| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Staining as a Surrogate for PAX8/PPAR Fusion Oncogene Expression in Follicular Neoplasms: Clinicopathological Correlation and Histopathological Diagnostic Value
Departments of Medicine (M.S., B.L.A., J.G.P., M.Y., X.-L.W., I.D.H., Y.Z., S.K.G.G., N.L.E., B.M.), Laboratory Medicine and Pathology (J.R.G., T.J.S., S.K.G.G.), and Biochemistry and Molecular Biology (N.L.E.), Mayo Clinic and Foundation, Rochester, Minnesota 55905
Address all correspondence and requests for reprints to: Stefan Grebe, Department of Laboratory Medicine and Pathology, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905. Email: grebs{at}mayo.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(PPFP) fusion-oncogene is moderately specific for follicular thyroid carcinomas (FTC). It remains unknown whether this can be translated into improved diagnosis, classification, or outcome prediction.
We studied a cohort of well-characterized follicular adenomas (FA), FTC, and Hürthle cell carcinomas (HCC) from patients with complete clinical follow-up, to determine whether PPAR immunohistochemistry (as a surrogate of PAX8/PPAR expression) helps to distinguish FA from FTC and to assess its diagnostic accuracy as an adjunct to frozen section. We also correlated PPAR staining with clinical outcomes to assess its role as a prognostic marker.
PPAR staining was more common in FTC (31 of 54; 57%) than in HCC (one of 23; 4%) or FA (four of 31; 13%) (P < 0.000001). Adjunctive use of PPAR immunohistochemistry improved diagnostic sensitivity of intraoperative frozen section from 84% to 96% (P < 0.05) but reduced specificity from 100% to 90% (P < 0.05). PPAR staining was associated with favorable prognostic indicators (female gender, better tumor differentiation, and lesser risk of metastases).
PPAR staining may be helpful in the differential diagnosis of FA, FTC, and HCC, particularly when diagnostic sensitivity of histomorphology is reduced (e.g. during intraoperative frozen section). PPAR staining also shows an association with favorable prognosis and may have a role in risk stratification.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(PPFP) FUSION gene is a putative oncogene found in follicular cell-derived thyroid tumors, which is created by a somatic tumor genetic translocation between chromosome arms 2q and 3p. This t(2;3)(q12–13;p24–25) translocation fuses the entire coding region of the nuclear transcription factor gene PPAR in-frame with the first six to nine exons of the PAX8 gene, which encodes a thyroid-specific paired box transcription factor (1). The resulting fusion protein exhibits oncogenic properties, leading to increased cell-cycle transit, inhibition of apoptosis, loss of contact inhibition, and matrix-independent growth (2). These effects appear to be mediated, at least in part, by wild-type PPAR inhibition (1, 2). In the original description by Kroll et al. (1), five of eight follicular thyroid carcinoma (FTC) specimens contained the PPFP fusion gene. In contrast to this high prevalence in FTC, the authors found that neither hyperplastic nodules (n = 10), follicular adenomas (FA; n = 20), nor papillary thyroid carcinomas (PTC; n = 10) showed evidence of the rearrangement (1). Since this first description of PPFP, several papers have been published that have examined the prevalence of this fusion gene in 411 additional thyroid tumor specimens of different morphotypes (3, 4, 5, 6, 7, 8). These studies, together with the original Kroll et al. study, have confirmed that an average of 50% of FTC (n = 116; range, 29–78%) express PPFP, whereas PTC are almost never PPFP positive (one in 130). Most, but not all, studies have also shown that FA are less likely than FTC to harbor PPFP, the average rate being 17% (n = 115; range, 0–55%). Finally, a limited number of Hürthle cell adenomas (n = 25), Hürthle cell carcinomas (HCC; n = 31), and anaplastic thyroid cancers (ATC; n = 16) have also been examined and found not to express PPFP (3, 7, 8).
It therefore appears that PPFP is specific to the follicular morphotype; and within follicular tumors, it is mainly, but not completely, restricted to malignant neoplasms. What remains largely unknown is whether these facts can be translated into improved tumor diagnosis, classification, or outcome prediction. Only two studies have attempted to address the first two questions in a limited fashion (3, 7), whereas no data about patient outcomes are currently available. These issues are of significant practical relevance because important clinical diagnostic problems center on the differential diagnosis of FA vs. FTC in fine-needle aspiration biopsy specimens and surgical frozen sections. With regard to outcome prediction, correlation of tumor genetic changes with patient outcomes might assist in risk stratification and staging, thereby ultimately influencing follow-up strategies and decisions on adjuvant therapy.
