European Journal of Endocrinology (2004) 151 779–786

ISSN 0804-4643

EXPERIMENTAL STUDY

Thyroid follicular adenomas may display features of follicular carcinoma and follicular variant of papillary carcinoma Vasily V Vasko1,5, Jean Gaudart2, Claude Allasia3, Victoria Savchenko4, Julie Di Cristofaro1, Motoyasu Saji5, Matthew D Ringel5,6 and Catherine De Micco1,3 1 INSERM U 555, 2Department of Bio Statistic, 3Laboratory of Pathology, Faculty of Medicine, Mediterranean University, Marseilles, France, 4Center for Endocrine Surgery, Kiev, Ukraine, 5MedStar Research Institute/Washington Hospital Center, Washington, DC, USA and 6The Ohio State University and Arthur G. James Cancer Center, Columbus, Ohio, USA

(Correspondence should be addressed to Vasily V Vasko, Washington Hospital Center/MedStar Research Institute, 108 Irving Street, NW Room 205, Washington, DC 20010, USA; Email: [email protected])

Abstract Thyroid follicular adenomas (FA) are encapsulated tumors lacking vascular, capsular or lymphatic invasion and the typical nuclear features of papillary carcinoma (PC). However, some FA demonstrate nuclear atypia reminiscent of either follicular carcinomas (FC) or follicular variant of papillary carcinomas (FVPC), suggesting they may represent precursors of malignant transformation. We hypothesized that an objective evaluation of nuclear chromatin patterns could be used to define atypical follicular tumors (AFT) that are likely to be premalignant. To test this hypothesis, we used a computer-aided image analysis system to define the chromatin pattern of nuclei from thyroid tumors. To validate the system, we analyzed 3000 nuclei from 10 FA, 10 FC, and 10 FVPC samples and accurately distinguished between these classes of tumors. Then, we analyzed nine AFT and, in parallel, we analyzed the tumors for activating mutations of N2-RAS and over-expression of RET. The predominant chromatin pattern of AFT was of FA type in two cases, FC type in two cases, and PC type in three cases. One case contained similar numbers of FC and PC nuclei and one was comprised of a mixture of the three nuclear types. Neither RAS mutation nor RET overexpression were detected in FA. N2-RAS mutations were found in 33% of AFT, 20% of FC and 20% of FVPC without correlation with chromatin pattern. Over-expression of RET was detected in 45% of AFT, 20% of FC and 50% of FVPC and was correlated with PC nuclei. These results show that AFT are a heterogenous group of tumors, containing genuine benign tumors and tumors that share morphological and molecular features with follicular and papillary carcinomas that might be precursors of both types of thyroid carcinomas. European Journal of Endocrinology 151 779–786

Introduction Thyroid follicular adenomas (FA) are encapsulated thyroid neoplasms lacking the characteristic invasiveness of follicular carcinoma, and the typical nuclear features of papillary cancers (1, 2). The term ‘atypical adenoma’ was used by Hazard and Kenyon (3) to describe FA that departed from the norm by a solid architecture but did not fulfil the criteria for carcinoma after examination of multiple tissue blocks (3 – 5). The WHO’s committee later extended the term of atypical adenoma to include FA with unusual cellular density, mitoses, and a less regular cytological pattern (1). In practice, these atypical adenomas were frequently confused with follicular variant of papillary cancer (FVPC) in which encapsulated tumors have follicular architecture and evidence of papillary nuclear changes with or without tumor invasion. In an attempt to alleviate this confusion, and because the ‘typical nuclear features of

q 2004 Society of the European Journal of Endocrinology

papillary carcinoma (PC)’ are less distinctive in tumors with follicular architecture (6), the term atypical adenoma was further extended to include atypical follicular tumors (AFT) possessing only limited nuclear features of PC (LNFPC) that did not appear to meet the criteria for FVPC. Finally, as stated in the Third Series Fascicle of Tumors of the Thyroid Gland (2), the term atypical adenoma became ‘a wastebasket term in which any adenoma that looks worrisome may be included’. To avoid the swelling of such an imprecise diagnostic class and relevant therapeutic wavering, the authors advocated its replacement by hypercellular adenoma, thus compelling pathologists to arbitrarily place AFT into a benign class. However, some encapsulated follicular tumors with LNFPC may show distant metastasis and many pathologists are concerned about overlooking a malignant tumor (7). This concern may explain the tendency to over-diagnose FVPC, and the existence of considerable

