The utility of PAX8 and IMP3 immunohistochemical stains in the differential diagnosis of benign, premalignant, and malignant endocervical glandular lesions Richard Danialan a , Margaret Assaad a , Jason Burghardt b , Pamela Newcomb b , Richard W. Cartun a , Srinivas Mandavilli a, ⁎ a Hartford Hospital, Department of Pathology and Laboratory Medicine, Hartford, CT, USA b Clinical Laboratory Partners, Newington, CT, USA HIGHLIGHTS • PAX8 stains strong and diffusely in benign endocervical glandular lesions, and is progressively lost in EC AIS and ECA. • IMP3 antibody shows near absent staining in a cohort of benign endocervical glandular lesions • IMP3 extent and intensity of staining gradually increase in carcinoma in situ to conventional invasive endocervical adenocarcinoma abstract article info Article history: Received 29 January 2013 Accepted 12 April 2013 Available online 23 April 2013 Keywords: PAX8 IMP3 Ki-67 Cervical adenocarcinoma Cervical adenocarcinoma in situ Immunohistochemistry Glandular lesions of the endocervix can be diagnostically challenging and occasionally the differential diag- nosis includes endocervical adenocarcinoma in situ (EC AIS) and well-differentiated endocervical adenocar- cinoma (ECA). PAX8 and IMP3 are two markers which have not been well studied in the endocervix. Our aim was to evaluate their immunohistochemical (IHC) expression in benign and malignant endocervical glandular lesions as well as to compare them to the traditionally used panel (Ki-67, p16, CEA). Design. We searched our surgical pathology files for a cohort of benign endocervical glandular lesions as well as premalignant and malignant groups including EC AIS and ECA. An IHC panel consisting of PAX8, IMP3, Ki-67, p16, and CEA was performed on all cases. Immunoreactivity was scored on a degree of positivity (S0 = no immunoreactivity, S1 = up to 10% cells, S2 = between 10 and 50% cells, S3 => 50% cells) and intensity (Int0 — absent, Int1 — mild/faint, Int2 — moderate, Int3 — strong). Results. PAX8 showed diffuse positivity (S3) with at least a moderate intensity of staining (Int2) in the benign group. PAX8 was focal (S1) in ECA and faint (Int1), compared to EC AIS, which was moderate (S2) and faint (Int1). IMP3 expression was focal in the benign group (S1), moderate (S2) in EC AIS and moderate-to-diffuse (S2-3) in ECA. IMP3 intensity was faint (Int1) in benign lesions, moderate (Int2) in EC AIS, and strong (Int3) in ECA. Significant Ki-67, p16, and CEA expression was noted in the premalignant/malignant cohort. Conclusion. PAX8 and IMP3 can be helpful in the differential diagnosis of benign vs. malignant endocervical glandular lesions. Our study, however, shows that there is some degree of overlap of staining in both the benign and malignant group. As such, PAX8 and IMP3 should always be interpreted with caution and in combination with the histomorphology. Published by Elsevier Inc. Introduction There is a broad spectrum of benign glandular proliferations and changes of the endocervix including tunnel clusters (TC), endocervical laminar hyperplasia (ELH), microglandular hyperplasia (MGH), meso- nephric hyperplasia (MNH) and tuboendometrioid metaplasia (TEM). While the separation from such commonly found benign proliferations from EC AIS and ECA is usually straightforward, in some cases this can be extremely challenging. A variety of IHC markers including CEA, Ki-67, p16 and p53 [1,2] have been reported in the literature to aid in this distinction. However, there are limitations to the utility of such an IHC panel. Benign lesions such as MGH have been reported to express CEA and p53 [1,3], while p16 expression has been reported in metapla- sias of the endocervix [4]. PAX8 (paired box protein 8) is a recently described transcription factor on chromosome 2p13 that is involved in the development of the kidneys, eye, thyroid and reproductive system [5,6]. Recent liter- ature has demonstrated PAX8 expression in benign endocervical and endometrial epithelium as well as in a majority of conventional endo- metrial carcinomas [6,7]. However there is limited literature evaluat- ing PAX8 IHC expression in the endocervix [5,8–10]. Gynecologic Oncology 130 (2013) 383–388 ⁎ Corresponding author at: Hartford Hospital, Department of Pathology and Laboratory Medicine, 80 Seymour Street, Hartford, CT 06102, USA. E-mail address: [email protected](S. Mandavilli). 0090-8258/$ – see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.ygyno.2013.04.020 Contents lists available at SciVerse ScienceDirect Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Transcript
