Sonography depicts the internal structure of the thyroid gland and the regional anatomy and pathology without using ionizing radiation or iodine containing contrast medium. [3,4] Rather, high frequency sound waves in the megahertz range (ultrasound), are used to produce an image. The procedure is safe, does not cause damage to tissue and is less costly than any other imaging procedure. The patient remains comfortable during the test, which takes only a few minutes, does not require discontinuation of any medication, or preparation of the patient. The procedure is usually done with the patient reclining with the neck hyperextended but it can be done in the seated position. A probe that contains a piezoelectric crystal called a transducer is applied to the neck but since air does not transmit ultrasound, it must be coupled to the skin with a liquid medium such a gel. This instrument rapidly alternates as the generator of the ultrasound and the receiver of the signal that has been reflected by internal tissues. The signal is organized electronically into numerous shades of gray and is processed electronically to produce an image instantaneously (real-time). Although each image is a static picture, rapid sequential frames are processed electronically to depict motion. Two-dimentional images have been standard and 3-dimentional images are an improvement in certain circumstances. [4A] There is considerable potential for improving ultrasound images of the thyroid by using ultrasound contrast agents. These materials are gas-filled microbubbles with a mean diameter less than that of a red blood corpuscle and are injected intravenously. [5]

Dynamic information such as blood flow can be added to the signal by employing a physics principle called the Doppler effect. The Doppler signals, which are superimposed on real time gray scale images, are extremely bright in black and white images and may be color coded to reveal the velocity (frequency shift) and direction of blood flow (phase shift) as well as the degree of vascularity of an organ. [6,7] Flow in one direction is made red and in the opposite direction, blue. The shade and intensity of color can correlate with the velocity of flow. Thus, in general terms, venous and arterial flow can be depicted by assuming that flow in these two kinds of blood vessels is parallel, but in opposite directions. Since portions of blood vessels may be tortuous, modifying orientation to the probe, different colors are displayed within the same vessel even if the true direction of blood flow in that vessel has not changed. Thus, an analysis of flow characteristics requires careful observations and cautious interpretations. The absence of flow in a fluid-filled structure can differentiate a cystic structure and a blood vessel.

The ultrasound is treated differently by the various tissues. [1,4] The air-filled trachea does not transmit the ultrasound. Calcified tissues such as bone and sometimes cartilage and calcific deposits in other anatomic structures block the passage of ultrasound resulting in a very bright signal and a linear echo-free shadow distally. Most tissues transmit the ultrasound to varying degrees and interfaces between tissues reflect portions of the sound waves. Fluid-filled structures have a uniform echo-free appearance whereas fleshy structures and organs have a ground glass appearance that may be uniform or heterogeneous depending on the characteristics of the structure.

The depth penetration and resolving power of ultrasound depends greatly on frequency. [3] Depth penetration is inversely related and spatial resolution is directly related to the frequency of the ultrasound. For thyroid, a frequency of 7.5 to 10 or 14 megahertz is generally optimal for all but the largest goiters. Using these frequencies, nodules as small as two to three millimeters can be identified.

Routine protocols for sonography are not adequate. Although some technologists become extremely proficient after specific training and experience, supervision and participation by a knowledgeable and interested physician-sonographer is usually required to obtain a precise and pertinent answer to a specific problem that has been posed by the clinician. Standard sonographic reports may provide considerable information about the anatomy, but are suboptimal unless the specific clinical concern is explored and answered. Indeed, because some radiologists cannot address the clinical issue adequately, and for convenience, numerous thyroidologists perform their own ultrasound examinations, in which case it is essential that they have state-of-the-art equipment (that might not be cost-effective) and that they are willing to expend a considerable amount of time for a complete study. Technical ingenuity, electronic enhancements such as Doppler capability, and even artistry are frequently required. Special maneuvers, various degrees of hyperextension of the neck, swallowing to the facilitate elevation of the lower portions of the thyroid gland above the clavicles, swallowing water to identify the esophagus, and a Valsalva maneuver to distend the jugular veins may enhance the value of data. Nevertheless, sonography is rather difficult to interpret in the upper portion in of the jugular region and in the areas adjacent to the trachea. Sonography is generally not useful below the clavicles.

It is informative for orientation to survey the entire thyroid gland with a low-energy transducer before proceeding to 10-14 megahertz equipment to delineate the fine anatomy. Protocols have been devised to assemble a montage of images to encompass an unusually large lobe or goiter. For an overview, panoramic ultrasound, which is a variation of conventional ultrasound has been reported to produce images with a large anatomic field of view, displaying both lobes of the thyroid gland on a single image.[5A]

There may be considerable differences between sonologists in estimating the size of large goiters or nodules. One investigation has reported that curved-array transducers avoid significant inter-observer variation that may occur when linear-array equipment is employed, especially when the gland is enlarged. [5B] The inter-observer variation may be almost 50% among experienced ultrasonographers for the determination of the volume of thyroid nodules, because it is difficult to reproduce a two-dimensional image plane for multiple studies. [5C] Accuracy in volume estimation becomes most important when one uses ultrasound measurements to calculate an isotope dose or to compare changes over time in the size of a nodule or a goiter. Using planimetry from three-dimensional images reportedly has lower intra-observer variability (3.4%) and higher repeatability (96.5%) than the standard ellipsoid model for nodules and lobes, with 14.4% variability and 84.8% repeatability (p < 0.001). [5D]