Online version via http://www.eje.org

780

V V Vasko and others

inter-observer variation in the interpretation of nuclear features of papillary carcinomas (8, 9). Facing this confusion, the Chernobyl Pathologists Group suggested more recently that AFT be referred to as ‘well differentiated tumors of uncertain malignant potential’ (10). This updating of the old concept of atypical adenoma recognizes the existence of tumors that share a number of characters with thyroid carcinomas and may be premalignant in nature, although there is at the present time no direct proof for that concept (11 – 15). The development of papillary carcinomas (PC) is correlated with activation of the RET oncogene, mostly due to the presence of RET/PTC gene rearrangements (16 – 18), whereas RAS oncogene mutations, more specifically in codon 61 of N-RAS (N2-RAS), are incriminated in the development of follicular carcinomas (FC) (19, 20). RET activation was shown to parallel the nuclear changes of papillary cancers (PC-NC) on histology (21). However, the interpretation of the chromatin pattern on histology is highly subjective (22, 23). Computerized morphometry can significantly improve the analysis of nuclear characters of various subtypes of follicular tumors, and, when analyzed carefully, can provide objective criteria to aid in the correct diagnosis of thyroid tumors (24 –29). To gain insights into the biological meaning of nuclear alterations in atypical thyroid follicular tumors, we performed a computer-assisted morphometric analysis of digitalized images of nuclei from cells of typical and atypical follicular tumors. Then we compared the results to the presence of N2-RAS mutation and RET activation.

Materials and methods Patients and tissues Tumors were classified according to WHO criteria (1) by three independent pathologists: paraffin-embedded specimens of 10 follicular adenomas (FA), 9 atypical follicular tumors (AFT), 10 follicular carcinomas (FC) and 10 follicular variant of papillary carcinomas (FVPC) were obtained from 39 patients operated on in the Ukrainian Center of Endocrine Surgery (Kiev). None of the nine patients with AFT had any clinical features of thyroid malignancy. The information on 5 years follow-up was available in 4 cases of AFT, and no local recurrences or metastases were detected in these patients. The Institutional Review Board approved the protocol for this study and informed consents were obtained from all patients enrolled.

Histomorphometry Image analysis of nuclei was carried out on a SAMBA2000 cell image processor (System for Analytical Microscopy in Biological Application, SAMBA INC, www.eje.org

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

Grenoble, France). The fields of vision were selected randomly and approximately 100 nuclei of follicular cells were measured on each slide through a £ 100 objective. Sixteen features were computed from each nucleus (Table 1). The exact mathematical definition of these indices is provided in the SAMBA-2000 manual and the image processor software. These features were subdivided into three groups of parameters: (i) size and shape features (calculated from a previously found object contour), (ii) densitometric features (calculated from a previously computed histogram and restricted to the optical density distribution of the nuclear image points) and (iii) textural features (calculated by means of the gray level co-occurrence matrix method or computed by means of the run-length section method). Typical tumors (FA, FC and FVPC) where used in a first step to calibrate the statistical model of analysis. AFT where then analyzed with this system.

Statistical analysis The data were analyzed by the SPSS statistical package (SPSS, Chicago, IL, USA). The values of all quantitative variables were examined for normality by the nonparametric Kolmogorov– Smirnov test. The variables that did not demonstrate normal distribution were subjected to log transformation. The assumption of normality permitted application of parametric methods in the next step. A database was first constructed cumulating the parameters of 3000 nuclei, including 100 nuclei from thirty cases: ten typical FA (FAc), FC (FCc) and FVPC (FVPCc) samples. A discriminant analysis with forward stepwise selection of variables based on minimization of Wilk’s lambda was performed in order to determine variables able to distinguish the nuclei from FA, FC Table 1 Morphometric parameters computed on each cell image. The features were divided into three groups: 1 –5, size and shape; 6 and 7, densitometry; 8 –16, texture. N

Parameters

1 2 3 4 5

NA Nuclear area SD Smallest diameter GD Greatest diameter FF Form factor BE Bending energy

6 7

IOD Integrated optical density MOD Mean optical density

8 9 10 11 12 13 14 15 16

LM ENE ENT INE SRE LRE RLD RLP GLD

Local mean Energy Entropy Inertia Short run-length emphasis Long run-length emphasis Run-length distribution Run-length percentage Grey level distribution

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

and FVPC. Linear discrimination functions using these variables were computed in 3 formulae defining the three types of cells: FAc ¼ 22723:580 2 22:589 £ SD 2 67:9 £ log GD 2 38:956 £ log FF þ 55:464 £ log BE þ 227:329 £ log IOD 2 0:911 £ MOD þ 88:257 £ ENT þ 170:456 £ log ENE þ 189:181 £ log LRE 2 120:119 £ log RLD þ 16:290 £ RLP 2 184:505 £ log INE FCc ¼ 22760:200 2 22:570 £ SD 2 62:813 £ log GD 2 37:918 £ log FF þ 52:794 £ log BE þ 229:210 £ log IOD 2 0:905 £ MOD þ 90:090

