Diagnostic and prognostic markers for thyroid cancer
Loading...
Date
2014
Authors
Pathak, Kumar Alok
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Statement of the problem: The objectives of our study were to evaluate the debatable
diagnostic role of FDG-PET/CT in predicting the risk of malignancy in follicular neoplasm
and to build prognostic models to predict the risks of relapse and death from thyroid cancer, as there is no universally acceptable model valid for different histological types of thyroid cancers.
Methods: The efficacy of FDG-PET/CT scan for predicting the risk of malignancy was
assessed in a prospective cohort of 50 follicular neoplasms. Disease specific and relapse
free survivals of a 2306-patient Manitoba thyroid cancer cohort were estimated by the
Kaplan-Meier method and the independent influence of various prognostic factors was
assessed by Cox Proportional Hazard models. Cumulative incidence of deaths and relapses from thyroid cancer was calculated by competing risk analysis, and was used to develop and validate prognostic nomograms, using R version 2.13.2 (www.r-project.org). A web-based prognostic model was developed to predict the disease specific survival, and validated internally and externally on an independent patient cohort from London, Ontario.
Results: FDG PET/CT had an overall accuracy of 81% in predicting risk of malignancy in
non-Hürthle follicular cell neoplasms and 87% accuracy in distinguishing follicular and
Hürthle cell adenomas. The age standardized incidence of thyroid cancer in Manitoba
increased by 373% from 1970 to 2010, with the proportion of papillary cancers increasing
from 58% to 85.9%, and that of anaplastic cancer falling from 5.7% to 2.1% (p<0.001).
The disease specific survival was adversely influenced by anaplastic histology, male
gender, stage IV disease, incomplete surgical resection and age at diagnosis, during a
median follow-up of 11.5 years. Prognostic nomograms were designed to predict the
individualized 10-year risks of death and relapse from thyroid cancer, with their respective
concordance indices of 0.92 and 0.76. A web-based model was successfully developed and externally validated with excellent discrimination. It compared favorably with the existing staging/risk stratification systems.
Conclusions: FDGPET/CT has a very good accuracy of predicting risk of malignancy in
non-Hürthle follicular cell neoplasms. We have successfully developed and validated
prognostic nomograms and a web model for predicting oncological outcome of thyroid
cancer.
Description
Keywords
Thyroid Cancer, Prognostic, Diagnosis