Value of Thyroid Specific Peroxidase and Ki-67 Stains in Preoperative Cytology for Thyroid Follicular Tumors Junko Maruta, C.T., (I.A.C.),1,2* Hironobu Hashimoto, C.T., (J.S.C.),1 Hiroto Yamashita, M.D., Ph.D.,1 Hitoshi Noguchi, M.D., Ph.D.,3 Shiro Noguchi, M.D., Ph.D.,4 Tadao K. Kobayashi, Ph.D.,5 Hidehiro Tsuneoka, Ph.D.,2 and Mutsuo Takahashi, M.D., Ph.D.6
Background: The aim of this study was to elucidate immunocyto-
chemically whether thyroid specific peroxidase (TPO) and Ki-67 can complement fine-needle aspiration (FNA) cytology as useful markers in order to distinguish between follicular adenoma (FA) and follicular carcinoma (FC). Methods: We studied 40 FAs and 68 FCs obtained by surgical resection. FNA cytology smears were divided into two groups: Cytology-A (Cy-A) (94 cases) with typical benign cytology and Cytology-B (Cy-B) (14 cases) with atypical cytology. FCs were divided into two groups: FC-I (42 cases) without any poorly differentiated structures and FC-II (26 cases) with some poorly differentiated structures. Cytology smears and histology from FAs and FCs were studied immunocytochemically for thyroid specific peroxidase (TPO) and Ki-67. Results: TPO expression was negative in 12.5% FAs, 21.4% FC-I, and 46.2% FC-II. In 68 FC cases, Cy-B were more frequently observed in TPO-negative cases (38.1%) than in TPOpositive cases (12.8%). The mean Ki-67 LI was 0.46 in FAs, 0.53 in FC-I, and 1.13 in FC-II. The high Ki-67 LI was correlated with Cy-B. Moreover, higher Ki-67 LI showed a close
This article was published online on 08 September 2014. An error was subsequently identified. This notice is included in the online and print versions to indicate that both have been corrected on 12 September 2014. 1 Department of Pathology, The Noguchi Thyroid Hospital and Foundation, Oita, Japan 2 Department of Laboratory Science, Faculty of Health Sciences, Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan 3 Department of Internal Medicine, The Noguchi Thyroid Hospital and Foundation, Oita, Japan 4 Department of Surgery, The Noguchi Thyroid Hospital and Foundation, Oita, Japan 5 Cancer Education and Research Center, Osaka University Graduate School of Medicine and Health Science, Osaka, Japan 6 Department of Administration, Mine City Hospital, Yamaguchi, Japan. *Correspondence to: Junko Maruta, C.T., (I.A.C.), Department of Pathology, The Noguchi Thyroid Hospital and Foundation, Buppu, Oita, Japan. E-mail: [email protected] Received 11 December 2013; Accepted 17 July 2014 DOI: 10.1002/dc.23204 Published online 8 September 2014 in Wiley Online Library (wileyonlinelibrary.com).
202
Diagnostic Cytopathology, Vol. 43, No 3
relationship with distant metastasis. In 94 Cy-A cases, 54 cases were FCs. When 38 cases with negative TPO or Ki-67 LI over 0.62 were extracted from them, as many as 28 cases were FCs, the rate of FCs were significantly higher than the rest. Conclusion: Therefore, addition of TPO stain and Ki-67 stain to routine Papanicolaou stain could improve the diagnostic reliability of FNA cytology for FC with high degree of malignancy. Diagn. Cytopathol. 2015;43:202–209. VC 2014 Wiley Periodicals, Inc. Key Words: thyroid; thyroid specific peroxidase; Ki-67; follicular tumor; immuocytochemistry
Follicular carcinoma (FC) represents 10 to 15% of malignant tumors derived from thyroid follicular epithelium,1 while follicular adenoma (FA) is the most common benign tumor. Histological criteria to distinguish FC from FA consist in capsular and/or vascular invasion of tumor cells. Not all thyroid FCs show cytological atypia or structural abnormalities and the definitive diagnosis of FC can be made only at the histological level after a consistent number of sections of the capsule, according to the size of the neoplasm, have been examined.2 This prevents fine needle aspiration (FNA) cytology from achieving a sufficient diagnostic accuracy for follicular tumors. In fact, while FNA cytology can identify follicular tumors, in most of the cases it cannot distinguish FA from FC nor does it evaluate the degree of malignancy in carcinomas. Many immunocyto(histo)chemical studies have been performed in the effort to distinguish FA from FC. Reportedly, some antibodies were found to be useful markers to differentiate between these two neoplasms. One of them is thyroid specific peroxidase (TPO), which belongs to the group of thyroid-specific proteins that lose expression during thyroid cell dedifferentiation.3,4 It was documented that TPO immunoreactivity was significantly reduced or negative in malignant thyroid tumors C 2014 WILEY PERIODICALS, INC. V
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY
Fig. 1. Cytology-A, Papanicolaou stain (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
Fig. 2. Cytology-A, Papanicolaou stain (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
compared to benign ones.5–10 Another candidate ancillary test was represented by the proliferation marker Ki-67 antigen. It is reported as a useful marker for discriminating FA from FC,11–14 evaluating the degree of malignancy, and as a prognostic indicator.15 The aim of this study was to elucidate immunochemically whether TPO and Ki-67 antigen can complement FNA cytology as useful markers in order to distinguish between FA and FC.
FC-I, and 8:18 and 64.9 6 15.5 years in FC-II, respectively. The averages (6SD) of the maximum diameter of the tumors were 35.9 6 13.8 mm in FAs and 28.1 6 15.8 mm in FCs; 23.2 6 13.1 mm in FC-I, and 35.9 6 16.8 mm in FC-II. The values of preoperative thyroglobulin (Tg) (ng/mL) were analyzed after logarithmic transformation. The means (6SD) of log10 (Tg) were 2.1 6 0.7 in FAs and 2.5 6 0.9 in FCs; 2.3 6 0.8 in FC-I and 2.8 6 1.0 in FC-II.
Cytological Features Materials and Methods Patients From January 2003 through December 2009, 40 FAs and 68 FCs underwent preoperative FNA cytology at the Noguchi Thyroid Clinic and Hospital Foundation (Beppu, Oita, Japan). All patients were informed about the procedure and signed a consent form approved by the ethical committee of the Foundation. FCs were divided into two groups: 42 cases of group I (FC-I) and 26 cases, including four Hurtle cell variants, of group II (FC-II). FC-I were mostly composed of follicular structures without insular, trabecular, and/or solid patterns, whereas FC-II were partly, but not mostly, composed of insular, trabecular and/or solid patterns. Atypical adenomas, follicular variant papillary carcinomas and poorly differentiated carcinomas, which are composed mostly of poorly differentiated component,1 were not included in this study. Histological classification was performed by H.Y. All cases are being followed-up for recurrence and metastasis. Up to date, 39 (92.9%) FC-I and 19 (73.1%) FC-II are free from recurrence, with a mean follow-up period (6SD) of 6.6 6 1.3 years. The sex ratios (males: females) and mean ages (6SD) at surgery were 10:30 and 50.7 6 14.4 years in FAs, and 14:54 and 56.2 6 16.4 years in FCs; 6:36 and 50.8 6 14.7 years in
FNA smears were stained with traditional Papanicolaou stain. Cytological samples were divided into two groups retrospectively: 94 cases of group A (Cy-A), that showed regular arrangement without multistratification of follicular cells, low nuclear-cytoplasmic ratio, round nuclei, uniformly distributed chromatin without nuclear atypia (Figs.1 and 2); and 14 cases of group B (Cy-B) that showed crowding or irregular arrangement of follicular cells, high nuclear-cytoplasmic ratio, nuclear atypia, coarse granular or dense chromatin and/or prominent nucleoli (Figs.3 and 4). Cytological classification was performed before surgery by J.M. All of Cy-B cases had follicular cells in altered architectural patterns or cellular atypia and were confirmed to fulfill at least the criteria for “suspicious for follicular neoplasm” of Bethesda system.
Immunochemistry Immunochemical study was retrospectively performed by J.M. Immunocytochemical study was carried out using 95% alcohol fixed smears from FNA. Some of the smears having been stained for cytological diagnosis was decolorized before the study. Immunohistochemical study was performed on 4 lm serial sections taken from formalinfixed, paraffin-embedded tissue blocks. ImmunocytoDiagnostic Cytopathology, Vol. 43, No 3
203
Diagnostic Cytopathology DOI 10.1002/dc
MARUTA ET AL.