Thyroid adenomas as precursors of carcinomas

781

forward 50 -TCT TAC AGA AAA CAA GTG GT-30 ; reverse 50 -GTA GAG GTT AAT ATC CGC AA-30 . The PCR was performed in 50 ml of mixture containing 0.5 mg genomic DNA, 1.5 ml MgCl2 PCR buffer concentrated (Qiagen), 200 nmol/l deoxyribonucleoside triphosphate, 200 nmol/l of each primer and 1.25 units HotStar Taq DNA Polymerase (Qiagen). HotStar Taq DNA Polymerase was activated by incubation for 15 min at 95 8C followed by 35 cycles of PCR comprising, for each cycle, 1 min denaturation at 94 8C, 1 min annealing at 53 8C and 1 min extension at 72 8C. PCR products were electrophoresed on 3% agarose gel and stained by ethidium bromide. Direct sequencing was performed with an Applied Biosystem 373XL sequencer (PE Applied Biosystems, Foster City, CA, USA), according to the manufacturer’s protocol.

£ ENT þ 171:846 £ log ENE þ 191:743 £ log LRE 2 119:204 £ log RLD þ 16:385 £ RLP 2 183:277 £ log INE FVPCc ¼ 22753:358 2 21:927 £ SD 2 55:480 £ log GD 2 39:142 £ log FF þ 52:837 £ log BE þ 227:216 £ log IOD 2 0:890 £ MOD þ 85:277 £ ENT þ 168:589 £ log ENE þ 188:177 £ log LRE 2 122:363 £ log RLD þ 16:299 £ RLP 2 185:294 £ log INE where SD, smallest diameter; GD, greatest diameter; FF, form factor; BE, bending energy; IOD, integrated optical density; MOD, mean optical density; ENT, entropy; ENE, energy; LRE, long run-length emphasis; RLD, run-length distribution; RLP, run-length percentage; INE, inertia. These formulae were used to classify the nuclei from the cumulated database. A cross validation test was carried out afterwards on the 30 individual tumors (10 FA, 10 FC and 10 FVPC) to check the goodnessof-fit of the method. Then, we used this statistical model to study the whole population of 900 nuclei derived from all 9 AFT, and then to classify the 9 tumors individually using the Mann – Whitney U-test to compare data.

DNA extraction and PCR amplification of exon 2 of N-RAS After microdissection of tumors on slides, tissues were suspended in 400 ml extraction buffer containing 100 mmol/l Tris– HCl, pH 8.0, 100 mmol/l NaCl, 20 mmol/l EDTA, 2% sodium dodecyl sulfate and 200 mg/ml proteinase K (Boeringer Mannheim). DNA was extracted using QI-Amp DNA mini kit (Qiagen S.A.) according to the manufacturer’s instructions. N-RAS gene was amplified using the following primers:

Immunohistochemical staining of RET protein Sections were dewaxed, soaked in alcohol and incubated in 3% hydrogen peroxide for 15 min to inactivate endogeneous peroxidase activity after microwave treatment in antigen unmasking solution (Vector Lab, Burlingame, CA, USA). Then, sections were incubated overnight at 4 8C with polyclonal anti-RET antibody (C 19) from Santa Cruz Biotechnology (Santa Cruz, CA, USA) at a dilution of 1:100. This antibody recognizes the tyrosine kinase domain of RET which is expressed by both RET/PTC and wild-type RET genes. Immunostaining was performed with Vectastain Universal Quick kit (Vector Lab) according to the manufacturer’s instructions. Peroxidase activity was revealed in 3,3 diaminobenzidine. Negative control was performed by omission of anti-RET antibodies.

Results Morphometry on typical tumors Discriminant analysis selected 12/16 most discriminative variables: 4 size and shape variables (SD, GD, FF, BE), 2 densitometric variables (IOD, MOD) and 6 texture variables (ENT, ENE, LRE, RLD, RLP, INE). On the whole population of 3000 cells the model allowed discrimination between 3 types, i.e. cells typical for FA (FAc), cells typical for FC (FCc) and cells typical for PC (PCc). The Wilk’s lambda was 0.317, which corresponded to a significance level of P , 0.0001. The application of the computed formulae defining the three cellular types to the 3000 pooled cells gave a correct reclassification of 73% of them into predicted types (Fig. 1): 1000 cells derived from FA were made up of 75.5% FAc, 15.6% FCc, and 8.9% PCc; in 1000 cells derived from FC, 56.8% were FCc, 27% were FAc, and 16.2% were PCc; among 1000 cells derived from FVPC, 88% were PCc, 11% were FCc, and 1% were FAc. The morphological features of nuclei typical for FCc or PCc are shown in Fig. 2 (A and B respectively). www.eje.org

