Thyroid volume estimation from ultrasound extend field of view images using star algorithm with green's theorem

• Main Author: Esmaile, Maher Faik
• Format: Thesis
• Language: English
• Published: 2013
• Online Access: http://psasir.upm.edu.my/id/eprint/48169/1/FK%202014%2049R.pdf; http://psasir.upm.edu.my/id/eprint/48169/

Area calculation is regarded as an important step in many applications (e.g. industry, medicine, biology). Greens’ theorem is one of the methods that can be applied to calculate the area with the advantages: it evaluates a double integral over the region of an object by a simple integration along the boundary of the object, and it can result in high accuracy for the area calculations by increasing the edge’s point of the object. In this work a modified Star algorithm is incorporated with Greens’ theorem (SAGT) to calculate the area of an arbitrary shape. As an application to this modification, it has been applied in medical field to calculate the actual area of thyroid gland from ultrasound extend field of view (EFOV) images and for several slices along the neck. Then area summation technique will be used to estimate the volume of thyroid gland. Diagnosis of thyroid gland diseases and correct radioiodine dosage depend on the correct thyroid gland volume estimation. Different approaches used to determine thyroid size including ultrasound, scintigraphy, SPECT, CT scan, and MRI. Recently, ultrasonography is the preferred aid in clinical diagnosis. The use of ultrasound to estimate thyroid volume involves several disadvantages. First; the application of ellipsoid formula with thyroid gland ultrasound images to evaluate thyroid volume which is applied to irregular shape is reported to be inaccurate and may result in underestimation or overestimation. Second; the thyroid lobe length is sometimes difficult to be measured, because most modern transducers are available with a footpad of 4 cm or less which is less than the adult’s thyroid lobe length. Third; the accuracy of ultrasound thyroid gland volume estimation depends on operator skill. A new methodology is proposed to estimate the volume of thyroid gland using SAGT and ultrasound (EFOV) images through four steps: first; modifying Star algorithm with Green’s theorem (SAGT) to calculate the actual thyroid area instead of measuring the thyroid dimensions manually that avoids using the ellipsoid formula. Second; the neck scanning is systemised in order to minimize the operator experience effect and to create the EFOV image during the registration process. Third; a new edge detection filter suitable to this application is designed that’s based on the extraction and registration of the ultrasound image features to be used with SAGT to calculate the thyroid actual area from US EFOV images. Fourth: applying the summation of area technique to calculate the volume of thyroid gland. The experimental results conducted for the phantom neck showed that the proposed method could perform the volume estimation with accuracy between (96-98) % as compared to the ellipsoid formula that has accuracy between (83-92) %. It’s concluded that, the area calculation of thyroid gland using SAGT, with systemising the neck scanning has shown an improving accuracy for the thyroid gland volume estimation.