Fig. 3. Cytology-B, Papanicolaou stain (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
Fig. 4. Cytology-B, Papanicolaou stain (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
histochemistry was performed in an automated immunostainer, the Ventana Benchmark XT device (Roche, Ventana Medical System, Inc.), using a streptavidin-biotinperoxidase kit with 3,3’-diamino-benzidine (LSAB; Ventana / VIEW DAB Universal Kit, Roche). The primary mouse monoclonal antibodies were anti-TPO antibody (clone TPO47, Biocytex, Marseilles, France) at a dilution of 1:50 and anti-Ki-67 (MIB-1) antibody (clone IM505, Immunotech, Marseille, France) at a dilution of 1:100, respectively.
(15.4%) FC-II had local lymph node metastasis and 3 (11.5%) were found to have distant metastasis to the lung in whole body nuclear scan, including 1 (3.4%) having both local and distant metastasis. Of 23 completely excised FC-II, 4 (17.4%) gave distant metastasis to bone or lung during the follow-up period. Among 54 FC cases in Cy-A 22 (40.7%) cases had wide invasions, while 8 (57.1%) cases of 14 FC cases in Cy-B had wide invasions; the difference was not significant (P 5 0.27, v2 test). Five (9.3%) cases in Cy-A showed distant metastases at surgery or during the follow-up period, whereas 5 (35.7%) cases in Cy-B did them; the difference of the rate was significant (P 5 0.0128, v2 test).
Evaluation of TPO Reactivity The proportion of immunoreactive cells for TPO was scored as 0 (no reactive cells), 11 (less than 20% of reactive cells), 21 (21–50% of reactive cells), 31 (51– 80% of reactive cells), 41 (more than 80% of reactive cells). According to the previous studies,5–9,16–19 cases with more than 80% staining were regarded as positive TPO and cases with less than 80% as negative TPO.
Ki-67 Labeling Index (LI) Ki-67 LI was calculated as the number of positive cells per 100 follicular epithelial cells. At least 1,000 cells were evaluated for each specimen.
Invasion and Metastasis Of 42 FC-I, 25 (59.5%) had minimal invasion and 17 (40.5%) had wide invasion. Of 26 FC-II, 13 (50.0%) had minimal invasion and 13 (50.0%) had wide invasion. Two (4.8%) FC-I had local lymph node metastasis and 2 (4.8%) were found to have distant metastasis to bone in whole body nuclear scans after surgery. Of 40 FC-I with radical surgical excision, only 1 (2.5%) gave distant metastasis to bone during the follow-up period. Four 204
Diagnostic Cytopathology, Vol. 43, No 3
Statistical Analysis TPO reactivity, Ki-67 LI, cytological features, and biological features were recorded and analyzed in all 108 specimens. The degree of agreement between immunocytochemistry and immunohistochemistry of TPO was evaluated by the rate of agreement with kappa statistic, and that of Ki-67 LI by the correlation coefficient. The positive predictive value (PPV) of TPO was calculated as the number of FCs with negative TPO divided by the total number of negative TPO, and the negative predictive value (NPV) as the number of FAs with positive TPO divided by the total number of positive TPO. A cut-off value of Ki-67 LI was determined so as to maximize the sum of sensitivity and specificity. Samples with Ki-67 LI higher than or equal to the cut-off value are regarded as positive, the others as negative. The PPV of Ki-67 LI was calculated as the number of positive FCs divided by the total number of the positives, and the NPV as the number of negative FAs divided by the total number of the negatives. Student’s t-test, analysis of variance, pair wise
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY
Fig. 5. Follicular carcinoma-I, immunohistochemistry with thyroid specific peroxidase (TPO) antibody. TPO was positive in most tumor cells. The positive rate was over 80% (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
comparisons with Tukey-Kramer’s honestly significant difference (HSD), Pearson’s v2 test, Tukey-Welsch’s closed test procedure for multiple comparison of contingency tables, logistic regression analysis, linear regression analysis, log-rank test in Kaplan-Meier analysis were applied when they were appropriate using SAS JMP 8.0 (SAS Institute Inc., Cary, NC.).
Results Correlation of Immunocytochemistry with Immunohistochemistry TPO reactivities of cytological and histological samples showed the agreement rate of as high as 77.8% with kappa statistic of 0.571 which indicate a moderate agreement between them. Ki-67 LI of cytological and histological samples showed the correlation coefficient of as high as 0.803 (P < 0.0001) indicating an excellent agreement between them. Further analyses were carried out for immunohistochemical results.
TPO Expression and Clinicopathological Features Gender, tumor size, and age at surgery showed no relation with TPO expression in all follicular tumors examined (P 5 0.34, v2 test; P 5 0.26, t-test; P 5 0.44, t-test, respectively). Cases with higher Tg level showed lower TPO expression in all follicular tumors (P 5 0.02, logistic regression). TPO expression was negative in 12.5% (5/40) of FA, in 21.4% (9 of 42) of FC-I (Fig. 5), and in 46.2% (12 of 26) of FC-II (Fig. 6) (Table I). PPV was 80.8% (21 of 26) and NPV was 42.7% (35 of 82). TPO expression was significantly lower in malignant than in benign lesions, and also significantly lower in lesions with poorly differentiated component than those without (P 5 0.28 between
Fig. 6. Follicular carcinoma-II, immunohistochemistry with TPO antibody. TPO was negative in most tumor cells. The positive rate was far lower than 80% (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
FA and FC-I, P 5 0.002 between FA and FC-II, P 5 0.03 between FC-I and FC-II, v2). The TPO expression showed no relation with the degree of invasion (P 5 0.15, v2 test), the number of local lymph node metastases at surgery (P 5 0.07, logistic regression) or the existence of distant metastasis (P 5 0.6, v2 test). Among 63 cases with complete surgery, TPO reactivity showed no relation with rate of recurrence of FC (P 5 0.69, log-rank test in KaplanMeier analysis).
Ki-67LI and Clinicopathological Features Gender and tumor size showed no relation with Ki-67 LI (P 5 0.67 logistic regression, P 5 0.45 linear regression, respectively). Age at surgery in all cases had no significant relation with Ki-67 LI (P 5 0.14, linear regression). Cases with higher Tg levels showed higher Ki-67 LI (P 5 0.002, linear regression). The means (6SE) of Ki-67 LI were 0.46 6 0.07 in FA, 0.51 6 0.10 in FC-I, and 1.27 6 0.28 in FC-II, showing significant differences not only between FA and FC-II (P 5 0.0007, HSD), but also between FC-I and FC-II (P 5 0.0014, HSD) (Table II). The cut-off value of Ki-67 LI was 0.63. PPV was 75.8% (25 of 33) and NPV was 42.7% (32 of 75). Ki-67 LI showed no relation with the degree of invasion (P 5 0.69, logistic regression). Ki-67 LI showed a marginal positive correlation with the number of local lymph node metastases at surgery (P 5 0.053, linear regression). A higher Ki-67 LI had a significant positive relation with distant metastasis at surgery (P 5 0.0018, logistic regression). Among 63 cases of FC with complete surgery, those with low Ki-67 LI (0.0–4.9) had lower rate of recurrence Diagnostic Cytopathology, Vol. 43, No 3
205
Diagnostic Cytopathology DOI 10.1002/dc
MARUTA ET AL. Table I. TPO Expression in Follicular Tumors Positive
Negative
Follicular adenoma (n 5 40)
35 (87.5%)
5 (12.5%)
Follicular carcinoma-I (n 5 42)
33 (78.6%)
9 (21.4%)
Follicular carcinoma-II (n 5 26)
14 (53.8%)
12 (46.2%)
a
Significance levela
2
Pearson’s v test followed by Tukey-Welsh’s closed test procedure for multiple comparison.
Table II. Ki-67 Labeling Index in Follicular Tumors Significance levela
Mean (6SE) Follicular adenoma (n 5 40)
0.46 (60.07)
Follicular carcinoma-I (n 5 42)
0.51 (60.10)
Follicular carcinoma-II (n 5 26)
1.27 (60.28)
a
Tukey-Kramer’s honestly significant difference (HSD).
of FC than those with high Ki-67 LI (0.5–4.28) (P 5 0.033, log-rank test in Kaplan-Meier analysis). If all the distant metastases were lumped together regardless of the time when they were found at surgery or during follow-up period, the relation with Ki-67 LI became more significant (P 5 0.0007, v2 test) (Table III). Among 40 cases with the Ki-67 LI lower than 0.5, 39 (97.5%) were free of distant metastasis at surgery and during the follow-up period. On the other hand, among 28 cases with Ki-67 LI higher than 0.5, only 19 (67.9%) were free of distant metastasis and as many as 9 (32.1%) showed distant metastasis at surgery or during the follow-up period.