782

V V Vasko and others

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

RET immunohistochemistry Immunostaining with anti-RET antibody was positive in 0/10 FA (0%), 2/10 FC (20%), 5/10 FVPC (50%) and in 4/9 AFT (45%). Staining was moderate or low, located in the cytoplasm. Mann–Whitney U-test revealed a significant correlation between a positive RET immunostaining and the number of PCc in examined tumors (P , 0.001). Conversely, the number of FAc was inversely correlated with RET immunostaining (P , 0.007). Three tumors (1 AFT, 1 FC and 1 FVPC) carried both the N2-RAS mutation and positive RET immunostaining. In these cases, only the areas with clear nuclei were positive for RET. Figure 1 Results of morphometric reclassification of cells derived from follicular adenomas (FA), follicular carcinomas (FC), follicular variant of papillary carcinomas (FVPC) and atypical adenomas (AA). PCc, cells typical for follicular variant of papillary carcinomas; FCc, cells typical for follicular carcinomas; FAc, cells typical for follicular adenomas.

The study of individual tumors showed that 9/10 FA demonstrated predominance of FAc (range: 65% – 97%). One FA had 5% FAc, 35% FCc and 55% PCc. In 7/10 FC, the predominance of FCc was evident (range: 53% –83%). In 2 FC about equal rates of FCc and PCc were observed, and in one minimally invasive FC the FAc were predominant (89%). In 10/10 FVPC, the PCc predominated (range: 64% – 100%).

Morphometry on AFT In the pooled 900 cells derived from all nine AFT, approximately equal rates of FAc (27.9%), FCc (33.8%) and PCc (38.3%) were observed. Among the nine AFT, 2 cases exhibited a predominance of FAc (74% and 93%), 2 cases a predominance of FCc (52% and 98%) and 3 cases a predominance of PCc (from 52% to 82%). In one case, similar rates of FCc and PCc were observed (46% and 49% respectively). The last case contained 24% FAc, 30% FCc and 46% PCc. The typical morphological features of nuclei derived from ATF with a predominance of FCc (C) or PCc (D) are shown in Fig. 2.

N2-RAS mutations N2-RAS mutations of codon 61 were found in 0/10 FA (0%), 2/10 FC (20%), 2/10 FVPC (20%) and in 3/9 AFT (33.3%). In all cases, mutations were transitions from A to G. Histological re-examination did not reveal special morphological features in tumors containing mutations. Mann– Whitney U-test did not show a significant relationship between the prevalence of N2-RAS mutations and the proportion of FAc, FCc and PCc in the tumors. www.eje.org

Discussion As stressed by half a century of debate regarding the proper terminology for AFT, the actual meaning of nuclear atypia in encapsulated follicular thyroid tumors remains enigmatic. Nevertheless, nuclear atypia is correlated with malignancy in thyroid cytology (30). As a consequence, the increasing use of fine needle aspiration cytology (FNA) for the exploration of thyroid nodules has resulted in an increased proportion of both carcinomas and AFT in surgical specimens. Indeed, AFT accounted for 2.7% to 9.8% of follicular adenomas in the earliest studies (3, 4), and 4% in our institution before the introduction of FNA. Since the use of FNA in our surgical series, 14% of thyroid adenomas exhibit atypical features upon histological examination; thus the need for a better understanding of the actual meaning of nuclear atypia in follicular adenomas is growing. In a previous study we demonstrated that N2-RAS mutation in codon 61, which is significantly correlated with follicular carcinogenesis, is absent in common adenomas but is detected at similar rates in atypical adenomas (23%) and in follicular carcinomas (17.6%) (20). On the other hand, Fusco et al. demonstrated that, in adenomatous nodules, microscopic foci with LNFPTC showed positive RET immunoreactivity (21). More recently, mutations of p53 were detected in the bizarre cells of some atypical adenomas but not in the bland-looking cells (31). Taken together, these observations suggest that AFT may precede the development of thyroid carcinomas. As the distinction between follicular and papillary thyroid carcinomas relies on the analysis of nuclear features, we postulated that the nuclei of atypical adenomas should possess – depending on the cases - the ‘papillary’ (fine), or the ‘follicular’ (coarse) nuclear chromatin phenotype (23). In order to explore this hypothesis we performed a computer-aided image analysis of nuclear chromatin in 39 typical and atypical follicular tumors and we looked, in parallel, at the presence of N2-RAS mutations at codon 61, and of RET oncogene activation.

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

Thyroid adenomas as precursors of carcinomas

783

Figure 2 Morphological features of cells derived from follicular carcinomas (A), follicular variant of papillary carcinomas (B) and atypical adenomas with a predominance of FCc (C) or PCc (D).

