cuts. Recommendation 2 is to use p16 if a diagnosis of cervical intraepithelial neoplasia grade 2 (CIN2) is considered, which morphologically falls between a low-grade lesion and a precancerous lesion. Recommendation 3 is to employ p16 in the case of professional disagreement. Recommendation 4a is for the use of p16 in cases of high-risk colposcopic referral situations where the H&E biopsy specimen is interpreted as LSIL or lower. The aim of this study is to determine if clinical implementation of the LAST Project recommendations for p16 use results in improved patient care. We hypothesize that 844 Am J Clin Pathol 2015;144:

2 AJCP / Original Article following the LAST Project recommendations for p16 will enhance pathologists ability to reliably and accurately differentiate between precancerous lesions and mimics. This will result in more appropriate treatment, decreasing the risk of cervical cancer, as well as the morbidity associated with unnecessary intervention. Materials and Methods After obtaining approval from our institutional review board, the CoPath (Sunquest Information Systems, Tucson, AZ) records of our institution were searched for all cervical biopsy cases from May 1, 2011, through October 31, All glandular lesions of the cervix were excluded from our study. The following data were recorded for each patient: (1) the pathologic diagnosis, (2) recommendation category for p16 use, and (3) interpretation of p16 immunostain. The recommendation categorization of p16 use was per the LAST guidelines as follows and will be referred to as recommendations 1, 2, 3, and 4a throughout this article: Recommendation 1: HSIL vs mimics Recommendation 2: Possible CIN2 Recommendation 3: Professional disagreement Recommendation 4a: High-risk colposcopic referral situations with H&E biopsy specimens initially LSIL or lower For recommendation 4a, we used HSIL or atypical squamous cells, cannot rule out HSIL (ASC-H), as the criterion for the high-risk colposcopic referral situations. We calculated the total number of cervical biopsy specimens for two time periods. The 12-month time period before implementation of the LAST guidelines from May 1, 2011, through April 30, 2012, was designated period A. We implemented the LAST guidelines in our laboratory in May 2012, but we allowed for a 6-month transition period. The 12-month time period after implementation of the LAST guidelines from November 1, 2012, through October 31, 2013, was designated period B. In addition, the frequency of p16 use, category of p16 use, and the number of cases in which the HSIL diagnosis was made with the assistance of p16 were calculated for each time period. For period B, the reason for the p16 immunostain was indicated at the time ordered by the sign-out pathologist per the LAST guidelines. For period A, the cases with p16 were identified from CoPath using a keyword search, and the glass slides were pulled from the slide archives and reviewed to designate a reason for the p16 immunostain. According to our standard histopathologic protocol for cervical biopsy specimens and endocervical curettage, an initial slide with at least three 4-μm thick serial sections was prepared from formalin-fixed, paraffin-embedded (FFPE) blocks followed by three deeper levels (D1, D2, and D3). Each level, D1 to D3, was approximately 100 μm deeper in the FFPE block and contained at least three 4-μm thick serial sections. The slides were subsequently stained with H&E. A blank serial slide was collected after D1 and was held for potential p16 immunostain if deemed necessary and thus ordered by the sign-out pathologist. Immunostaining for p16 (clone JC8; Santa Cruz Biotechnology, Dallas, TX) was performed at our laboratory using the Bond III immunostainer (Leica Microsystems, Buffalo Grove, IL). The antibody was diluted 1:800 with 20 minutes of heatinduced epitope retrieval. Positive controls were run with each sample. Strong and diffuse block staining with p16 was interpreted as positive (ie, p16-positive HSIL diagnosis), and patchy, incomplete p16 staining was interpreted as negative. We obtained follow-up data by searching CoPath for loop electrosurgical excision procedure (LEEP) or cervical conization specimens for the HSIL cervical biopsy cases. For those patients with an identifiable follow-up LEEP or cone specimen, we recorded the excision diagnoses and calculated the positive predictive values of a cervical biopsy HSIL diagnosis before and after implementation of the LAST guidelines. All statistical calculations were