ORIGINAL ARTICLE

PARAFIBROMIN IMMUNOHISTOCHEMICAL STAINING TODIFFERENTIATE PARATHYROID CARCINOMA FROMPARATHYROID ADENOMA

Hee Kyung Kim, MD,1 Young Lyun Oh, MD, PhD,2 Seok-Hyung Kim, MD, PhD,2

Dong Youn Lee, MD, PhD,3 Ho-Cheol Kang, MD, PhD,4 Ji In Lee, MD,1 Hye Won Jang, MD,1

Kyu Yeon Hur, MD, PhD,1 Jae Hyeon Kim, MD, PhD,1 Yong Ki Min, MD, PhD,1

Jae Hoon Chung, MD, PhD,1 Sun Wook Kim, MD, PhD1

1Division of Endocrinology and Metabolism, Department of Medicine, Samsung Medical Center,Sungkyunkwan University School of Medicine. E-mail: swkimmd@skku.edu

2Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine3Department of Dermatology, Samsung Medical Center, Sungkyunkwan University School of Medicine4Department of Internal Medicine, Chonnam University Hwasun Hospital, Chonnam University Medical School

Accepted 11 November 2010Published online 29 June 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/hed.21716

Abstract: Background. Parafibromin is a protein encoded

by the HRPT2 oncosuppressor gene, and the expression is

reported to be decreased or absent in parathyroid

carcinomas.

Methods. A total of 26 tumor specimens from 18 patients

with adenoma and 8 patients with carcinoma were immune-

stained with an antibody against parafibromin.

Results. Parafibromin immunostaining showed strong posi-

tivity in 17 of 18 adenomas. Negative staining was noted in 3

of 8 carcinomas, and weak positivity was found in 3 of 8 carci-

nomas. The remaining 2 cases of carcinoma showed strong

positivity. The loss of parafibromin expression (negative or

weak positivity) demonstrated 94.4% specificity in the diagno-

sis of parathyroid carcinomas. Relapses or distant metastases

of carcinoma occurred only in cases in which there was a loss

of parafibromin immunostaining.

Conclusions. Loss of parafibromin immunostating showed

promising results in the differential diagnosis of parathyroid

carcinoma from adenoma and may also serve as a prognostic

marker. VVC 2011 Wiley Periodicals, Inc. Head Neck 34: 201–

206, 2012

Keywords: parathyroid carcinoma; parathyroid adenoma;

parafibromin; HRPT2; prognosis

Parathyroid carcinoma is a rare cause of primary hy-perparathyroidism, accounting for less than 1% of pri-mary hyperparathyroidism.1 The diagnosis ofparathyroid malignancy is based on histologic evi-dence (vascular invasion, perineural space invasionand growth into adjacent tissues) or clinical evidenceof recurrence and distant metastases. However, othermicroscopic features of malignancy such as trabeculargrowth pattern, increased mitotic activity, thick fi-

brous bands, and capsular invasion were also foundin atypical parathyroid adenoma and therefore arenot considered to be pathognomonic for parathyroidcarcinomas.2 These histopathologic uncertainties of-ten result in underrecognition and suboptimal ther-apy for parathyroid carcinomas. There is therefore aneed for an adjunctive marker to assist in the diagno-sis of parathyroid carcinomas. Several immunohisto-chemical markers such as Ki-67, galectin-3,retinoblastoma protein, p27Kip1, bcl-2, cyclin D1,mdm-2, and CaSR have been suggested for the differ-ential diagnosis of parathyroid tumors, but studieshave not demonstrated enough diagnostic value forits routine use in clinical practice.3–9