Using nuclei of typical follicular tumors as a reference group we calibrated a method of analysis that distinguished the nuclei of cells from follicular adenomas (FAc), follicular carcinomas (FCc) and papillary carcinomas (PCc). This system successfully diagnosed malignancy in 19/20 carcinomas and was able to recognize the histological subtypes of 26/30 typical follicular tumors. Its sensitivity for the diagnosis of malignancy (95%) matched previous results (26). The accuracy for the diagnosis of tumor type was 86.6%, confirming studies showing that it is possible to distinguish nuclei of FA, FC and PC by morphometric analysis

(29). Interestingly, when applied to the analysis of typical follicular tumors, the three types of nuclei (FAc, FCc, PCc) were all present in varying proportions in the three classes of tumors (FA, FC and FVPC), although with a majority of corresponding nuclei in 26/30 cases. This finding underlines the heterogeneity of follicular tumors, even when they are considered ‘typical’ by histology. We postulated that this heterogeneity underlies the overlapping of nuclear features between various subtypes of benign and malignant thyroid follicular tumors. The results obtained from the study of 9 AFT confirmed this hypothesis. www.eje.org

784

V V Vasko and others

Indeed, we found that only two AFT had a phenotype of true adenoma with a predominance of FAc, whereas 7 cases were made predominantly of FCc and/or PCc, thus exhibiting nuclear characteristics of malignant tumors: FCc were predominant in two cases and PCc in three. Interestingly, in two cases of AFT and in two cases of FC we observed an approximately equal percentage of FCc and PCc cells and thus the ‘follicular’ or ‘papillary’ phenotype remained undetermined. These tumors are reminiscent of the recently described ‘hybrid’ follicular and papillary thyroid carcinomas, thus corroborating the potential existence of tumors sharing characters with both classes of carcinoma (23). N2-RAS mutations were not found in FA, but occurred with a similar rate in AFT (33%), FC (20%) and FVPC (20%). This corroborates our previous results on AFT supporting the idea that some of them may be precursors of thyroid cancer with follicular architecture (20). The identification that some AFT are comprised of mostly FCc and PCc by morphometry, and the presence of a similar rate of N2-RAS mutations in AFT, FC, and FVPC are consistent with two recent reports suggesting a correlation between the follicular pattern of growth in papillary cancers and RAS activation (32, 33). Indeed, RAS mutations are uncommon in classical papillary cancers and usually involve mutations in the gene encoding K-RAS rather than N2-RAS (20). The presence of the same activating N2-RAS mutation in atypical follicular tumors and clearly malignant tumors with follicular architecture suggests that the activity of this gene might sustain follicular architecture independent of the nuclear phenotype. In this study, we found RET expression in a similar proportion of AFT and FVPC (45% and 50%). RET expression was previously identified by immunohistochemistry in approximately 40 to 86% of papillary cancers with antibodies directed against its tyrosine-kinase (TK) domain (34 –38). RET immunoreactivity was found to correlate with the expression of the TK domain of RET by RT-PCR (39 – 41). This expression may be due to the presence of activating RET/PTC rearrangements or may correspond to the activity of wild-type RET (34, 42, 43). Alternatively spliced RET transcripts were also found in PC (44). Differences in the frequency of RET activation are also related to geographical and epidemiological parameters (45). Indeed the highest rates of RET immunoreactivity were found in tumors occurring after external or internal exposure to radiation (38). RET/PTC rearrangements were also found in post-Chernobyl benign nodules and in benign thyroid nodules exposed to external radiation (45, 46). The high frequency of RET expression in FVPC in our series may thus be explained by the inclusion of tumors from patients living in areas of Ukraine contaminated after the Chernobyl accident. In benign tumors we did not find RET activation in FA devoid of nuclear atypia, whereas AFT had a similar www.eje.org