Table III. Ki-67 Labeling Index and Distant Metastasis in Follicular Carcinomas
Relationship Between TPO Expression and Ki-67 LI
negative TPO (n 5 21) positive TPO (n 5 47)
TPO expression had no relation with Ki-67 LI in all cases examined (P 5 0.18, t-test). In FA, the mean (6SE) Ki67 LI was 0.85 6 0.25 in negative cases for TPO expression; this was slightly higher than 0.40 6 0.07 in positive cases for TPO reactivity (P 5 0.043, t-test). There was no significant differences in FC-I (P 5 0.32, t-test) or in FC-II (P 5 0.89, t-test).
Cytological Features, TPO Expression, and Ki-67 LI All cases of FA had Cy-A in the preoperative FNA cytology. Fifty-four of 68 FCs had Cy-A, and 14 cases had Cy-B. Among FCs, Cy-B was more frequent in FC-II (38.5%; 10 of 26) than in FC-I (9.5%; 4 of 42). Among the 68 cases of FCs, Cy-B were more frequently observed in TPO-negative cases (38.1%) than in TPO-positive cases (12.8%) (P 5 0.017, v2 test) (Table IV) (Fig. 7). Among the 18 cases that had Cy-A and negative TPO, 13 cases (72%) were FCs. 206
Diagnostic Cytopathology, Vol. 43, No 3
Ki-67 labeling index
No distant metastasisa
Distant metastasisa
0–0.49 (n 5 40)
39 (97.5%)
1 (2.5%)
0.5–4.28 (n 5 28)
19 (67.9%)
9 (32.1%)
Significance levelb
a
At operation or during follow-up period. Pearson’s v2 test.
b
Table IV. TPO Expression and Cytological Features in Follicular Carcinomas Cytology A
Cytology B
13 (61.9%) 41 (87.2%)
8 (38.1%) 6 (12.8%)
P 5 0.017, Pearson’s v2 test.
In FCs, the higher was the Ki-67 LI, the higher was the rate of Cy-B appearance (P 5 0.029, logistic regression) (Fig. 8). In 94 Cy-A cases, 54 cases (57%) were FCs. When 38 cases with negative TPO or Ki-67 LI over 0.62 were extracted from them, as many as 28 cases (74%) were FCs, the rate of FCs was significantly higher than the rest in this group (P 5 0.0087, v2 test).
Discussion TPO is a hemo-containing glycoprotein that plays a key role in thyroid hormone biosynthesis.5 Since TPO reactivity disappears in parallel with de-differentiation of the follicular cell, immunocyto(histo)logical studies of TPO expression have been carried out for the diagnosis of follicular tumors.4–10,16–22 Increased de-differentiation,
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY
Fig. 7. Cytology-B, immunocytochemistry with TPO antibody. TPO was negative in most tumor cells. The positive rate was far lower than 80% (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Fig. 8. Cytology-B, immunocytochemistry with Ki-67 antibody (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
including loss of tissue-specific cell proteins such as TPO, is a well-known hallmark of thyroid tumor progression.23 High immunoreactivity for TPO was present in adenomatous goiter and follicular tumors, less reactivity in follicular or papillary carcinomas, and no reactivity in undifferentiated carcinomas.16 The reported results for TPO expression were various: some investigators underlined that immunocyto(histo)chemical study of TPO was useful for the diagnosis of follicular tumors,5–7,10,19 whereas some others stated it was not enough,8,9 or not useful for the diagnosis of those tumors.16–22 In the current study with 68 FCs and 40 FAs, FCs showed a significantly higher rate (30.9%) of TPO negatives than FAs (18.5%). The PPV of 80.8% was not high enough to provide reliable diagnosis, and the NPV was as low as 42.7%. TPO alone failed to discriminate FCs from FAs, which was concordant with the previous report.18 The current study shows that there are not only some cases that have negative TPO expression in FAs, but also some cases that show positive TPO expression in FCs. The WHO classification defines poorly differentiated carcinoma as mostly consisting of poorly differentiated components, such as trabecular, solid, and/or insular pattern.1 Therefore, tumors with only partial expression, but not prevalent of these components are classified into follicular carcinoma. In the present study, FCs were divided into two groups; FC-I have mostly follicular structure without such poorly differentiated components, and FC-II have partial but not prevalent expression of poorly differentiated components. The TPO reactivity was compared between those two FCs. The negative rates of TPO expression was 46.2% (12 of 26) in FC-II; this value was significantly higher than 21.4% (9 of 42) in FC-I (P 5 0.03) and FAs (P 5 0.002). The negative rate of FC-I
was close to the negative rate of FAs. Thus, FCs may be considered as consisting of two groups having different reactivity for TPO that could be expected to differ from each other in their biological behaviors. The FC-II had higher rate of distant metastasis and recurrence than FC-I, and TPO expression was inversely related with the progressive potential of FCs. There was a striking inverse relation between TPO reactivity and degree of cellular atypia as foci not expressing TPO corresponded to areas of high cellular or architectural atypia.7,10 In Bethesda system for thyroid follicular tumors, cytology should be categorized into “follicular tumor” or “suspect of follicular tumor” without benign/malignant judgment.24 In general, high cellularity, irregular cell stratification, high nuclear-cytoplasmic ratio, nuclear atypia, increased amount of chromatin, evident and/or prominent nucleoli are regarded as cytologically malignant-appearing features. Thus, thyroid FNA cytology can be divided into two groups; one with malignantappearing features (Cy-B), and the other one without (Cy-A). Looking at the cytological features of FCs, 38.1% (8 of 21) of negative cases for TPO expression were Cy-B, whereas, 12.8% (6 of 47) of positive cases for TPO expression were Cy-B. All the eight cases with negative TPO expression and Cy-B were FC-II. Thus negative TPO expression is significantly reflected in an atypical cell morphology and in a higher degree of malignancy. Determination of cell proliferative activity is one of the important factors for assessing the biological behavior of carcinomas. One of the useful markers to evaluate cell proliferative activity is Ki-67, which is expressed in almost all cells in the body except those in the G0 phase of cell cycle.25,26 The use of Ki-67 in the differential Diagnostic Cytopathology, Vol. 43, No 3
207
Diagnostic Cytopathology DOI 10.1002/dc
MARUTA ET AL. Table V. Clinical and Immunocyto(histo)chemical Features of Histological Types
Cytology-B group TPO negative Ki-67 labeling index Distant metastasisa a
Follicular adenoma, N 5 40
Follicular carcinoma-I, N 5 42
Follicular carcinoma-II, N 5 26
0% 12.5% 0.46 6 0.07 NA
9.5% 21.4% 0.51 6 0.10 7.1%
38.5% 46.2% 1.27 6 0.28 26.9%
At surgery or during follow-up period.
diagnosis of thyroid nodules is thought to be of limited usefulness, because Ki-67 LI is generally low in thyroid carcinomas, except that in undifferentiated (anaplastic) carcinoma and poorly differentiated carcinoma.27,28 Although the usefulness of Ki-67 LI has also been reported in thyroid follicular tumors,11,14,29 the values of Ki-67 LI were overlapping between FAs and FCs.12,13,30–32 In the current study, the mean Ki-67 LI (6SE) were 0.46 (60.07) in FAs and 0.80 (60.13) in all FCs. The statistic difference was marginal (P 5 0.06, ttest). The PPV was a moderate value of 75.8%, and the NPV was as low as 42.7%. When FCs were divided into two groups, i.e., FC-I and FC-II, the statistic difference became significant. Ki-67 LI were 0.51 (61.0) in FC-I and 1.27 (60.28) in FC-II. That is, FC-II showed significantly higher Ki-67 LI than FA (P 5 0.0007) and FC-I (P 5 0.014). Since the value of Ki-67 LI represents the proliferation activity of tumor cells, the above results indicate that many of the FC-II have high proliferative activity. Moreover, it was documented that Ki-67 expression was significantly higher in patients with FC with metastasis than in those without metastasis,29 Ki-67 LI also positively correlated with pathological stage in differentiated thyroid carcinomas15 and with distant metastasis at surgery in papillary carcinoma.31,33 In this study, 40 cases of FC showed a Ki-67 LI lower than 0.5. Only one (2.5%) of them gave distant metastasis. In contrast, among 28 cases that showed Ki-67 LI higher than 0.5, 9 (32.1%) cases gave distant metastasis. The odds ratio, (9 of 19)/(1 of 39), is as high as 18.5. These data indicate that Ki-67 LI strongly reflects the biological malignancy, especially in the propensity for distant metastasis. In the present study, high Ki-67 LI was correlated with Cy-B. Cytological features for Cy-B can be regarded to have a relation with biological malignancy. In 94 Cy-A cases, 54 cases (57%) were FCs. When 38 cases with negative TPO or Ki-67 LI over 0.62 were extracted from them, as many as 28 cases (74%) were FCs, the rate of FCs was significantly higher than the rest. In this series, FCs were divided into FC-I and FC-II. Our results indicated FC-II are prone to have cellular atypia, negative TPO expression, high Ki-67 LI and distant metastasis compared to FC-I (Table V). These facts suggest that the FC-II can be regarded as a distinct category from FC-I. 208
Diagnostic Cytopathology, Vol. 43, No 3
Although FCs and FAs cannot be separated by TPO or Ki-67 LI, FCs with more malignant features, such as cytological atypia, poorly differentiated components, distant metastasis, are prone to show the negative TPO and the higher Ki-67 LI. Identifying these FCs should be focused on in the future research. Addition of TPO stain and Ki-67 stain to routine Papanicolaou stain will improve the diagnostic reliability of FNA cytology for FC with high degree of malignancy.