Gynecologic Oncology 130 (2013) 383–388
Contents lists available at SciVerse ScienceDirect
Gynecologic Oncology
j ourna l homepage: www.e lsev ie r .com/ locate /ygyno
The utility of PAX8 and IMP3 immunohistochemical stains in thedifferential diagnosis of benign, premalignant, and malignantendocervical glandular lesions
Richard Danialan a, Margaret Assaad a, Jason Burghardt b, Pamela Newcomb b,Richard W. Cartun a, Srinivas Mandavilli a,⁎a Hartford Hospital, Department of Pathology and Laboratory Medicine, Hartford, CT, USAb Clinical Laboratory Partners, Newington, CT, USA
H I G H L I G H T S
• PAX8 stains strong and diffusely in benign endocervical glandular lesions, and is progressively lost in EC AIS and ECA.• IMP3 antibody shows near absent staining in a cohort of benign endocervical glandular lesions• IMP3 extent and intensity of staining gradually increase in carcinoma in situ to conventional invasive endocervical adenocarcinoma
0090-8258/$ – see front matter. Published by Elsevier Ihttp://dx.doi.org/10.1016/j.ygyno.2013.04.020
a b s t r a c t
a r t i c l e i n f o
Article history:Received 29 January 2013Accepted 12 April 2013Available online 23 April 2013
Keywords:PAX8IMP3Ki-67Cervical adenocarcinomaCervical adenocarcinoma in situImmunohistochemistry
Glandular lesions of the endocervix can be diagnostically challenging and occasionally the differential diag-nosis includes endocervical adenocarcinoma in situ (EC AIS) and well-differentiated endocervical adenocar-cinoma (ECA). PAX8 and IMP3 are two markers which have not been well studied in the endocervix. Ouraim was to evaluate their immunohistochemical (IHC) expression in benign and malignant endocervicalglandular lesions as well as to compare them to the traditionally used panel (Ki-67, p16, CEA).
Design. We searched our surgical pathology files for a cohort of benign endocervical glandular lesions as wellas premalignant and malignant groups including EC AIS and ECA. An IHC panel consisting of PAX8, IMP3, Ki-67,p16, and CEA was performed on all cases. Immunoreactivity was scored on a degree of positivity (S0 = noimmunoreactivity, S1 = up to 10% cells, S2 = between 10 and 50% cells, S3 => 50% cells) and intensity (Int0— absent, Int1 — mild/faint, Int2 — moderate, Int3 — strong).
Results. PAX8 showed diffuse positivity (S3)with at least amoderate intensity of staining (Int2) in the benigngroup. PAX8 was focal (S1) in ECA and faint (Int1), compared to EC AIS, which was moderate (S2) and faint(Int1). IMP3 expression was focal in the benign group (S1), moderate (S2) in EC AIS and moderate-to-diffuse(S2-3) in ECA. IMP3 intensity was faint (Int1) in benign lesions, moderate (Int2) in EC AIS, and strong (Int3) inECA. Significant Ki-67, p16, and CEA expression was noted in the premalignant/malignant cohort.
Conclusion. PAX8 and IMP3 can be helpful in the differential diagnosis of benign vs. malignant endocervicalglandular lesions. Our study, however, shows that there is some degree of overlap of staining in both the benignand malignant group. As such, PAX8 and IMP3 should always be interpreted with caution and in combinationwith the histomorphology.
Published by Elsevier Inc.
Introduction
There is a broad spectrum of benign glandular proliferations andchanges of the endocervix including tunnel clusters (TC), endocervicallaminar hyperplasia (ELH), microglandular hyperplasia (MGH), meso-nephric hyperplasia (MNH) and tuboendometrioid metaplasia (TEM).While the separation from such commonly found benign proliferationsfrom EC AIS and ECA is usually straightforward, in some cases this can
nt of Pathology and Laboratory
lli).
nc.
be extremely challenging. A variety of IHC markers including CEA,Ki-67, p16 and p53 [1,2] have been reported in the literature to aid inthis distinction. However, there are limitations to the utility of such anIHC panel. Benign lesions such as MGH have been reported to expressCEA and p53 [1,3], while p16 expression has been reported in metapla-sias of the endocervix [4].
