| | Parathyroid neoplasms: Clinical, histopathological, and tissue microarray-based molecular analysisAccepted 16 April 2002. Abstract We studied 45 patients with typical and 8 with atypical parathyroid adenomas as well as 20 with parathyroid carcinomas. Clinical, pathological, and molecular analyses were conducted on all adenomas. Clinical data were analyzed for 20, histopathologic slides for 16, and tissue specimens for 8 patients with carcinoma. Molecular expression profiles were investigated by immunohistochemistry (IHC) for Ki-67, p53, mdm2, p21, Bcl-2, cyclin D1, and p27 on paraffin-embedded tissues arrayed on tissue microarrays. Trabecular growth and vascular, capsular, and soft-tissue invasion were characteristic of parathyroid carcinomas but not of typical adenomas. No adenomas recurred. Seventy-four percent of carcinomas recurred, most in the neck. Seventy-nine percent of patients with such illness died of disease after an indolent, multiply recurrent course responsive to repeated resections; the 5-year survival rate was 50%. High Ki-67 proliferative index was seen in 2% of adenomas and 25% of carcinomas, whereas p27 expression was present in 80% of adenomas and 18% of carcinomas. The molecular phenotype, p27(+)Bcl-2(+)Ki-67(−)mdm2(+), was observed in 76%, 29%, and 0% of typical and atypical adenomas and carcinomas, respectively. The complexity of molecular phenotypes increased with tumor aggressiveness. Parathyroid carcinoma is an aggressive disease with a propensity for multiple recurrences. It is characterized by capsular, vascular, and soft-tissue invasion. Recurrence portends poor outcome. Molecular markers, Ki-67 and p27, may distinguish parathyroid carcinoma from adenoma. The molecular phenotype, p27(+)Bcl-2(+)Ki-67(−)mdm2(+), appears to be unique to nonmalignant parathyroid tumors, and multimarker phenotypes are more complex in carcinomas. HUM PATHOL 34:54-64. Copyright 2003, Elsevier Science (USA). All rights reserved. Address correspondence and reprint requests to Ronald A. Ghossein, MD, Department of Pathology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021.
Parathyroid neoplasms are comprised of adenomas and carcinomas. Parathyroid adenomas are benign neoplasms that account for 85% of patients with primary hyperparathyroidism.1 In the majority of cases, adenomas involve a single gland. Parathyroid carcinoma is a rare disease that accounts for 1% to 3% of cases of primary hyperparathyroidism.1 These two neoplasms have disparate natural histories, but they can be difficult to differentiate on the basis of histopathologic findings alone. Although thick fibrous bands, mitotic activity, trabecular growth pattern, and capsular, vascular, and adjacent soft-tissue invasion are considered characteristic of parathyroid carcinoma, morphological features—such as fibrous bands, mitotic activity, and trabecular growth—have been identified in parathyroid adenomas as well.2, 3, 4 To complicate the morphological diagnosis further, a group of neoplasms referred to as “atypical adenomas” exhibit many of these features, but they lack indisputable evidence of malignancy such as angioinvasion and metastases.4 The clinical behavior of atypical adenomas is unpredictable, and the presence or absence of local recurrence or distant metastases is the only reliable feature that differentiates these neoplasms from small parathyroid carcinomas.
In an effort to establish more reliable diagnostic methods, investigators have studied the expression of cell-cycle regulatory proteins in parathyroid neoplasms. Proliferative activity, as evaluated by staining with MiB-1 for Ki-67 antigen, appears to be valuable in distinguishing benign from malignant parathyroid tumors.5, 6 Nuclear immunostaining for the cyclin-dependent kinase inhibitor p27 in parathyroid tissue appears to be a useful marker in differentiating between benign and malignant neoplasms.7 Other tumor markers, such as p53, Bcl-2, and cyclin D1, have not proven to be useful in this regard.6, 8
Unlike other human malignancies, alterations in p53 gene expression appear to have no role in parathyroid tumorigenesis.9, 10, 11, 12 Some studies have investigated Bcl-2, an important regulator of p53-mediated apoptosis, but they did not consider other molecular components of the p53 pathway such as mdm2 and p21.9, 13 The importance of the p53 pathway for parathyroid neoplasia remains undefined because important components of this pathway—mdm2, p21, and Bcl-2—have not been collectively studied. Parathyroid tumors may express multiple protooncogenes and tumor-suppressor genes; thus, to further characterize molecular changes in the p53 pathway and other important cell-cycle regulators for parathyroid tumorigenesis, the patterns of molecular expression must be studied collectively in the same tissues.
We recently validated the tissue microarray technique that allows large-scale molecular profiling of multiple cell-cycle regulatory proteins in the same paraffin-embedded tumor specimen.14 This study presents a critical histopathologic appraisal of parathyroid adenomas and carcinomas and extends the characterization of these tumors in the form of a comprehensive molecular analysis utilizing immunohistochemistry carried out on tissue microarrays. We have identified a molecular phenotype that may facilitate the differential diagnosis between benign and malignant parathyroid tumors.