References 1. Sobrinho-Simones M, Asa SL, Kroll TG, et al. Follicular carcinoma. In: DeLellis RA, Lloyd RV, Heitz PU, Eng C, editors. WHO pathology & genetics tumours of endocrine organs. Lyon, France: IARC Press; 2004. p 67–72. 2. Goldstein RE, Netterville JL, Burkey B, Johnson JE. Implications of follicular neoplasms atypia and lesions suspicious for malignancy diagnosed by fine-needle aspiration of thyroid nodules. Ann Surg 2002;235:656–664. 3. Taurog A, Dorris ML, Doerge DR. Mechanism of simultaneous iodination and coupling catalyzed by thyroid peroxidase. Arch Biochem Biophys 1996;330:24–32. 4. De Micco C, Ruf J, Chrestian MA, Gros N, Henry JF, Carayon P. Immunohistochemical study of thyroid peroxidase in normal, hyperplastic, and neoplastic human thyroid tissues. Cancer 1991;67: 3036–3041. 5. De Micco C, Zoro P, Garcia S, et al. Thyroid peroxidase immunodetection as a tool to assist diagnosis of thyroid nodules on fineneedle aspiration biopsy. Eur J Endocrinol 1994;131:474–479. 6. Henry JF, Denizot A, Porelli A, et al. Thyroperoxidase immunodetection for the diagnosis of malignancy on fine-needle aspiration of thyroid nodules. World J Surg 1994;18:529–534. 7. De Micco C, Vasko V, Garcia S, Zoro P, Denizot A, Henry JF. Fine-needle aspiration of thyroid follicular neoplasm: Diagnostic use of thyroid peroxidase immunocytochemistry with monoclonal antibody 47. Surgery 1994;116:1031–1035. 8. Pluot M, Faroux MJ, Flament JB, Patey M, Theobald S, Delisle MJ. Quantitative cytology and thyroperoxidase immunochemistry: new tools in evaluating thyroid nodules by fine-needle aspiration. Cancer Detect Prev 1996;20:285–293. 9. Faroux MJ, Theobald S, Pluot M, Patey M, Menzies D. Evalution of the monoclonal antibody antithyroperoxidase MoAb47 in the diagnostic decision of cold thyroid nodules by fine-needle aspiration. Pathol Res Pract 1997;193:705–712. 10. Garcia S, Vassko V, Henry JF, De Micco C. Comparison of thyroid peroxidase expression with cellular proliferation in thyroid follicular tumors. Thyroid 1998;8:745–749. 11. Pujani M, Arora B, Pujani M, Singh SK, Tejwani N. Role of Ki-67 as a proliferative marker in lesions of thyroid. Indian J Cancer 2001;47:304–307. 12. Cui W, Lu X, Zheng S, Ma Y, Liu X, Zhang W. The use of a combination of Ki-67, Galection-3, and PTTG can distinguish the benign and malignant thyroid tumor. Clin Lab 2012;58:419–426. 13. Heikkil€a A, Siironen P, Hagstr€ om J, et al. Follicular thyroid neoplasm: Clinicopathologic features suggesting malignancy. APMIS 2010;118:846–854. 14. Lacoste-Collin L, d’Aure D, Berard E, Rouquette I, Delisle MB, Courtade-Sa€ıdi M. Improvement of the cytological diagnostic accuracy of follicular thyroid lesions by the use of the Ki-67 proliferative index in addition to cytokeratin-19 and HBME-1 immunomarkers: A study of 61 cases of liquid-based FNA cytology with histological controls. Cytopathology 2014;25:160–169.
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY 15. M€ussig K, Wehrmann T, Dittmann H, et al. Expression of the proliferation marker Ki-67 associates with tumour staging and clinical outcome in differentiated thyroid carcinomas. Clin Endocrinol 2012;77:139–145. 16. Lima MA, Gontijo VA, Santos MC, Schmitt FCL. Thyroid peroxidase expression in diseased human thyroid glands. Endocr Pathol 1999;10:223–228. 17. Savin S, Cvejic D, Isic T, Paunovic I, Tatic S, Havelka M. The efficacy of the thyroid peroxidase marker for distinguishing follicular thyroid carcinoma from follicular adenoma. Exp Oncol 2006;28: 70–74. 18. Savin S, Cvejic D, Isic T, et al. Thyroid peroxidase immunohistochemistry in differential diagnosis of thyroid tumors. Endocr Pathol 2006;17:53–60. 19. De Micco C, Savchenko V, Giorgi R, Sebag F, Henry JF. Utility of malignancy markers in fine-needle aspiration cytology of thyroid nodules: Comparison of Hector Battifora mesothelial antigen-1, thyroid peroxidase and dipeptidyl aminopeptidase IV. Br J Cancer 2008;98:818–823. 20. Weber KB, Shroyer KR, Heinz DE, Nawaz S, Said MS, Haugen BR. The use of a combination of galection-3 and thyroid peroxidase for the diagnosis and prognosis of thyroid cancer. Am J Clin Pathol 2004;122:524–531. 21. Savin S, Cvejic D, Isic T, Paunovic I, Tatic S Havelka M. Thyroid peroxidase and galectin-3 immunostaining in differentiated thyroid carcinoma with clinicopathologic correlation. Hum Pathol 2008;39: 1656–1663. 22. Paunovic I, Isic T, Havelka M, Tatic S, Cvejic D, Savin S. Combined immunohistochemistry for thyroid peroxidase, galection-3, CK19 and HMBE-1 in differential diagnosis of thyroid tumors. APMIS 2012;120:368–379. 23. Di Cristofaro J, Silvy M, Lanteaume A, Marcy M, Carayon P, De Micco C. Expression of tpo mRNA in thyroid tumors: Quantitative PCR analysis and correlation with alterations of ret, Braf, ras and pax8 genes. Endocr Relat Cancer 2006;13:485–495.
24. Ali CZ, Cibas ES. The Bethesda system for reporting thyroid cytopathology: Definitions, criteria, and explanatory notes. New York: Springer-Verlag; 2010:59–73. 25. Brown DC, Gatter KC. Monoclonal antibody Ki-67. Its use in histopathology. Histopathology 1990;17:489–503. 26. Cattoretti G, Becker MH, Key G, et al. Monoclonal antibodies against recombinant parts of the Ki-67 antigen (MIB 1 and MIB 3) detect proliferating cells in microwave-processed formalin-fixed paraffin sections. J Pathol 1992;168:357–363. 27. Aiad HA, Bashandy MA, Abdou AG, Zahran AA. Significance of AgNORs and Ki-67 proliferative markers in differential diagnosis of thyroid lesions. Pathol Oncol Res 2013;19:167–175. 28. Wallin G, B€ackdahl M, Christensson B, Grimelius L, Auer G. Nuclear protein content and Ki-67 immunoreactivity in nonneoplastic and neoplastic thyroid cells. Anal Quant Cytol Histol 1992;14:296–303. 29. Erickson LA, Jin L, Wollan PC, Thompson GB, van Heerden J, Lloyd RV. Expression of p27kip1 and Ki-67 in benign and malignant thyroid tumors. Modern Pathol 1998;11:169–174. 30. Kr olicka A, Kobierzycki C, Pula B, et al. Comparison of metallothionein (MT) and Ki-67 antigen expression in benign and malignant thyroid tumors. Anticancer Res 2010;30:4945–4949. 31. Sofiadis A, Tani E, Foukakis T, et al. Diagnostic and prognostic potential of MIB-1 proliferation index in thyroid fine needle aspiration biopsy. Int J Oncol 2009;35:369–374. 32. Tan A, Etit D, Bayol U, Altinel D, Tan S. Comparison of proliferating cell nuclear antigen, thyroid transcription factor-1, Ki-67, p63, p53 and high-molecular weight cytokeratin expressions in papillary thyroid carcinoma, follicular carcinoma, and follicular adenoma. Ann Diagn Pathol 2011;15:108–116. 33. Ito Y, Miyauchi A, Kakudo K, Hirokawa M, Kobayashi K, Miya A. Prognostic significance of Ki-67 labeling index in papillary thyroid carcinoma. World J Surg 2010;34:3015–3021.
