Fax: (011) 81 552739534
RET gene rearrangements (RET/PTC1 and RET/PTC3) in papillary thyroid carcinomas from an iodine-rich country (Japan)
Article first published online: 13 JUL 2005
Copyright © 2005 American Cancer Society
Volume 104, Issue 5, pages 943–951, 1 September 2005
How to Cite
Nakazawa, T., Kondo, T., Kobayashi, Y., Takamura, N., Murata, S.-i., Kameyama, K., Muramatsu, A., Ito, K., Kobayashi, M. and Katoh, R. (2005), RET gene rearrangements (RET/PTC1 and RET/PTC3) in papillary thyroid carcinomas from an iodine-rich country (Japan). Cancer, 104: 943–951. doi: 10.1002/cncr.21270
- Issue published online: 17 AUG 2005
- Article first published online: 13 JUL 2005
- Manuscript Accepted: 4 MAY 2005
- Manuscript Received: 17 FEB 2005
- papillary thyroid carcinoma;
- RET rearrangement;
- patient age;
- reverse transcriptase-polymerase chain reaction
The frequency of RET rearrangements (RET/PTC) in papillary thyroid carcinomas varies significantly according to geographic area, with the greatest incidence reported in the Belarus region, which is iodine-deficient and was contaminated severely after the Chernobyl reactor accident, and with the lowest incidence in iodine-rich, nonirradiated Japan. The authors investigated the prevalence of RET/PTC in a large number of thyroid tumors from Japanese patients.
Fresh and paraffin embedded tumor tissues from 215 Japanese patients were examined for RET rearrangements (RET/PTC1 and RET/PTC3) by means of reverse transcriptase-polymerase chain reaction analysis, with primers flanking the chimeric region, followed by direct-sequence analysis.
RET/PTC was found only in papillary carcinomas and was not observed in other histologic types of thyroid tumors. The overall frequency of RET/PTC in papillary carcinomas was 28.4%, with a greater frequency in younger patients, including 41.9% of younger patients age < 20 years, 27.6% of patients age 20–40 years, and 24.8% of patients age > 40 years. Among the patients in these 3 age groups, the prevalence rate of RET/PTC1 was similar, but RET/PTC3 was observed most frequently among patients age < 20 years. When the tumors were grouped further according to histologic subtypes, the prevalence of RET/PTC3 was greater in solid/solid-follicular papillary carcinomas than in classic papillary carcinomas.
The results indicated that RET/PTC may be useful as a specific molecular marker for papillary thyroid carcinomas. Furthermore, its incidence in such tumors was not low in Japanese patients, and it seemed to be associated with patient age. Therefore, the current results raise questions regarding the belief that the frequency of RET/PTC differs geographically and is especially low in Japan. Cancer 2005. © 2005 American Cancer Society.
Geographically, Japan is comprised of four main islands scattered in an arc northeast of the Asian continent. Nearly all Japanese belong to a single race, with similar diet patterns and preference for seaweed, such as sea tangles, in which iodine content is high. Among the histologic subtypes of thyroid carcinoma, the rate of papillary thyroid carcinoma is very high in Japan compared with other countries, and especially compared with iodine-deficient areas. It has been proposed that the large amount of dietary iodine may be affecting the high rate of papillary thyroid carcinoma in the Japanese population.1
Somatic rearrangements of the RET protooncogene (RET/PTC) represent the most common genetic alterations and are regarded as genetic markers of papillary thyroid carcinoma.2, 3 The prevalence of RET/PTC in papillary thyroid carcinomas varies significantly among patients from different geographic regions, with the reported frequencies ranging from 2.5% to 85%.4–12 Compared with other countries, a relatively low RET/PTC prevalence of 0–30%13–17 has been reported in Japan. However, the number of Japanese patients examined in these reports was limited, and the actual prevalence rate of RET/PTC for Japanese patients has not been established firmly.
A high proportion of RET/PTC has been reported in papillary thyroid carcinomas from irradiated children and young patients living in iodine-deficient areas, such as Belarus, Ukraine, and western regions of Russia,18–24 which suggests a causal link between radiation exposure and RET/PTC. However, the genetic characteristics of thyroid carcinomas that occur in children and young patients without radiation exposure have not been studied extensively.