In the current study, we examined PPFP expression in a cohort of histologically well-characterized FA, FTC, and HCC specimens from patients with complete clinical follow-up. We assessed the ability of PPAR immunohistochemistry, as a surrogate marker for PPFP expression, to distinguish FA from FTC, and correlated PPAR immunohistochemistry with clinical outcomes. Finally, we applied PPAR staining to a group of frozen section specimens and assessed retrospectively its effect on diagnostic accuracy
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
These studies were approved and monitored by the Mayo Clinic Institutional Review Board.
We examined 108 follicular neoplasms, including 31 FA, 23 HCC, and 54 FTC, for which full clinical data and sufficient archival tissue were available. FA were defined as follicular neoplasms without capsular or vascular invasion. By contrast, tumors displaying invasiveness but, at the same time, lacking any papillary or anaplastic elements, were classified as FTC or, if they contained more than 75% oxyphilic cells, as HCC. Tissues were fixed in 10% phosphate-buffered formalin and embedded in paraffin by standard methods. In addition, fresh-frozen tissue was also available for 10 of the FTC and 10 of the HCC specimens. We sectioned paraffin-embedded blocks of formalin-fixed tissues at 4-µm thickness and mounted several sections onto microscope slides. For each specimen, one of these slides was stained with hematoxylin and eosin (H&E), and the remaining were left unstained. Experienced thyroid pathologists (J.R.G., T.J.S.) reviewed the stained slides to verify the tumor morphotype, degree of differentiation, and other relevant histopathological features.
The FTC included in this study represented the entire set of paraffin-embedded tissue blocks that were available to us. All of these tumors were derived from patients included in the Mayo Clinic Thyroid Cancer database, maintained by one of us (I.D.H.), and all had complete demographic, clinical, and follow-up data available. These 54 tumors represent approximately 35% of the total number of FTC included in the clinical database and are representative (in demographic, clinical, and outcome terms) of the entire clinical dataset. The HCC and FA studied were selected arbitrarily from the available tissue samples. All 54 FTC cases and all FA and HCC cases were used in the assessment of clinical pathological correlations.
To assess the potential value of adjunctive PPAR immunochemistry on frozen section diagnosis of follicular neoplasms, we excluded 15 of the 54 FTC patients that had obvious clinical malignancy at the time of surgery, including widely metastatic disease, malignant nodes, or gross capsular invasion, which makes intraoperative frozen section unnecessary. The tumors from the remaining 39 FTC patients, as well as those from all 31 FA patients, had been subjected to frozen section at the time of surgery, as is standard practice in our institution. The final diagnosis for each tumor was based on permanent sections of formalin-fixed, paraffin-embedded, and H&E-stained slides. To allow calculations of test-predictive values for our patient population, we used the statistics for thyroid surgery performed for follicular neoplasms between 1996 and 2000 (inclusive) from the surgical pathology database.
PPAR expression in archival tumors as a surrogate marker of PPFP expression
We deparaffinized the unstained sections in xylene, followed by rehydration through graded ethanol washes. Antigen retrieval was performed by incubation at 98 C for 30 min in 1 mmol/liter EDTA buffer, adjusted to pH 8.0 with NaOH. We probed the rehydrated sections with a primary mouse monoclonal antibody, raised against the C terminus of PPAR (Santa Cruz Biotechnology, Santa Cruz, CA). PPAR is expressed only at low levels in the normal adult thyroid gland and therefore serves as a good surrogate marker for PPFP expression. Color development was accomplished using the EnVision nonbiotin detection system (DakoCytomation, Carpinteria, CA). Two independent investigators (M.S., B.M.), blinded to the histological and clinical information, graded the nuclear staining intensity on each slide as both strong and specific, or negative. Weak nonnuclear staining was considered negative. They also classified the staining pattern as uniform or patchy, depending on whether all or only portions of the tumor tissue displayed immunoreactivity. Regions of normal follicular tissue on each slide served as internal negative controls.
