Date of Award

January 2013

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

David L. Rimm

Subject Area(s)



Thyroid fine-needle aspiration (FNA) biopsy, the preoperative diagnostic standard of care for thyroid nodules, has limitations. Spectral imaging captures visible light information beyond the human eye's capability, potentially increasing the accuracy of FNA biopsy. We sought to demonstrate the feasibility of using spectral imaging in combination with computer-based spatial analysis for thyroid FNA classification. To do this, we developed two classifiers: one that distinguishes between spectral images of papillary thyroid carcinoma (PTC) and benign goiter (BG), and a second that distinguishes between follicular carcinoma (FC) and follicular adenoma (FA). Spectral images representing 100 cases of PTC and goiter and 14 cases of FC and FA were taken with a multispectral camera. Used in conjunction with commercial software (InForm), 10 PTC and goiter cases were used as a training set to develop a PTC/BG "classifier," a classification algorithm that segments digitized multispectral images into regions of PTC, benign goiter, and "non-feature." This algorithm was used to generate a screening test and a diagnostic test which were validated on an independent set of images representing 30 PTC and 30 benign goiter cases. The FC and FA cases were also used to train a separate FC/FA classifier which was tested on distinct images taken from the same FC and FA cases used for training. The area under the Receiver Operator Curve (ROC) for the PTC/BG classifier was 0.90. The screening test had a sensitivity of 0.93 and a specificity of 0.73. The diagnostic test had a sensitivity of 0.70 and a specificity of 0.90. The area under the ROC of the FC/FA classifier was 0.76. These data demonstrate the potential value of spatial spectral imaging as an adjunct test for classification of thyroid FNAs.


This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.