PAX8 (paired box protein 8) is a recently described transcriptionfactor on chromosome 2p13 that is involved in the development ofthe kidneys, eye, thyroid and reproductive system [5,6]. Recent liter-ature has demonstrated PAX8 expression in benign endocervical andendometrial epithelium as well as in a majority of conventional endo-metrial carcinomas [6,7]. However there is limited literature evaluat-ing PAX8 IHC expression in the endocervix [5,8–10].
384 R. Danialan et al. / Gynecologic Oncology 130 (2013) 383–388
IMP3 (insulin-like growth factor-II mRNA-binding protein 3), alsoknown as KOC (K-homologous domain containing protein overexpressedin cancer), is composed of a 580 amino acid protein encoded by 4350 nu-cleotides on an mRNA. The IMP3 gene is located on 7p11.5 and isexpressed during embryogenesis but not in adult tissue [4]. Aberrant ex-pressionof IMP3has been observed in a variety of carcinomas anddyspla-sias including the pancreas [11], cervix [12], colon [13], liver [14], headand neck [15], and the breast [16]. Recently, Cuizhen et al. [4] reportedp16 and IMP3 expression in benign endocervical glands, TEM and EC AIS.
The aim of this study was to compare the IHC expression of PAX8and IMP3 along with currently established markers (Ki-67, p16, andCEA) in benign endocervical glandular proliferations with that in ECAIS and ECA.
Materials and methods
Case selection
The surgical pathology files at Hartford Hospital were searched forcases of both benign and premalignant/malignant cervical lesionsbetween 2005 and 2012. There were 23 cases in the benign cohort(15 hysterectomy, 8 biopsy specimens). The premalignant/malignantcohort consisted of 21 cases (12 hysterectomy, 9 biopsy specimens).All cases were processed routinely using 10% buffered formalin and cutat 4 μm. The benign cohort included: MGH (5 cases), ELH (5 cases),TEM (2 cases), cervical endometriosis (CEM) (5 cases), tunnel clusters(TC) (5 cases), and MNH (1 case). The premalignant and malignantcohorts included: EC AIS (8 cases) and ECA (11). One additional case ofminimal deviation adenocarcinoma (MDA)metastatic to the small intes-tine and a poorly differentiated endocervical adenocarcinoma with sig-net ring features were also included for comparison.
Immunohistochemistry
Selected paraffin blocks from both cohorts were retrieved fromour surgical pathology archives and unstained slides were cut at4 μm in thickness. One unstained slide from each case was incubatedwith PAX8 polyclonal antibody. Another unstained slide was incubat-ed with KOC polyclonal antibody. Benign kidney and a case of pancre-atic adenocarcinoma were used as positive controls for the PAX8 andIMP3 antibodies, respectively. In addition, all cases were stained withcurrently utilized markers Ki-67, p16, and CEA antibodies for compar-ative study (Table 1). Each slide was examined for proper positiveand negative internal controls. The extent (S) and intensity (Int) ofimmunoreactivity were scored independently by two pathologists(SM, RD) with excellent interobserver agreement (>80%) utilizingthe following criteria: S0 = no immunoreactivity (absent), S1 = upto 10% of cells showing immunoreactivity (focal), S2 = between 10and 50% of cells showing immunoreactivity (moderate), S3 = >50%cells showing immunoreactivity (diffuse), Int0 = no staining, Int1 =mild intensity, Int2 = moderate intensity, Int3 = strong intensity.PAX8 nuclear and IMP3 cytoplasmic immunoreactivitywere consideredpositive, respectively. Only minimal background nonspecific stainingwas observed.
Table 1PAX8 and IMP3 IHC antibody information.
Antibody Source Dilution Antigenretrieval
Positive control
PAX8 ProteinTech, clone >1PAX8 ab (polyclonal)
1:100 LeicaMicrosystemsBond Max
Benign kidney
IMP3 DAKO, clone 69.1(polyclonal)
1:50 LeicaMicrosystemsBond Max
Pancreaticadenocarcinoma
Results
The results of the IHC patterns for both benign andmalignant lesionsare listed in Tables 2 and 3.
