In terms of correlation between physical exam findings and ultrasound findings, a number of studies have shown a weak correlation. In a sample of 2441 patients, physical exam was able to palpate nodules in 169 patients, while ultrasounds located nodules in 249 patients. Of the nodules located on ultrasonography, only 21% were palpable, and of those 169 palpable nodules, 115 were not confirmed by ultrasound. It was notable that only 6.4% of nodules less than 0.5 cm were identified by physical exam, while only about half (48%) of nodules larger than 2 cm were identified with palpation [18]. Additionally, inter-physician variability for palpation of thyroid nodules has been shown to be relatively high [19]. Thus, ultrasound is the preferred and reliable method for evaluation of nodules as palpation can be elusive and provide inaccurate information. This does not imply that a physical exam is not necessary, but rather that findings need to be confirmed by further testing.

Characterization of the nodule is important as there are features that increase the suspicion for malignancy and help the clinician assess for the need for a fine needle aspiration (FNA). Features to account for during imaging include the structural makeup of the nodule (solid vs. cystic vs. spongiform), presence of calcifications (none vs. micro vs. macro), invasion into surrounding structures, vascularity, lymphadenopathy, and location of nodule [20, 21]. In a study at the Mayo Clinic, 360 consecutively surgically removed thyroids showed that the majority (88%) were solid with only 9% having up to 50% cystic component and 3% having more than 50% cystic component [22].

In a case control study, the three characteristics most associated with presence of cancer were microcalcifications, size greater than 2 cm, and an entirely solid composition. The suggestion was made to only perform biopsies if at least two of those characteristics were met to decrease the number of unnecessary biopsies and false-positive results. Biopsy criteria based on only one characteristic would have a low positive predictive value, and those based on all three characteristics would miss diagnosis of nodules which justifiably should be biopsied [20].

One of the most important features to assess is the size of the nodule. Although size of the thyroid nodule has not been shown to be associated with the presence of malignancy in a thyroid nodule in numerous studies, it is still a major criterion for determination for the need of an FNA [23]. The rationale for the use of nodule size during the evaluation is centered around the concept of prognosis and overall survival. Nodules of small size have a low rate of metastasis and overall mortality, even when containing cancer [12]. It has been shown that the probability of distant metastasis does not seem to be significant until tumor size is >20 mm, although spread to lymph nodes and extrathyroidal growth do occur with smaller nodules of PTC [24]. This rationale is supported by findings on autopsies of patients with no previously known thyroid disease in which small thyroid nodules with malignancy were found. In the United States, over half a million thyroid ultrasounds are performed annually. If every nodule, regardless of size or characteristics, was biopsied, the number of unnecessary procedures and false-positive results would lead to a great number of unnecessary surgeries.

A more recent emerging evaluation technique is ultrasound elastography. The evaluation is performed by using the B-mode of the ultrasound to evaluate changes in the nodule with applied pressure. Nodules with partial or no elasticity are associated with higher levels of malignancy [25].

CT has a limited role in the workup of thyroid nodules. The main limitation is the inability to help distinguish benign from malignant nodules, as well as the associated cost increase. The main utility is for evaluation of nodules that are substernal and inaccessible to ultrasound evaluation [26].

MRI is not a commonly used modality for evaluation of a thyroid nodule, but studies have shown that it can be used to help differentiate a malignant nodule from a benign nodule fairly accurately [27].

PET/CT has been shown to have a high level of sensitivity when evaluating thyroid nodules. In a small study, it has been shown to be efficient in cases where the FNA has shown FLUS or AUS [28]. While theoretically it might add diagnostic information, insufficient sensitivity and specificity have limited its routine use. In the majority of cases, thyroid nodule is noted on PET/CT incidentally while evaluating a different malignancy. At the current time, a major limitation to a PET/CT is the high cost and difficulty of access [28].