Abstract
All observations “depend on an unequivocal relation between the observation and the physical phenomena on which it is based”.1
When thyroid nodules are suspected by palpation, their presence must be confirmed, they must be more precisely localized, and their nature discerned. The first of these tasks is not a problem due to the superb detection capabilities of conventional US. These are emphasized in recent guideline papers (1 –4). Thus, “ultrasound” or “sonography” appears in 13 of the 80 recommendations of the 2009 American Thyroid Association (ATA) guidelines for patients with thyroid nodules and differentiated thyroid cancer (3) and is found throughout “The European Consensus for the Management of Patients with Differentiated Thyroid Carcinoma of the Follicular Epithelium” (2). All (1 –4) advocate thyroid US as the most accurate imaging technique for detecting thyroid nodules. The ATA guidelines state, “Thyroid US should be performed in all patients with a suspected thyroid nodule, nodular goiter, or radiographic abnormality” (3), while the European consensus statement states “currently, thyroid US is the most accurate imaging technique for the detection of thyroid nodules and this procedure is mandatory when a nodule is discovered at palpation” (2). These recommendations are made not only to verify the presence of a palpable nodule, but also to determine if other nodules or “suspicious cervical nodes” are present (2).
US is also important during the first year of follow-up and later to detect lesions that might represent residual, recurrent, or metastatic thyroid cancer. The ATA guidelines explicitly state that “following surgery, cervical US to evaluate the thyroid bed and central and lateral cervical nodal compartments should be performed at 6–12 months and then periodically, depending on the patient's risk for recurrent disease and thyroglobulin status” (3). The European consensus states, in reference to long-term follow-up, that “neck US is performed either routinely or in patients with suspicious clinical findings” (2).
Another role of US is to guide invasive diagnostic procedures (1 –4). The ATA guidelines state “US guidance for fine needle aspiration (FNA) is recommended for those nodules that are nonpalpable, predominantly cystic, or located posteriorly in the thyroid lobe” (3). The European consensus advocates US as “a guide to fine needle aspiration cytology” (2). A key recommendation of The American Association of Clinical Endocrinologists, Associazione Medici Endocrinologi, and European Thyroid Association (AACE/AMA/ETA) guidelines is that “cytologic diagnosis is more reliable and the nondiagnostic rate is lower when FNA biopsy is performed with US guidance” (4). Recommendation 6 of the ATA guidelines states that “US guidance should be used when repeating the FNA procedure for a nodule with an initial nondiagnostic cytology result” (3). Elsewhere, in recommendation 21 of the ATA guidelines, it is stated that “US-guided FNA of sonographically suspicious lymph nodes should be performed to confirm malignancy if this would change management” (3). In at least one guideline paper, that of the Society of Radiologists (SOR), only US-guided biopsy procedures are considered (1). In the SOR guidelines another role for US is discussed (1). This is to determine which thyroid nodules should undergo FNA. The SOR US consensus conference states, for example, that “the maximum diameter should be used when considering whether or not US-guided FNA should be performed” and later “the recommended minimal size for US-guided FNA is lower for solid or predominantly solid nodules than the recommended minimal size for mixed solid and cystic nodules” (1). Similar considerations are found in the ATA and AACE/AME/ETA guidelines (3,4).
To summarize current guidelines, it is evident that conventional US is well established for detecting and localizing thyroid nodules, determining their size, distinguishing solid from cystic nodules, and determining the need for an FNA. The question to be considered is whether US or related procedures have an additional role in the management of thyroid nodules. Can their readings, in themselves, be used to decide the probability of a thyroid nodule being malignant?
As noted later, many characteristics of thyroid nodules as viewed by US have been described (3,5,6). Some are suggestive of malignancy and others of a benign nodule. In the November 2009 issue of Thyroid, Park et al. (6) propose an equation, based on the analysis of 12 aspects of thyroid US readings, for predicting the probability of a nodule being malignant. Their scheme includes five US categories: benign, probably benign, indeterminate, probably malignant, and malignant. They report a significant correlation between these US readings and FNA cytology, concluding that “the usefulness of this requires confirmation by a prospective study” (6).