PPFP RT-PCR
We used RT-PCR followed by DNA sequencing to confirm the presence of, and to characterize, PPFP fusion genes in 10 FTC- and 10 HCC fresh-frozen samples. We isolated RNA by TRIZOL (Invitrogen, Carlsbad, CA) extraction according to the manufacturer’s instructions and reverse transcribed 1 µg of each RNA sample using Moloney murine leukemia virus reverse-transcriptase and an oligo-dT17 primer (Molecular Biology Core Laboratory, Mayo Clinic). The cDNAs were PCR-amplified using a sense primer derived from PAX8 exon 9 and an antisense primer derived from PPAR exon 1. Direct sequencing of PCR products was accomplished by automated fluorescent dye-terminator method (Molecular Biology Core Laboratory, Mayo Clinic), using the same primers.
Clinicopathological correlation
For the entire dataset of 54 FTC, 31 FA, and 23 HCC patients, we related PPAR immunostaining to tumor histopathology (morphotype, degree of differentiation/grade, minimal or extensive angioinvasion) and clinical features (tumor size, local and distant tumor spread, patient age and gender, tumor-related death). Most FTC and HCC in this study were well differentiated, corresponding to grade 1 in a four-tier grading system. Tumor grade 1 was defined as showing folliculogenesis with relative uniformity to nuclear size and shape, whereas higher grade tumors display more nuclear heterogeneity, angulation, and size variability, as well as increasing amounts of solid growth.
Statistical analysis
We compared continuous data between groups using two-tailed t tests, if necessary after data transformation to ensure normality. Categorical data were compared between groups by 
2 tests with appropriate degrees of freedom. Yates correction was applied to 2 testing with a single degree of freedom, which was used for comparisons of angoinvasion, locoregional spread, and gender frequency in PPAR-positive vs. PPAR-negative FTC. Test sensitivity, specificity, and positive and negative predictive values were calculated in standard fashion.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Table 1
summarizes the relevant patient demographics and tumor-specific characteristics of the specimens studied. The median duration of clinical follow-up was 5.8 yr (range, 1.2 months to 49.2 yr).
|
and PPFP expression in archival tumors
Thirty-one (57%) of the 54 FTC showed strong nuclear staining for PPAR (Table 2). Of the 31 FA, only four (13%) exhibited strong staining, and only one (4%) of the 23 HCC stained PPAR-positive. The staining pattern was uniform in 13 of 31 PPAR-positive FTC and patchy in the remaining 18. The four FA and the single HCC that stained positive for PPAR displayed uniform nuclear staining. Figure 1 shows examples of PPAR staining. The difference in the prevalence of PPAR-positive staining among HCC, FA, and FTC was highly significant (P < 0.000001). The predicted sensitivity and specificity of PPAR staining to distinguish FTC from FA were 57% (95% CI, 43–71%) and 87% (95% CI, 70–96%), respectively. This corresponded to a positive predictive value for FTC in PPAR-positive samples of 89% (95% CI, 72–98%) and a negative predictive value for FTC of PPAR-negative samples of 54% (95% CI, 41–67%).
|
|
positive. All ten of the frozen HCC specimens were negative for PPAR staining. RT-PCR and sequencing confirmed the presence of the expected PPFP fusion transcripts (fusion between exon 9 of PAX8 and exon 1 of PPAR) in all eight PPAR-positive FTC specimens but none of the ten HCC specimens (Fig. 2).
|
immunohistochemistry and intraoperative frozen section analysisDuring 1996 through 2000, 540 patients underwent surgery at Mayo Clinic for a diagnosis of follicular neoplasm. The final diagnoses were FA for 450 specimens and FTC for 90 specimens.