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

proportion of RET immunoreactive cells as FVPC. This observation is in accordance with the recently reported presence of RET/PTC translocations in follicular tumors exhibiting incomplete features of papillary carcinomas (21). It may explain some discordant results in previous studies on RET activation in follicular adenoma depending on whether or not tumors with nuclear atypia or LNFPC were included (36, 37). In AFT as in PC, RET immunoreactivity was specifically correlated with the rate of PCc, in accordance with previous demonstrations that nuclear morphology of papillary carcinomas with clearing of the chromatin and nuclear envelope irregularity may be induced in thyroid cells by RET activation (47, 48). The presence of populations of PCc expressing the TK domain of RET in some AFT reasonably suggests that some follicular variant of papillary cancers may arise within pre-existing adenoma. Interestingly, in three cases (1 AFT, 1 FC and 1 FVPC) both RAS mutation and RET expression were found within the same tumor, suggesting that the molecular events underlying the ‘follicular’ and the ‘papillary’ pathway of thyroid carcinogenesis may not be entirely exclusive. This may also explain why, in some cases, the follicular or papillary subtype of a thyroid tumor cannot be asserted: the term ‘differentiated carcinomas, no other specification’ (NOS) was proposed by the Armed Force Institute of Pathology (AFIP) as a name for these tumors (2). We conclude from this study that the group of thyroid follicular adenomas is heterogeneous and includes genuine benign tumors devoid of any features of carcinomas, and atypical tumors that possess both morphological and genetic features of pre-invasive malignancies. These atypical tumors can share cellular morphology and oncogene expression with either follicular carcinoma or papillary carcinoma, suggesting that they can precede either tumor type. This interpretation is consistent with the classical clinical observation of follicular carcinoma occurring in a setting of longstanding thyroid nodules. It may also explain the nodular and encapsulated character of most FVPC as well as their slow growth. In view of the ongoing controversies concerning the terminology of follicular tumors (7 –10), our study brings forward objective arguments for the existence of an intermediate step between adenoma and carcinoma. Whatever the name that will be finally chosen to call these tumors there is a need to separate them from common adenomas in order to analyze predictive factors and to collect follow-up data that will help physicians and surgeons to manage them in the most appropriate way.

Acknowledgement This work was funded by a clinical research contract from Marseille Public Hospital System.

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

References 1 Hedinger C. Histological Typing of Thyroid Tumours, 2nd edition. Berlin: Springer-Verlag, 1988. 2 Rosai J, Carcangiu ML & De Lellis RA. Tumors of the thyroid gland. In Atlas of Tumor Pathology, Fascicle 5, Series 3, pp 161–182. Washington, DC: Armed Force Institute of Pathology, 1992. 3 Hazard JB & Kenyon R. Atypical adenoma of the thyroid. American Medical Association Archives of Pathology 1954 58 554– 563. 4 Lang W, Georgii G, Stauch G & Kienzle E. The differentiation of atypical adenomas and encapsulated follicular carcinomas in the thyroid gland. Virchows Archive: A Pathological Anatomy and Histopathology 1980 385 125–141. 5 Baloch ZW & Livolsi VA. Follicular-patterned lesions of the thyroid: the bane of the pathologist. American Journal of Clinical Pathology 2002 117 143–150. 6 De Micco C, Vassko V & Henry JF. Interest of thyroid peroxidase immunocytochemistry in the preoperatory fine needle aspiration diagnosis of follicular variant of thyroid papillary cancers. Surgery 1999 126 1200–1204. 7 Baloch ZW & LiVolsi VA. Encapsulated follicular variant of papillary thyroid carcinoma with bone metastases. Modern Pathology 2000 13 861–865. 8 Renshaw AA & Gould EW. Why there is the tendency to ‘overdiagnose’ the follicular variant of papillary thyroid carcinoma. American Journal of Clinical Pathology 2002 117 19– 21. 9 Chan JK. Strict criteria should be applied in the diagnosis of encapsulated follicular variant of papillary thyroid carcinoma. American Journal of Clinical Pathology 2002 117 16–18. 10 Williams ED. Guest Editorial. Two proposals regarding the terminology of thyroid tumors. International Journal of Surgery and Pathology 2000 8 181 –183. 11 Hemmer S, Wasenius VW, Knuutila S, Joensuu H & Franssila K. Comparison of benign and malignant follicular thyroid tumours by comparative genomic hybridization. British Journal of Cancer 1998 78 1012–1017. 12 Zedenius J, Wallin G, Svensson A, Bovee J, Hoog A, Backdahl M & Larsson C. Deletions of the long arm of chromosome 10 in progression of follicular thyroid tumors. Human Genetics 1996 97 299– 303. 13 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. 14 Coli A, Bigotti G, Zucchetti F, Negro F & Massi G. Galectin-3, a marker of well- differentiated thyroid carcinoma, is expressed in thyroid nodules with cytological atypia. Histopathology 2002 40 80 –87. 15 Vassko V, Garcia S, Henry JF & De Micco C. Expression of proliferating cell nuclear antigen in follicular thyroid tumors: correlation with clinicopathological findings. Oncology Reports 1999 6 359– 364. 16 Grieco M, Santoro M, Berlingieri MT, Melillo RM, Donghi R, Bongarzone I, Pierotti MA, Della Porta G, Fusco A & Vecchio G. PTC is a novel rearranged form of the ret protooncogene and is frequently detected in vivo in human thyroid papillary carcinomas. Cell 1990 60 557–563. 17 Santoro M, Melillo RM, Carlomagno F, Fusco A & Vecchio G. Molecular mechanisms of RET activation in human cancer. Annals of the New York Academy of Sciences 2002 963 116 –121. 18 Kimura ET, Nikiforova MN, Zhu Z, Knauf JA, Nikiforov YE & Fagin JA. High prevalence of BRAF mutations in thyroid cancer: genetic evidence for constitutive activation of the RET/PTC-RAS-BRAF signaling pathway in papillary thyroid carcinoma. Cancer Research 2003 63 1454–1457. 19 Lemoine NR, Mayall ES, Wyllie FS, Williams ED, Goyns M, Stringer B & Wynford-Thomas D. High frequency of ras oncogene activation in all stages of human thyroid tumorigenesis. Oncogene 1989 4 159 –164.

