Intraoperative Consultation of Thyroid and Parathyroid Lesions

General Clinical Introduction

Thyroid nodules are frequent in the general population and can be detected by ultrasound in up to 50% of adults. They are also often incidentally found on other imaging studies of the neck such as computed tomography (CT) scans and magnetic resonance imaging (MRI). Only 5% to 10% of these nodules are clinically palpable and less than 5% of them harbor malignancy (Caron and Clark, 2006; Clark, 2010).

The initial investigation of a patient with a thyroid nodule requires biochemical assessment of thyroid function by measuring thyroid-stimulating hormone (TSH) level. Euthyroid or hypothyroid patients are usually evaluated with fine-needle aspiration (FNA) to exclude a malignancy. In contrast, the presence of a suppressed TSH level suggests the possibility of a hyperfunctioning nodule that is likely benign; FNA should be carried out selectively in these patients because the biopsy may be reported as “atypical” because of associated hypercellularity within the autonomously functioning nodule(s) (Baloch et al, 2008; Ogilvie et al, 2006). Screening of thyroid nodules with FNA has substantially reduced the number of patients undergoing unnecessary operation for definitive diagnosis, with a sensitivity approaching 98% and false-negative rates between 5% and 6%. False negative rates are higher in nodules greater than 4 cm (McCoy et al, 2007; Meko, 1995). FNA is particularly valuable in the diagnosis of papillary, medullary, and anaplastic carcinomas but less so in the evaluation of follicular lesions such as follicular carcinoma and follicular variant of papillary carcinoma because benign lesions such as follicular adenoma or nodular adenomatous goiter may show morphologic overlap with follicular malignancies.

Ultrasound is very useful for preoperative assessment. The presence of microcalcifications, abnormal lymph nodes, nodules with irregular borders and nodule dimensions taller more than wide favors thyroid malignancy. The presence of 1 and 2 of these sonographic features predicts malignancy with odds ratios of 2.0, and 4.0 respectively. All patients with three or more malignant ultrasound features have thyroid cancer on final pathology and probably can have surgery without preoperative FNA (Goldfarb et al, 2012; Alexander et al, 2004). Presence of multiple nodules, a cystic component, and echogenicity does not predict malignancy. Presence of nodules with regular borders favors a benign diagnosis.

Preoperative molecular testing of cells collected by FNA is emerging as another diagnostic modality for thyroid cancer that may be particularly helpful in the evaluation of follicular lesions. Gene expression classifiers (microRNA) that are suspicious can raise the malignancy risk of a follicular lesion of unknown significance (5% to 10%) or follicular neoplasm (20% to 30%) to up to 40%, supporting a recommendation for total thyroidectomy, avoiding a possible second lobectomy and obviating the need for intraoperative frozen section (Alexander, 2004). Molecular testing is perhaps not as useful at this time for the diagnosis of thyroid cancer cell type. For example, a positive test for BRAF will be 85% to 99% indicative of papillary carcinoma, but most surgeons will perform a total thyroidectomy, even if the FNA of a papillary lesion is only 50% to 75% suspicious for cancer.

The surgeon arrives in the operating room with a substantial amount of preoperative information but may still benefit from intraoperative consultation (IOC) with frozen section in deciding whether to perform a lobectomy or total thyroidectomy, and whether to perform lymph node dissections. Operative planning is important because the risks associated with total thyroidectomy and central node dissection are significant, and yet returning to the operating room for completion thyroidectomy and possible additional lymph node dissection also involves risk, not to mention increased anesthesia time and inconvenience for the patient. Intraoperative diagnosis and frozen section are particularly important in older patients and those with larger thyroid nodules. Assessment of frozen section is less critical for a young patient with a mass smaller than 1 cm because lobectomy will frequently be sufficient in most such cases.

It is important to recognize that total thyroidectomy should be avoided whenever possible because it renders the patient dependent on lifelong hormone replacement therapy. Nevertheless, if it can be done safely, it is most often the preferred surgical treatment for thyroid malignancies because it allows the option of supplemental radioactive iodine (RAI) treatment, and it lowers baseline thyroglobulin to near zero, facilitating subsequent monitoring. Lifelong dependence on thyroid hormones after total thyroidectomy is a secondary consideration because patients with thyroid cancer will always need thyroid medication to suppress TSH, even if one thyroid lobe remains in the neck.