Of the 70 follicular tumors included in this portion of the study, frozen section alone correctly identified 33 of 39 as malignant, a sensitivity of 84%, similar to previous reports from this institution (9). None of the 31 FA were misclassified as malignant, and therefore specificity in this group of tumors was 100%, giving an overall test accuracy of 91%. Extrapolating these numbers to the entire cohort of 540 tumors, the positive and negative predictive values of frozen section would be 100% and 96%, respectively. These values are in keeping with an earlier, more comprehensive review of the diagnostic accuracy of frozen section of follicular neoplasms at our institution, which had found positive and negative predictive values of 90% and 98%, respectively. PPAR staining alone identified 24 of 39 FTC, a sensitivity of 62%. PPAR expression was also present in four of 31 tumors classified as FA by H&E permanent section, yielding a specificity of PPAR staining of 90% in the diagnosis of FTC. Adding PPAR expression status to frozen section identified an additional five patients (from 33 to 38) of the total 39 FTC, thereby increasing the sensitivity for detection of FTC significantly from 84% to 96% (P < 0.05). However, specificity decreased from 100% to 90%, because PPAR immunostaining introduced four false positives (P < 0.05). Combined frozen section and PPAR staining in our patient population had positive and negative predictive values of 72% and 99%, respectively, with a test accuracy of 93%.
Clinicopathological correlation
Within the malignant tumors, there were no significant differences in PPAR immunostaining with regards to patient age, primary tumor size, locoregional spread, or angioinvasion. However, a significantly higher proportion of patients with PPAR-negative tumors were male (P < 0.03), and their tumors were less well differentiated (P < 0.04). As shown in Table 3, there was a relationship between the pattern of immunohistochemical staining and the degree of differentiation (P < 0.006) across the three strata (PPAR-positive, -patchy positive, and -negative). Tumors that stained uniformly PPAR-positive were better differentiated than tumors with patchy staining (2, df = 1, Yates-corrected 3.9, P < 0.05), or PPAR-negative tumors (2, df = 1, Yates-corrected 7.6, P < 0.006). A nonsignificant trend (P < 0.076) was also observed between staining pattern and metastatic disease across the three strata. However, because there were only 10 patients with metastatic disease, the power to detect a significant difference across the three strata was only 36% and similarly low for each possible paired comparison.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
, occurs in a significant subset of FTC but few FA and almost none of the HCC at rates similar to those reported in most other studies (1, 3, 4, 7, 8). Although it is difficult to compare the various studies of PPFP expression in thyroid tumors directly, because they have often used different methodologies for PPFP detection, including PPAR-staining, PPAR Western blotting, RT-PCR, and fluorescent in situ hybridization, it appears that PPFP expression is confined to FA and FTC and occurs with at least 3-fold higher frequency in FTC than in FA, with virtually no PPFP expression in HCC, PTC, and ATC (Table 4).
|
staining was shown in our study to be capable of potentially enhancing identification of FTC at the time of frozen section. Because PPAR-expression occurs in approximately half of all FTC, it is not a particularly sensitive marker when used alone. However, when used in a situation where diagnostic sensitivity of conventional histomorphology is reduced, for example when studying frozen sections, the addition of PPAR staining increases diagnostic sensitivity. Conversely, the specificity of frozen section in our study was slightly reduced when combined with PPAR staining data, as some tumors that were classified later as FA also stained positive. Intraoperative frozen section of follicular thyroid neoplasms is practiced at only a few institutions, because most authors report that the procedure is error prone, with some reported test sensitivities of less than 60% (10, 11, 12, 13, 14, 15). Consequently, most endocrine surgeons and pathologists in the United States do not advocate, nor use routinely, intraoperative frozen section to differentiate FA from FTC. All patients either undergo near-total or total thyroidectomies for all follicular neoplasms, or, if the initial procedure is a lobectomy, around 20% of patients will require reoperation at a later stage, with a consequent increased complication rate (16). By contrast, intraoperative frozen section of follicular thyroid neoplasms has been employed routinely in our institution for over 50 yr (9, 17). Our pathologists have developed considerable expertise in this technique, making it possible for many patients to have a local excision if their tumor is deemed benign, rather than a total or near-total thyroidectomy, procedures associated with higher morbidity (18). Only a small number of these patients, later diagnosed on permanent section to have a FTC rather than a FA, require completion thyroidectomy. It would be ideal if this group could be reduced further, while still maintaining high diagnostic specificity. The combination of PPAR staining with frozen section may increase test accuracy; and further exploration of this approach, including optimization for rapid use during frozen section, appears worthwhile. However, in our study, specificity was reduced, possibly negating some of the benefits of increased sensitivity.