Thyroid adenomas as precursors of carcinomas

785

20 Vasko V, Ferrand M, Di Cristofaro J, Carayon P, Henry JF & De Micco C. Specific pattern of ras oncogene mutations in follicular thyroid tumors. Journal of Clinical Endocrinology and Metabolism 2003 88 2745– 2752. 21 Fusco A, Chiappetta G, Hui P, Garcia-Rostan G, Golden L, Kinder BK, Dillon DA, Giuliano A, Cirafici AM, Santoro M, Rosai J & Tallini G. Assessment of RET/PTC oncogene activation and clonality in thyroid nodules with incomplete morphological evidence of papillary carcinoma: a search for the early precursors of papillary cancer. American Journal of Pathology 2002 160 2157–2167. 22 Hirokawa M, Carney JA, Goellner JR, DeLellis RA, Heffess CS, Katoh R, Tsujimoto M & Kakudo K. Observer variation of encapsulated follicular lesions of the thyroid gland. American Journal of Surgical Pathology 2002 26 1508–1514. 23 Mai KT, Bokhary R, Yazdi HM & Thomas J. Hybrid thyroid carcinoma with a coarse chromatin pattern and nuclear features of papillary thyroid carcinoma. Pathology Research and Practice 2002 198 253–260. 24 Nagashima T, Suzuki M, Oshida M, Hashimoto H, Yagata H, Shishikura T, Koda K & Nakajima N. Morphometry in the cytologic evaluation of thyroid follicular lesions. Cancer 1998 84 115– 118. 25 Tseleni-Balafouta S, Kavantzas N, Paraskevakou H & Davaris P. Computerized morphometric study on fine needle aspirates of cellular follicular lesions of the thyroid. Analytical and Quantitative Cytology and Histology 2000 22 323–326. 26 Tsybrovskyy O, Vassilenko I, Mannweiler S & Klimpfinger M. Multivariate karyometric approach in differential diagnosis of follicular thyroid neoplasms. A study of 31 cases. Virchows Arch 1998 433 135–143. 27 Kirillov VA, Yuschenko YP, Paplevka AA & Demidchik EP. Thyroid carcinoma diagnosis based on a set of karyometric parameters of follicular cells. Cancer 2001 92 1818–1827. 28 Eldar S, Sabo E, Cohen A, Misselevich I, Abrahamson J, Cohen O, Kelner J & Boss JH. The value of histomorphometric nuclear parameters in the diagnosis of well differentiated follicular carcinomas and follicular adenomas of the thyroid gland. Histopathology 1999 34 453–461. 29 Murata S, Mochizuki K, Nakazawa T, Kondo T, Nakamura N, Yamashita H, Urata Y, Ashihara T & Katoh R. Detection of underlying characteristics of nuclear chromatin patterns of thyroid tumor cells using texture and factor analyses. Cytometry 2002 49 91–95. 30 Kelman AS, Rathan A, Leibowitz J, Burstein DE & Haber RS. Thyroid cytology and the risk of malignancy in thyroid nodules: importance of nuclear atypia in indeterminate specimens. Thyroid 2001 11 271– 277. 31 Tzen CY, Huang YW & Fu YS. Is atypical follicular adenoma of the thyroid a preinvasive malignancy? Human Pathology 2003 34 666–669. 32 De Micco C, Vasko V, Ferrand M, Henry JF & Carayon P. N-Ras oncogene mutations in the follicular variant of papillary thyroid carcinomas. Cancer Detection and Prevention 2000 24 (Suppl 1) S74. 33 Zhu Z, Gandhi M, Nikiforova MN, Fischer AH & Nikiforov YE. Molecular profile and clinical-pathologic features of the follicular variant of papillary thyroid carcinoma. An unusually high prevalence of ras mutations. American Journal of Clinical Pathology 2003 120 71 –77. 34 Tallini G, Santoro M, Helie M, Carlomagno F, Salvatore G, Chiappetta G, Carcangiu ML & Fusco A. RET/PTC oncogene activation defines a subset of papillary thyroid carcinomas lacking evidence of progression to poorly differentiated or undifferentiated tumor phenotypes. Clinical Cancer Research 1998 4 287–294. 35 Basolo F, Molinaro E, Agate L, Pinchera A, Pollina L, Chiappetta G, Monaco C, Santoro M, Fusco A, Miccoli P, Elisei R, Capezzone M & Pacini F. RET protein expression has no prognostic impact on the long-term outcome of papillary thyroid carcinoma. European Journal of Endocrinology 2001 145 599–604.