Benign endocervical glandular lesions
All benign endocervical glandular lesions (Images 1a, d, g, j, m, p)showed diffuse (S3) nuclear positivity for PAX8. The intensity of PAX8expression was moderate to strong (Int2-3) in all cases (Images 1b, e, h,k, n, and q).
IMP3 showed positive cytoplasmic staining in all cases of benignendocervical glandular lesions, but the extent of positivity was focal(S1) and the intensity faint or mild (Int1) (Images 1c, f, i, l, o, r).
Ki-67 showed a range of staining. Therewas either absent (S0/Int0) orfocal staining with up to moderate intensity (S1/Int2). CEA immunoreac-tivity was absent in all cases (S0/Int0). P16 was focal (S1) in all cases, butwith variable intensity of staining (ranging from Int1 to Int3).
Endocervical adenocarcinoma in situ (EC AIS)
All cases of EC AIS (Image 2a) were positive for PAX8, but with 50%or less of tumor cells staining (S2) and additionally the intensity of im-munoreactivity was mild (Int1) (Image 2b). IMP3 showed moderateimmunoreactivity in both extent and intensity (S2/Int2) (Image 2c).
The extent of Ki-67 staining was moderate (S2), while the intensi-ty was moderate to strong (Int2-3). P16 was diffuse (S3) with moder-ate intensity of staining (Int2) in EC AIS. CEA was moderate, both inextent and intensity of staining (S2/Int2).
Invasive endocervical adenocarcinoma (ECA)
Ten of twelve cases (83%) of ECA (Image 2d) showed focal (S1) andfaint (Int1) intensity of staining for PAX8 (Image 2e). IMP3 immunoreac-tivity ranged frommoderate to diffuse (S2-3) albeitwith strong intensity(Int3) (Image 2f). The single case ofmetastatic,minimal deviation ECA tothe small intestine (Image 2 g) showed focal extent of PAX8 staining(S1) (Image 2 h) with mild (Int1) intensity and moderate to diffuse(S2-3) extent of staining with strong (Int3) IMP3 intensity (image 2i).In contrast, the single case of poorly differentiated ECA with signet ring
Table 3Cohort of malignant endocervical lesions and their immunoreactivity for PAX8 andIMP3.
Endocervical lesion (no. of cases) PAX8 IMP3
EC AIS (8) S2/Int1 S2/Int2ECA (11)a S1/Int1 S2-3/Int3Metastatic MDA (1) S1/Int1 S2-3/Int3PD ECA with signet ring features S2-3/Int2-3 S1/Int1
EC AIS, adenocarcinoma in situ; ECA, invasive adenocarcinoma, including minimal devia-tion adenocarcinoma (MDA); PDECA, poorly differentiated endocervical adenocarcinoma.
a Two cases of ECA showing diffuse PAX8 immunoreactivity.
385R. Danialan et al. / Gynecologic Oncology 130 (2013) 383–388
features (image 2j) showed moderate to diffuse (S2-3) immunoreactiv-ity with moderate to strong (Int2-3) intensity of staining for PAX8(image 2 k), while IMP3 (image 2 l) showed focal (S1) positivity withmild (Int1) intensity of staining.
The Ki-67 showed diffuse expression (S3) with moderate intensityof staining (Int2) in ECA. P16 was also diffuse (S3) and moderate inintensity (Int2). CEA showed moderate extent of staining (S2) butwith strong intensity (Int3) in all cases of ECA.