Patients and methods  Patient search Patients with parathyroid tumors were identified from the Departments of Surgery and Pathology data bases and the hospital tumor registry. Because patients with parathyroid neoplasms constitute the majority of patients with parathyroid disease seen at our institution, the relatively few patients with parathyroid hyperplasia were not included as part of this study. This is a retrospective review of patients whose parathyroid carcinoma was diagnosed (n = 20) and who were consecutively treated and followed-at Memorial Sloan-Kettering Cancer Center from 1953 to 1995. For the purpose of histopathologic and molecular comparison, this study also includes patients (n = 53) who recently (1993 to 2000) underwent surgical treatment for a single parathyroid adenoma, either typical or atypical. Patients with parathyroid hyperplasia were excluded from the analysis. Pathology reports were reviewed for all patients to confirm the diagnosis of typical and atypical parathyroid adenoma and parathyroid carcinoma. Pathologic review: Inclusion criteria A critical review of all available histologic slides was conducted by a single member [RAG] of the pathology department. A median of 7 (range, 3 to 44) hematoxylin-eosin (H&E) slides per patient, along with their pathologic records were reviewed without knowledge of clinical characteristics or outcome. All confirmed single typical or atypical parathyroid adenomas and parathyroid carcinomas were included in the analysis. Histopathological definitions A primary tumor was defined as a localized lesion without regional or distant metastases that was previously untreated or biopsied (needle aspiration, incisional, or inadequate excisional biopsy) before definitive surgical therapy was undertaken. Parathyroid adenomas were defined as encapsulated neoplasms that were composed primarily of chief cells, that possessed a thin surrounding capsule, and that typically showed an uninvolved rim of adjacent normal parathyroid tissue. Adenomas could also include cell populations with oncocytic or clear-cell cytoplasm. The rim of normal parathyroid gland characteristically contained parenchymal fat. The absence of this adjacent rim of normal or suppressed parathyroid tissue does not exclude the diagnosis of parathyroid adenoma.4 Histologic confirmation of at least one normal-sized parathyroid gland at time of surgery was required. Degenerative adenomas showing chronic inflammatory infiltrates, hemosiderin deposition, and/or fibrosis—but without the principle features of carcinoma—were classified as adenomas. Adenomas showing some features of malignancy but not unequivocal morphological findings of carcinoma, such as vascular or adjacent soft-tissue invasion, were defined as atypical adenomas. These features include the presence of fibrous bands, capsular invasion, mitotic activity, and trabecular growth pattern.4 The presence of fibrous bands, mitotic activity, and large atypical cell clusters were not regarded as absolute criteria for malignancy because these morphological features also can be seen in some cases of typical and atypical adenomas.4 We defined a parathyroid tumor as carcinoma if it displayed vascular, perineural, and adjacent soft-tissue invasion. The presence of fibrous bands, high mitotic activity, and, of course, capsular invasion were considered highly suggestive of malignancy. The presence of incomplete capsular invasion or entrapped tumor cells within the capsule alone was insufficient for the diagnosis of parathyroid carcinoma. The latter finding was classified as pseudoinvasion. Vascular invasion was evident when a vessel located within or immediately external to the capsule contained tumor cells within it and these cells were found to adhere to the vascular wall. Only characteristic parenchymal mitotic figures were included in the appraisal of mitotic activity reported as number of mitoses per 10 high-power fields. These definitions of typical and atypical parathyroid adenoma and parathyroid carcinoma were established before slide review was conducted. Presence of metastasis was considered an unequivocal diagnosis of malignancy. Histopathologically confirmed parathyroid neoplasms Clinical and pathological records were available for 73 patients. Of these, we were able to obtain histopathologic slides for 69 (95%) patients. The 4 patients with parathyroid carcinoma for whom slides could not be retrieved were included in the analysis because their diagnosis was confirmed by their clinical course; there was recurrence in all 4 after resection of the primary tumor, and they each died of disease. Clinicopathologic categories Clinical data included patient age, gender, and presenting symptoms and signs (palpable neck mass or not). During the histopathologic review, a number of morphological findings were recorded. Gross tumor data included primary tumor size and weight. Histomorphological data included capsular, vascular, perineural, and adjacent soft-tissue invasion; pseudoinvasion; fibrous bands; mitotic activity; atypical mitoses and cell clusters; and nuclear pleomorphism. Findings suggestive of degenerative change also were recorded and included cystic degeneration, chronic inflammatory cell infiltrate, hemosiderin deposition, and fibrosis. Tissues, array construction, and immunohistochemistry Tissue from parathyroid adenomas and carcinomas were embedded in paraffin. Five-micron sections stained with H&E were obtained to confirm the diagnosis and to identify viable, representative areas of the specimen. From these defined areas, core biopsies specimens were taken with a precision instrument (Beecher Instruments, Silver Spring, MD) as previously described.15 From each specimen, tissue cores with a diameter of 0.6 mm were punched and arrayed in triplicate on a recipient paraffin block.14 Five-micron sections of these tissue array blocks were cut and placed on charged polylysine-coated slides. These sections were used for immunohistochemical analysis. Tissues and cell lines known to express the antigens under study were utilized as positive controls. Paraffin embedded tissue from 65 parathyroid neoplasms was used for the microarray. The median number of cores lost during processing was 4 tissue cores (6%) per molecular marker (range, 2 to 7 cores [3% to 10%]). Sections from tissue arrays were deparaffinized, rehydrated in graded alcohols, and processed using the avidin-biotin immunoperoxidase method. Briefly, sections were submitted to antigen retrieval by microwave treatment for 15 minutes in 0.01 molar citrate buffer at pH 6.0. This procedure was performed for all antibodies under study. For Ki-67 antibody, an additional step of incubation in preheated 0.05% trypsin, 0.05% CaCl2 in Tris-HCl (pH 7.6) for 5 minutes at 37°C before microwave treatment was carried out. Slides were subsequently incubated in 10% normal horse serum for 30 minutes. The slides were then incubated overnight at 4°C in appropriately diluted primary antibody. Mouse antihuman monoclonal antibodies to p53, mdm2, p21, p27, cyclin D1, Ki-67, and Bcl-2 were used for immunohistochemistry (IHC). For each molecular marker, the antibody, clone, dilution, and source are as follows:
•p53 (Ab-2, clone 1801, 1:500, Calbiochem, Cambridge, MA),
•mdm-2 (clone 2A10, 1:500, provided by Dr. A. Levine, Rockefeller University, NY),
•p21 (Ab-1, clone EA10, 1:20, Calbiochem),
•p27 (Ab-2, clone DCS72, 1:500, Oncogene Research Products, Cambridge, MA),
•cyclin D1 (Ab-3, clone DCS-6, 1:500, Calbiochem),
•Ki-67 (Mib-1, 1:50, Immunotech, Marseille, France), and
•Bcl-2 (clone 124, 1:72, Dako, Glostrup, Denmark).
Samples were then incubated with biotinylated antimouse immunoglobulins at 1:500 dilution (Vector, Burlingame, CA) followed by avidin-biotin peroxidase complexes (1:25, Vector) for 30 minutes. Diaminobenzidine was used as the chromogen and hematoxylin as the nuclear counterstain. Immunoreactivity was classified as continuous data (undetectable levels or 0% to homogeneous staining or 100%) for all markers. Several investigators (RAG, AH, AS) reviewed and scored slides independently by estimating the percentage of tumor cells showing characteristic staining. The percentage of tumor cell staining was determined by consensus among the reviewing investigators. The cut-off values for tumor cell staining used in the present study were defined based on previously established cut-off values used in clinicopathologic studies of different solid tumors employing identical reagents.11, 14 These established cut-offs were used as the basis for the present analysis and were modified according to clinicopathologic characteristics of parathyroid tumors. The cut-off values for tumor cell staining used in this study were defined as follows: (1) p53 nuclear overexpression if >5% tumor nuclei stained; (2) mdm2 overexpression if >50% tumor nuclei stained; (3) p21 overexpression if >10% tumor nuclei stained; (4) p27 overexpression if >30% tumor nuclei stained; (5) cyclin D1 overexpression if >5% of tumor nuclei stained; (6) high Ki-67 proliferative index if >5% tumor nuclei stained; and (7) Bcl-2 overexpression if >50% of tumor cells showed cytoplasmic staining. Tumors were then grouped into two categories defined as follows: normal expression (neoplasms below defined cut-off values of immunoreactivity in tumor cells) and abnormal expression (neoplastic tissues above defined cut-off values of immunoreactivity). All histopathologic and IHC information—along with clinical, treatment, and follow-up data—was entered into a computerized data base. Patient follow-up was obtained using clinical chart review, tumor registry information, private physician records, patient correspondence, and telephone interview. Study cohort The study consisted of 45 patients with typical and 8 patients with atypical single parathyroid adenomas (Table 1).