www.eje.org

786

V V Vasko and others

36 Martinez I, Mantilla A, Medrano ME, Hernandez R, Hernandez DM, Lazos M, Santiago H, Gonzalez B, Hidalgo A & Salcedo M. High prevalence of RET tyrosine kinase activation in Mexican patients with papillary thyroid carcinomas. Endocrine Pathology 2001 12 113 –123. 37 Cerilli LA, Mills SE, Rumpel CA, Dudley TH & Moskaluk CA. Interpretation of RET immunostaining in follicular lesions of the thyroid. American Journal of Clinical Pathology 2002 118 186 –193. 38 Collins BJ, Chiappetta G, Schneider AB, Santoro M, Pentimalli F, Fogelfeld L, Gierlowski T, Shore-Freedman E, Jaffe G & Fusco A. RET expression in papillary thyroid cancer from patients irradiated in childhood for benign conditions. Journal of Clinical Endocrinology and Metabolism 2002 87 3941–3946. 39 Sugg SL, Ezzat S, Rosen IB, Freeman JL & Asa SL. Distinct multiple RET/PTC gene rearrangements in multifocal papillary thyroid neoplasia. Journal of Clinical Endocrinology and Metabolism 1998 83 4116–4122. 40 Thomas GA, Bunnell H, Cook HA, Williams ED, Nerovnya A, Cherstvoy ED, Tronko ND, Bogdanova TI, Chiappetta G, Viglietto G, Pentimalli F, Salvatore G, Fusco A, Santoro M & Vecchio G. High prevalence of RET/PTC rearrangements in Ukrainian and Belarussian post-Chernobyl thyroid papillary carcinomas: a strong correlation between RET/PTC3 and the solid-follicular variant. Journal of Clinical Endocrinology and Metabolism 1999 84 4232–4238. 41 Santoro M, Melillo RM, Carlomagno F, Fusco A & Vecchio G. Molecular mechanisms of RET activation in human cancer. Annals of the New York Academy of Sciences 2002 963 116 –121. 42 Kjellman P, Learoyd DL, Messina M, Weber G, Hoog A, Wallin G, Larsson C, Robinson BG & Zedenius J. Expression of the RET protooncogene in papillary thyroid carcinoma and its correlation with clinical outcome. British Journal of Surgery 2001 88 557 –563.

www.eje.org

EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

43 Learoyd DL, Messina M, Zedenius J, Guinea AI, Delbridge LW & Robinson BG. RET/PTC and RET tyrosine kinase expression in adult papillary thyroid carcinomas. Journal of Clinical Endocrinology and Metabolism 1998 83 3631–3635. 44 Fluge O, Haugen DR, Akslen LA, Marstad A, Santoro M, Fusco A, Varhaug JE & Lillehaug JR. Expression and alternative splicing of c-RET RNA in papillary thyroid carcinomas. Oncogene 2001 20 885– 892. 45 Elisei R, Romei C, Vorontsova T, Cosci B, Veremeychik V, Kuchinskaya E, Basolo F, Demidchik EP, Miccoli P, Pinchera A & Pacini F. RET/PTC rearrangements in thyroid nodules: studies in irradiated and not irradiated, malignant and benign thyroid lesions in children and adults. Journal of Clinical Endocrinology and Metabolism 2001 86 3211 –3216. 46 Sadetzki S, Calderon-Margalit R, Modan B, Srivastava S & Tuttle RM. Ret/PTC activation in benign and malignant thyroid tumors arising in a population exposed to low-dose externalbeam irradiation in childhood. Journal of Clinical Endocrinology and Metabolism 2004 89 2281–2289. 47 Fischer AH, Bond JA, Taysavang P, Battles OE & Wynford-Thomas D. Papillary thyroid carcinoma oncogene (RET/PTC) alters the nuclear envelope and chromatin structure. American Journal of Pathology 1998 153 1443– 1450. 48 Fischer AH, Taysavang P & Jhiang SM. Nuclear envelope irregularity is induced by RET/PTC during interphase. American Journal of Pathology 2003 163 1091–1100.

Received 23 June 2004 Accepted 7 September 2004