Diagnostic Cytopathology, Vol. 43, No 3
209
Background: The aim of this study was to elucidate immunocyto-
chemically whether thyroid specific peroxidase (TPO) and Ki-67 can complement fine-needle aspiration (FNA) cytology as useful markers in order to distinguish between follicular adenoma (FA) and follicular carcinoma (FC). Methods: We studied 40 FAs and 68 FCs obtained by surgical resection. FNA cytology smears were divided into two groups: Cytology-A (Cy-A) (94 cases) with typical benign cytology and Cytology-B (Cy-B) (14 cases) with atypical cytology. FCs were divided into two groups: FC-I (42 cases) without any poorly differentiated structures and FC-II (26 cases) with some poorly differentiated structures. Cytology smears and histology from FAs and FCs were studied immunocytochemically for thyroid specific peroxidase (TPO) and Ki-67. Results: TPO expression was negative in 12.5% FAs, 21.4% FC-I, and 46.2% FC-II. In 68 FC cases, Cy-B were more frequently observed in TPO-negative cases (38.1%) than in TPOpositive cases (12.8%). The mean Ki-67 LI was 0.46 in FAs, 0.53 in FC-I, and 1.13 in FC-II. The high Ki-67 LI was correlated with Cy-B. Moreover, higher Ki-67 LI showed a close
This article was published online on 08 September 2014. An error was subsequently identified. This notice is included in the online and print versions to indicate that both have been corrected on 12 September 2014. 1 Department of Pathology, The Noguchi Thyroid Hospital and Foundation, Oita, Japan 2 Department of Laboratory Science, Faculty of Health Sciences, Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan 3 Department of Internal Medicine, The Noguchi Thyroid Hospital and Foundation, Oita, Japan 4 Department of Surgery, The Noguchi Thyroid Hospital and Foundation, Oita, Japan 5 Cancer Education and Research Center, Osaka University Graduate School of Medicine and Health Science, Osaka, Japan 6 Department of Administration, Mine City Hospital, Yamaguchi, Japan. *Correspondence to: Junko Maruta, C.T., (I.A.C.), Department of Pathology, The Noguchi Thyroid Hospital and Foundation, Buppu, Oita, Japan. E-mail: [email protected] Received 11 December 2013; Accepted 17 July 2014 DOI: 10.1002/dc.23204 Published online 8 September 2014 in Wiley Online Library (wileyonlinelibrary.com).
202
Diagnostic Cytopathology, Vol. 43, No 3
relationship with distant metastasis. In 94 Cy-A cases, 54 cases were FCs. When 38 cases with negative TPO or Ki-67 LI over 0.62 were extracted from them, as many as 28 cases were FCs, the rate of FCs were significantly higher than the rest. Conclusion: Therefore, addition of TPO stain and Ki-67 stain to routine Papanicolaou stain could improve the diagnostic reliability of FNA cytology for FC with high degree of malignancy. Diagn. Cytopathol. 2015;43:202–209. VC 2014 Wiley Periodicals, Inc. Key Words: thyroid; thyroid specific peroxidase; Ki-67; follicular tumor; immuocytochemistry
Follicular carcinoma (FC) represents 10 to 15% of malignant tumors derived from thyroid follicular epithelium,1 while follicular adenoma (FA) is the most common benign tumor. Histological criteria to distinguish FC from FA consist in capsular and/or vascular invasion of tumor cells. Not all thyroid FCs show cytological atypia or structural abnormalities and the definitive diagnosis of FC can be made only at the histological level after a consistent number of sections of the capsule, according to the size of the neoplasm, have been examined.2 This prevents fine needle aspiration (FNA) cytology from achieving a sufficient diagnostic accuracy for follicular tumors. In fact, while FNA cytology can identify follicular tumors, in most of the cases it cannot distinguish FA from FC nor does it evaluate the degree of malignancy in carcinomas. Many immunocyto(histo)chemical studies have been performed in the effort to distinguish FA from FC. Reportedly, some antibodies were found to be useful markers to differentiate between these two neoplasms. One of them is thyroid specific peroxidase (TPO), which belongs to the group of thyroid-specific proteins that lose expression during thyroid cell dedifferentiation.3,4 It was documented that TPO immunoreactivity was significantly reduced or negative in malignant thyroid tumors C 2014 WILEY PERIODICALS, INC. V
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY
Fig. 1. Cytology-A, Papanicolaou stain (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
Fig. 2. Cytology-A, Papanicolaou stain (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
compared to benign ones.5–10 Another candidate ancillary test was represented by the proliferation marker Ki-67 antigen. It is reported as a useful marker for discriminating FA from FC,11–14 evaluating the degree of malignancy, and as a prognostic indicator.15 The aim of this study was to elucidate immunochemically whether TPO and Ki-67 antigen can complement FNA cytology as useful markers in order to distinguish between FA and FC.
FC-I, and 8:18 and 64.9 6 15.5 years in FC-II, respectively. The averages (6SD) of the maximum diameter of the tumors were 35.9 6 13.8 mm in FAs and 28.1 6 15.8 mm in FCs; 23.2 6 13.1 mm in FC-I, and 35.9 6 16.8 mm in FC-II. The values of preoperative thyroglobulin (Tg) (ng/mL) were analyzed after logarithmic transformation. The means (6SD) of log10 (Tg) were 2.1 6 0.7 in FAs and 2.5 6 0.9 in FCs; 2.3 6 0.8 in FC-I and 2.8 6 1.0 in FC-II.
Cytological Features Materials and Methods Patients From January 2003 through December 2009, 40 FAs and 68 FCs underwent preoperative FNA cytology at the Noguchi Thyroid Clinic and Hospital Foundation (Beppu, Oita, Japan). All patients were informed about the procedure and signed a consent form approved by the ethical committee of the Foundation. FCs were divided into two groups: 42 cases of group I (FC-I) and 26 cases, including four Hurtle cell variants, of group II (FC-II). FC-I were mostly composed of follicular structures without insular, trabecular, and/or solid patterns, whereas FC-II were partly, but not mostly, composed of insular, trabecular and/or solid patterns. Atypical adenomas, follicular variant papillary carcinomas and poorly differentiated carcinomas, which are composed mostly of poorly differentiated component,1 were not included in this study. Histological classification was performed by H.Y. All cases are being followed-up for recurrence and metastasis. Up to date, 39 (92.9%) FC-I and 19 (73.1%) FC-II are free from recurrence, with a mean follow-up period (6SD) of 6.6 6 1.3 years. The sex ratios (males: females) and mean ages (6SD) at surgery were 10:30 and 50.7 6 14.4 years in FAs, and 14:54 and 56.2 6 16.4 years in FCs; 6:36 and 50.8 6 14.7 years in
FNA smears were stained with traditional Papanicolaou stain. Cytological samples were divided into two groups retrospectively: 94 cases of group A (Cy-A), that showed regular arrangement without multistratification of follicular cells, low nuclear-cytoplasmic ratio, round nuclei, uniformly distributed chromatin without nuclear atypia (Figs.1 and 2); and 14 cases of group B (Cy-B) that showed crowding or irregular arrangement of follicular cells, high nuclear-cytoplasmic ratio, nuclear atypia, coarse granular or dense chromatin and/or prominent nucleoli (Figs.3 and 4). Cytological classification was performed before surgery by J.M. All of Cy-B cases had follicular cells in altered architectural patterns or cellular atypia and were confirmed to fulfill at least the criteria for “suspicious for follicular neoplasm” of Bethesda system.
Immunochemistry Immunochemical study was retrospectively performed by J.M. Immunocytochemical study was carried out using 95% alcohol fixed smears from FNA. Some of the smears having been stained for cytological diagnosis was decolorized before the study. Immunohistochemical study was performed on 4 lm serial sections taken from formalinfixed, paraffin-embedded tissue blocks. ImmunocytoDiagnostic Cytopathology, Vol. 43, No 3
203
Diagnostic Cytopathology DOI 10.1002/dc
MARUTA ET AL.
Fig. 3. Cytology-B, Papanicolaou stain (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
Fig. 4. Cytology-B, Papanicolaou stain (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
histochemistry was performed in an automated immunostainer, the Ventana Benchmark XT device (Roche, Ventana Medical System, Inc.), using a streptavidin-biotinperoxidase kit with 3,3’-diamino-benzidine (LSAB; Ventana / VIEW DAB Universal Kit, Roche). The primary mouse monoclonal antibodies were anti-TPO antibody (clone TPO47, Biocytex, Marseilles, France) at a dilution of 1:50 and anti-Ki-67 (MIB-1) antibody (clone IM505, Immunotech, Marseille, France) at a dilution of 1:100, respectively.