To date, at least 10 different types of RET/PTC have been reported, of which RET/PTC1 and RET/PTC3, by far, are the most common.25RET/PTC1 results from the fusion between the tyrosine kinase domain of RET and H4 (D10S170) located on chromosome 10q2126; whereas RET/PTC3 forms during the fusion of the tyrosine kinase domain of RET with the ELE1, also known as RFG or ARA70, located on chromosome 10q11.2.27 It is interesting to note that several studies observed a high prevalence of RET/PTC3 in the solid or solid/follicular variant of papillary thyroid carcinomas from children who were exposed to radiation after the Chernobyl accident.18, 19, 23 However, for patients without radiation exposure, little is known regarding the histologic features of papillary thyroid carcinomas harboring RET/PTC3.
In the current study, to determine the actual prevalence of RET/PTC in thyroid tumors from nonirradiated Japanese patients on iodine-rich diets, and to elucidate the correlation between patient age and RET/PTC frequency, we examined RET/PTC1 and RET/PTC3 in a large series of thyroid tumors using a reverse transcriptase-polymerase chain reaction (RT-PCR) method with microdissection.
MATERIALS AND METHODS
We studied 215 surgical specimens (62 frozen tissues and 153 paraffin embedded tissues) from Japanese patients with thyroid tumors. They were collected from the files of the University of Yamanashi Hospital, Kofu Municipal Hospital, Tokyo Women's Medical University Hospital, and Ito Hospital. Informed consent had been obtained before resection in all patients. A thorough review of clinical data revealed nothing to suggest radiation exposure of the neck in any patient. The materials consisted of follicular adenomas (32 specimens), follicular carcinomas (10 specimens), papillary carcinomas (169 specimens), an undifferentiated carcinoma (1 specimens), and medullary carcinomas (3 specimens). In these specimens, diagnoses were made on the basis of the second edition of the Histological Typing of Thyroid Tumours (World Health Organization),28 and all specimens were reviewed by two pathologists (T.N. and R.K.). The patients with papillary thyroid carcinomas (169 patients) were divided into 3 groups according to their age (ages < 20 years, 20–40 years, and > 40 years) (Tables 1 and 2).
|Age||No. of patients||Positive for RET/PTC: no. (%)||Subtype of RET/PTC: No. (%)|
|PTC1||PTC3||PTC1 + PTC3|
|<20 yrs||31||13 (41.9)||9 (29.0)||3 (9.7)||1 (3.)|
|20–40 yrs||29||8 (27.6)||7 (24.1)||0 (0.0)||1 (3.4)|
|>40 yrs||109||27 (24.8)||24 (22.0)||2 (1.8)||1 (0.9)|
|Age/gender||No. of patients||Positive for RET/PTC: no. (%)|
Microdissection and RNA Extraction from Paraffin Embedded Tissue
Four serial sections, 10 μm thick, were cut from routinely processed, formalin fixed and paraffin embedded tissue blocks. These four sections were stained with hematoxylin after deparaffinization. Microscopically comparing the sections for orientation, each tumor tissue sample was microdissected with a disposable syringe needle under a stereomicroscope, as described previously.29 The tumor tissues collected by centrifugation were rehydrated with graded ethanol (100%, 80%, and 50%) and diethyl pyrocarbonate (DEPC)-treated water. The DEPC-treated water was removed, and the pellet was resuspended in 0.5 mL digestion buffer (0.1 mol/L Tris-HCl, 25 μmol/L ethylenediamine tetra-acetate, and 1% sodium dodecyl sulfate, pH 7.3) containing 500 μg proteinase K. Total RNA was extracted using the acid guanidinium-phenol-chloroform system (Isogen; Nippon Gene Company, Ltd., Toyama, Japan) according to the manufacturer's instructions.
RNA Extraction from Frozen Tissue
A fresh fragment (50 mg) of thyroid tissue was obtained after resection, immediately frozen, and stored at − 80 °C. The materials were homogenized with PT 10-35 homogenizer (Kinematica AG, Switzerland). Total RNA was prepared from the homogenized thyroid tissue using the acid guanidinium-phenol-chloroform system (Isogen; Nippon Gene Company) according to the manufacturer's instructions.