There were 20 cases of parathyroid carcinoma. All patients were treated with surgical resection of the abnormal parathyroid gland. Those patients with carcinoma were treated with en bloc resection of the tumor and a margin of normal surrounding tissue. Involved adjacent tissue, such as the strap musculature or thyroid lobe, was resected in continuity. Formal modified radical or radical neck dissections were performed for grossly positive nodal disease. For the group with parathyroid carcinoma, clinical data were available for 20 patients, histopathologic slides for 16 (slides for 4 of the patients were lost at the time of the study) patients, and 11 paraffin-embedded tissue specimens for 8 patients. For 2 of these, tumor specimens were available from both the primary and recurrent tumor. Because of the rarity of disease, all available tissues were used for microarray analysis. | | |  | Characteristic | Adenoma (n = 45) | Atypical Adenoma (n = 8) | Carcinoma (n = 20) | |
|---|
| F-to-M ratio | 33:12 | 4:4 | 11:9 | |
| Median age (y) | 58 | 63 | 47 | |
| No. palpable neck mass (%) | 0 | 2 (25) | 16 (75) | |
| Median size in cm (range) | 1.6 (1.0-3.3) | 2.9 (1.8-3.5) | 2.8 (2.0-6.7) | |
| Weight range (g) | 0.2-4.8 | 1.4-17.0 | 1.6-14.5 | |
| Median preop Ca++ in mg/dL* (range) | 11.2 (10.3-12.7) | 12 (7.9-13.2) | 14 (11-15.5) | |
| *Normal Ca++ range: 8.5 to 10.5 mg/dL. | | | | |
Median age for the study population was 56 years [range, 22 to 80 years]; 48 (66%) patients were female, and 25 (34%) patients were male. There was a preponderance of female patients in the adenoma group (female-to-male ratio = 2.7:1), whereas the gender distribution was similar among patients with atypical adenomas and parathyroid carcinomas (Table 1). Patients with carcinoma were significantly younger than those with typical or atypical adenomas (47 versus 58 to 63 years; P = .01). Median follow-up for patients with adenomas, atypical adenomas, and carcinomas was 20, 25, and 59 months, respectively. Follow-up of patients with carcinoma exceeded that of patients with benign disease because those who underwent surgical treatment for hypercalcemia attributable to a single parathyroid adenoma, either typical or atypical, were recently (1993 to 2000) treated at our institution. Statistical analysis Summary statistics were obtained utilizing established methods.16 Associations between categoric variables were evaluated using Fisher exact test.17 Hypothesis testing was carried out using X2 test with Yates correction when variable size or frequency was large enough to justify its use.16, 18 Nonparametric comparison of median values across groups was performed for continuous variables using the Wilcoxon rank sum test and the Kruskal-Wallis test. Outcome was classified according to sites of first disease recurrence. Time to recurrence and tumor-related mortality were calculated from the date of primary surgery. Deaths resulting from disease were treated as an end point for disease-specific survival (DSS). Those patients who died of other causes were censored in the analysis of DSS. Disease-free interval was calculated from time of surgery to any local, regional nodal or distant disease recurrence. The rate of recurrence or death was estimated using the Kaplan-Meier product limit method.19, 20 In all statistical analyses, a two-tailed P value ≤ .05 was considered statistically significant. All analyses were carried out using JMP statistical software (SAS Institute, Cary, NC). Results Clinical features Patients for whom a detailed history was available, ie, 67% of patients with either typical or atypical adenoma and 75% of those with parathyroid carcinoma, presented with symptomatic hypercalcemia. No adenoma was palpable, whereas 2 of 8 (25%) patients with atypical adenoma and 16 of 20 (75%) with carcinoma had a palpable neck mass at time of initial evaluation (P < .001). Median primary tumor size of atypical adenoma and carcinoma significantly exceeded that of parathyroid adenoma (Table 1; P < .001). The adenomas in this study are larger compared with those encountered in the general hospital practice. This reflects the fact that the study was not population based, and referral bias may have influenced the results. We do not know whether this fact has any influence on the molecular analysis in this study. The median preoperative serum calcium level increased going from typical adenomas to atypical adenomas and was the highest in carcinomas (Table 1). Histopathology Tumor cells arranged in a trabecular pattern of growth were identified in 75% and 94% of atypical adenomas and carcinomas, respectively (Fig 1a).
No typical adenoma in this study showed trabecular growth (Table 2).