Discussion
The distinction between benign andmalignant endocervical glandu-lar lesions can be histologically challenging. The overdiagnosis of a
benign endocervical glandular proliferation as “potentially malignant”can lead to unnecessary surgery with implications for child-bearing inyoung women, while the underdiagnosis of ECA as a benign prolifera-tion has potentially grave clinical implications. The vast majority ofbenign endocervical glandular proliferations can be accurately catego-rized and the histology of TC, ELH, MGH, TEM, MNH and CEM are welldescribed in the literature [3,19,20]. However, as is borne out in the lit-erature [2], careful use of histologic features alone cannot reliably dis-tinguish benign and malignant endocervical glandular lesions/tumorsin all cases. An IHC panel comprised of CEA, p16 and Ki67 has been sug-gested [1,2] to aid in this differential diagnosis. However, p16 IHC ex-pression can be seen in some benign lesions such as TEM [4]. Ki-67expression can be variable in endocervical glandular lesions, thereby
386 R. Danialan et al. / Gynecologic Oncology 130 (2013) 383–388
limiting its utility [17]. CEA is occasionally expressed in TEM [18] andp53 positivity can be observed in MGH [1]. Hence, PAX8 and IMP3 IHCstainswere evaluated in this study to evaluate a set of lesions in a prob-lematic area in gynecologic pathology.
When applied to a cohort of benign endocervical glandular lesions,PAX8 was largely diffusely positive with at least a moderate intensityof staining (S3/Int2). In contrast EC AIS cases showed less PAX8 ex-pression (S2/Int1), both in extent and intensity. The ECAs showedonly focal and mild PAX8 staining (S1/Int1). Of note, there weretwo cases of ECA (2/12, or 16.7%) that did not conform to this stainingpattern and instead showed diffuse PAX8 staining with moderate in-tensity. One such case was from a LEEP specimen, the other from ahysterectomy. Both cases were well-differentiated cervical adenocar-cinomas without mucinous or poorly differentiated components andlacked significant cellular pleomorphism or necrosis. Their originfrom the cervix was confirmed by their location, positive IHC stainingfor p16 and CEA, and negative staining for ER. Our results show thatPAX8 is progressively lost in a majority (10/12, 83%) of ECA whencompared with benign endocervical glandular lesions (Figs. 1 and 2).
There is very limited literature evaluating PAX8 IHC expression incervical lesions [5,8–10]. Laury et al. [8] reported moderate to strongPAX8 expression in 5 of 5 EC AIS and in 1 of 2 ECA. Tacha et al. [9]reported 5 of 6 adenocarcinomas of the endocervix showing positivityfor PAX8. However, they considered >5% positivity and/or focalstrong immunoreactivity to be positive. Woodard et al. [10] showedPAX8 to be positive in 12 of 19 typical ECA. Our results are in agree-ment with those of Tong et al. [5], who showed absent PAX8 immuno-reactivity in 5 cases of ECA (0/5, 0%) (Table 4).
A casual review of literature suggests PAX8 expression to be com-monly expressed in EC AIS and ECA. However, these results are some-what confounded by a small number of cases [8] and using >5% ofimmunoreactivity for positivity [9], a low cutoff for interpreting aPAX8 stain as “positive.” Woodard et al. [10] used an H scoring meth-od to quantify the extent of immunoreactivity, which was scored be-tween 0 (absent) to 300 (diffusely positive). Cases with an H scoregreater than >10 were considered positive. It is our view that an Hscore of 10 or above is too low of a cutoff to interpret any IHC stainas “positive.” According to the H scoring system, 10% of cells stainingfocally (S1 or Int1), 5% of cells staining moderately (S2 or Int2), andslightly greater than 3% of cells staining strongly (S3 or Int3) would
receive the identical H score of 10. The scoring method we used inthis study allowed for easier interpretation of staining extent and inten-sity using a 4-tier system (absent, focal, moderate, diffuse). As such, thelow cutoff employed by Woodard et al. [10] may account for the highpercentage of PAX8 “positive” ECA cases, which would otherwise havebeen interpreted as “focal” utilizing our scoring system.
In the benign endocervical glandular lesions, IMP3 is only focallyexpressed. In comparison, as endocervical glandular epithelium tran-sitions to dysplasia and frank invasive carcinoma, IMP3 expression isgradually increased. This is opposite to the pattern of expressionnoted with regard to PAX8. According to recent literature [4,11–16],IMP3 expression appears to be a marker of carcinogenesis and inthe set of cases we evaluated it does appear that when a significant pro-portion of lesional cells (>50%) show strong IMP3 expression, there is agreater likelihood of the lesion being neoplastic. This is concordant withthe literature that suggests IMP3 to be a marker of carcinogenesis[4,14–16].