performed by using SAS statistical software version 9.3 (SAS Institute, Cary, NC) in collaboration with a biostatistician. A P value of less than.05 was considered significant. Results Implementation of LAST Guidelines We reviewed 1,829 and 1,623 cervical biopsy cases in periods A and B, respectively. The rate of p16 use significantly increased from 2.79% to 6.16% in periods A to B (P <.001) Figure 1A. In addition, there were significant shifts with respect to the recommendation category of p16 use. There was no change in the use of p16 per recommendation 1, but there was a marked increase in the utilization of recommendation 2, when the pathologist suspected CIN2, from 0.16% to 1.42% of all cervical biopsy specimens (; Figure 1A). There was also an increase in recommendation 4a from 0% to 1.91% of all cervical biopsy specimens (P <.0001; Figure 1A). Both before and after implementation of the LAST guidelines, approximately half of the cases in recommendation 1 were p16 positive and diagnosed as HSIL, representing 1.37% of all cases in period A and 1.29% in period B Figure 1B. Prior to implementation of the LAST guidelines, approximately one-third of the cases in recommendation 2 were p16 positive and diagnosed as HSIL, representing 0.05% of all cervical biopsy specimens in this time period. After implementation of the LAST guidelines, approximately one-fourth of the cases in recommendation Am J Clin Pathol 2015;144:

3 Clinton et al / p16 LAST Guidelines in Clinical Practice A B Overall Rate of p16 Use 8% 6% 4% 2% 0% Figure 1 A, Overall rate of p16 use from period A to period B per Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions (LAST) recommendation category. B, Rate of p16+ high-grade squamous intraepithelial lesion (HSIL) from period A to period B per LAST recommendation category. CIN2, cervical intraepithelial neoplasia grade 2. 2 were p16 positive and diagnosed as HSIL, a significant increase to 0.37% of all cervical biopsy specimens (P <.05; Figure 1B). Recommendation 4a did not exist prior to the LAST guidelines, but approximately one-third of cases in this category were p16 positive, representing 0.68% of all cervical biopsy specimens. We never cited recommendation 3, professional disagreement, as a reason for p16 use. Our practice does not employ a formal routine review of cases, and thus the opportunity for overt professional disagreement is minimal. In cases with intradepartmental consultation, the reason for p16 use could be assigned to recommendation category 1, 2, or 4a. There was a significant increase in the total incidence of p16-positive HSIL from 1.42% in period A to 2.34% in period B (P <.05; Figure 1B). Representative H&E photomicrographs with corresponding positive p16 stains for recommendations 1, 2, and 4a are shown in Image 1. CIN2: Biopsies The rate of CIN2 in the p16-positive HSIL biopsy specimens increased from 34.6% in period A to 68.4% in period B (P <.001) Figure 2, and the rate of CIN2 in the non p16-positive HSIL biopsy specimens decreased significantly from 24.1% in period A to 4.32% in period B (; Figure 2). The overall rate of CIN2 decreased significantly from 25.1% in period A to 16.5% in period B (P <.05; Figure 2). Follow-up 2.6% 2.8% HSIL vs Mimic Period A Period B 0.16% Consider CIN2 1.4% Professional Disagreement High-Risk Colposcopic Referral P <.001 There was no significant difference in the LEEP or cone frequency for patients with HSIL biopsy specimens between 1.9% 2.8% Total 6.2% Rate of p16+ HSIL 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% period A, at 70.7%, and period B, at 65.0% Table 1. However, excision specimens that resulted from the p16-positive HSIL biopsy specimens illustrated a significant increase in the frequency of HSIL diagnosis, from 47.6% in period A to 75.9% in period B (P <.05; Table 1). Nearly 50% of the additional p16-positive HSIL diagnoses were due to the use of p16 per recommendations 2 and 4a. The overall rate of HSIL detection on the excision specimens trended upward from 64.5% in period A to 70.0% in period B. Of note, 85% of the patients who underwent excisional procedures for HSIL were between ages 21 and 45 years, potential childbearing years. Discussion 1.4% 1.3% HSIL vs Mimic Period A Period B P < % 0.37% Consider CIN2 Professional Disagreement High-Risk Colposcopic Referral 0.68% P <.05 Implementation of the p16 LAST guidelines in our practice resulted in a significant increase in p16 use, more than doubling the rate from 2.8% to 6.2%. The LAST guidelines predicted a p16 use rate of approximately 20%. 