Fibrous bands consisting of acellular collagenous tissue were observed in all carcinomas in this study ( Fig 1b). The presence of thick, fibrous bands was not a specific finding for malignancy because 9% of typical and all atypical adenomas manifested fibrous banding. Although no adenoma in this study had histological evidence of invasion, 3 (38%) atypical adenomas were adherent to the adjacent thyroid capsule. One adenoma and 4 (50%) atypical adenomas were found to have intracapsular tumor cell clusters defined as pseudoinvasion. Adjacent soft-tissue (69%, Fig 1c), capsular (94%, Fig 1e), and vascular (81%, Fig 1d) invasion were common findings for parathyroid carcinomas. Perineural invasion was evident in 3 of 16 (19%) cancers. Mitotic figures were seen in the majority (81%) of cancers as well as in 3 each of typical (7%) and atypical adenomas (38%). Although not specific for malignancy, mitoses in excess of 1/10 high-power fields and nuclear pleomorphism were significantly more frequent in the setting of parathyroid carcinoma ( Table 2). Atypical mitoses, albeit infrequently (25%), were identified exclusively in parathyroid carcinomas. Degenerative changes were typically seen with adenomas. | | | | Characteristic | Adenoma (n = 45) | Atypical Adenoma (n = 8) | Carcinoma (n = 16) | Difference Between Groups* (P) | |
|---|
| No. trabecular growth (%) | 0 (0) | 6 (75) | 15 (94) | <.001 | |
| No. fibrous bands (%) | 4 (9) | 8 (100) | 16 (100) | <.001 | |
| No. capsular invasion (%) | 0 (0) | 0 (0) | 15 (94) | <.001 | |
| No. vascular invasion (%) | 0 (0) | 0 (0) | 13 (81) | <.001 | |
| No. perineural invasion (%) | 0 (0) | 0 (0) | 3 (19) | .002 | |
| No. adjacent tissue invasion (%) | 0 (0) | 0 (0) | 11 (69) | <.001 | |
| No. pseudoinvasion (%) | 1 (2) | 4 (50) | 1 (6) | .002 | |
| No. nuclear pleomorphism (%) | 1 (2) | 1 (12) | 12 (75) | <.001 | |
| No. atypical mitoses (%) | 0 (0) | 0 (0) | 4 (25) | .001 | |
| No. mitosis/10 HPF (%) | | | | <.001 | |
|  0 | 42 (93) | 5 (62) | 3 (19) | | |
| 1 | 3 (7) | 2 (25) | 9 (56) | | |
| 2-5 | 0 (0) | 1 (13) | 4 (25) | | |
| No. cystic degeneration (%) | 26 (56) | 8 (100) | 0 (0) | <.001 | |
| No. hemosiderin deposition (%) | 1 (2) | 3 (37) | 1 (6) | .42 | |
| No. chronic inflammation (%) | 3 (6) | 0 (0) | 1 (6) | .61 | |
| No. atypical cell clusters (%) | 1 (2) | 0 (0) | 1 (6) | .59 | |
| *Comparison refers to adenoma (typical and atypical) versus carcinoma. | | | | |
IHC profiling of cell-cycle regulatory proteins IHC analysis of p53 expression showed an absence of nuclear staining in all parathyroid adenomas and carcinomas. More diverse findings were observed for other molecular components of the p53 pathway. The majority of parathyroid tumors studied showed nuclear staining for mdm2 (Table 3).
Overexpression of mdm2 was more frequent among parathyroid adenomas than carcinomas (63% to 93% v 18%; P < .001). Eighty-two percent of carcinomas showed <30% nuclear mdm2 expression. Nuclear p21 expression was present in all three types of parathyroid tumors, although a significantly lower proportion of atypical adenomas manifested the p21-positive phenotype (14% v 55% to 58%; P < .02). | | | | Molecular Marker | Adenoma (n = 45) | Atypical Adenoma (n = 8) | Carcinoma (n = 11) | Difference Between Groups* (P) | |
|---|
| p21 | | | | 0.88 | |
| No. negative (%) | 17 (42) | 6 (86) | 5 (45) | | |
| No. positive (%) | 23 (58) | 2 (14) | 6 (55) | | |
| p53 | | | | N/A | |
| No. negative (%) | 44 (100) | 7 (100) | 11 (100) | | |
| p27 | | | | <.001 | |
| No. negative (%) | 9 (20) | 4 (57) | 9 (82) | | |
| No. positive (%) | 36 (80) | 3 (43) | 2 (18) | | |
| Bcl-2 | | | | .003 | |
| No. negative (%) | 1 (2) | 1 (12) | 5 (45) | | |
| No. positive (%) | 41 (98) | 7 (88) | 6 (55) | | |
| Ki-67 | | | | .03 | |
| No. negative (%) | 40 (98) | 7 (100) | 8 (73) | | |
| No. positive (%) | 1 (2) | 0 (0) | 3 (27) | | |
| Cyclin D1 | | | | .76 | |
| No. negative (%) | 40 (91) | 7 (87) | 9 (82) | | |
| No. positive (%) | 4 (9) | 1 (13) | 2 (18) | | |
| mdm2 | | | | <.001 | |
| No. negative (%) | 3 (7) | 3 (37) | 9 (82) | | |
| No. positive (%) | 39 (93) | 5 (63) | 2 (18) | | |
| *Comparison refers to adenoma (typical and atypical) versus carcinoma. | | | | |
Overexpression of the Bcl-2 protein, defined as >50% positive tumor cells, was identified in nearly every (48 of 50 [96%]) adenoma. Decreased immunoreactivity for Bcl-2 was evident in 45% (5 of 11) of carcinomas. The frequency of nuclear expression of cyclin D1 was uniformly low among the study groups (adenoma, 9% [4 of 44]; atypical adenoma, 13% [1 of 8]; carcinoma, 18% [2 of 11]) and did not distinguish between the tumor types. Nuclear immunostaining of the cell proliferation marker, Ki-67, was observed in 1 of 48 typical and atypical adenomas. The Ki-67–positive phenotype was present in 27% of patients with parathyroid carcinoma. Evaluation of differential expression of p27 identified strong expression in all tumor types. The p27-positive phenotype, defined as >30% of tumor cell nuclear immunoreactivity, was most frequently observed in the adenoma group (80%). Along the continuum of parathyroid neoplasia—including adenomas, atypical adenomas, and invasive carcinomas—a progressive decrease in p27 overexpression was identified (80%, 43%, and 18%, respectively; P < .001). Thus, there was statistically significant differential expression of the following cell-cycle regulatory proteins among benign, indeterminate, and malignant parathyroid tumors: p27, Bcl-2, Ki-67, and mdm2 (Table 3). Because the patient number in the carcinoma group was small, survival analysis based on molecular phenotype was not feasible. Recurrence, follow-up, and outcome data No patients with typical adenoma developed a malignant parathyroid neoplasm. To our knowledge, no patient with atypical adenoma (n = 8) developed carcinoma; however, follow-up is limited because these patients were treated within the past 3 years. Median follow-up for the patients with parathyroid carcinoma was 59 months. Nineteen patients were treated for localized parathyroid cancer, and 1 presented with widespread metastatic disease and died 2 months later. One patient presented with localized parathyroid carcinoma 14 months after undergoing resection of atypical parathyroid adenoma on the same side. However, we were not able to review the slides of this atypical adenoma. Median DSS was 86 months, and overall 5-year survival rate was 50%. Five patients are alive and free of recurrent disease 3 to 229 months (median, 18 months) after undergoing curative resection (Table 5).