Elsewhere in this issue Lee et al. (7) describe 10 patients with 0.6- to 4.2-cm hyalinizing trabecular tumors (HTT) of the thyroid, an entity that almost always runs a benign course. In the preoperative US, the lesion usually appeared to be a follicular neoplasm. Consistent with this, nine were read as indeterminate and one as benign. In contrast to the US readings, the FNA diagnosis was papillary thyroid carcinoma (PTC) in six of the patients, in two it was suspicious for PTC, and in two patients it was PTC vs. HTT. All of the patients had thyroid surgery with a final diagnosis of HTT on histopathology. The study supports others (8,9) showing that HTT is often misjudged as PTC on FNA. If the Lee et al. study (7) is confirmed, HTT would be one of the thyroid lesions that is more reliably evaluated by US than by FNA.
Thyroid cancers have a harder consistency than benign thyroid nodules. Real-time elastography (RTE) provides an image output that reflects the firmness of thyroid nodules. Some of the more expensive US devices can also perform RTE. A paper by Bojunga et al. (10) published in last year's October issue of Thyroid is a meta-analysis of the ability of RTE to differentiate benign from malignant thyroid nodules. In their analysis of 639 thyroid nodules, the authors calculated a sensitivity of 92% and a specificity of 90% for the diagnosis of malignant thyroid nodules by RTE. Taken at face value, these numbers compare favorably with FNA.
The recent ATA guidelines and other sources include excellent summaries of the US characteristics of benign and malignant thyroid nodules (3). These relate to nodule symmetry (tallness), calcification, echogenicity, echotexture, margin characteristics, size, and composition. Interestingly, there are differences between the US characteristics of different types of differentiated thyroid carcinoma. As noted in The ATA guidelines, “a PTC is generally solid or predominantly solid and hypoechoid, often with infiltrative irregular margins and increased nodular vascularity (3). Microcalcifications, if present, are highly specific for PTC, but may be difficult to distinguish from colloid. Conversely, follicular cancer is more often iso- to hyperechoic and has a thick and irregular halo, but does not have microcalcifications” (3). A predominant aspect of benign nodules is a spongiform appearance. Pure cysts, as has been recognized for many years, are rarely malignant.
Although ATA guidelines discuss the US features of malignant and benign thyroid nodules, they and other guidelines do not incorporate this information into formal recommendations as they do for FNA. Recommendation 4 of the ATA Guidelines states, “FNA is the procedure of choice in the evaluation of thyroid nodules” (3). In the European guidelines FNA is referred to as the “gold standard for the differential diagnosis of thyroid nodules” (2). The SOR Consensus Conference Statement speaks of FNA as “the accepted method for screening a thyroid nodule for cancer” (1).
In the same 2009 issue of Thyroid as the ATA guidelines (3) there were two papers on FNA classification systems, one of which was a summary of the “Bethesda System” for reporting thyroid cytopathology (11) and the other, by Theoharis et al., which summarizes 1 year's experience of an academic institution with the “Bethesda System” (12). In the latter paper it was stated that FNA “can be considered as either a diagnostic test or a screening tool.” Can the same be said for US, RTE, and related procedures? Can these procedures be used to decide the probability of a thyroid nodule being malignant? Should evidence-based guideline recommendations be formulated regarding using these procedures to decide if thyroid lobectomy or thyroidectomy be performed for thyroid nodules? A straightforward answer to this question is that data regarding US and RTE are far less abundant and perhaps compelling than they are for thyroid cytopathology. This has serious consequences for “evidence-based” recommendations.
Although FNA is currently considered the gold standard preoperative diagnostic procedure for thyroid nodules, it too, has limitations (11,12). Further efforts will no doubt be made to refine FNA, US, RTE, and other methodologies to improve their diagnostic capabilities. In this regard, analysis of FNA samples for molecular markers of thyroid cancer seems quite promising. A fundamental question at this juncture is whether future studies should continue to focus on only one modality, as most have done to date. It seems more likely that the best evidence-based guidelines will be realized by prospective studies that simultaneously include and evaluate disparate methods and technologies, including US, RTE, and related noninvasive procedures.
Footnotes
1
Adapted from Werner Heisenberg in On Modern Physics, Collier Books, a division of The Crowell-Collier Publishing Company, New York, NY, 1961. Originally published by Paolo Boringhieri Editore, Turin with English Translation by M. Goodmand and J.W. Binns.