3 Although our rate of p16 use was much lower than that predicted by the LAST project, we detected significantly more p16-positive HSIL biopsy specimens when we implemented the LAST p16 guidelines. Unifying both the terminology of cervical dysplasia and p16 use in cervical biopsy specimens may also help decrease discrepant cytohistologic results. Intrainstitutional variability in p16 use can be quite high, ranging from 0% to 21% at one institution, and using p16 in more than 10% of cervical biopsy specimens resulted in improved cytohistologic correlation rates and lower variability in the frequencies of histologic diagnoses. 4 Moreover, 1.4% Total 2.3% 846 Am J Clin Pathol 2015;144:

4 AJCP / Original Article Percentage of CIN2 80% 70% P <.001 Period A Period B 68.4% 60% 50% 40% 34.6% 30% 24.1% 20% 16.5% 10% 0% P < % 4.32% p16+ HSIL Non-p16+ HSIL Overall CIN2 Figure 2 Cervical intraepithelial neoplasia grade 2 (CIN2) frequency on cervical biopsy specimens. HSIL, high-grade squamous intraepithelial lesion. p16 has been shown to reduce interobserver variability, particularly for CIN2+ compared with H&E alone.5 As the cost of medicine becomes an increasingly important topic, detecting more p16-positive HSILs on cervical biopsy specimens with only a modest increase in p16 use may be advantageous. Our increase in p16 use was primarily due to increased use for recommendations 2 (suspect CIN2) and 4a (high-risk Papanicolaou [Pap]). Prior to the LAST guidelines, cases with high-risk Pap smears and cervical biopsy specimens that were LSIL or lower did not undergo further investigation. The addition of p16 use in this recommendation category is designed to capture potential false-negative biopsy specimens for patients with a previous high-risk Pap. We performed p16 immunohistochemistry on the endocervical curettings (ECCs) and cervical biopsy specimens for these cases; approximately 30% of these cases would have been underdiagnosed as LSIL or lower without implementation of the LAST guidelines. In other words, when patients had a high-risk Pap with a cervical biopsy specimen diagnosed as LSIL or lower, p16 staining often demonstrated small fragments of tissue with HSIL. The positive p16 staining in these cases was typically in the ECC, and the lesion was extremely difficult to appreciate on H&E alone. This group of patients harbored occult HSIL detectable only by p16 stain. The photomicrographs in Image 1 illustrate how minute, inconspicuous fragments of tissue in an ECC are easily missed on H&E and strongly highlighted with p16. Prior to the LAST guidelines, diagnoses of CIN2 were made based on evaluation of H&E sections alone. CIN2 is a biologically equivocal lesion with morphologic features intermediate between both low- and high-grade lesions.6,7 It is a poorly reproducible diagnosis with uncertain biologic implications7; thus, decreasing the frequency of this Am J Clin Pathol 2015;144: Image 1 Representative photomicrographs from the recommendation categories: high-grade squamous intraepithelial lesion (HSIL) vs mimic (recommendation 1) (H&E and p16 immunostain, 200); consider cervical intraepithelial neoplasia grade 2 (CIN2; recommendation 2) (H&E and p16 immunostain, 200); and high-risk colposcopic referral (recommendation 4a) (H&E and p16 immunostain, 100).

5 Clinton et al / p16 LAST Guidelines in Clinical Practice Table 1 LEEP/Cone Excisions From Patients With p16+ HSIL Cervical Biopsy Specimens Biopsy Diagnosis diagnosis as shown in our study may be advantageous. The LAST guidelines recommend using p16 in all cases where CIN2 is considered, to differentiate HSIL from LSIL. In our practice, applying p16 in this manner resulted in an increased frequency of p16-positive HSIL diagnoses. With an increase in recommendation 2, one might expect a higher percentage of CIN2 diagnoses on cervical biopsy specimens. However, when we evaluated the breakdown of HSIL (CIN2 vs CIN3), there was only an increase in the p16 cases, with an overall decrease in the rate of CIN2 from 25.1% to 16.5%. One of the most interesting findings of this study was the increase in the positive predictive value of a p16-positive cervical biopsy HSIL diagnosis after implementation of the LAST guidelines, with a significantly higher percentage of the patients in period B harboring HSIL in their excision specimens. Indeed, nearly half of the HSIL diagnoses on the excision specimens in period B were due to additional p16-positive HSIL biopsy diagnoses in recommendation categories 2 and 4a, cases that may have been missed prior to implementation of the LAST guidelines. Furthermore, we found a trend toward improved overall detection of HSIL on excisions. Treatment options for high-grade cervical dysplasia such as LEEP and conization should be used appropriately, since they are not without short- and long-term complications. Short-term morbidities include infection, bleeding, and pain, and long-term potential complications include preterm labor, preterm premature rupture of membranes, and cervical stenosis. 