We have also observed that a single case from our archives of met-astatic MDA shows nearly identical PAX8 and IMP3 immunoreactivityas the primary ECAs. This finding highlights the potential utility ofPAX8 in the metastatic setting. The single case of poorly differentiatedendocervical adenocarcinoma with signet ring features, however,showed staining which was opposite to that of the conventional ECAs inour study. This suggests that poorly differentiated forms of endocervicaladenocarcinomamay have a different underlying biology with preservedPAX8 protein expression and focal IMP3 expression.
The expression of Ki-67, p16, and CEA in our benign andpremalignant/malignant cohorts was as described in literature [2,3].Generally, the benign cohort showed absent to focal Ki-67 expression,focal p16 staining with a somewhat variable intensity of staining, andno CEA expression. While the IHC expression of Ki-67, p16, and CEAdid not vary significantly between EC AIS and ECA, there was appre-ciably increased staining of all three markers in the premalignant/malignant group.
Our particular series of cases is the largest to date reported of ECAand EC AIS in which the PAX8 IHC results have been scored for extentand intensity of immunoreactivity. Additionally, we have a cohort ofbenign and proliferative endocervical glandular lesions for compari-son. It appears that the combination of PAX8 and IMP3 IHC stains ispotentially useful additions to a panel that might include Ki-67, p16,
a b c
d e f
g h i
j k l
Fig. 2. PAX8 and IMP3 IHC in premalignant andmalignant endocervical lesions. Cohort of premalignant andmalignant cervical lesions. (a) Endocervical adenocarcinoma in situ (EC AIS), H&E;(b) EC AIS, moderate PAX8 immunoreactivity; (c) EC AIS, moderate IMP3 immunoreactivity; (d) endocervical adenocarcinoma (ECA), H&E; (e) ECA, focal to near absent PAX8 immunoreactiv-ity; (f) ECA, diffuse IMP3 immunoreactivity; (g) minimal deviation adenocarcinoma (MDA), metastatic to small intestine, H&E; (h) metastatic MDA, focal to absent PAX8 immunoreactivity;(i) metastatic MDA, diffuse IMP3 immunoreactivity; (j) poorly differentiated (PD) ECA, H&E; (k) PD ECA, strong PAX8 immunoreactivity; (l) PD ECA, focal and faint IMP3 immunoreactivity.
387R. Danialan et al. / Gynecologic Oncology 130 (2013) 383–388
and CEA in challenging endocervical glandular lesions. However, asnoted in our study, there is a degree of overlap of staining in individ-ual cases. As such, the results of PAX8 and IMP3 IHC should always be
Table 4Studies of PAX8 expression in EC AIS and ECA.
Authors (year of publication) Numberof cases
Lesionof study
Proportion of cases withpositive PAX8 staining
Laury et al. (2011) 5 EC AIS 100% (5 of 5)Laury et al. (2011) 2 ECA 50% (1 of 2)Tong et al. (2011) 5 ECA 0% (0 of 5)Tacha et al. (2011) 6 ECA 83% (5 of 6)Woodard et al. (2011) 19 ECA 63% (12 of 19)
EC AIS, adenocarcinoma in situ; ECA, invasive adenocarcinoma.
interpreted with caution and in the context of the clinical picture,histomorphology, and other ancillary studies. While our results sug-gest a clinical utility of PAX8 and IMP3, a larger cohort of cases ofECA (both conventional and subtypes) would be helpful in assessingthe role of this panel in diagnostic practice.
Conflict of interest statement
We wish to confirm that there are no known conflicts of interest associated with thispublication and that no significant financial support for this work has influenced itsoutcome.
Acknowledgements
The authors acknowledge the work of laboratory staff in theImmunopathology Laboratory at Clinical Laboratory Partners.
388 R. Danialan et al. / Gynecologic Oncology 130 (2013) 383–388
References
[1] Cina SJ, Richardson MS, Austin RM, Kurman RJ. Immunohistochemical immunore-activity for Ki-67 antigen, carcinoembryonic antigen, and p53 in the differentialdiagnosis of glandular lesions of the cervix. Mod Pathol 1997;10(3):176–80.
[2] Liang J, Mittal KR, Wei JJ, Yee H, Chiriboga L, Shukla P, et al. Utility of p16INK4a, CEA,Ki67, P53 and ER/PR in the differential diagnosis of benign, premalignant, andmalignant glandular lesions of the uterine cervix and their relationship withSilverberg scoring system for endocervical glandular lesions. IJGP 2006;26:71–5.
