Format

Send to

Choose Destination
Endocr Pathol. 2019 Nov 16. doi: 10.1007/s12022-019-09592-3. [Epub ahead of print]

Using a Novel Diagnostic Nomogram to Differentiate Malignant from Benign Parathyroid Neoplasms.

Author information

1
Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Unit 1484, 1515 Holcombe Boulevard, Houston, TX, 77030, USA.
2
Division of Surgery, Universidad Finis Terrae, Santiago, Chile.
3
Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
4
Division of Endocrine Surgery, Hospital Universitari de Bellvitge, Barcelona, Spain.
5
Division of Surgical Oncology, Medical College Wisconsin, Milwaukee, WI, USA.
6
Department of Endocrine Neoplasia and Hormonal Disorders, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
7
Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Unit 1484, 1515 Holcombe Boulevard, Houston, TX, 77030, USA. nperrier@mdanderson.org.
8
Department of Pathology, The University of Texas MD Anderson Cancer Center, Unit 85, 1515 Holcombe Boulevard, Houston, TX, 77030, USA. mdwillia@mdanerson.org.

Abstract

We sought to develop an immunohistochemical (IHC) tool to support the diagnosis of parathyroid carcinoma (PC) and help differentiate it from atypical parathyroid neoplasms (atypical) and benign adenomas. Distinguishing PC from benign parathyroid neoplasms can be challenging. Many cases of PC are histopathologically borderline for definitive malignancy. Recently, individual IHC biomarkers have been evaluated to aid in discrimination between parathyroid neoplasms. PC, atypical parathyroid neoplasms, and parathyroid adenomas treated at our institution from 1997 to 2014 were studied retrospectively. IHC analysis was performed to evaluate parafibromin, retinoblastoma (RB), protein gene product 9.5 (PGP9.5), Ki67, galectin-3, and E-cadherin expression. Receiver operating characteristic (ROC) analysis and multivariable logistic regression model for combinations of biomarkers were evaluated to classify patients as PC or atypical/adenoma. A diagnostic nomogram using 5 biomarkers was created for PC. Sixty-three patients were evaluated. The percent staining of parafibromin (p < 0.0001), RB (p = 0.04), Ki67 (p = 0.02), PGP9.5 (p = 0.04), and Galectin-3 (p = 0.01) differed significantly in the three diagnostic groups. ROC analysis demonstrated that parafibromin had the best performance in discriminating PC from atypical/adenoma; area under the curve (AUC) was 81% (cutoff, 92.5%; sensitivity rate, 64%; specificity rate, 87%). We created a diagnostic nomogram using a combination of biomarkers; AUC was 84.9% (95% confidence interval, 73.4-96.4%). The optimism-adjusted AUC for this model was 80.5% (mean absolute error, 0.043). A diagnostic nomogram utilizing an immunoexpression, a combination of immunohistochemical biomarkers, can be used to help differentiate PC from other parathyroid neoplasms, thus potentially improving diagnostic classification.

KEYWORDS:

Biomarkers; Nomograms; Parathyroid cancer; Parathyroid neoplasms

PMID:
31734935
DOI:
10.1007/s12022-019-09592-3

Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center