The hereditary hyperparathyroidism type 2(HRPT2) tumor suppressor gene encodes a ubiqui-tously expressed 531–amino acid protein, which isnamed as parafibromin. A germline mutation in theHRPT2 gene is responsible for hereditary hyperpara-thyroidism–jaw tumor syndrome, a rare autosomaldominant cause of parathyroid tumors with anincreased risk of parathyroid cancer, ossifying fibro-mas of the mandible or maxilla and various cysticand neoplastic renal abnormalities.10,11 A somaticmutation in the HRPT2 gene is also found in 67% to100% of sporadic parathyroid carcinomas.12–14 How-ever, the prevalence of HRPT2 mutation in sporadicparathyroid adenomas is rare, ranging from 0% to 4%of cases.14,15 Accordingly, the most parathyroid carci-nomas in western countries are characterized by inac-tivating mutations of the HRPT2 gene that lead to aloss of parafibromin expression. Thus studies haveused parafibromin immunostaining for the differen-tial diagnosis of parathyroid carcinoma, instead ofHRPT2 gene analysis in parathyroid tumor speci-mens, because of the lower costs to patients and

Correspondence to: S. W. Kim

VVC 2011 Wiley Periodicals, Inc.

Parafibromin Immunostaining for Diagnosis of Parathyroid Carcinoma HEAD & NECK—DOI 10.1002/hed February 2012 201

greater accessibility for physicians.16–18 However,there are no reports on the expression pattern of par-afibromin in parathyroid neoplasm of Asians. Theaim of this study was to evaluate the diagnostic valueof parafibromin immunostaining in the diagnosis ofparathyroid carcinomas in Asians, particularlyKoreans.

MATERIALS AND METHODS

Patients and tissue samples. We obtained forma-lin-fixed paraffin-embedded tissue blocks from 26patients who underwent parathyroid surgery betweenNovember 1994 and May 2009, either at SamsungMedical Center (Seoul, Korea) or Chonnam NationalUniversity Hospital (Gwangju, Korea). Eighteen para-thyroid adenomas were randomly selected. Eight tis-sue blocks from sporadic parathyroid carcinomas (6from primary parathyroid carcinomas and 2 fromlocally relapsed tissues) were the only cases diag-nosed as parathyroid carcinomas during the study pe-riod at 2 tertiary referral hospitals. Theclinicopathologic features of cases with parathyroidcarcinomas are listed in Table 1. No patient had afamily history of primary hyperparathyroidism or hy-perparathyroidism–jaw tumor syndrome. This studywas reviewed and approved by the Samsung SeoulHospital Institutional Review Board.

Pathologic diagnosis. All pathologic specimenswere reviewed by 2 pathologists (Y.L. Oh and S. Kim)to assess malignancy of parathyroid tumors on thebasis of the recent World Health Organizationclassification.19

Immunohistochemistry. Paraffin sections of para-thyroid tumors were deparaffinized in xylene, rehy-drated in alcohol, and immersed in preheated Tris-ethylenediamine tetraacetic acid buffer pH 9.0 at97�C for 20 minutes. The sections were incubated for2 hours with primary mouse monoclonal antiparafi-bromin antibody (Clone 2H1, SC-33638; Santa CruzBiotechnology, Santa Cruz, CA) at a dilution of 1:50.

This commercially available antibody was designed totarget amino acid positions 87-100 of parafibromin.

The sections were incubated with anti-mouseEnvison PO (Dako, Carpinteria, CA) for 40 minutes,diaminobenzidine tetrahydrochloride for 5 minutesand counterstained in hemtoxylin for 30 seconds.Finally, the sections were dehydrated and mounted.

The staining pattern of each specimen was classi-fied into 3 categories following to Gill et al.18 Diffusenuclear staining greater than 95% with strong inten-sity was considered ‘‘diffusely strong positive.’’ ‘‘Nega-tive staining’’ was defined as the absence of nuclearstaining in all (>99%) of the tumor tissue. ‘‘Weakstaining’’ was indicated by all other staining patterns.Negative and weak positive parafibromin stainingwere considered to represent loss of parafibrominexpression. Tumors were classified as negative only ifthere were convincing internal positive controlsdetected throughout the slide and adjacent areas ofnegative staining. All parafibromin-stained slideswere independently reviewed by 2 pathologists with-out blinding (Y.L. Oh and S. Kim).

Measurement of Intact Parathyroid Hormone. In-tact parathyroid hormone (iPTH) was measured byimmunoradiometric assay for the quantitative deter-mination of the intact biological chain of 84 aminoacids of parathyroid hormone (CIS Bio International,Paris, France). The measurement range was between0.7 and 1500 pg/mL, and the normal range wasbetween 10 and 65 pg/mL.

Statistical Analysis. Statistical computations wereperformed with the PASW 17.0 software (SPSS, Inc.,Chicago, IL). The iPTH levels between parathyroidcarcinoma and adenoma were compared by use of aStudent t test. We performed Fisher’s exact test toassess the relationship between parafibromin immu-nostaining patterns and prarthyroid carcinomas.Diagnostic values, including sensitivity, specificity,positive predictive value, negative predictive value,and 95% confidence interval (CI) for proportions,were calculated by use of standards methods for

Table 1. Clinicopathologic data and evaluation of immunohistochemical staining for carcinoma specimens.

Case

Age at

surgery, y/sex

Tumor

location

Parafibromin

IHC

Preoperative

iPTH (pg/ml)

Presence of

metastasis at

initial surgery

Follow-/up

duration, mo

Time to

recurrence, mo Clinical progression

1 51/M Primary Strong positive 2784 — 26 No relapse

2 46/M Primary Strong positive 1151.2 — 8 No relapse

3 62/F Primary Negative 301 — 14 Unknown

4 12/F Primary Weak positive 556.4 — 106 No relapse

5 30/M Metastasis, local Negative 772 — 83 21 Local and distant meta

6 70/M Primary Weak positive 983.9 — 25 19 Local relapse

7 48/F Primary Negative 123.7 — 118 109 Local relapse

8 73/F Metastasis, local Weak positive 810.1 — 53 19 Local relapse Expired

Abbreviations: M, Male; F, female.

202 Parafibromin Immunostaining for Diagnosis of Parathyroid Carcinoma HEAD & NECK—DOI 10.1002/hed February 2012

binominal distribution. The p value of <.05 was con-sidered statistically significant.

RESULTS

Patient characteristics. The mean ages of patientswith parathyroid adenoma and parathyroid carci-noma were 50.6 � 13.7 and 49.0 � 20.5 years, respec-tively. These cases included 11 women (61%) of 18cases of parathyroid adenomas and 4 women (50%) of8 cases of parathyroid carcinomas. The mean iPTHsof the patients with parathyroid carcinomas werehigher than that of patients with parathyroid ade-noma, but there were no statistically significant dif-ferences between the groups (513.9 � 684.9 pg/mL[range, 78–2301] vs 935.3 � 820.6 pg/mL (range,123.7–2784], p ¼ .232).

Parafibromin Immunohistochemistry. Two patholo-gists reviewed parafibromin-immunostained slides in-dependently but the blinding was not performed.Seventeen (94.4%) of 18 parathyroid adenomas

showed diffuse strong nuclear parafibromin immuno-staining and only 1 adenoma showed negative stain-ing (Figure 1). In contrast, 6 (75%) of 8 parathyroidcarcinomas showed absolute or relative loss of parafi-bromin immunostaining (3 negative and 3 weak posi-tive results), and only 2 cases showed diffusely strongpositivity for parafibromin (Figure 2). Thus loss ofparafibromin immunostaining was significantly asso-ciated with parathyroid carcinomas (p < .001) (Table2). Furthermore, all recurred parathyroid carcinomasshowed the loss of parafibromin immunostaining(Table 1).

Diagnostic Values of Parafibromin Immunohistoche-

mistry. The sensitivity of ‘‘loss of parafibrominexpression’’ by immunohistochemistry (IHC) for diag-nosing parathyroid carcinoma was 75% (95% CI; 36%to 96%) and specificity was 94.4% (95% CI; 71% to100%). We presented estimated positive and negativepredictive values according to the reported prevalenceof parathyroid carcinoma among parathyroid neo-plasms in previous literatures (Table 3).1,20

FIGURE 1. Parathyroid adenoma. (A) Strong positive parafibromin staining. The majority of cells showed diffuse nuclear immunoreac-

tivity for parafibromin (original magnification, �400). (B) Negative parafibromin staining. The majority of cells were not stained by par-

afibromin (original magnification, �200). The arrows indicate the internal positive staining of endothelial cells. [Color figure can be

viewed in the online issue, which is available at wileyonlinelibrary.com.]

FIGURE 2. Parathyroid carcinoma. (A) Strong positive parafibromin staining (original magnification, �400). (B) Weak positive parafi-

bromin staining. Less than 95% of cells were stained by parafibromin (original magnification, �400). (C) Negative parafibromin staining

(original magnification, �200). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Parafibromin Immunostaining for Diagnosis of Parathyroid Carcinoma HEAD & NECK—DOI 10.1002/hed February 2012 203

DISCUSSION

Our data support the usefulness of parafibromin im-munostaining in the differential diagnosis of parathy-roid carcinoma in Asians, just as it is in subjects ofwestern population, and it may serve as a reliablemarker for poor prognosis of parathyroid carcinomas.Parafibromin is a member of the polymerase-associ-ated factor 1 complex associated with RNA polymer-ase II, which regulates transcription elongation,histone modification, and cell proliferation.11 Parafi-bromin overexpression is reported to inhibit colonyformation, anchorage-dependent cell growth, and cel-lular proliferation and to induce cell cycle arrest inthe G1 phase.21 These findings suggest the potentialroles of parafibromin as a tumor suppressor in carci-nogenesis. Wild-type parafibromin immunostaining islocalized to the nucleus, whereas deletions or muta-tions of amino acid residues result in the loss of nu-clear expression.22

The use of parafibromin immunostaining in thedifferential diagnosis of parathyroid tumors was firstintroduced by Tan et al16 who reported a diagnosticvalue of 96% for sensitivity and 99% for specificity,which are similar or superior to the values that weobtained in this study (75% for sensitivity, 94.4 % forspecificity). This study also agrees with previousreports that almost all sporadic adenomas of theparathyroid show strong positive parafibromin stain-ing, which renders a high specificity of the loss of par-afibromin expression for the diagnosis of parathyroidcarcinomas.

Complete en bloc resection of all tumors at thetime of initial surgery represents the best opportunityfor cure.23–25 However, the diagnosis of parathyroidcarcinoma is often made retrospectively, after tumorrecurrence or metastasis, as in the cases of the 8patients in Table 1 who might have undergone subop-timal therapy and surveillance before the diagnosis ofparathyroid carcinoma via local recurrence. If theintraoperative assessment of parafibromin immunore-activity by ultrarapid immunostaining had been per-formed, it would have been helpful for determiningthe surgical plan and follow-up plan to prevent dis-ease recurrence.

Loss of parafibromin expression (negative orweakly positive immunostaining) also appears prom-ising as a possible prognostic marker of parathyroidcarcinomas. Although there were only 8 cases withparathyroid carcinomas in our study subjects, all 4cases with recurrence showed loss of parafibrominstaining. Similarly, Juhlin et al15 reported that 83.3%(10/12) of recurrences from 22 parathyroid carcinomashad negative staining results for parafibromin. Like-wise, parafibromin in breast, gastric, and colon cancerappears to play a role as a tumor suppressor, andtherefore loss of protein expression is reported to beassociated with adverse clinical and pathologic pa-rameters.26–28

