Proceedings Article | 23 February 2010
KEYWORDS: Tissues, Raman spectroscopy, Collagen, Proteins, Injuries, Diagnostics, Cancer, Ultrasonography, Tissue optics, Spectroscopy
Thyroid gland is a small gland in the neck consisting of two lobes connected by an isthmus. Thyroid's main function is to
produce the hormones thyroxine (T4), triiodothyronine (T3) and calcitonin. Thyroid disorders can disturb the production of
these hormones, which will affect numerous processes within the body such as: regulating metabolism and increasing
utilization of cholesterol, fats, proteins, and carbohydrates. The gland itself can also be injured; for example, neoplasias,
which have been considered the most important, causing damage of to the gland and are difficult to diagnose. There are
several types of thyroid cancer: Papillary, Follicular, Medullary, and Anaplastic. The occurrence rate, in general is between 4
and 7%; which is on the increase (30%), probably due to new technology that is able to find small thyroid cancers that may
not have been found previously. The most common method used for thyroid diagnoses are: anamnesis, ultrasonography, and
laboratory exams (Fine Needle Aspiration Biopsy- FNAB). However, the sensitivity of those test are rather poor, with a high
rate of false-negative results, therefore there is an urgent need to develop new diagnostic techniques. Raman spectroscopy
has been presented as a valuable tool for cancer diagnosis in many different tissues. In this work, 27 fragments of the thyroid
were collected from 18 patients, comprising the following histologic groups: goitre adjacent tissue, goitre nodular tissue,
follicular adenoma, follicular carcinoma, and papillary carcinoma. Spectral collection was done with a commercial FTRaman
Spectrometer (Bruker RFS100/S) using a 1064 nm laser excitation and Ge detector. Principal Component Analysis,
Cluster Analysis, and Linear Discriminant Analysis with cross-validation were applied as spectral classification algorithm.
Comparing the goitre adjacent tissue with the goitre nodular region, an index of 58.3% of correct classification was obtained.
Between goitre (nodular region and adjacent tissue) and papillary carcinoma, the index of correct classification was 64.9%,
and the classification between benign tissues (goitre and follicular adenoma) and malignant tissues (papillary and follicular
carcinomas), the index was 72.5%.