One microgram of total RNA was reverse-transcribed using Moloney murine leukemia virus reverse transcriptase (Wako Company, Ltd., Osaka, Japan) primed with random hexamers (Roche, Tokyo, Japan). All reverse-transcribed (RT) reactions were performed at 25 °C for 10 minutes, at 42 °C for 45 minutes, and at 95 °C for 5 minutes. After RT reactions, amplification of cDNA corresponding to RET/PTC1 and RET/PTC3 was performed by PCR using a thermal cycler (Zymoreactor II; ATTO Corporation, Tokyo, Japan). The primary and nested amplifications were performed as described previously.11 The sequences of nucleotide primers for RET/PTC 1 and RET/PTC 3 are provided in Table 3.
|Primer sequence||Positiona||Size (bp)|
|5′-GCT GGA GAC CTA CAA ACT GA-3′||279–298||165|
|5′-GTT GCC TTG ACC ACT TTT C-3′||425–443|
|5′-GCA CTG CAG GAG GAG AAC C-3′||307–325||85|
|5′-CCA AGT TCT TCC GAG GGA AT-3′||372–391|
|5′-ACC TGC CAG TGG TTA TCA AGC T-3′||693–714||154|
|5′-TTC GCC TTC TCC TAG AGT TTT TCC-3′||823–846|
|5′-CCA GGA CTG GCT TAT CCA AA-3′||738–757||80|
|5′-CCA AGT TCT TCC GAG GGA AT-3′||798–817|
|5′-GAA GGTGAA GGT CGG AGT C-3′||66–84||226|
|5′-GAA GAT GGT GAT GGG ATT TC-3′||272–291|
The bands of RET/PTC1 (6 tumors) and RET/PTC3 (2 tumors) in low-melting-temperature agarose gels were cut and purified by phenol/chloroform extraction and ethanol precipitation. The purified PCR products were sequenced by a Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems Division, Foster City, CA). The sequencing reaction mixture was incubated for 10 seconds at 96 °C, for 5 seconds at 50 °C, and for 4 minutes at 60 °C for 1 cycle, with a total of 25 cycles performed. Products were purified by spin column (Applied Biosystems) and analyzed on an ABI PRISM 310 genetic analyzer (Applied Biosystems).
Frequency of RET/PTC in Thyroid Tumors
We analyzed RET/PTC1 and RET/PTC3 in 215 thyroid tumors. Figures 1 and 2 show that, among the histologic types of thyroid tumors, RET/PTC was demonstrated only in papillary carcinomas and not in other histologic types of thyroid neoplasms. Neither RET/PTC1 nor RET/PTC3 was detected in nontumorous thyroid tissues, including 10 normal thyroids and 10 adenomatous goiters (data not shown). The overall prevalence rate of RET/PTC in papillary carcinomas was 28.4% (48 of 169 tumors), including 40 tumors with RET/PTC1 (23.7%), 5 tumors with RET/PTC3 (3.0%), and 2 patients with both RET/PTC1 and RET/PTC3 in 3 tumors (1.8%) (Table 4).
|Histologic type||No. of patients||Positive for RET/PTC: no. (%)||Subtype of RET/PTC: No. (%)|
|PTC1||PTC3||PTC1 + PTC3|
|Follicular adenoma||32||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
|Follicular carcinoma||10||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
|Papillary carcinoma||169||48 (28.4)||40 (23.7)||5 (3.0)||3 (1.8)|
|Undifferentiated carcinoma||1||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
|Medullary carcinoma||3||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
We performed direct sequencing studies in 8 RET/PTC-positive specimens (6 RET/PTC1-positive tumors and 2 RET/PTC3-positive tumors). Direct sequence studies confirmed these 2 rearrangements in all tumors and also revealed that 4 of 6 RET/PTC1-positive tumors had the T-to-G mutation in the H4 gene, 7 base pairs from the fusion point (Fig. 3).
Age Groups and RET/PTC in Papillary Carcinomas
We divided the 169 patients with papillary thyroid carcinomas into 3 age groups: age < 20 years (31 patients), age 20–40 years (29 patients), and age > 40 years (109 patients). The clinicopathologic features of the 31 patients age < 20 years with papillary carcinoma are summarized in Table 5. Patient age in this group ranged from 11 years to 19 years (mean, 17.2 years; 9 male patients and 22 female patients), and tumor size ranged from 7 mm to 86 mm (mean, 25.3mm). All but two patients had lymph node metastases, one patient had multiple metastases to the bilateral lungs, and there was local recurrence after surgery in two patients. The follow-up in 25 patients ranged from 19 months to 199 months (mean, 52.2 months), and all patients were alive, with or without disease.
|Patient no.||Age (yrs)||Gender||Tumor size (mm)||Histopathologic subtype||Lymph node metastasis||Distant metastasis||Recurrence||Follow-up (mos)||RET/PTC|
|26||18||Male||34||SV||−||−||−||32||− (PTC1 and PTC3)|
The RT-PCR method with microdissection revealed that the group of patients age < 20 age had greater proportions of tumors that contained RET/PTC1 and/or RET/PTC3 (41.9%), including RET/PTC1 in 9 tumors (29.0%), RET/PTC3 in 3 tumors (9.7%), and simultaneous detection of RET/PTC1 and RET/PTC3 in 1 tumor (3.2%). We detected RET/PTC1 and/or RET/PTC3 in 27.6% of tumors among patients age 20–40 years and in 24.8% of tumors among patients age > 40 years. Among the 3 age groups, the prevalence of RET/PTC3, either alone or with RET/PTC1, was greatest in the group age < 20 years (12.9%), whereas the prevalence of RET/PTC1 did not differ significantly.
Figure 4 shows the age distribution of RET/PTC in 31 papillary thyroid carcinomas from patients age < 20 years. RET/PTC1 was not identified in any of the 8 tumors from patients age < 17 years, whereas it was detected in 10 of 23 tumors (43.5%) from patients age 17–19 years. In contrast, RET/PTC3 was found in 2 of 8 tumors (25.0%) from patients age < 17 years and in 2 of 23 tumors (8.7%) from patients age 17–19 years.
Gender and RET/PTC in Papillary Carcinomas
The correlation between gender and RET/PTC is shown in Table 2. The frequency of RET/PTC tended to be higher in male patients than in female patients in the groups age 20–40 years and age > 40 years, whereas the frequency was nearly equal between males and females in the group age < 20 years. Statistical analysis failed to demonstrate significant differences in these age groups.
Histology and RET/PTC in Papillary Carcinomas
We classified the 169 papillary carcinomas from the current study into 5 histologic subtypes: the classic type (146 tumors), the follicular variant (8 tumors), the tall cell variant (4 tumors), the solid or solid-follicular variant (9 tumors), and the cribriform-morular variant (2 tumors) (Fig. 5). A summary of the prevalence of RET/PTC in these histologic subtypes of papillary thyroid carcinoma is provided in Table 6. The frequency of RET/PTC1 was 25.3% (37 of 146 tumors) in the classic type, 50.0% (4 of 8 tumors) in the follicular variant, 50.0% (2 of 4 tumors) in the tall cell variant, 55.6% (5 of 9 tumors) in the solid or solid-follicular variant, and 0% (0 of 2 tumors) in the cribriform-morular variant. Of the 9 papillary thyroid carcinomas of the solid or solid-follicular variant, RET/PTC1 was evident in 2 of 6 patients (33.3%) age < 20 years and in 1 of 3 patients (33.3%) age ≥ 20 years. In contrast, the prevalence of RET/PTC3 in this variant was high (3 of 6 patients; 50.0%) in the group age < 20 years, whereas none of the 3 solid or solid-follicular variant tumors in patients age ≥ 20 years were positive for RET/PTC3 (Table 7).
|Histologic subtype||No. of patients||Positive for RET/PTC: no. (%)||Subtype of RET/PTC: no. (%)|
|PTC1||PTC3||PTC + PTC3|
|Classic type||146||37 (25.3)||33 (22.6)||2 (2.1)||2 (2.1)|
|Follicular variant||8||4 (50.0)||4 (50.0)||0 (0.0)||0 (0.0)|
|Tall cell variant||4||2 (50.0)||1 (25.0)||1 (25.0)||0 (0.0)|
|Solid or solid-follicular variant||9||5 (55.6)||2 (22.2)||2 (22.2)||1 (11.1)|
|Cribriform-molular variant||2||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)|
|Histologic subtype||No. of Patients||Age <20 yrs||No. of patients||Age >20 yrs|
|PTC1||PTC3||PTC1 + PTC3||PTC1||PTC3||PTC1 + PTC3|
|Classic type||20||7 (35.0)||1 (5.0)||0 (0.0)||126||26 (20.6)||1 (0.8)||2 (1.6)|
|Follicular variant||3||1 (33.3)||0 (0.0)||0 (0.0)||5||3 (60.0)||0 (0.0)||0 (0.0)|
|Tall cell variant||1||0 (0.0)||0 (0.0)||0 (0.0)||3||1 (33.3)||1 (33.3)||0 (0.0)|
|Solid or solid-follicular variant||6||1 (16.7)||2 (33.3)||1 (16.7)||3||1 (33.3)||0 (0.0)||0 (0.0)|
|Cribriform-molular variant||1||0 (0.0)||0 (0.0)||0 (0.0)||1||0 (0.0)||0 (0.0)||0 (0.0)|
The dietary intake of iodine in the Japanese population is very high, possibly the highest in the world because of a preference for seaweed with a high iodine content. There is a possibility that dietary iodine may relate to genetic alterations. When we also considered that there is a high proportion of papillary carcinoma and a low proportion of follicular and undifferentiated thyroid carcinoma in Japan, we hypothesized that the difference in dietary iodine may relate to genetic alterations of the thyroid. Shi et al. reported that dietary iodine modulates the ras oncogene mutation and that, in iodine-deficient areas, ras oncogene activation may play a more important role in the initiation and/or maintenance of follicular tumors.32 Compared with other countries, Japan's reported prevalence rate of 0–30% for RET/PTC13–17 is low. However, the number of patients examined in these reports was small, and the actual prevalence rate of RET/PTC in Japanese is equivocal. In the current study, we examined a large number of thyroid tumors from Japanese patients and demonstrated RET/PTC only in papillary carcinomas. The RET/PTC was not present in other histologic types of thyroid tumors, including follicular adenoma, follicular carcinoma, undifferentiated carcinoma, and medullary carcinoma. The presence of RET/PTC strictly within papillary carcinoma supports the hypothesis that RET/PTC is a distinguishing and characteristic finding in genetic alterations of papillary thyroid carcinomas. Among 169 papillary carcinomas from Japanese patients, the prevalence rate of RET/PTC was 28.4%, which is consistent with the reported prevalence rates of RET/PTC in papillary carcinomas among Europeans and Americans (20–40%).3–6, 33, 34 From this information, we suggest that high iodine intake may not play an important role in RET rearrangements in the carcinogenesis of papillary thyroid carcinoma.
Ethnic and endemic background may influence the frequency of RET/PTC3 in papillary thyroid carcinomas, as evidenced by the wide variation of reported frequencies among individuals from different countries.4, 11, 12 The prevalence of RET/PTC3 in papillary thyroid carcinomas was reported as 3% in New Caledonia,11 9% in China,12 12% in Italy,4 and 15% in Australia.11 Radiation exposure affects the frequency of RET/PTC3: Nikiforov et al. investigated RET/PTC3 in thyroid carcinomas from children who were exposed to radiation as a result of the Chernobyl nuclear accident and reported an extremely high incidence (58%) of RET/PTC3.18 In the current study, the prevalence of RET/PTC3 in papillary carcinomas from Japanese patients without a history of irradiation was not high (4.8%).
We detected the simultaneous expression of RET/PTC1 and RET/PTC3 in 3 papillary thyroid carcinomas. Several studies have suggested a similar phenomenon.6, 11, 24, 31, 35 Sugg et al. suggested that normal thyrocytes that are subjected to the same genetic or environmental influences concurrently express different forms of RET/PTC,35 and Eng et al. reported a heterogeneous mutation of the RET protooncogene in subpopulations of medullary thyroid carcinoma.36 The simultaneous expression of RET/PTC1 and RET/PTC3 suggests that the additional variant of RET/PTC can arise as a progression within a single tumor or that papillary thyroid carcinoma can be of polyclonal origin. Although the former hypothesis is more likely, further studies to determine RET/PTC variants in subpopulations of papillary thyroid carcinoma will be necessary.
We performed a direct-sequence study in 6 specimens that showed RET/PTC1 and found a T-to-G mutation in the H4 region in 4 specimens. This mutation, which was described first in a papillary thyroid carcinoma cell line (TPC-1), is silent and causes no change in the amino acid.30 Subsequently, the same mutation also has been found in papillary thyroid carcinoma tissues.11, 12 Although it has been suggested that this mutation may be regarded as a genetic polymorphism, its significance remains unclear.
Papillary thyroid carcinomas in children or young adults are extremely rare in areas that are not associated with radiation exposure; therefore, it is difficult to study the frequency of RET/PTC in papillary thyroid carcinomas from young patients. To our knowledge, there have been only five studies on RET/PTC in pediatric and/or young adult patients with papillary thyroid carcinoma in areas without radiation exposure.17, 18, 31, 33, 37 Three studies from the U.S. and Italy suggested a greater prevalence of RET/PTC in children and young adult patients with papillary thyroid carcinomas compared with adult patients.18, 31, 33 In contrast, Williams et al. showed that 13 of 22 adult patients (59%) and 10 of 21 pediatric patients (48%) with papillary thyroid carcinoma in the U.K. showed evidence of RET activation.37 More recently, Motomura et al. described the frequency of RET/PTC in papillary thyroid carcinomas from Japanese children and found that it did not differ significantly the frequency observed in adult patients: 3 of 10 pediatric patients (30%) and 4 of 11 adult patients (36%) with adult papillary thyroid carcinomas.17 We examined the 31 papillary thyroid carcinomas from patients age < 20 years and found that this group had the highest prevalence rate (41.9%) of RET/PTC compared with the older age groups (27.6% for patients age 20–40 years and 24.8% for patients age > 40 years), which supports the concept that age is a factor in the thyroid-specific carcinogenic process.
In the current study, the prevalence of RET/PTC3 was greater in the younger patients, occurring in 12.9% of patients age < 20 years and in 2.7% of patients age ≥ 40 years. In contrast, there were no differences in the prevalence rates of RET/PTC1 among age groups. Thus, it is reasonable to suggest that RET/PTC3 may be related to patient age and may be important in the carcinogenesis of papillary thyroid carcinomas in young patients.
Patient gender is a significant factor that affects the survival of patients with papillary thyroid carcinoma. Several reports have revealed that male patients demonstrate a relatively poor survival rate.38, 39 In the current study, the prevalence rate of RET/PTC tended to be greater in male patients than in female patients in the older age groups, whereas there was no significant difference in the frequency of RET/PTC between males and females in the younger age group.
It has been suggested by some investigators that RET/PTC may be associated with histologic architecture in papillary thyroid carcinomas.6, 18, 19, 23 In the current study, therefore, papillary thyroid carcinomas were classified into five subtypes: classic type, follicular variant, tall cell variant, solid or solid-follicular variant, and cribriform-morular variant. In the classic type, the prevalence of RET/PTC1 and RET/PTC3 did not differ significantly between patients age < 20 years and patients age ≥ 20 years, and RET/PTC1 predominated in both age groups. It is interesting to note that a large proportion (50.0%) of the 6 solid or solid-follicular variant papillary thyroid carcinomas from patients age < 20 years demonstrated RET/PTC3, whereas no tumors of this type in patients age ≥ 20 years demonstrated RET/PTC3. Conversely, RET/PTC1 was represented equally (33.3%) in the 2 age groups. These results indicate that the classic type of papillary thyroid carcinoma may be associated with RET/PTC1 regardless of patient age. The solid or solid-follicular variant may be related to RET/PTC3 in tumors that occur in young patients, although the study included only a limited number of this type of papillary thyroid carcinoma.
The current study results indicate that the frequency of RET/PTC in papillary thyroid carcinoma is not low and presumably is age-related in iodine-rich, nonirradiated Japanese patients. Therefore, it is conceivable that a high iodine intake may not play an important role in RET rearrangements in the carcinogenesis of papillary thyroid carcinoma. In addition, these results raise questions regarding the belief that the frequency of RET/PTC differs geographically and is especially low in Japan.
The authors thank Ms. Miyuki Ito and Ms. Mikiko Yoda for their technical help.
- 28Histological classification of thyroid tumours. In: HedingerC, WilliamsED, SobinLH, editors. Histological typing of thyroid tumours. World Health Organization international histological classification of tumours, 2nd edition. Paris: Springer-Verlag, 1988: 3–18., , .