8,9 These complications are of particular importance since 85% of our patients with HSIL who underwent LEEP/cone were of childbearing age. Our findings illustrate the positive clinical impact of following the p16 LAST guidelines and how the significant increase in p16-positive HSILs (without an increase in total HSIL diagnoses) is of great clinical importance. In summary, we found a significant increase in overall p16 use with the addition of recommendations 2 and 4a. Despite higher use of recommendation 2, our overall rate HSIL LEEP/Cone Diagnosis, No./Total No. (%) of CIN2 diagnosis actually decreased. With overall p16 use only one-fourth of what was predicted by the LAST project, we showed significantly improved detection of p16-positive HSIL. Therefore, there was a trend toward improved HSIL detection on cervical excisions without excessive use of p16 and without overtreatment. As pathologists become more familiar with and begin to implement the LAST guidelines, more frequent use of p16 may further improve HSIL detection, leading to more appropriate follow-up treatment. Corresponding author: Lani K. Clinton, MD, PhD, University of Hawaii Residency Program, 651 Halo St #411 E, Honolulu, HI 96813; lclinton@hawaii.edu. Biostatistical support was partially supported by grants from the National Institute on Minority Health and Health Disparities (U54MD and G12MD007601) from the National Institutes of Health. The views expressed in this article do not necessarily represent those of the Queen s Medical Center. References Rate of LEEP/Cone A B P Value A B P Value p16+ HSIL Recommendation category 1. HSIL vs mimic 10/20 (50.0) 10/13 (76.9) NS 20/25 (80.0) 13/15 (86.7) NS 2. Consider CIN2 0/1 (0) 6/8 (75.0) NA 1/1 (100.0) 8/12 (66.7) NS 3. Professional disagreement 0 0 NA 0 0 NA 4. High-risk colposcopic referral 0 6/8 (75.0) NA 0 8/11 (72.7) NA Total 10/21 (47.6) 22/29 (75.9) <.05 21/26 (80.8) 29/38 (76.3) NS Non-p16+ HSIL 110/165 (66.7) 69/101 (68.3) NS 165/237 (69.6) 101/162 (62.3) NS Total HSIL 120/186 (64.5) 91/130 (70.0) NS 186/263 (70.7) 130/200 (65.0) NS A, period A; B, period B; CIN2, cervical intraepithelial neoplasia grade 2; HSIL, high-grade squamous intraepithelial lesion; LEEP, loop electrosurgical excision procedure; NA, not applicable; NS, not significant. 1. Doorbar J. Papillomavirus life cycle organization and biomarker selection. Dis Markers. 2007;23: Doorbar J. The papillomavirus life cycle. J Clin Virol. 2005;32(suppl 1):S7-S Darragh TM, Colgan TJ, Thomas Cox J, et al. The Lower Anogenital Squamous Terminology Standardization project for HPV-associated lesions: background and consensus recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. Int J Gynecol Pathol. 2013;32: Singh C, Manivel JC, Truskinovsky AM, et al. Variability of pathologists utilization of p16 and Ki-67 immunostaining in the diagnosis of cervical biopsies in routine pathology practice and its impact on the frequencies of cervical intraepithelial neoplasia diagnoses and cytohistologic correlations. Arch Pathol Lab Med. 2014;138: Reuschenbach M, Wentzensen N, Dukstra MG, et al. p16ink4a immunohistochemistry in cervical biopsy specimens: a systematic review and meta-analysis of the interobserver agreement. Am J Clin Pathol. 2014;142: Am J Clin Pathol 2015;144:

6 AJCP / Original Article 6. Galgano MT, Castle PE, Atkins KA, et al. Using biomarkers as objective standards in the diagnosis of cervical biopsies. Am J Surg Pathol. 2010;34: Castle PE, Stoler MH, Solomon D, et al. The relationship of community biopsy-diagnosed cervical intraepithelial neoplasia grade 2 to the quality control pathology-reviewed diagnoses: an ALTS report. Am J Clin Pathol. 2007;127: Samson SL, Bentley JR, Fahey TJ, et al. The effect of loop electrosurgical excision procedure on future pregnancy outcome. Obstet Gynecol. 2005;105: Heinonen A, Gissler M, Riska A, et al. Loop electrosurgical excision procedure and the risk for preterm delivery. Obstet Gynecol. 2013;121: Am J Clin Pathol 2015;144: