CA125 and thyroglobulin staining in papillary carcinomas ofthyroid and ovarian origin is not completely specific for site oforigin

C E Keen, S Szakacs, E Okon, J S Rubin* & B M Bryant†Departments of Histopathology and *ENT Surgery, University Hospital Lewisham, London and †Department of MedicalOncology, Queen Elizabeth Hospital, Woolwich, London, UK

Date of submission 3 April 1998Accepted for publication 6 October 1998

Keen C E, Szakacs S, Okon E, Rubin J S & Bryant B M

(1999) Histopathology 34, 113–117

CA125 and thyroglobulin staining in papillary carcinomas of thyroid and ovarian origin is notcompletely specific for site of origin

Aims: A 70-year-old woman presented with metastaticpsammoma body-rich papillary carcinoma in a supra-clavicular lymph node. No primary site was evident.The tumour showed strong staining for CA125 andweak staining for thyroglobulin. Prompted by this casewe aimed to assess the reliability of immunostaining forCA125 and thyroglobulin in making the distinctionbetween thyroid and ovarian papillary carcinoma.Methods and results: Nine papillary carcinomas of thethyroid and 17 serous papillary carcinomas of the ovarywere stained for CA125 and thyroglobulin, as well asCAM 5.2, LP 34, carcinoembryonic antigen (CEA),S100 and diastase/periodic acid–Schiff. Nine of ninethyroid carcinomas stained for thyroglobulin; in

addition CA125 was positive in four of nine. Normalsurrounding thyroid also showed some reaction. Seven-teen of 17 ovarian serous carcinomas were positive forCA125; in addition one case showed moderately strongstaining for thyroglobulin. Mucin stains were positive in14/17 ovarian serous carcinomas, but negative in allthyroid carcinomas. The other antibodies assessedshowed no useful differences in staining frequency.Conclusion: Many cases of papillary carcinoma of thethyroid show CA125 staining, and this feature thereforehas little positive predictive value for an ovarian origin.Occasional cases of ovarian papillary carcinoma mayshow staining for thyroglobulin, and this result shouldtherefore be interpreted cautiously.

Keywords: CA125, thyroglobulin, immunohistochemistry, D/PAS, papillary carcinoma, thyroid, ovary

Introduction

Papillary carcinoma of the thyroid gland and serouspapillary carcinoma of the ovary share some morpho-logical features, including the presence, in many cases,of psammoma bodies. In cases of metastatic papillarycarcinoma where the primary site is occult, it might beexpected that this could be resolved by the application ofimmunohistochemical markers against thyroglobulinand CA125. CA125 is a tumour marker usuallyassociated with ovarian epithelial malignancy, althoughrecognized to stain benign ovarian epithelial tumours,

and occasionally other tumours such as seminal vesiclecarcinoma1 and even anaplastic lymphoma.2

We were prompted to study the specificity of reactionfor CA125 and thyroglobulin in ovarian serous carcin-oma and papillary thyroid carcinoma after encounteringa case presenting with metastatic papillary carcinoma inlymph nodes in the left supraclavicular fossa, and noobvious primary tumour. In addition we applied a smallpanel of other commonly used antibodies and a mucinstain to assess possible diagnostic usefulness.

Materials and methods

C A S E D E TA I L S

A 70-year-old woman presented with metastatic

Histopathology 1999, 34, 113–117

q 1999 Blackwell Science Limited.

Address for correspondence: Dr C E Keen, Department ofHistopathology, Royal Devon and Exeter Hospital, Church Lane,Heavitree, Exeter EX2 5AD, UK.

114 C E Keen et al.

q 1999 Blackwell Science Ltd, Histopathology, 34, 113–117.

Table 1. Antibodies used in this study

Clone no./Antibody Type Antigen Supplier code no. Antigen unmasking Dilution

NCL-CA125 MC CA125 Novocastra Laboratories Ov185:1 200701 2 min P/C 1:100

Anti-thyroglobulin MC Thyroglobulin Dako Dak Tg6 M781 Nil 1:800

Anti-thyroglobulin PC Thyroglobulin Dako A 0251 Lot 096 Nil 1:500

CAM 5.2 MC Cytokeratins 8, 18 Becton-Dickinson CAM 5.2 12 min trypsin 0.1% 1:40

LP 34 MC High molecular weight cytokeratins Dako LP 34 M717 12 min trypsin 0.1% 1:60

Anti-CEA MC CEA Dako 11-7 M7072 12 min trypsin 0.1% 1:400

Anti-S100 PC S100 protein Dako Z 311 Lot 129 5 min trypsin 0.1% 1:8000

MC, Monoclonal; PC, polyclonal; CEA, carcinoembryonic antigen; P/C, pressure cooker.

Table 2. Immunostaining of thyroid and ovarian tumours

CA125 Thyroglobulin (monoclonal) CAM 5.2 LP 34 CEA S100 D/PAS

Thyroid1 ¹ þþþ þþþ þ þ þ ¹

2 þ þþ þþþ þ/¹ ¹ þþþ ¹

3 þþ þþ þþþ ¹ ¹ þ ¹

4 þ þþþ þþþ þ þ þþþ ¹

5 þþ þþ þþþ þ/¹ þ/¹ þ ¹

6 ¹ þþþ þþþ ¹ ¹ þ ¹

7 ¹ þþþ þþþ þþ þþ þþ ¹

8 ¹ þþ þþþ ¹ ¹ ¹ ¹

9 ¹ þþþ þþþ ¹ ¹ ¹ ¹

Ovary10 þþ ¹ þþþ þþþ ¹ ¹ þ/¹

11 þþ ¹ þþþ þ ¹ ¹ þ/¹

12 þþ ¹ þþþ þ ¹ þ þþ

13 þþ ¹ þþþ þþ ¹ þ þþ

14 þþþ ¹ þþþ þþ þþþ ¹ þþþ

15 þþ ¹ þþþ þþ ¹ ¹ ¹

16 þþþ ¹ þþþ ¹ ¹ ¹ þ/¹

17 þþþ þþ þþþ þþ ¹ ¹ þ

18 þþþ ¹ þþþ þþ ¹ þþ þ

19 þþþ ¹ þþþ þ/¹ ¹ ¹ þþ

20 þþþ ¹ þ ¹ ¹ ¹ ¹

21 þþþ ¹ þþþ ¹ ¹ ¹ ¹

22 þþþ ¹ þþþ ¹ ¹ ¹ þþþ

23 þ ¹ þþþ þ þþþ þ þþ

24 þ ¹ þþþ ¹ ¹ ¹ þ

25 þþþ ¹ þþ ¹ þ/¹ ¹ þ/¹

26 þþþ ¹ þþþ þ þ ¹ þ

papillary carcinoma in a left supraclavicular lymphnode, initially diagnosed on fine needle aspiration. Thechest X-ray was clear at presentation. Technetiumthyroid scan showed a cold area in the lower half of theleft lobe. Excision biopsy of a 50 mm diameter mass oflymph nodes confirmed poorly differentiated papillarycarcinoma with numerous psammoma bodies. Neithercolloid nor follicular structures were identified, butthere were some nuclear grooves and occasional intra-nuclear cytoplasmic inclusions. The primary site wasuncertain: there was no palpable thyroid mass; abdom-inal CT scan showed calcified para-aortic lymph nodeenlargement and a 15 mm diameter pelvic cyst butthere was no pelvic lymphadenopathy or ovarian mass.The metastatic tumour showed strong staining forCA125 and carcinoembryonic antigen (CEA) and weakstaining for thyroglobulin. The patient was treatedinitially with radio-iodine but because of diseaseprogression in the neck and the right hilar region shewas later treated with carboplatin as for ovariancarcinoma. She died approximately 1 year after

diagnosis. Post-mortem examination showed fibroticatrophy of the thyroid gland. There was focal pulmon-ary infarction on the left, and a mass of hilar nodes onthe right, 60 mm in diameter. There was involvement ofthe right main bronchus. The para-aortic nodes weremassively enlarged. The ovaries were grossly normalbut histology showed some psammoma bodies and onlyone or two papillary structures. The site of originremained unclear despite autopsy and, in fact, thedifferential diagnosis widened to thyroid, ovary orrespiratory tract.

S T U DY G RO U P

Nine cases of papillary carcinoma of the thyroid and 17cases of papillary serous carcinoma of the ovary wereretrieved from the surgical files. Thyroid papillarymicrocarcinomas were not included. Immunostainingwas performed by a routine avidin-biotin complexmethod with the antibodies as detailed in Table 1(polyclonal thyroglobulin was used in case 17 only).

CA125 and thyroglobulin in papillary carcinomas 115

q 1999 Blackwell Science Ltd, Histopathology, 34, 113–117.

Figure 1. Papillary carcinoma of thyroid staining positive for CA125; note positivity of surrounding non-neoplastic thyroid follicles and colloidin b.

Appropriate positive and negative controls wereincluded. In addition, a diastase/periodic acid–Schiff(D/PAS) mucin stain was performed. Staining wasassessed semiquantitatively on a scale from 0 to þ þ þ.

Results

The staining results are shown in Table 2. Four thyroidpapillary carcinomas stained for CA125. Normalsurrounding thyroid tissue also showed some reaction.The dominant staining pattern was on the luminalsurface of the thyroid epithelial cells, and some stainingwas also present in the colloid (Figure 1).

One case of ovarian carcinoma showed moderatelystrong focal staining for thyroglobulin (Figure 2).Immunostaining in this case was repeated with apolyclonal antithyroglobulin antiserum (Table 1) andshowed similar staining (Figure 3).

The other antibodies used showed no usefuldifferences in staining frequency. Diastase/PAS stain-ing was frequently observed in the serous papillary

carcinomas of the ovary, but was not observed in thethyroid carcinomas. In the thyroid tumours there wassome D/PAS staining in the colloid and psammomabodies, but the cytoplasm and luminal borders werenegative.

Discussion

Almost half of our cases of papillary carcinoma of thethyroid showed at least some staining with CA125. Thiswas on average not as strong as the consistent positivestaining in the ovarian cases, but there is some overlap.We conclude that CA125 staining is not specific forovarian carcinoma in this context. It is uncertainwhether this phenomenon of staining in thyroidtumours, and indeed in non-neoplastic thyroid tissue,is a cross-reaction with thyroglobulin or some otherthyroid protein, or whether there is genuine expressionof CA125 in thyroid tissue and neoplasms. The functionand molecular nature of CA125 are poorly understood.Recent isolation using a new monoclonal antibody

116 C E Keen et al.

q 1999 Blackwell Science Ltd, Histopathology, 34, 113–117.

Figure 2. Serous adenocarcinoma of ovary showing focal positivestaining for thyroglobulin (monoclonal antibody).

Figure 3. Serous adenocarcinoma of ovary showing focal positivestaining for thyroglobulin (polyclonal antiserum).

showed a very high molecular weight mucin-typeglycoprotein.3 The molecule is highly glycosylated andrich in serine, threonine and proline. The elevation ofserum CA125 is better related to factors influencing itsrelease into the circulation than to its synthesis,4 and itis therefore conceivable that CA125 may be synthesizedin thyroid tissue but not secreted into the circulation. Itwould be of interest to measure CA125 levels in thyroidtissue and in the serum of patients with disseminatedthyroid cancer. We note that the product data sheetrelating to the commercially available anti-CA125 usedfor the immunostaining in this study cites, as a legalconsideration, that the antibody is recommended forresearch purposes only. We do not lend diagnostic weightto CA125 positivity in our surgical histology practice.

Conversely, occasional ovarian carcinomas may showfocal positivity, at least, for thyroglobulin. The tumourshowing moderate staining in our study was nodifferent histologically from the others. There was noquestion of a thyroidal or other endocrine component tothis tumour. One might, of course, expect thyroglobulinstaining in struma ovarii or strumal carcinoid tumourof the ovary, or indeed in other germ cell tumours; thisfinding in a straightforward serous papillary carcinomais, however, unexpected. We used a different reagent, apolyclonal antithyroglobulin, to confirm thyroglobulinexpression in the one positive case of ovarian carcinoma,but molecular methods would be required to conclusivelyconfirm cross-expression of these molecules, or to assessany possible regions of homology. The thyroglobulinstaining in the thyroid carcinomas was uniformlymoderate to strong, and the weak thyroglobulinstaining in the index case is therefore perhaps more inkeeping with a tumour of ovarian origin.

Our findings with D/PAS staining might be read assomewhat unexpected. Chan and Tse5 have shown thatstains for mucin can be positive in the colloid, luminalborder or cytoplasm in papillary carcinoma of thyroid.We found some staining in the colloid and ofpsammoma bodies, but cytoplasm and luminal borders

were negative, influencing us to score these tumours asnegative for epithelial mucin by D/PAS staining.Ovarian mucinous tumours are reported to be ‘morepositive’ for mucin than serous tumours.6 However,positivity for epithelial mucin (defined as intracyto-plasmic, luminal or surface staining with Alcian blue-Hand D/PAS) has been demonstrated in a high proportion(nine of 10) of ovarian serous tumours;7 indeed Aalto8,9

has also shown D/PAS positivity in many serouscarcinomas; quantification of the positivity relates todifferentiation and prognosis. In our hands, D/PASpositivity was a more specific discriminant than theimmunohistochemistry in distinguishing papillarycarcinomas of thyroid and ovary.

References1. Ohmori T, Okada K, Tabei R et al. CA125-producing adenocarci-

noma of the seminal vesicle. Pathol. Int. 1994; 44; 333–337.2. Rodriguez JM. Elevated CA125 levels in association with Ki-1

anaplastic lymphoma. Clin. Oncol. 1994; 6; 137.3. Lloyd KO, Yin BWT, Kudryashov V. Isolation and characterization of

ovarian cancer antigen CA 125 using a new monoclonal antibody(VK-8): Identification as a mucin-type molecule. Int. J. Cancer1997; 71; 842–850.

4. Jacobs I, Bast RC. The CA 125 tumour-associated antigen: a reviewof the literature. Hum. Reprod. 1989; 4; 1–12.

5. Chan JKC, Tse CCH. Mucin production in metastatic papillarycarcinoma of the thyroid. Hum. Pathol. 1988; 19; 195–200.

6. Aalto M-L. Mucosubstances in classification of serous andmucinous ovarian tumors: a morphometrical study. Eur. J. Obstet.Gynecol. Reprod. Biol. 1986; 22; 139–144.

7. Khoury N, Raju U, Crissman JD, Zarbo RJ, Greenawald KA. Acomparative immunohistochemical study of peritoneal and ovarianserous tumors, and mesotheliomas. Hum. Pathol. 1990; 21; 811–819.

8. Aalto M-L, Collan Y. Periodic acid-Schiff stain as a prognosticindicator in serous and mucinous ovarian tumors. Int. J. Gynaecol.Obstet. 1986; 24; 27–34.

9. Aalto M-L, Selkainaho K, Collan Y. Evaluation of carcinoembryonicantigen and periodic acid-Schiff stains in the diagnosis of serousand mucinous ovarian tumors: morphometrical study. Appl. Pathol.1986; 4; 83–89.

CA125 and thyroglobulin in papillary carcinomas 117

q 1999 Blackwell Science Ltd, Histopathology, 34, 113–117.