PPFP has potent, multifaceted oncogenic properties (2), leading to speculation that those FA expressing PPFP may represent FTC in situ. Consequently, such tumors may have an increased likelihood of progression to invasive cancer, compared with other FA. Unlike PTC, where no clear progenitor lesion is known, an adenoma to carcinoma progression sequence has long been proposed for FTC (19, 20, 21). FA can appear aberrant on cytological grounds and may display features suggesting aggressive biological potential, whereas some minimally invasive FTC may look relatively benign (22). At the same time, FA often display both generalized and specific genomic abnormalities, both of which are also increased in FTC, roughly paralleling the degree of histological differentiation (20, 23). Alternatively, some of the PPAR-positive FA might, in fact, be FTC in which invasion was not seen because of tissue-sectioning limitations, or secondary to sampling error. Interestingly, in the one study that showed a high rate of PPFP positivity in FA, two of the examined FA, one of which was PPFP positive by PPAR staining and RT-PCR, showed incomplete capsular invasion, arguably suggesting that they could, in fact, represent FTC rather than FA (6). Finally, it is interesting to speculate whether some FA that stain positive for PPAR may harbor a distinct genetic rearrangement that involves a different fusion partner than PAX8. Kroll et al. (1), in their original description, has hinted at the possible existence of such additional rearrangements; and more recently, Dwight et al. (8) have provided data that support this concept. Regardless of the exact mechanism of PPAR-positivity in some FA, it could be that these patients benefit from a more complete surgical intervention. This interpretation, of course, must be stated cautiously, because the mortality risk of minimally invasive FTC is negligible. Nonetheless, if it can be confirmed that PPAR-positive FA are indeed FTC in situ, then PPAR staining might become an obligatory adjunct not only to standard frozen section but also to definitive histomorphological diagnosis. This approach might result in improved differential diagnosis of FA vs. FTC, particularly if PPAR histochemistry is combined with one of several other promising histochemical markers that have been developed during the last 10 yr, including telomerase, galectin 3, and p53 (24, 25, 26).
The almost complete absence of PPFP expression in any HCC specimens studied to date suggests that another application of PPAR immunohistochemistry may be that of differential diagnosis of typical FTC with oxyphilic features vs. HCC. Based on histological appearance and unique biological behavior, it has been argued that HCC should be regarded as a separate entity (19). The presence of follicular elements in HCC might be interpreted as a biologically irrelevant morphological feature, similarly to the way in which follicular elements in PTC are currently viewed. One of the reasons this reclassification step has not already been taken is that the traditional pathological diagnosis of HCC is not based on unequivocal nuclear features but on an arbitrarily defined proportion of cells (>75%) that display oxyphilic staining characteristics (27). Assessment of PPFP status by PPAR staining or by RT-PCR might facilitate more definitive differential diagnoses of FTC vs. HCC, particularly if combined with assessment of PTC/RET oncogene expression status. This PTC-associated oncogene may be expressed in at least some HCC but is usually not found in typical FTC (28).
Our data also suggest that PPAR immunohistochemistry, as a surrogate marker of PPFP expression, might play a role in tumor classification and outcome prediction. The number of patients included in our study, their length of follow-up, and the number of deaths were too low to draw any definitive conclusions, but it nonetheless appeared that PPAR expression was inversely correlated with biological, morphological, and demographic features that are associated with an adverse outcome. Patients with PPAR-negative tumors were more often male, had less well-differentiated tumors, and were marginally more likely to have metastatic disease at the time of diagnosis. Moreover, it appears that there is a graduated inverse relationship between PPAR staining pattern and degree of differentiation and metastases risk. The most differentiated tumors displayed predominately uniform and strong staining, whereas the less differentiated tumors tended to exhibit patchy staining, and the most aggressive neoplasms were mainly PPAR-negative.