[3] Zaino RJ. Glandular lesions of the uterine cervix. Mod Pathol 2000;13(3):261–74.[4] Li C, Rock KL, Woda BA, Jiang Z, Fraire AE, Dresser K, et al. IMP3 is a novel biomarker
for adenocarcinoma in situ of the uterine cervix: an immunohistochemical study incomparison with p16INK4a expression. Mod Pathol 2007;20:242–7.
[5] Tong GX, Devaraj K, Hamele-Bena D, Yu WM, Turk A, Chen X, et al. PAX8: a markerfor carcinoma of Müllerian origin in serous effusions. Diagn Cytopathol 2011;39(8):567–74.
[6] Ozcan A, Shen SS, Hamilton C, Anjana K, Coffey D, Krishnan B, et al. PAX8 expres-sion in non-neoplastic tissues, primary tumors, and metastatic tumors: a compre-hensive immunohistochemical study. Mod Pathol 2011;24(6):751–64.
[7] Ordonez Nelson G. Value of PAX8 immunoreactivity in tumor diagnosis: a reviewand update. Adv Anat Pathol 2012;19(3):140–51.
[8] Laury AR, Perets R, Piao H, Krane JF, Barletta JA, French C, et al. A comprehensiveanalysis of PAX8 expression in human epithelial tumors. Am J Surg Pathol2011;35(6):816–26.
[9] Tacha D, Ding Z, Liang C. Expression of PAX8 in normal and neoplastic tissues:a comprehensive immunohistochemical study. Appl Immunohistochem MolMorphol 2011;19(4):293–9.
[10] Woodard AH, Yu J, Dabbs DJ, Beriwal S, Florea AV, Elishaev E, et al. NY-BR-1 andPAX8 immunoreactivity in breast, gynecologic tract, and other CK7+ carcinomas.Am J Clin Pathol 2011;136:428–35.
[11] Yantiss RK, Cosar E, Fischer AH. Use of IMP3 in identification of carcinoma in fineneedle aspiration biopsies of pancreas. Acta Cytol 2008;52(2):133–8.
[12] Lu D, Yang X, Jiang NY, Woda BA, Liu Q, Dresser K, et al. IMP3, a new biomarker topredict progression of cervical intraepithelial neoplasia into invasive cancer. Am JSurg Pathol 2011;35(11):1638–45.
[13] Li D, Yan D, Tang H, Zhou C, Fan J, Li S, et al. IMP3 is a novel prognostic marker thatcorrelates with colon cancer progression and pathogenesis. Ann Surg Oncol2009;16(12):3499–506.
[14] Wachter DL, Kristiansen G, Soll C, Hellerbrand C, Breuhahn K, Fritzsche F, et al.Insulin-like growth factor II mRNA-binding protein 3 (IMP3) expression inhepatocellular carcinoma. A clinicopathological analysis with emphasis on diagnos-tic value. Histopathology 2012;60(2):278–86.
[15] Clauditz TS, Wang CJ, Gontarewicz A, Blessmann M, Tennstedt P, Borgmann K,et al. Expression of insulin-like growth factor II mRNA-binding protein 3 in squa-mous cell carcinomas of the head and neck. J Oral Pathol Med 2012, doi:10.1111/j.1600-0714.2012.0117.
[16] Walter O, Prasad M, Lu S, Quinlan RM, Edmiston KL, Khan A. IMP3 is a novel bio-marker for triple negative invasive mammary carcinoma associated with a moreaggressive phenotype. Hum Pathol 2009;40(11):1528–33.
[17] Pirog EC, Isacson C, Szabolcs MJ, Kleter B, Quint W, Richart RM. Proliferative activ-ity of benign and neoplastic endocervical epithelium and correlation with HPVDNA detection. Int J Gynecol Pathol 2002;21(1):22–6.
[18] Marques T, Andrade LA, Vassallo J. Endocervical tubal metaplasia and adeno-carcinoma in situ: role of immunohistochemistry for carcinoembryonic anti-gen and vimentin in differential diagnosis. Histopathology 1996;28(6):549–50.