| | | | | | | | 1st Rec | | 2nd Rec | | 3rd Rec | | |
|---|
| Patient No. | Rec | Status | DFI (mo) | Site | Tx | DFI (mo) | Site | Tx | DFI (mo) | Site | Tx | Survival (mo) | |
|---|
| 1 | 0 | NED | 18 | — | | | | | | | | 18 | |
| 2 | 0 | NED | 3 | — | | | | | | | | 3 | |
| 3 | 0 | NED | 229 | — | | | | | | | | 229 | |
| 4 | 0 | NED | 7 | — | | | | | | | | 7 | |
| 5 | 0 | NED | 61 | — | | | | | | | | 61 | |
| 6 | 1 | NED | 10 | Neck | R | — | | | | | | 137 | |
| 7 | 3 | NED | 5 | Neck/bone | R | 14 | Neck | R | 8 | Neck | R | 31 | |
| 8 | >3 | NED | 84 | Neck | R | 70 | Neck/med | R | 22 | Neck | R | 220 | |
| 9 | 1 | AWD | 16 | Lung | — | | | | | | | 25 | |
| 10 | 1 | AWD | 50 | Neck | R | | | | | | | 52 | |
| 11 | 2 | DOD | 12 | Neck | R + RT | 43 | Bone | — | | | | 71 | |
| 12 | 3 | DOD | 13 | Bone | R | 13 | Neck | R + CRT | 3 | Bone | C | 32 | |
| 13 | 3 | DOD | 16 | Neck | R | 5 | Neck | R | 10 | Neck/bone | R + RT | 34 | |
| 14 | >3 | DOD | 14 | Neck | R | 45 | Neck | R | 52 | Neck | R | 153 | |
| 15 | >3 | DOD | 19 | Lung | R | 4 | Lung | R + C | 4 | Lung/bone | R + C | 57 | |
| 16 | >3 | DOD | 16 | Lung | R + C | 24 | Lung | R | 17 | Lung/bone | R | 86 | |
| 17 | >3 | DOD | 36 | Neck | R | 12 | Neck/med | R | 24 | Neck | R | 178 | |
| 18 | >3 | DOD | 35 | Neck/med/lung | R + C | 24 | Neck | R | 8 | Lung | R + C | 109 | |
| 19 | 2 | DOD | 23 | Neck/lung | R | 14 | Neck/med/lung | R | | | | 58 | |
| 20 | 0 | DOD | — | Widespread metastatic disease at initial presentation | | | | 2 | |
| | | | | |
Seventy-four percent (14 of 19) of patients developed recurrent disease after undergoing resection of the primary tumor. Median disease-free interval after initial surgery was 16 months (range, 5 to 84 months). In all, 11 (79%) of those who relapsed had more than one recurrence. Of these, 9 patients were treated for three or more recurrences. Median time to second recurrence after resection of initial recurrent disease was 14 months (range, 4 to 70 months). Median disease-free interval after treatment of the second recurrent tumor(s) was 10 months (range, 3 to 52 months). In patients with multiply recurrent disease, median time to tumor-related mortality from treatment of third recurrence was 21 months (range, 3 to 106 months). A total of twelve (60%) patients with parathyroid carcinoma developed distant metastases during the course of disease. The most frequent sites of metastasis were bone (7 of 12 [58%]), lung (6 of 12 [50%]), and mediastinum (5 of 12 [42%]). One patient developed metastatic disease to the adrenal gland. Recurrence of disease was associated with a significant decrease in DSS (5-year DSS, 60% versus 100%). Of 14 patients who relapsed, 3 (16%) became disease-free and are alive without evidence of relapse at 26, 127, and 136 months, respectively, after resection during the initial recurrence. Two of these patients had multiple recurrent parathyroid cancers with mediastinal and distant bone metastases that were completely resected. Discussion This study was undertaken with the aim of conducting a critical analysis of the clinical, histopathologic, and molecular characteristics of parathyroid carcinoma for patients with this disease who were treated and followed-up at Memorial Sloan-Kettering Cancer Center from 1953 to 1995. For the purpose of morphological and molecular comparison, we also analyzed 53 recently (1993 to 2000) treated patients with single parathyroid adenomas, 8 of whom had adenomas belonging to the “atypical” category. We employed tissue microarray technology with IHC to profile multiple cell-cycle regulatory proteins in parathyroid neoplasia. The clinical features of patients with parathyroid carcinoma were characteristic: The majority of patients presented with a palpable neck mass (75%). The latter finding is attributable to the significantly larger tumor size for parathyroid carcinoma relative to typical adenoma. The presence of a palpable mass in the setting of hypercalcemia is not specific for malignancy because this finding was evident in only 2 of 8 atypical adenomas. Our histopathologic classification was based on the criteria set in the parathyroid tumor fascicle of the Armed Forces Institute of Pathology by DeLellis.4 We found that broad, fibrous bands crossing the tumor, as well as a trabecular growth pattern, are statistically more frequent in carcinomas than typical adenomas (P < .001). However, the latter features are not entirely specific for carcinomas. Fibrous bands were seen in all of the “atypical adenomas” and 4 of 46 (9%) typical adenomas. Trabecular growth pattern was present in 6 of 8 (75%) of the atypical adenomas. As expected, there was a statistically significant increase in nuclear pleomorphism, mitotic rate, and atypical mitoses along the spectrum of parathyroid neoplasia (Table 2). The presence of mitosis is by no means an absolute criterion for malignancy because we observed mitosis in 3 of 45 (7%) of the benign typical adenomas. Other investigators have found mitotic activity in parathyroid adenomas at even higher rates than those identified here.21, 22 There is, however, general agreement that in cases involving more than 1 mitosis/10 high-power fields, the specimen should be meticulously searched for other features of malignancy.4 Indeed, we found more than 1 mitosis/10 high-powered fields in 4 of 16 (25%) carcinomas, and 1 of 8 (12.5%) atypical adenomas but in none of the typical benign tumors. In this study, atypical mitoses were identified only in parathyroid carcinomas. The finding of a clearly atypical mitosis is worrisome and for some investigators virtually diagnostic of malignancy.4 Vascular, perineural, and soft-tissue invasion were present only in cases of parathyroid carcinoma because these were used as definite criteria of malignancy. Despite stringent criteria for vascular invasion (ie, the presence of a tumor inside the vessel and attached to its wall), 13 of 16 (81%) carcinomas in this study showed this phenomenon, a much higher rate than the 10% to 15% incidence reported in other studies.3, 4 This difference could be explained by the presence of a much higher percentage of aggressive carcinomas in our patient population reflecting the referral patterns of our institution. Capsular invasion is considered highly predictive of malignancy, but it has been reported in some atypical adenomas.4, 6 In this study, capsular invasion, defined as tongue-like protrusion of tumor cells through the capsule, was found only in carcinomas. The presence of clusters of parathyroid cells entrapped within the capsule was observed in 4 of 8 (50%) atypical and none of the typical parathyroid adenomas. This phenomenon, called “pseudoinvasion,” may superficially resemble true capsular invasion and could be responsible for the cases of “capsular invasion” in atypical