There are two possible explanations for the apparent association of positive PPAR-staining with less aggressive behavior. First, it is possible that there exist two distinct pathways of FTC development, one dependent on PPFP expression, leading to a well-differentiated tumor phenotype, and the other arising through a yet unknown mechanism, leading to a more aggressive phenotype. Nikiforova et al (7) have shown that RAS mutations and PPFP expression appear to be mutually exclusive, lending support to such a concept. With RAS being a relatively weak thyroid oncogene (29, 30, 31), mutant RAS-1 carrying FTC may be associated with even less aggressive tumors than PPFP (7). Alternatively, the absence of PPAR-staining may simply be another indicator of a poorly differentiated tumor, even if the tumor originally expressed PPFP. Our identification of an intermediate group of tumors that demonstrate a patchy staining pattern for PPAR, along with an intermediate degree of differentiation, may support this latter hypothesis. At least some of these tumors may have progressed down a common PPFP-driven pathway but have lost PAX8, and consequently PPFP, expression in their ongoing tumor genetic evolution. In this model, PPFP represents an early oncogene, promoting the transition of FA to FTC. The resultant FTC are initially well differentiated and continue to express PPFP. However, subsequent tumor genetic events lead to increasing dedifferentiation, a hallmark of tumor progression. Tissue-specific cell functions are progressively lost, including loss of thyroid-specific PAX8 transcription, and consequently of PPFP transcription, which is driven by the PAX8 promoter (Fig. 1
). Although speculative, the fact that PPFP expression was not detected in any of the 16 ATC studied by other groups (Table 4) might lend further support to such a model, because at least a proportion of ATC is thought to arise from a preexisting follicular neoplasm.
In conclusion, PPAR-immunohistochemistry represents a convenient way to determine PPFP expression status indirectly, but a small percentage of tumors may have different rearrangements involving PPAR or may overexpress PPAR for other reasons. PPFP appears to be involved in approximately 50% of FTC and is either completely specific to FTC and FA that may represent FTC in situ, or occurs with at least 3-fold lesser frequency in FA, making it a potential marker in FTC vs. FA differential diagnosis. This applies particularly in situations where the diagnostic sensitivity or specificity of histopathology is limited, such as in the case of FNA or frozen section. However, future prospective application of PPAR-immunostaining to follicular neoplasms will be necessary before definitive recommendations can be made. PPAR-staining also shows an association with better-differentiated tumors and a more favorable prognosis, and may have a role in risk-stratification. In this latter context, it remains to be determined whether PPFP expression, assessed by PPAR-staining or RT-PCR, identifies a distinct pathogenic subgroup of tumors or whether its expression is merely a surrogate marker for differentiation status by reflecting PAX8 expression.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online October 13, 2004
Abbreviations: ATC, Anaplastic thyroid cancers; FA, follicular adenoma(s); FTC, follicular thyroid carcinoma(s); HCC, Hürthle cell carcinoma(s); H&E, hematoxylin and eosin; PPFP, PAX8/PPAR; PTC, papillary thyroid carcinoma(s).
Received June 23, 2004.
Accepted October 4, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-1 fusion oncogene in human thyroid carcinoma. Science 289:1357–1360 fusion oncoprotein transforms immortalized human thyrocytes through a mechanism probably involving wild-type PPAR inhibition. Oncogene 23:3634–3641[CrossRef][Medline]
rearrangement in thyroid tumors: RT-PCR and immunohistochemical analyses. Am J Surg Pathol 26:1016–1023[CrossRef][Medline]
1 rearrangements in both follicular thyroid carcinomas and adenomas. J Clin Endocrinol Metab 87:3947–3952 fusion oncogene in both follicular thyroid carcinomas and adenomas. J Clin Endocrinol Metab 88:354–357 rearrangement in thyroid tumors: evidence for distinct molecular pathways in thyroid follicular carcinoma. J Clin Endocrinol Metab 88:2318–2326 rearrangement in follicular thyroid tumors. J Clin Endocrinol Metab 88:4440–4445This article has been cited by other articles:
 |
|
 |
|
  H. V. Reddi, B. McIver, S. K. G. Grebe, and N. L. Eberhardt The Paired Box-8/Peroxisome Proliferator-Activated Receptor-{gamma} Oncogene in Thyroid Tumorigenesis Endocrinology, March 1, 2007; 148(3): 932 - 935. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Kristiansen, J. Jacob, A.-C. Buckendahl, R. Grutzmann, I. Alldinger, B. Sipos, G. Kloppel, M. Bahra, J. M. Langrehr, P. Neuhaus, et al. Peroxisome Proliferator-Activated Receptor {gamma} Is Highly Expressed in Pancreatic Cancer and Is Associated With Shorter Overall Survival Times. Clin. Cancer Res., November 1, 2006; 12(21): 6444 - 6451. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Xing BRAF mutation in thyroid cancer Endocr. Relat. Cancer, June 1, 2005; 12(2): 245 - 262. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