(15.4%) FC-II had local lymph node metastasis and 3 (11.5%) were found to have distant metastasis to the lung in whole body nuclear scan, including 1 (3.4%) having both local and distant metastasis. Of 23 completely excised FC-II, 4 (17.4%) gave distant metastasis to bone or lung during the follow-up period. Among 54 FC cases in Cy-A 22 (40.7%) cases had wide invasions, while 8 (57.1%) cases of 14 FC cases in Cy-B had wide invasions; the difference was not significant (P 5 0.27, v2 test). Five (9.3%) cases in Cy-A showed distant metastases at surgery or during the follow-up period, whereas 5 (35.7%) cases in Cy-B did them; the difference of the rate was significant (P 5 0.0128, v2 test).
Evaluation of TPO Reactivity The proportion of immunoreactive cells for TPO was scored as 0 (no reactive cells), 11 (less than 20% of reactive cells), 21 (21–50% of reactive cells), 31 (51– 80% of reactive cells), 41 (more than 80% of reactive cells). According to the previous studies,5–9,16–19 cases with more than 80% staining were regarded as positive TPO and cases with less than 80% as negative TPO.
Ki-67 Labeling Index (LI) Ki-67 LI was calculated as the number of positive cells per 100 follicular epithelial cells. At least 1,000 cells were evaluated for each specimen.
Invasion and Metastasis Of 42 FC-I, 25 (59.5%) had minimal invasion and 17 (40.5%) had wide invasion. Of 26 FC-II, 13 (50.0%) had minimal invasion and 13 (50.0%) had wide invasion. Two (4.8%) FC-I had local lymph node metastasis and 2 (4.8%) were found to have distant metastasis to bone in whole body nuclear scans after surgery. Of 40 FC-I with radical surgical excision, only 1 (2.5%) gave distant metastasis to bone during the follow-up period. Four 204
Diagnostic Cytopathology, Vol. 43, No 3
Statistical Analysis TPO reactivity, Ki-67 LI, cytological features, and biological features were recorded and analyzed in all 108 specimens. The degree of agreement between immunocytochemistry and immunohistochemistry of TPO was evaluated by the rate of agreement with kappa statistic, and that of Ki-67 LI by the correlation coefficient. The positive predictive value (PPV) of TPO was calculated as the number of FCs with negative TPO divided by the total number of negative TPO, and the negative predictive value (NPV) as the number of FAs with positive TPO divided by the total number of positive TPO. A cut-off value of Ki-67 LI was determined so as to maximize the sum of sensitivity and specificity. Samples with Ki-67 LI higher than or equal to the cut-off value are regarded as positive, the others as negative. The PPV of Ki-67 LI was calculated as the number of positive FCs divided by the total number of the positives, and the NPV as the number of negative FAs divided by the total number of the negatives. Student’s t-test, analysis of variance, pair wise
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY
Fig. 5. Follicular carcinoma-I, immunohistochemistry with thyroid specific peroxidase (TPO) antibody. TPO was positive in most tumor cells. The positive rate was over 80% (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
comparisons with Tukey-Kramer’s honestly significant difference (HSD), Pearson’s v2 test, Tukey-Welsch’s closed test procedure for multiple comparison of contingency tables, logistic regression analysis, linear regression analysis, log-rank test in Kaplan-Meier analysis were applied when they were appropriate using SAS JMP 8.0 (SAS Institute Inc., Cary, NC.).
Results Correlation of Immunocytochemistry with Immunohistochemistry TPO reactivities of cytological and histological samples showed the agreement rate of as high as 77.8% with kappa statistic of 0.571 which indicate a moderate agreement between them. Ki-67 LI of cytological and histological samples showed the correlation coefficient of as high as 0.803 (P < 0.0001) indicating an excellent agreement between them. Further analyses were carried out for immunohistochemical results.
TPO Expression and Clinicopathological Features Gender, tumor size, and age at surgery showed no relation with TPO expression in all follicular tumors examined (P 5 0.34, v2 test; P 5 0.26, t-test; P 5 0.44, t-test, respectively). Cases with higher Tg level showed lower TPO expression in all follicular tumors (P 5 0.02, logistic regression). TPO expression was negative in 12.5% (5/40) of FA, in 21.4% (9 of 42) of FC-I (Fig. 5), and in 46.2% (12 of 26) of FC-II (Fig. 6) (Table I). PPV was 80.8% (21 of 26) and NPV was 42.7% (35 of 82). TPO expression was significantly lower in malignant than in benign lesions, and also significantly lower in lesions with poorly differentiated component than those without (P 5 0.28 between
Fig. 6. Follicular carcinoma-II, immunohistochemistry with TPO antibody. TPO was negative in most tumor cells. The positive rate was far lower than 80% (310). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
FA and FC-I, P 5 0.002 between FA and FC-II, P 5 0.03 between FC-I and FC-II, v2). The TPO expression showed no relation with the degree of invasion (P 5 0.15, v2 test), the number of local lymph node metastases at surgery (P 5 0.07, logistic regression) or the existence of distant metastasis (P 5 0.6, v2 test). Among 63 cases with complete surgery, TPO reactivity showed no relation with rate of recurrence of FC (P 5 0.69, log-rank test in KaplanMeier analysis).
Ki-67LI and Clinicopathological Features Gender and tumor size showed no relation with Ki-67 LI (P 5 0.67 logistic regression, P 5 0.45 linear regression, respectively). Age at surgery in all cases had no significant relation with Ki-67 LI (P 5 0.14, linear regression). Cases with higher Tg levels showed higher Ki-67 LI (P 5 0.002, linear regression). The means (6SE) of Ki-67 LI were 0.46 6 0.07 in FA, 0.51 6 0.10 in FC-I, and 1.27 6 0.28 in FC-II, showing significant differences not only between FA and FC-II (P 5 0.0007, HSD), but also between FC-I and FC-II (P 5 0.0014, HSD) (Table II). The cut-off value of Ki-67 LI was 0.63. PPV was 75.8% (25 of 33) and NPV was 42.7% (32 of 75). Ki-67 LI showed no relation with the degree of invasion (P 5 0.69, logistic regression). Ki-67 LI showed a marginal positive correlation with the number of local lymph node metastases at surgery (P 5 0.053, linear regression). A higher Ki-67 LI had a significant positive relation with distant metastasis at surgery (P 5 0.0018, logistic regression). Among 63 cases of FC with complete surgery, those with low Ki-67 LI (0.0–4.9) had lower rate of recurrence Diagnostic Cytopathology, Vol. 43, No 3
205
Diagnostic Cytopathology DOI 10.1002/dc
MARUTA ET AL. Table I. TPO Expression in Follicular Tumors Positive
Negative
Follicular adenoma (n 5 40)
35 (87.5%)
5 (12.5%)
Follicular carcinoma-I (n 5 42)
33 (78.6%)
9 (21.4%)
Follicular carcinoma-II (n 5 26)
14 (53.8%)
12 (46.2%)
a
Significance levela
2
Pearson’s v test followed by Tukey-Welsh’s closed test procedure for multiple comparison.
Table II. Ki-67 Labeling Index in Follicular Tumors Significance levela
Mean (6SE) Follicular adenoma (n 5 40)
0.46 (60.07)
Follicular carcinoma-I (n 5 42)
0.51 (60.10)
Follicular carcinoma-II (n 5 26)
1.27 (60.28)
a
Tukey-Kramer’s honestly significant difference (HSD).
of FC than those with high Ki-67 LI (0.5–4.28) (P 5 0.033, log-rank test in Kaplan-Meier analysis). If all the distant metastases were lumped together regardless of the time when they were found at surgery or during follow-up period, the relation with Ki-67 LI became more significant (P 5 0.0007, v2 test) (Table III). Among 40 cases with the Ki-67 LI lower than 0.5, 39 (97.5%) were free of distant metastasis at surgery and during the follow-up period. On the other hand, among 28 cases with Ki-67 LI higher than 0.5, only 19 (67.9%) were free of distant metastasis and as many as 9 (32.1%) showed distant metastasis at surgery or during the follow-up period.
Table III. Ki-67 Labeling Index and Distant Metastasis in Follicular Carcinomas
Relationship Between TPO Expression and Ki-67 LI
negative TPO (n 5 21) positive TPO (n 5 47)
TPO expression had no relation with Ki-67 LI in all cases examined (P 5 0.18, t-test). In FA, the mean (6SE) Ki67 LI was 0.85 6 0.25 in negative cases for TPO expression; this was slightly higher than 0.40 6 0.07 in positive cases for TPO reactivity (P 5 0.043, t-test). There was no significant differences in FC-I (P 5 0.32, t-test) or in FC-II (P 5 0.89, t-test).