tumors reported by others. Perineural invasion, although highly suggestive of carcinoma, is infrequently seen; it was apparent in only 3 of 16 (19%) carcinomas in these series. Although the majority of parathyroid tumors can be classified by morphological examination, histological diagnosis of parathyroid carcinomas remains challenging. Histological parameters identified in this study as being highly specific for malignancy include vascular and perineural invasion and soft-tissue extension. The appraisal of these findings, however, depends on the diligent scrutiny of multiple permanent sections by an experienced pathologist and is somewhat subjective in nature. Even soft-tissue extension can be difficult to assess in certain circumstances. For example, benign parathyroid tissue that has been “spilled” during surgery may regrow within fibrous tissue, mimicking soft-tissue extension and therefore simulating carcinoma at the gross and microscopic levels. There also remains the problem of atypical adenomas. These lesions have some morphological features of malignancy, such as fibrous bands and trabecular growth pattern, but they lack unequivocal evidence of carcinoma such as vascular invasion and soft-tissue extension. At the present time, we cannot predict the behavior of these atypical tumors. From the above discussion, it is clear that there is a need for more reliable diagnostic criteria based on the analysis of the primary tumor. A number of investigators have studied the expression of cell-cycle regulatory proteins in parathyroid tumors.5, 6, 7, 8, 9, 23, 24, 25, 26 Several of these studies analyzed a single molecular marker. Because parathyroid neoplasms may have multiple molecular alterations in their cellular machinery, the characterization of tumor biology requires investigation of various molecules in the same tissues. Some investigators have studied alterations in p53 gene expression, cellular proliferation (Ki-67), and apoptosis (Bcl-2), but other molecules that play pivotal roles in the p53 pathway, such as mdm2 and p21, have yet to be studied in parathyroid tumors.9, 13 Our multimolecular profiling focused on proteins involved in cell-cycle regulation and proliferation. As part of this cellular machinery, we studied important components of the p53 pathway. p53 overexpression was not detected in any parathyroid neoplasms. This is consistent with previous studies and suggests that p53 abnormalities do not play a role in parathyroid tumorigenesis.13, 23, 27 The mdm2 protein inhibits the function of p53 and is overexpressed in human cancers. We found a statistically significant decrease in mdm2 expression with adverse histology. This phenomenon cannot be adequately explained on the basis of our current understanding of mdm2 function and is further evidence against the role of p53 in parathyroid neoplasia. p21 is a cyclin-dependent kinase inhibitor that can be induced by p53. Overexpression of this protein triggers cell-cycle arrest in the G1 phase in proliferating tumor cells.28 We found no significant correlation between p21 expression and the three types of parathyroid neoplasia (Table 3). The absence of p21 overexpression, however, has been correlated with poor prognosis in gastric, colorectal, and bladder carcinomas.29, 30, 31 p53 mediates apoptosis through the Bcl-2/Bax pathway. In this balanced system of apoptosis, Bcl-2 has an antiapoptotic function that may promote tumor growth.32 In our series, the expression of Bcl-2 was significantly higher in adenomas than carcinomas. This curious finding also has been reported by others and represents a phenotype that has been correlated with adverse outcome in colorectal carcinoma.9, 12 This cannot be explained mechanistically outside the complex system of apoptosis-regulating factors, but it reflects the variability of molecular expression profiles in human cancer. The cyclin-dependent kinase inhibitor, p27 protein, regulates cell-cycle progression from the G1 phase to the S phase of the cell cycle. Abnormalities in p27 have caused multiorgan hyperplasia in murine models, which supports its role as a tumor-suppressor gene.33 In this study, typical parathyroid adenomas had the highest p27 positivity rate (36 of 45, 80%), followed by atypical adenoma (3 of 7, 43%) and parathyroid carcinomas (2 of 11, 18%). This statistically significant decrease in p27 positivity is consistent with its role as a potent tumor suppressor and corroborates the results of Erikson et al7 who found a similar p27 pattern of expression in parathyroid neoplasia. The cyclin D1 oncogene product is another regulator of the G1-S transition of the cell cycle. Cyclin D1 gene amplification has been implicated in the progression of many neoplasms. This gene can be overexpressed through a translocation that juxtaposes the parathyroid hormone gene to cyclin D1 in a subset of parathyroid adenomas.34 Hsi et al25 found cyclin D1 overexpression in 18% of parathyroid adenomas and in 2 of 3 parathyroid carcinomas, thereby suggesting that this phenomenon may occur in parathyroid malignancy. Our findings do not support this hypothesis. We found mainly low expression of cyclin D1 among parathyroid neoplasms, suggesting a marginal relevance of this molecule for parathyroid carcinogenesis. Cellular proliferation can be viewed as an end point of several cell-cycle regulatory pathways. Different studies have assessed tumor cell proliferative activity by immunostaining against the Ki-67 antigen that is expressed in all cells that are not in G0 phase of the cell cycle.5, 6, 7, 9, 13, 26 All have shown significantly increased Ki-67 expression in carcinomas relative to adenomas. In this study, proliferative activity—as shown by nuclear immunostaining of Ki-67—was observed in 1 of 41 (2%) parathyroid adenomas and 3 of 11 (27%) carcinomas, consistent with previous reports correlating Ki-67 expression with parathyroid malignancy. Taken together, these findings show that abnormal expression of few critical factors in the cell-cycle regulatory machinery can contribute to unbalanced cellular growth in parathyroid neoplasia. The importance of particular factors may vary between tumor types and between individual tumors. The minor role of the p53 pathway in parathyroid tumors distinguishes these neoplasms from the majority of other solid tumors. Although the expression of some of these cell-cycle regulatory proteins correlates with the histological type of parathyroid tumors, none is specific for carcinoma. We therefore analyzed multimolecular phenotypes using those cell-cycle markers with significant differential expression across parathyroid tumor types. One phenotype was highly associated with typical parathyroid adenoma, p27(+)bcl-2(+)Ki-67(−)mdm2(+); it was present in 29 of 38 (76%) cases and in none of the carcinomas. This phenotype is not pathognomonic for typical adenomas because it was also identified in atypical adenomas; however, it may be useful in differentiating benign from malignant parathyroid tumors. Two other phenotypes, each found in only 2% of cases, were entirely specific but not sensitive for typical adenoma (Table 4). Three molecular phenotypes—p27(+) bcl-2(−)Ki-67(+)mdm2(−), p27(−)bcl-2(−)Ki-67(−) mdm2(−), and p27(−)bcl-2(+)Ki-67(+)mdm2(−)—were identified only in parathyroid carcinomas and found in 1 of 11 (9%), 3 of 11 (27%), and 2 of 11 (18%) of the malignant cases, respectively. Other phenotypes were not tumor specific and were regarded as indeterminate (Table 4). The phenotypic characterization is based on few patients, and comparison with histological criteria cannot be made definitively. Specific phenotypes may support the diagnosis of benign parathyroid nodules; however, they do not allow diagnosis of malignancy in the absence of absolute morphological criteria. The results of this study, as well as our current understanding of the biology of these molecular markers, do not allow definitive explanation as to why certain markers are upregulated or downregulated within a specific phenotypic cluster. For example, one of the multimolecular phenotypes identified in 3 of 11 (27%) carcinomas, p27(−)bcl-2(−)Ki-67(−)mdm2(−), was only observed with this tumor type. If one considers their function, one would expect the last three markers to be upregulated in tumorigenesis and not downregulated as it is the case here. Along the continuum of parathyroid neoplasia that includes benign typical adenomas, atypical adenomas, and invasive carcinomas, one finds marked phenotypic heterogeneity that is most pronounced among malignant tumors. An interesting finding is the variation in phenotypes between primary and recurrent carcinomas within a specific patient and among individuals. This raises obvious concern as to any specificity of IHC molecular phenotyping for parathyroid carcinoma. Thus, molecular phenotypes in these neoplasms are complex, and this complexity increases with progression from benign to malignant phenotype such that formulating targeted therapy based on the primary tumor will be challenging.35 The specific molecular phenotypes identified in this study may be clinically relevant in that they may facilitate differentiation between benign and malignant neoplasms. The phenotype, p27(+)bcl-2(+)Ki-67(−)mdm2(+), generally was unique to adenomas and specifically characterized 29 of 38 (76%) of typical parathyroid adenomas. However, one should not underestimate the limitations of IHC multimolecular marker phenotyping. First, the reading of the immunohistochemically stained arrays is subjective and dependent on the observer. Second, even the most specific phenotype for carcinoma, p27(−)bcl-2(+)Ki-67(+) mdm2(−), is not as sensitive (3 of 11, 27%) as classic morphological parameters. For example, soft-tissue extension is 100% specific and was present in 11 of the 16 (69%) parathyroid carcinomas cases in our study. Our view is that presently these phenotypes will complement but not replace a careful morphological evaluation in parathyroid tumors. We believe that morphology remains the “gold standard” for the diagnosis of parathyroid neoplasms at the present time. Nine of 14 (64%) carcinoma patients whose disease recurred experienced recurrence in the neck. Recurrence of parathyroid carcinoma was associated with unfavorable outcome; only 3 of 14 (21%) patients with recurrent disease were subsequently cured, and 9 of 14 (64%) died of disease by the time of last clinical follow-up. Median disease-free interval after initial resection ranged from 5 to 84 months, and time to second recurrence after treatment of the first recurrence was 3 to 52 months. In patients who died of disease after multiple local recurrences, overall survival ranged from 32 to 178 months. These findings thereby typify the natural history of parathyroid carcinoma marked by an indolent, multiply recurrent course.36 Moreover, the importance of lifelong follow-up of patients with this disease is evident. Because the majority of patients with parathyroid carcinoma die secondary to metabolic abnormalities rather than direct effects of the tumor, control of hypercalcemia is the principal long-term goal of treatment. Aggressive, often repeated resection of functional local and distant recurrences is the most effective means of achieving this end.37 Our findings support the principle that reoperation for recurrent parathyroid carcinoma is often performed with palliative but seldom curative intent. The present study is a critical appraisal of the parathyroid neoplasms on the basis of clinical features, morphology, and multimolecular tissue-array—based IHC profiling. Although our findings improve the understanding of these tumors in general, parathyroid carcinoma remains an enigmatic disease, the biological features of which are incompletely defined. Nevertheless, the new techniques of molecular profiling hold potential to further enhance our understanding of this rare tumor. Acknowledgements The authors are grateful to Denis H. Y. Leung for his contributions to the statistical analyses and to David Kuo for his expert technical assistance. References 1.
1
Koea JB, Shaw JH.
Parathyroid cancer: Biology and management.
Surg Oncol. 1999;8:155–165. Abstract | Full Text |
Full-Text PDF (230 KB)
|
CrossRef
2.
2
Ghandur-Mnaymneh L, Kimura N.
The parathyroid adenoma: A histopathologic definition with a study of 172 cases of primary hyperparathyroidism.
Am J Pathol. 1984;115:70–83. MEDLINE 3.
3
Schantz A, Castleman B.
Parathyroid carcinoma: A study of 70 cases.
Cancer. 1973;31:600–605. 4.
4
DeLellis R.
Tumors of the parathyroid gland.
In:
Rosai J, Sobin L editor.
Atlas of Tumor Pathology (3rd series, Fascicle-6). Washington, DC: Armed Forces Institute of Pathology:; 1993;. 5.
5
Abbona GC, Papotti M, Gasparri G, et al.
Proliferative activity in parathyroid tumors as detected by Ki-67 immunostaining.
Hum Pathol. 1995;26:135–138. Abstract |
Full-Text PDF (443 KB)
|
CrossRef
6.
6
Vargas MP, Vargas HI, Kleiner DE, et al.
The role of prognostic markers (MIB-1, RB, and bcl-2) in the diagnosis of parathyroid tumors.
Mod Pathol. 1997;10:12–17. MEDLINE 7.
7
Erickson LA, Jin L, Wollan P, et al.
Parathyroid hyperplasia, adenomas, and carcinomas: Differential expression of p27 protein.
Am J Surg Pathol. 1999;23:288–295. MEDLINE |
CrossRef
8.
8
Vasef MA, Brynes RK, Sturm M, et al.
Expression of cyclin D1 in parathyroid carcinomas, adenomas, and hyperplasias: A paraffin immunohistochemical study.
Mod Pathol. 1999;12:412–416. MEDLINE 9.
9
Naccarato AG, Marcocci C, Miccoli P, et al.
Bcl-2, p53 and MIB-1 expression in normal and neoplastic parathyroid tissues.
J Endocrinol Invest. 1998;21:136–141. MEDLINE 10.
10
Levine AJ, Momand J, Finlay CA.
The p53 tumor suppressor gene.
Nature. 1991;351:453–456. MEDLINE |
CrossRef
11.
11
Cordon-Cardo C.
Mutation of cell cycle regulators: Biological and clinical implications for human neoplasia.
Am J Pathol. 1997;147:545–560. MEDLINE 12.
12
Buglioni S, D'Agnano I, Cosimelli M, et al.
Evaluation of multiple bio-pathological factors in colorectal adenocarcinomas: Independent prognostic role of p53 and Bcl-2.
Int J Cancer. 1999;84:545–552. MEDLINE |
CrossRef
13.
13
Wang W, Johansson H, Kvasnicka T, et al.
Detection of apoptotic cells and expression of Ki-67 antigen, Bcl-2, p53 oncoproteins in human parathyroid adenoma.
APMIS. 1996;104:789–796. MEDLINE |
CrossRef
14.
14
Hoos A, Urist MJ, Stojadinovic A, et al.
Validation of tissue microarrays for immunohistochemical profiling of cancer specimens using the example of human fibroblastic tumors.
Am J Pathol. 2001;158:1245–1251. MEDLINE 15.
15
Kononen J, Bubendorf L, Kallioniemi A, et al.
Tissue microarrays for high-throughput molecular profiling of tumor specimens.
Nat Med. 1998;4:844–847. MEDLINE |
CrossRef
16.
16
Altman DG.
Describing data.
In:
Altman DG editors.
Practical Statistics for Medical Research. London: Chapman and Hall; 1996;p. 404–408. 17.
17
Mehta C, Patel N.
A network algorithm for performing Fisher's exact test in rXc contingency tables.
J Am Stat Assoc. 1983;78:427–434. 18.
18
Fisher RA, Yates F.
Statistical Tables for Biological, Agricultural and Medical Research.
ed 6. Edinburgh: Oliver & Boyd; 1964;
. 19.
19
Kaplan E, Meier P.
Nonparametric estimation from incomplete observations.
J Am Statist Assoc. 1958;53:457–481. 20.
20
Peto R, Peto J.
Asymptotically efficient rank invariant procedures.
J R Stat Soc A. 1972;135:185–207. 21.
21
Snover DC, Foukar K.
Mitotic activity in benign parathyroid disease.
Am J Clin Pathol. 1981;75:345–347. MEDLINE 22.
22
San Juan J, Monteagudo C, Fraker D, et al.
Significance of mitotic activity and other morphologic parameters in parathyroid adenomas: Correlation with clinical behavior.
Am J Clin Pathol. 1989;92:523;
. 23.
23
Kishikawa S, Shan L, Ogihara K, et al.
Overexpression and genetic abnormality of p53 in parathyroid adenomas.
Pathol Int. 1999;49:853–857. MEDLINE |
CrossRef
24.
24
Karak AK, Sarkar S, Chumber S, et al.
MIB-1 proliferative index in parathyroid adenoma and hyperplasia.
Indian J Med Res. 1997;105:235–238. MEDLINE 25.
25
Hsi ED, Zuckerburg LR, Yang W, et al.
Cyclin D1/PRAD1 expression in parathyroid adenomas: An immunohistochemical study.
J Clin Endocrinol Metab. 1996;81:1736–1739.
CrossRef
26.
26
Kameyama K, Takami H, Umemura S, et al.
PCNA and Ki-67 as prognostic markers in human parathyroid carcinomas.
Ann Surg Oncol. 2000;7:301–305. MEDLINE |
CrossRef
27.
27
Hakim JP, Levine MA.
Absence of p53 point mutations in parathyroid adenomas and carcinomas.
J Clin Endocrinol Metab. 1994;78:103–106.
CrossRef
28.
28
Zhang H, Xeong Y, Beach D.
Proliferating cell nuclear antigen and p21 are components of multiple cell cycle kinase complexes.
Mol Biol Cell. 1993;4:897–906. MEDLINE 29.
29
Ogawa M, Maeda K, Onoda N, et al.
Loss of p21WAF1/DIP1 expression correlates with disease progression in gastric carcinoma.
Br J Cancer. 1997;75:1617–1620. MEDLINE 30.
30
Zirbes TK, Baldus SE, Moeing SP, et al.
Prognostic impact of p21/waf1/cip1 in colorectal cancer.
Int J Cancer. 2000;89:14–18. MEDLINE |
CrossRef
31.
31
Stein JP, Ginsberg DA, Grossfeld GD, et al.
Effect of p21WAF1/CIP1 expression on tumor progression in bladder cancer.
J Natl Cancer Inst. 1998;90:1072–1079. MEDLINE 32.
32
Chao DT, Korsmayer SJ.
Bcl-2 family: Regulators of cell death.
Ann Rev Immunol. 1998;16:395–419. 33.
33
Fero ML, Rivkin M, Tasch M, et al.
A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27 (Kip1)-deficient mice.
Cell. 1996;85:733–744. MEDLINE |
CrossRef
34.
34
Hemmer S, Wasenius VM, Haglund C, et al.
Deletion of 11q23 and cyclin D1 overexpression are frequent aberrations in parathyroid adenomas.
Am J Pathol. 2001;158:1355–1362. MEDLINE 35.
35
Hoos A, Lewis JJ, Antonescu CR, et al.
Characterization of molecular abnormalities in human fibroblastic neoplasms—A model for genotype-phenotype association in soft tissue tumors.
Cancer Res. 2001;61:3171–3175. MEDLINE 36.
36
Vetto JT, Brennan MF, Woodruff JM, et al.
Parathyroid carcinoma: Diagnosis and clinical history.
Surgery. 1993;114:882–892. MEDLINE 37.
37
Flye MW, Brennan MF.
Surgical resection of metastatic parathyroid carcinoma.
Ann Surg. 1981;193:425–435. MEDLINE Departments of Surgery and Pathology and the Laboratory of Epithelial Cancer Biology, Memorial Sloan-Kettering Cancer Center, New York, NY. PII: S0046-8177(03)00010-8 doi:10.1053/hupa.2003.55 © 2003 Elsevier Science (USA). All rights reserved. | |
|
FULL TEXT00010-8/journalimage?img=PIIS0046817703000108.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0046-8177&ishighres=false&allhighres=false&free=yes','journalimage');)
00010-8/journalimage?img=PIIS0046817703000108.gr2.lrg.gif&fig=&kwhquery=&issn=0046-8177&ishighres=false&allhighres=false&free=yes','journalimage');)