There are a few major limitations of this study.The first is the small number of cases of parathyroidcarcinoma in our sample. Only 8 cases of parathyroidcarcinoma were treated over the course of 15 years at2 large tertiary referral hospitals in Korea, whichreflects the rarity of parathyroid carcinomas. How-ever, the negative staining pattern of parafibromin inour study was observed in the presence of positive in-ternal and external controls. Furthermore, the preva-lence (75%) of negative immunostaining forparafibromin was similar to that in previously pub-lished reports, which ranged from 68% to 100% inparathyroid carcinoma.15,18,29 The second limitation isthat we could not investigate the mutation status ofthe HRPT2 gene in our study subjects because of thepoor quality of DNA available for mutation analysisfrom formalin-fixed paraffin-embedded samples,which might be caused by preservation and fixationof the samples over time. Positive staining for parafi-bromin can be found in parathyroid carcinomas,regardless of HRPT2 mutation status, not onlybecause missense mutations represent up to 10% ofall reported HRPT2 mutations, but also because up to30% of parathyroid carcinomas may not harbor anydocumented HRPT2 mutations.15,21,30 These findingsreduce the diagnostic value of parafibromin immuno-staining for the diagnosis of parathyroid carcinomaby reducing the sensitivity of the test. However,

Table 2. Summary of immunohistochemical parafibromin staining in

parathyroid tumors.

Parathyroid tumors

Total

patients

number

Parafibromin IHC

Strongly

positive

Loss of parafibromin

expression

Weakly

positive Negative

Sporadic carcinoma 8 2 3 3

Sporadic adenoma 18 17 0 1

Abbreviation: IHC, immunohistochemistry.Note: The p < .001 by Fisher’s exact test. There is a significant relationshipbetween parafibromin IHC and parathyroid tumor classification. In this statisticalanalysis, parafibromin IHC were classified into 2 categories (‘‘strong positive’’ and‘‘loss of parafibromin expression’’).

Table 3. Diagnostic value (%) of parafibromin immunostaining

(loss of parafibromin expression) for differentiating parathyroid

carcinomas from adenomas.

Loss of

parafibromin

expression

Sensitivity (95% CI) 75.0 (36–96)

Specificity (95% CI) 94.4 (71–100)

Positive predictive value at 1% prevalence (95% CI)* 12 (7–19)

Negative predictive value at 1% prevalence (95% CI)* 99.7 (99–100)

Positive predictive value at 5% prevalence (95% CI)* 41.5 (38–45)

Negative predictive value at 5% prevalence (95% CI)* 98.6 (98–99)

*Supposed prevalence of parathyroid carcinomas among parathyroid tumorsaccording to references 1, 2, 23, and 25 to estimate positive and negative predic-tive value of ‘‘loss of parafibromin expression’’ in the differential diagnosis of para-thyroid carcinomas from adenomas.

204 Parafibromin Immunostaining for Diagnosis of Parathyroid Carcinoma HEAD & NECK—DOI 10.1002/hed February 2012

although the mutation analysis of HRPT2 is expectedto slightly increase sensitivity for diagnosing parathy-roid carcinoma over parafibromin immunostaining,there is solid evidence that the concordance betweenHRPT2 mutations and loss of parafibromin expres-sion is high.29 Furthermore, immunostaining has itsvaluable merits over mutation analysis of HRPT2 inthat is costs much less for the patients and givesmore accessibility to the physicians.

Finally, for the interpretations of immunostaining,pathologists could not be blinded to the status ofpatients’ disease because they participated in theselection of parathyroid tumors for this study. Asstudied by Schulz et al,31,32 the investigator’s expecta-tions can bias the results by knowing group assign-ment, either by opinions about the intervention beingstudies, by knowing the context in which the datawere collected, or by increased vigilance in examiningthe data. Consequently, studies with inadequateblinding have been associated with larger interven-tion effects up to 41% compared with studies in whichauthors reported adequate allocation conceal-ment.31,32 This happens especially when the databeing collected are based on the judgment of theinvestigators. In this context, we should be cautiousin the interpretation of this study results. Althoughwe hope that independent reading by 2 pathologistswith a high level of expertise on endocrine tumorsand immunostaining made this problem less signifi-cant, the bias from not blinding the pathologists stillexists. Thus further studies with a larger dataset andadequate study design are needed. Recently, Howellet al30 introduced protein gene product 9.5 (PGP9.5)encoded by ubiquitin carboxyl-terminal esterase L1 asa new complementary immunostain-marker to parafi-bromin for the diagnosis of parathyroid carcinoma,because positive staining for PGP9.5 showed slightlysuperior sensitivity and similar specificity to that ofparafibromin. Further investigation is needed to com-bine these 2 valuable markers for immunostaining forthe diagnosis of pararthyroid carcinomas.

In conclusion, loss of parafibromin immunostain-ing showed promising results for the differential diag-nosis of parathyroid carcinomas from adenomas inAsians and also may serve as a prognostic marker inpatients with parathyroid carcinomas. Further stud-ies with a larger dataset are required to validate itsdiagnostic value.

REFERENCES

1. Fraser WD. Hyperparathyroidism. Lancet 2009;374(9684):145–158.

2. Shane E. Clinical review 122: Parathyroid carcinoma. J ClinEndocrinol Metab 2001;86:485–493.

3. Abbona GC, Papotti M, Gasparri G, Bussolati G. Proliferativeactivity in parathyroid tumors as detected by Ki-67 immuno-staining. Hum Pathol 1995;26:135–138.

4. Cryns VL, Thor A, Xu HJ, et al. Loss of the retinoblastoma tu-mor-suppressor gene in parathyroid carcinoma. N Engl J Med1994;330:757–761.

5. Erickson LA, Jin L, Wollan P, Thompson GB, van Heerden JA,Lloyd RV. Parathyroid hyperplasia, adenomas, and carcinomas:differential expression of p27Kip1 protein. Am J Surg Pathol1999;23:288–295.

6. Farnebo F, Auer G, Farnebo LO, et al. Evaluation of retino-blastoma and Ki-67 immunostaining as diagnostic markers ofbenign and malignant parathyroid disease. World J Surg1999;23:68–74.

7. Haven CJ, van Puijenbroek M, Karperien M, Fleuren GJ, Mor-reau H. Differential expression of the calcium sensing receptorand combined loss of chromosomes 1q and 11q in parathyroidcarcinoma. J Pathol 2004;202:86–94.

8. Vargas MP, Vargas HI, Kleiner DE, Merino MJ. The role of prog-nostic markers (MiB-1, RB, and bcl-2) in the diagnosis of para-thyroid tumors. Mod Pathol 1997;10:12–17.

9. Vasef MA, Brynes RK, Sturm M, Bromley C, Robinson RA.Expression of cyclin D1 in parathyroid carcinomas, adenomas,and hyperplasias: a paraffin immunohistochemical study. ModPathol 1999;12:412–416.

10. Szabo J, Heath B, Hill VM, et al. Hereditary hyperparathyroid-ism-jaw tumor syndrome: the endocrine tumor gene HRPT2maps to chromosome 1q21-q31. Am J Hum Genet 1995;56:944–950.

11. Carpten JD, Robbins CM, Villablanca A, et al. HRPT2, encodingparafibromin, is mutated in hyperparathyroidism-jaw tumor syn-drome. Nat Genet 2002;32:676–680.

12. Shattuck TM, Valimaki S, Obara T, et al. Somatic and germ-linemutations of the HRPT2 gene in sporadic parathyroid carci-noma. N Engl J Med 2003;349:1722–1729.

13. Howell VM, Haven CJ, Kahnoski K, et al. HRPT2 mutations areassociated with malignancy in sporadic parathyroid tumours. JMed Genet 2003;40:657–663.

14. Krebs LJ, Shattuck TM, Arnold A. HRPT2 mutational analysisof typical sporadic parathyroid adenomas. J Clin EndocrinolMetab 2005;90:5015–5017.