Cytological Features, TPO Expression, and Ki-67 LI All cases of FA had Cy-A in the preoperative FNA cytology. Fifty-four of 68 FCs had Cy-A, and 14 cases had Cy-B. Among FCs, Cy-B was more frequent in FC-II (38.5%; 10 of 26) than in FC-I (9.5%; 4 of 42). Among the 68 cases of FCs, Cy-B were more frequently observed in TPO-negative cases (38.1%) than in TPO-positive cases (12.8%) (P 5 0.017, v2 test) (Table IV) (Fig. 7). Among the 18 cases that had Cy-A and negative TPO, 13 cases (72%) were FCs. 206
Diagnostic Cytopathology, Vol. 43, No 3
Ki-67 labeling index
No distant metastasisa
Distant metastasisa
0–0.49 (n 5 40)
39 (97.5%)
1 (2.5%)
0.5–4.28 (n 5 28)
19 (67.9%)
9 (32.1%)
Significance levelb
a
At operation or during follow-up period. Pearson’s v2 test.
b
Table IV. TPO Expression and Cytological Features in Follicular Carcinomas Cytology A
Cytology B
13 (61.9%) 41 (87.2%)
8 (38.1%) 6 (12.8%)
P 5 0.017, Pearson’s v2 test.
In FCs, the higher was the Ki-67 LI, the higher was the rate of Cy-B appearance (P 5 0.029, logistic regression) (Fig. 8). In 94 Cy-A cases, 54 cases (57%) were FCs. When 38 cases with negative TPO or Ki-67 LI over 0.62 were extracted from them, as many as 28 cases (74%) were FCs, the rate of FCs was significantly higher than the rest in this group (P 5 0.0087, v2 test).
Discussion TPO is a hemo-containing glycoprotein that plays a key role in thyroid hormone biosynthesis.5 Since TPO reactivity disappears in parallel with de-differentiation of the follicular cell, immunocyto(histo)logical studies of TPO expression have been carried out for the diagnosis of follicular tumors.4–10,16–22 Increased de-differentiation,
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY
Fig. 7. Cytology-B, immunocytochemistry with TPO antibody. TPO was negative in most tumor cells. The positive rate was far lower than 80% (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Fig. 8. Cytology-B, immunocytochemistry with Ki-67 antibody (340). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
including loss of tissue-specific cell proteins such as TPO, is a well-known hallmark of thyroid tumor progression.23 High immunoreactivity for TPO was present in adenomatous goiter and follicular tumors, less reactivity in follicular or papillary carcinomas, and no reactivity in undifferentiated carcinomas.16 The reported results for TPO expression were various: some investigators underlined that immunocyto(histo)chemical study of TPO was useful for the diagnosis of follicular tumors,5–7,10,19 whereas some others stated it was not enough,8,9 or not useful for the diagnosis of those tumors.16–22 In the current study with 68 FCs and 40 FAs, FCs showed a significantly higher rate (30.9%) of TPO negatives than FAs (18.5%). The PPV of 80.8% was not high enough to provide reliable diagnosis, and the NPV was as low as 42.7%. TPO alone failed to discriminate FCs from FAs, which was concordant with the previous report.18 The current study shows that there are not only some cases that have negative TPO expression in FAs, but also some cases that show positive TPO expression in FCs. The WHO classification defines poorly differentiated carcinoma as mostly consisting of poorly differentiated components, such as trabecular, solid, and/or insular pattern.1 Therefore, tumors with only partial expression, but not prevalent of these components are classified into follicular carcinoma. In the present study, FCs were divided into two groups; FC-I have mostly follicular structure without such poorly differentiated components, and FC-II have partial but not prevalent expression of poorly differentiated components. The TPO reactivity was compared between those two FCs. The negative rates of TPO expression was 46.2% (12 of 26) in FC-II; this value was significantly higher than 21.4% (9 of 42) in FC-I (P 5 0.03) and FAs (P 5 0.002). The negative rate of FC-I
was close to the negative rate of FAs. Thus, FCs may be considered as consisting of two groups having different reactivity for TPO that could be expected to differ from each other in their biological behaviors. The FC-II had higher rate of distant metastasis and recurrence than FC-I, and TPO expression was inversely related with the progressive potential of FCs. There was a striking inverse relation between TPO reactivity and degree of cellular atypia as foci not expressing TPO corresponded to areas of high cellular or architectural atypia.7,10 In Bethesda system for thyroid follicular tumors, cytology should be categorized into “follicular tumor” or “suspect of follicular tumor” without benign/malignant judgment.24 In general, high cellularity, irregular cell stratification, high nuclear-cytoplasmic ratio, nuclear atypia, increased amount of chromatin, evident and/or prominent nucleoli are regarded as cytologically malignant-appearing features. Thus, thyroid FNA cytology can be divided into two groups; one with malignantappearing features (Cy-B), and the other one without (Cy-A). Looking at the cytological features of FCs, 38.1% (8 of 21) of negative cases for TPO expression were Cy-B, whereas, 12.8% (6 of 47) of positive cases for TPO expression were Cy-B. All the eight cases with negative TPO expression and Cy-B were FC-II. Thus negative TPO expression is significantly reflected in an atypical cell morphology and in a higher degree of malignancy. Determination of cell proliferative activity is one of the important factors for assessing the biological behavior of carcinomas. One of the useful markers to evaluate cell proliferative activity is Ki-67, which is expressed in almost all cells in the body except those in the G0 phase of cell cycle.25,26 The use of Ki-67 in the differential Diagnostic Cytopathology, Vol. 43, No 3
207
Diagnostic Cytopathology DOI 10.1002/dc
MARUTA ET AL. Table V. Clinical and Immunocyto(histo)chemical Features of Histological Types
Cytology-B group TPO negative Ki-67 labeling index Distant metastasisa a
Follicular adenoma, N 5 40
Follicular carcinoma-I, N 5 42
Follicular carcinoma-II, N 5 26
0% 12.5% 0.46 6 0.07 NA
9.5% 21.4% 0.51 6 0.10 7.1%
38.5% 46.2% 1.27 6 0.28 26.9%
At surgery or during follow-up period.
diagnosis of thyroid nodules is thought to be of limited usefulness, because Ki-67 LI is generally low in thyroid carcinomas, except that in undifferentiated (anaplastic) carcinoma and poorly differentiated carcinoma.27,28 Although the usefulness of Ki-67 LI has also been reported in thyroid follicular tumors,11,14,29 the values of Ki-67 LI were overlapping between FAs and FCs.12,13,30–32 In the current study, the mean Ki-67 LI (6SE) were 0.46 (60.07) in FAs and 0.80 (60.13) in all FCs. The statistic difference was marginal (P 5 0.06, ttest). The PPV was a moderate value of 75.8%, and the NPV was as low as 42.7%. When FCs were divided into two groups, i.e., FC-I and FC-II, the statistic difference became significant. Ki-67 LI were 0.51 (61.0) in FC-I and 1.27 (60.28) in FC-II. That is, FC-II showed significantly higher Ki-67 LI than FA (P 5 0.0007) and FC-I (P 5 0.014). Since the value of Ki-67 LI represents the proliferation activity of tumor cells, the above results indicate that many of the FC-II have high proliferative activity. Moreover, it was documented that Ki-67 expression was significantly higher in patients with FC with metastasis than in those without metastasis,29 Ki-67 LI also positively correlated with pathological stage in differentiated thyroid carcinomas15 and with distant metastasis at surgery in papillary carcinoma.31,33 In this study, 40 cases of FC showed a Ki-67 LI lower than 0.5. Only one (2.5%) of them gave distant metastasis. In contrast, among 28 cases that showed Ki-67 LI higher than 0.5, 9 (32.1%) cases gave distant metastasis. The odds ratio, (9 of 19)/(1 of 39), is as high as 18.5. These data indicate that Ki-67 LI strongly reflects the biological malignancy, especially in the propensity for distant metastasis. In the present study, high Ki-67 LI was correlated with Cy-B. Cytological features for Cy-B can be regarded to have a relation with biological malignancy. In 94 Cy-A cases, 54 cases (57%) were FCs. When 38 cases with negative TPO or Ki-67 LI over 0.62 were extracted from them, as many as 28 cases (74%) were FCs, the rate of FCs was significantly higher than the rest. In this series, FCs were divided into FC-I and FC-II. Our results indicated FC-II are prone to have cellular atypia, negative TPO expression, high Ki-67 LI and distant metastasis compared to FC-I (Table V). These facts suggest that the FC-II can be regarded as a distinct category from FC-I. 208
Diagnostic Cytopathology, Vol. 43, No 3
Although FCs and FAs cannot be separated by TPO or Ki-67 LI, FCs with more malignant features, such as cytological atypia, poorly differentiated components, distant metastasis, are prone to show the negative TPO and the higher Ki-67 LI. Identifying these FCs should be focused on in the future research. Addition of TPO stain and Ki-67 stain to routine Papanicolaou stain will improve the diagnostic reliability of FNA cytology for FC with high degree of malignancy.