During thyroidectomy, enlarged central neck nodes are typically removed. Some investigators recommend routine unilateral or bilateral central neck dissection due to increased incidence of microscopic metastases and data showing decreased rates of local recurrence when they are removed at initial operation (Cranshaw and Carnaille, 2008). However, these risks need to be balanced against the possibility of hypoparathyroidism associated with central neck dissection. Biopsy-proven lateral lymph node metastases should be managed with modified radical neck dissection because substantial nodal disease can be associated with invasion of nerves, vessels, and muscles. Prophylactic lateral neck dissection is not indicated in patients with papillary thyroid carcinoma (PTC) and grossly normal lymph nodes because micrometastases often can be ablated with RAI therapy or suppressed indefinitely by a low TSH. Furthermore, if there is subsequent clinical recurrence of lateral nodes, they can usually be excised at a later date without compromising survival because these cancers do not appear to metastasize systematically from lymph nodes (Cranshaw and Carnaille, 2008; Gradoni et al, 2011).

Indications for IOC with Frozen Section During Thyroid Surgery

The primary reason to perform an intraoperative consultation during thyroid operation is to provide a rapid diagnosis that will affect the extent of surgery being performed. Table 6-1 lists the most common indications for thyroid frozen section (FS). Although the decision to request a frozen section lies with the surgeon, effective communication between surgeon and pathologist is paramount for informed decision making.

Intraoperative consultation is most useful when there is no preoperative FNA diagnosis or when a previous FNA had revealed findings suspicious or indeterminate for PTC. The distinct architectural and cytomorphologic features of PTC, including fine nuclear details, can be adequately evaluated intraoperatively in the majority of cases using combined frozen section and cytologic assessments. An experienced pathologist should have little difficulty in recognizing most cases of nodular hyperplasia, classic PTC, anaplastic/undifferentiated carcinoma, poorly differentiated carcinoma, widely invasive follicular carcinoma, medullary carcinoma, lymphoma, and metastasis.

Intraoperative consultation with FS is also useful for identification of nodal metastasis or for the diagnosis of a highly malignant and unresectable tumor such as anaplastic carcinoma in an incisional biopsy.

Intraoperative Gross Evaluation of Thyroid Lesions

After measurement of the specimen, inking, and serial sectioning, the slices are laid out and carefully inspected for focal lesions ( Figure 6-1 ). Gross examination is critical in selecting the correct tissue or area of the lesion or tumor for cytologic evaluation and frozen sections. If serial sections demonstrate multiple well-circumscribed, colloid-rich nodules characteristic of multinodular goiter (MNG), a gross examination may suffice for IOC. PTC, the most common thyroid malignancy, varies considerably in size, ranging from microscopic to large, with average size of 2 cm to 3 cm. The tumor tends to have a solid, white, firm, sometimes gritty surface with an infiltrative appearance and poorly defined margins ( Figure 6-2A ). PTC may be multifocal (20%), encapsulated (10%), or exhibit marked cystic changes (10%).

Hyperplastic thyroid nodules consist of multifocal, often nonencapsulated lesions showing fibrous strands that subdivide their cut surface into smaller lobules and do not compress the adjacent thyroid tissue. In contrast, follicular neoplasms usually are solitary, well-encapsulated nodules which tend to distort and compress adjacent normal or nonneoplastic thyroid tissue. They show a homogenously solid cut surface and minimal amount of colloid ( Figure 6-3A ). The gross appearances of minimally invasive follicular carcinomas is similar to that of follicular adenomas, and capsular invasion is only rarely apparent grossly ( Figure 6-4A ). In contrast, widely invasive follicular carcinomas often exhibit obvious extracapsular tumor extension, and in some cases the capsule may be difficult to identify ( Figure 6-5A ). Hürthle cell neoplasm, a variant of follicular tumor, is usually brown ( Figure 6-3A ). Medullary carcinomas are typically nonencapsulated, solid, gray-tan-yellow, firm and may be infiltrative, varying in size from microscopic, small foci to large, widely invasive tumors with extrathyroidal extension into the neck soft tissue, often with associated hemorrhage and necrosis ( Figure 6-6A ). Anaplastic carcinomas are usually large, nonencapsulated tumors that often display extrathyroidal extension into the adjacent neck soft tissue and sometimes show intratumoral hemorrhage and necrosis ( Figure 6-7A ).