15. Juhlin CC, Villablanca A, Sandelin K, et al. Parafibromin immu-noreactivity: its use as an additional diagnostic marker for para-thyroid tumor classification. Endocr Relat Cancer 2007;14:501–512.

16. Tan MH, Morrison C, Wang P, et al. Loss of parafibromin immu-noreactivity is a distinguishing feature of parathyroid carci-noma. Clin Cancer Res 2004;10:6629–6637.

17. Cetani F, Pardi E, Ambrogini E, et al. Genetic analyses in fami-lial isolated hyperparathyroidism: implication for clinical assess-ment and surgical management. Clin Endocrinol (Oxf)2006;64:146–152.

18. Gill AJ, Clarkson A, Gimm O, et al. Loss of nuclear expressionof parafibromin distinguishes parathyroid carcinomas and hyper-parathyroidism-jaw tumor (HPT-JT) syndrome-related adenomasfrom sporadic parathyroid adenomas and hyperplasias. Am JSurg Pathol 2006;30:1140–1149.

19. Bondenson L, Grimelius L, DeLellis R, et al. Pathology andgenetics of tumours of endocrine organs. Lyon: IARC Press;2004.

20. Obara T, Okamoto T, Kanbe M, Iihara M. Functioning parathy-roid carcinoma: clinicopathologic features and rational treat-ment. Semin Surg Oncol 1997;13:134–141.

21. Zhang C, Kong D, Tan MH, et al. Parafibromin inhibits cancercell growth and causes G1 phase arrest. Biochem Biophys ResCommun 2006;350:17–24.

22. Bradley KJ, Bowl MR, Williams SE, et al. Parafibromin is a nu-clear protein with a functional monopartite nuclear localizationsignal. Oncogene 2007;26:1213–1221.

23. Rodgers SE, Perrier ND. Parathyroid carcinoma. Curr OpinOncol 2006;18:16–22.

24. Hundahl SA, Fleming ID, Fremgen AM, Menck HR. Two hun-dred eighty-six cases of parathyroid carcinoma treated in theU.S. between 1985–1995: a National Cancer Data BaseReport. The American College of Surgeons Commission onCancer and the American Cancer Society. Cancer 1999;86:538–544.

25. DeLellis RA. Parathyroid carcinoma: an overview. Adv AnatPathol 2005;12:53–61.

26. Selvarajan S, Sii LH, Lee A, et al. Parafibromin expression inbreast cancer: a novel marker for prognostication? J Clin Pathol2008;61:64–67.

27. Zheng HC, Takahashi H, Li XH, et al. Downregulated parafibro-min expression is a promising marker for pathogenesis, inva-sion, metastasis and prognosis of gastric carcinomas. VirchowsArch 2008;452:147–155.

Parafibromin Immunostaining for Diagnosis of Parathyroid Carcinoma HEAD & NECK—DOI 10.1002/hed February 2012 205

28. Zhao J, Yart A, Frigerio S, et al. Sporadic human renal tumorsdisplay frequent allelic imbalances and novel mutations of theHRPT2 gene. Oncogene 2007;26:3440–3449.

29. Cetani F, Ambrogini E, Viacava P, et al. Should parafibromin stainingreplace HRTP2 gene analysis as an additional tool for histologic diag-nosis of parathyroid carcinoma? Eur J Endocrinol 2007;156:547–554.

30. Howell VM, Gill A, Clarkson A, et al. Accuracy of combinedprotein gene product 9.5 and parafibromin markers for immuno-

histochemical diagnosis of parathyroid carcinoma. J Clin En-doc-rinol Metab 2009;94:434–441.

31. Schulz KF. Subverting randomization in controlled trials. JAMA1995;274:1456–1458.

32. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evi-dence of bias. Dimensions of methodological quality associatedwith estimates of treatment effects in controlled trials. JAMA1995;273:408–412.

206 Parafibromin Immunostaining for Diagnosis of Parathyroid Carcinoma HEAD & NECK—DOI 10.1002/hed February 2012