References 1. Sobrinho-Simones M, Asa SL, Kroll TG, et al. Follicular carcinoma. In: DeLellis RA, Lloyd RV, Heitz PU, Eng C, editors. WHO pathology & genetics tumours of endocrine organs. Lyon, France: IARC Press; 2004. p 67–72. 2. Goldstein RE, Netterville JL, Burkey B, Johnson JE. Implications of follicular neoplasms atypia and lesions suspicious for malignancy diagnosed by fine-needle aspiration of thyroid nodules. Ann Surg 2002;235:656–664. 3. Taurog A, Dorris ML, Doerge DR. Mechanism of simultaneous iodination and coupling catalyzed by thyroid peroxidase. Arch Biochem Biophys 1996;330:24–32. 4. De Micco C, Ruf J, Chrestian MA, Gros N, Henry JF, Carayon P. Immunohistochemical study of thyroid peroxidase in normal, hyperplastic, and neoplastic human thyroid tissues. Cancer 1991;67: 3036–3041. 5. De Micco C, Zoro P, Garcia S, et al. Thyroid peroxidase immunodetection as a tool to assist diagnosis of thyroid nodules on fineneedle aspiration biopsy. Eur J Endocrinol 1994;131:474–479. 6. Henry JF, Denizot A, Porelli A, et al. Thyroperoxidase immunodetection for the diagnosis of malignancy on fine-needle aspiration of thyroid nodules. World J Surg 1994;18:529–534. 7. De Micco C, Vasko V, Garcia S, Zoro P, Denizot A, Henry JF. Fine-needle aspiration of thyroid follicular neoplasm: Diagnostic use of thyroid peroxidase immunocytochemistry with monoclonal antibody 47. Surgery 1994;116:1031–1035. 8. Pluot M, Faroux MJ, Flament JB, Patey M, Theobald S, Delisle MJ. Quantitative cytology and thyroperoxidase immunochemistry: new tools in evaluating thyroid nodules by fine-needle aspiration. Cancer Detect Prev 1996;20:285–293. 9. Faroux MJ, Theobald S, Pluot M, Patey M, Menzies D. Evalution of the monoclonal antibody antithyroperoxidase MoAb47 in the diagnostic decision of cold thyroid nodules by fine-needle aspiration. Pathol Res Pract 1997;193:705–712. 10. Garcia S, Vassko V, Henry JF, De Micco C. Comparison of thyroid peroxidase expression with cellular proliferation in thyroid follicular tumors. Thyroid 1998;8:745–749. 11. Pujani M, Arora B, Pujani M, Singh SK, Tejwani N. Role of Ki-67 as a proliferative marker in lesions of thyroid. Indian J Cancer 2001;47:304–307. 12. Cui W, Lu X, Zheng S, Ma Y, Liu X, Zhang W. The use of a combination of Ki-67, Galection-3, and PTTG can distinguish the benign and malignant thyroid tumor. Clin Lab 2012;58:419–426. 13. Heikkil€a A, Siironen P, Hagstr€ om J, et al. Follicular thyroid neoplasm: Clinicopathologic features suggesting malignancy. APMIS 2010;118:846–854. 14. Lacoste-Collin L, d’Aure D, Berard E, Rouquette I, Delisle MB, Courtade-Sa€ıdi M. Improvement of the cytological diagnostic accuracy of follicular thyroid lesions by the use of the Ki-67 proliferative index in addition to cytokeratin-19 and HBME-1 immunomarkers: A study of 61 cases of liquid-based FNA cytology with histological controls. Cytopathology 2014;25:160–169.
Diagnostic Cytopathology DOI 10.1002/dc
TPO AND KI-67 IN FOLLICULAR TUMOR CYTOLOGY 15. M€ussig K, Wehrmann T, Dittmann H, et al. Expression of the proliferation marker Ki-67 associates with tumour staging and clinical outcome in differentiated thyroid carcinomas. Clin Endocrinol 2012;77:139–145. 16. Lima MA, Gontijo VA, Santos MC, Schmitt FCL. Thyroid peroxidase expression in diseased human thyroid glands. Endocr Pathol 1999;10:223–228. 17. Savin S, Cvejic D, Isic T, Paunovic I, Tatic S, Havelka M. The efficacy of the thyroid peroxidase marker for distinguishing follicular thyroid carcinoma from follicular adenoma. Exp Oncol 2006;28: 70–74. 18. Savin S, Cvejic D, Isic T, et al. Thyroid peroxidase immunohistochemistry in differential diagnosis of thyroid tumors. Endocr Pathol 2006;17:53–60. 19. De Micco C, Savchenko V, Giorgi R, Sebag F, Henry JF. Utility of malignancy markers in fine-needle aspiration cytology of thyroid nodules: Comparison of Hector Battifora mesothelial antigen-1, thyroid peroxidase and dipeptidyl aminopeptidase IV. Br J Cancer 2008;98:818–823. 20. Weber KB, Shroyer KR, Heinz DE, Nawaz S, Said MS, Haugen BR. The use of a combination of galection-3 and thyroid peroxidase for the diagnosis and prognosis of thyroid cancer. Am J Clin Pathol 2004;122:524–531. 21. Savin S, Cvejic D, Isic T, Paunovic I, Tatic S Havelka M. Thyroid peroxidase and galectin-3 immunostaining in differentiated thyroid carcinoma with clinicopathologic correlation. Hum Pathol 2008;39: 1656–1663. 22. Paunovic I, Isic T, Havelka M, Tatic S, Cvejic D, Savin S. Combined immunohistochemistry for thyroid peroxidase, galection-3, CK19 and HMBE-1 in differential diagnosis of thyroid tumors. APMIS 2012;120:368–379. 23. Di Cristofaro J, Silvy M, Lanteaume A, Marcy M, Carayon P, De Micco C. Expression of tpo mRNA in thyroid tumors: Quantitative PCR analysis and correlation with alterations of ret, Braf, ras and pax8 genes. Endocr Relat Cancer 2006;13:485–495.
24. Ali CZ, Cibas ES. The Bethesda system for reporting thyroid cytopathology: Definitions, criteria, and explanatory notes. New York: Springer-Verlag; 2010:59–73. 25. Brown DC, Gatter KC. Monoclonal antibody Ki-67. Its use in histopathology. Histopathology 1990;17:489–503. 26. Cattoretti G, Becker MH, Key G, et al. Monoclonal antibodies against recombinant parts of the Ki-67 antigen (MIB 1 and MIB 3) detect proliferating cells in microwave-processed formalin-fixed paraffin sections. J Pathol 1992;168:357–363. 27. Aiad HA, Bashandy MA, Abdou AG, Zahran AA. Significance of AgNORs and Ki-67 proliferative markers in differential diagnosis of thyroid lesions. Pathol Oncol Res 2013;19:167–175. 28. Wallin G, B€ackdahl M, Christensson B, Grimelius L, Auer G. Nuclear protein content and Ki-67 immunoreactivity in nonneoplastic and neoplastic thyroid cells. Anal Quant Cytol Histol 1992;14:296–303. 29. Erickson LA, Jin L, Wollan PC, Thompson GB, van Heerden J, Lloyd RV. Expression of p27kip1 and Ki-67 in benign and malignant thyroid tumors. Modern Pathol 1998;11:169–174. 30. Kr olicka A, Kobierzycki C, Pula B, et al. Comparison of metallothionein (MT) and Ki-67 antigen expression in benign and malignant thyroid tumors. Anticancer Res 2010;30:4945–4949. 31. Sofiadis A, Tani E, Foukakis T, et al. Diagnostic and prognostic potential of MIB-1 proliferation index in thyroid fine needle aspiration biopsy. Int J Oncol 2009;35:369–374. 32. Tan A, Etit D, Bayol U, Altinel D, Tan S. Comparison of proliferating cell nuclear antigen, thyroid transcription factor-1, Ki-67, p63, p53 and high-molecular weight cytokeratin expressions in papillary thyroid carcinoma, follicular carcinoma, and follicular adenoma. Ann Diagn Pathol 2011;15:108–116. 33. Ito Y, Miyauchi A, Kakudo K, Hirokawa M, Kobayashi K, Miya A. Prognostic significance of Ki-67 labeling index in papillary thyroid carcinoma. World J Surg 2010;34:3015–3021.
Diagnostic Cytopathology, Vol. 43, No 3
209