Intraoperative Cytologic Evaluation of Thyroid Lesions

Cytologic preparations should be performed routinely along with frozen sections during the intraoperative evaluation of thyroid lesions. Accurate assessment of fine nuclear details that can only be achieved on cytologic examination is of great value in confirming or excluding PTC because the cellular details of PTC are often severely compromised by FS artifacts. FS is useful to identify the growth features of a thyroid lesion, and a combination of both intraoperative techniques will yield the best possible information. Cytologic preparations include true touch preps or imprints, scrapings and smears. Scrape preparations are, in our experience, preferable to the other techniques, because they provide greater cellularity, especially when the lesions contain friable cystic contents or colloid or are fibrotic. Another advantage of scrape preparation over touch preparation is that the former method can be used to obtain lesional cells more precisely by scraping the exact area without the collection of too many nonlesional cells from the background. Scrape preparations are particularly helpful for the evaluation of small lesions. Smears and imprints are stained with hematoxylin and eosin (H&E) after alcohol fixation and air-dried smears or imprints are stained with Diff-Quik. Both preparations provide a very useful adjunct to frozen sections in distinguishing between a follicular lesion/neoplasm and PTC. Some pathologists prefer air-dried, Diff-Quik-stained smears or imprints over H&E-stained smears because the air-dried preparations show better the presence of multiple cell populations, provide more cytoplasmic details, and make more readily visible the presence of colloid (Papanicolaou Society of Cytopathology, 1996.

The key cytologic features of PTC include cellular smears with monolayer sheets of follicular cells, often with: three-dimensional papillary architecture (some with fibrovascular cores); enlarged overlapping nuclei; nuclear elongation and irregular nuclear contours; pale and powdery fine chromatin; nuclear grooves; multiple micronucleoli; intranuclear inclusions; no or minimal gummy colloid; psammoma bodies; cellular swirling, focally dense squamoid cytoplasm; and multinucleated giant cells. Although papillary three-dimensional overlapping sheets or clusters of cells are one of the hallmark features of PTC and its variants, they might be present very focally or even absent in many cases. It should be kept in mind that ground glass (optically clear or “Orphan Annie”) nuclei, a prominent feature of PTC in formalin-fixed paraffin sections, are fixation artifacts that are unapparent or absent in smears or frozen sections ( Figure 6-2B, C, D, E, F ). In contrast, smears from follicular lesions typically show uniform, flat, or sometimes three-dimensional sheets in a honeycomb arrangement of bland-appearing follicular cells and abundant background colloid with variable cellularity, depending upon the type of follicular lesions. Mildly enlarged uniform nuclei, with minimal crowding and overlapping, may be observed, but no apparent nuclear grooves, intranuclear inclusions, psammoma bodies, or papillary architecture are identified ( Figures 6-3B, 6-4B, 6-5B ).

Generally adenomatous and hyperplastic nodules tend to have more colloid and less cellularity, while follicular neoplasms (follicular adenoma, follicular carcinoma, and follicular variant of PTC) have more cellularity and less colloid. However, distinction between these two groups can be very difficult or impossible on cytologic evaluation alone because there are many overlapping features between them. Cellular follicular lesions with scant colloid and microfollicles should be evaluated carefully to determine if the lesion represents a cellular adenomatous nodule, follicular adenoma ( Figure 6-3B ), follicular carcinoma ( Figures 6-4B, 6-5B ) or follicular variant of PTC (FVPTC) ( Figure 6-8B, C ). It should be noted that no single feature by itself is diagnostic. With a constellation of findings, a more definitive diagnosis of PTC or follicular lesion may be achieved in the majority of cases.

Intraoperative cytologic preparation of medullary carcinoma typically tends to be highly cellular, consisting of many atypical cells singly or in small clusters with a great degree of discohesion. Tumor cells range from round to spindle-shaped and plasmacytoid cells with a neuroendocrine (salt-and-pepper) chromatin pattern, abundant cytoplasm, occasionally pink azurophilic granules, and a viable degree of pleomorphism. Sometimes the background may contain acellular material or amyloid, somewhat mimicking thick colloid ( Figure 6-6B, C ).

There are three histologic types of anaplastic carcinoma: giant cell pattern, spindled cell (sarcomatoid) pattern, and squamoid cells. Cytologic preparations typically show highly atypical cells, including: large, bizarre giant cells with prominent nucleoli; multinucleated giant cells; atypical spindle cells; squamoid cells with dense cytoplasm and hyperchromatic nuclei; and mixtures of all types of cells in variable degrees and sometimes necrotic debris ( Figure 6-7B ). Table 6-3 lists some of the key cytologic features for common thyroid lesions/neoplasms; key cytologic features of PTC are summarized in Table 6-5 .

Cytologic evaluation is critical in differentiating lymphomas from anaplastic carcinoma ( Figure 6-7B ), metastatic carcinoma ( Figure 6-9B ), and sometimes medullary carcinoma ( Figure 6-6B, C ), owing to similar clinical, gross, and frozen section features, with the former exhibiting markedly discohesive atypical lymphoid cells on either imprint or scrape preparation. Most lymphomas of the thyroid are of diffuse large B-cell type and occur in older patients.

Frozen Section Evaluation

Table 6-4 lists the various diagnoses to be considered during the intraoperative evaluation of a thyroid lesion with FS and cytologic preparations. The extent of surgery has been generally decided preoperatively in cases where a definitive diagnosis of carcinoma has been rendered by preoperative FNA. Conversely, if a preoperative FNA has revealed a follicular lesion or a Hürthle cell lesion/neoplasm, intraoperative consultation is rarely helpful for the reasons mentioned above. Diagnosis of PTC, medullary carcinoma, a neuroendocrine tumor derived from C cells, and sometimes anaplastic/undifferentiated carcinoma are often made by preoperative FNA, and IOC is usually not requested by the surgeon. An experienced pathologist should have little difficulty in recognizing most cases of adenomatous nodule, Hashimoto or lymphocytic thyroiditis, widely invasive follicular carcinoma, classic PTC, anaplastic/undifferentiated carcinoma, poorly differentiated carcinoma, medullary carcinoma, lymphoma, metastasis, and post-FNA–related changes during IOC. If there is no evidence of carcinoma identified intraoperatively, no additional surgery is performed. On the other hand, if the diagnosis of carcinoma is made during IOC, additional surgery involving the removal of the remainder of the thyroid gland and possibly the surrounding lymph nodes will be performed. The detailed morphologic features for each entity are beyond the scope of this chapter. Instead, several of the most commonly encountered scenarios will be discussed in greater detail.

Evaluation of Solitary Encapsulated Follicular Nodule

Encapsulated solitary follicular lesions/nodules often create difficult diagnostic challenges during IOC primarily because: 1) a rather broad differential diagnosis needs to be considered, including dominant adenomatous nodule, follicular adenoma, minimally invasive follicular carcinoma, PTC (follicular variant and encapsulated variant), and rarely a hyalinizing trabecular adenoma (HTA), and 2) evaluation is often limited by incomplete sampling of the capsule.

The sensitivity of IOC diagnosis of thyroid carcinoma depends on the type of carcinoma, the highest in classic PTC (94%), 27% in FVPTC, and the lowest in encapsulated or minimally invasive follicular carcinoma (17%), with overall sensitivity of 60% to 70% among all carcinomas (Leteurtre et al, 2001). The low sensitivity of IOC detection of minimally invasive follicular carcinoma is due to limited sampling of tumor capsule intraoperatively. Although more sampling of the tumor capsule may improve the sensitivity slightly, it will prolong the surgery and tie up resources, and it is certainly not cost-effective. In addition, more extensive sampling by performing multiple frozen sections will introduce histologic artifact on permanent sections which may modify the tumor capsule and consequently obscure or reduce sensitivity in establishing the final diagnosis (Bronner and LiVolsi, 1991). Taken together, IOC with FS is of limited value in the diagnosis of follicular carcinoma because evidence of malignancy (capsular and vascular invasions) cannot be visualized easily on frozen sections (Multanen et al, 1999). Therefore, extensive sampling of an encapsulated follicular lesion/mass during IOC should be strongly discouraged.

Although some advocate taking three or four sections from the tumor capsule thyroid interface to evaluate capsular and vascular invasions, most institutions and experts recommend taking one representative section from the tumor capsule thyroid interface for multiple reasons described previously so that capsular and/or vascular invasion can be evaluated. Attention should be paid at the thickened irregular area of the capsule where one representative section needs to be sampled. For lesions or nodules that lack PTC features and vascular and/or capsular invasions, it is a common practice to render a diagnosis of “follicular lesion, defer to permanent section.” Surgeons generally understand the limitations of sampling of follicular lesions during IOC, accept such a diagnosis, and perform a lobectomy. In cases where the follicular nodules may exhibit some, but not all, PTC features, especially on cytologic preparations, a diagnosis of “follicular nodule, follicular variant of PTC cannot be excluded” is also an appropriate intraoperative diagnosis; the surgeon will likely not perform a total thyroidectomy. FVPTC, the most troublesome of PTC variants, accounts for much of the false negative IOC diagnoses. In addition, preoperative FNA tends to underappreciate specific nuclear alterations because of their focal and random distribution in FVPTC. Lesions/nodules that exhibit multiple PTC features ( Figure 6-8B, C ), can be diagnosed as encapsulated or follicular variant PTC and treated with completion thyroidectomy. Similarly, if a diagnosis of minimally invasive follicular carcinoma is made intraoperatively based on the demonstration of capsular and/or vascular invasion ( Figure 6-4B, C ), most surgeons will pursue a completion thyroidectomy, although some may only perform a lobectomy.

Because of limited sampling during IOC, a majority of lesions diagnosed on preoperative FNA as follicular/Hürthle cell neoplasm or indeterminate for neoplasm will be deferred to final histopathologic examination. Studies have shown that among all solitary encapsulated lesions or nodules removed, the majority of them (80%) turned out to be follicular adenoma for which the standard surgery is lobectomy.

Figure 6-10 provides an algorithm for the intraoperative evaluation of thyroid lesions and lists key features of lesions to be considered in the differential diagnosis of encapsulated solitary follicular nodules.

Intraoperative Evaluation of a Nonencapsulated Infiltrative Thyroid Mass

The diagnostic considerations for nonencapsulated and infiltrative masses/nodules are also rather broad and include PTC and its variants, widely invasive follicular carcinoma, medullary carcinoma, poorly differentiated carcinoma and undifferentiated/anaplastic carcinoma of the thyroid, and metastatic or other lesions. A diagnosis of carcinoma can be reached readily if there are apparent cytologic and histologic atypia and infiltrative growth. Intraoperative diagnosis of the precise cell type of thyroid carcinoma has little impact on surgical management because all carcinomas other than undifferentiated/anaplastic carcinoma are appropriately treated with total thyroidectomy with probable local lymph node dissection. Surgeons usually do not perform a thyroidectomy in patients with undifferentiated/anaplastic carcinoma because of its frequent infiltrative features and extremely poor prognosis.

Papillary thyroid carcinoma is generally diagnosed by FNA. If intraoperative frozen-section examination of a lymph node or primary tumor confirms PTC, total thyroidectomy is performed. For patients who have minimal PTCs or microcarcinomas (less than 1 cm) confined to the thyroid gland without angioinvasion, there are data to support lobectomy alone, but many experts still favor total thyroidectomy and ipsilateral central node dissection. In addition to gross and cytologic features as described previously, histologically the typical PTC displays numerous true papillae which are usually complex, branching, and randomly oriented, with fibrovascular cores. The papillae are lined by cuboidal follicular cells exhibiting nuclear enlargement, elongation, and overlapping with finely dispersed optically clear chromatin (ground-glass or “Orphan Annie” nuclei, only seen on formalin-fixed permanent sections), thickened and irregular nuclear membrane, micronucleoli, nuclear longitudinal grooves, and intranuclear inclusions. The stroma of the papillae may be edematous, hyaline, or elastic and may contain foamy and hemosiderin-laden macrophages and lymphocytes. Psammoma bodies can be seen in 50% of tumors in fibrous stroma of papillary stalk ( Figure 6-2G, H ). The main gross, cytologic, and histologic features of PTC are summarized in Table 6-5 .

Intraoperative Evaluation of Multinodular Goiter with One Dominant Nodule

It is not uncommon for a pathologist to be requested to evaluate an enlarged multinodular thyroid or multinodular goiter (MNG) with one dominant nodule in either lobectomy or total thyroidectomy specimen. PTC (mostly microcarcinoma) is the most common malignancy in MNGs identified on permanent sections. A completion total thyroidectomy will be performed if a PTC is detected in a lobectomy specimen during IOC.

As described in a previous section, the nodules in MNG tend to be circumscribed but poorly encapsulated with a colloid-rich cut surface, focal hemorrhage, cystic change, and fibrosis. Careful gross inspection after serial sectioning is critical in identifying unusual areas such as white, tan, and firm areas suspicious for carcinoma such as PTC or encapsulated pale nodule suspicious for follicular nodule/neoplasm for cytologic preparations and frozen sections.

Most of these cases are benign with a dominant adenomatous nodule and focal fibrosis or FNA-related scaring in a background of multinodular goiter. It is important to be aware that some FNA-related changes including atypia can mimic PTC and minimally invasive follicular carcinoma, and therefore care must be taken to avoid overinterpretation.