Papillary thyroid cancer (PTC) accounts for approximately 85% of well differentiated thyroid cancers. PTC is generally indolent, with an associated 10-year survival of over 93% . Histologic variants of PTC are well recognized. They often display different degrees of aggressive behavior in the spectrum between well differentiated classic PTC and the undifferentiated anaplastic carcinoma, including higher rates of metastases, recurrence, resistance to radioactive iodine (RAI) therapy, and possibly compromised survival. The most aggressive types have included diffuse sclerosing, tall cell or columnar cell, and insular variants; solid, trabecular, oncocytic, microfollicular, pseudo-Warthin, and clear cell (among other rare) variants have variable prognoses, but many are closer to conventional PTC. This review will focus on the clinical and pathologic characteristics of the more aggressive variants and the associated patient outcomes. It will include diffuse sclerosing variant (DSV), tall cell variant (TCV), and insular thyroid cancer (ITC), and address novel diagnostic modalities and molecular changes.
MAKING THE DIAGNOSIS
Establishing the diagnosis of an aggressive variant of PTC or a poorly differentiated thyroid cancer has been inconsistent when examined from a historical perspective. Different diagnostic criteria have been applied in various geographic areas, resulting in wide discrepancies among pathologists and clinicians worldwide. In a landmark study in 2007, a panel of internationally recognized thyroid pathologists collected 83 cases from Europe, Japan, and the United States [2▪▪]. These cases were circulated and examined among the noted 12 pathologists, and a consensus meeting was held in Turin, Italy. Agreement was reached in diagnostic criteria for aggressive variants of PTC and poorly differentiated thyroid carcinoma. These included architectural changes and presence of solid or trabecular or insular pattern of growth; absence of conventional PTC nuclear features; and the presence of at least one feature of convoluted nuclei, high mitotic activity, and tumor necrosis. With regard to architecture, it was recognized that these changes may occur only focally. In order to call a tumor ‘aggressive’, the World Health Organization (WHO) recommends that the majority of the tumor display the high grade architecture, such as tall cell or columnar appearance . Nonetheless, this study found that for solid or trabecular or insular growth pattern, even focal changes may affect patient prognosis. This study laid the foundation for adopting unified pathologic diagnostics and became known as the Turin criteria.
Dettmer et al.  applied the Turin criteria in 42 patients who were diagnosed with PTC and who suffered adverse clinical outcomes. The authors compared them with 50 patients with follicular thyroid cancers (control group) in order to determine the proportion of a given carcinoma that would be required to be considered poorly differentiated or aggressive in order to affect prognosis. The authors compared the patients whose tumors demonstrated more than 10% aggressive or poorly differentiated characteristics to those whose tumors demonstrated more than 50% of these characteristics. They found that those patients whose tumors had more than 10% aggressive features had worse survival than those with follicular thyroid cancer, and had similar compromised survival as those patients whose tumors contained more than 50% aggressive tumor features. They concluded that PTC harboring even a low percentage of aggressive morphology has prognostic importance, and this percentage should be included in pathology reports.
DIFFUSE SCLEROSING VARIANT OF PAPILLARY THYROID CANCER
DSV is estimated to account up to 6% of all PTC . It was originally described by Vickery et al.  in 1985, and has typical clinical and histopathological characteristics. The morphology of DSV includes classic nuclear features of PTC, diffuse stromal fibrosis, dense lymphoid infiltration, squamous metaplasia, and abundant psammoma bodies. This tumor occurs slightly more frequently in women than men, with a female/male ratio of 1.75 : 1 .
The largest population-level study of DSV was by Kazaure et al. [8▪▪] who analyzed 261 DSV patients and compared them to 42 904 PTC patients in the Surveillance, Epidemiology, and End Result (SEER) database from 1988 to 2008. The mean follow-up for DSV and PTC patients was 4.0 and 5.4 years, respectively. The authors found that between the years 2001 and 2008, the incidence of classic PTC increased from 4.54 cases per 100 000 population to 7.3 cases per 100 000 population; this represents a 60.8% increase in incidence. The incidence of DSV increased by 126% (from 0.021 per 100 000 to 0.047 per 100 000). Patient demographics showed that DSV and PTC patients were similar in age, but more patients with DSV were women (82.8 vs. 76.4%, respectively, P = 0.003). Compared with patients with PTC, those with DSV underwent total thyroidectomy more often, had more cervical lymph nodes examined, and were administered postoperative RAI more often (all P < 0.001). Pathologically, compared with PTC tumors, most DSV tumors were less than 1 cm, multifocal, had extrathyroidal extension, and harbored lymph node metastases (all P < 0.001). The fact that most DSV tumors in their study were small suggests that DSV assumes aggressive behavior early in tumorigenesis.
Distant metastatic disease was found in 7.3% of patients with DSV compared with 4.3% of patients with PTC. Five-year disease-specific survival was lower for patients with DSV (96.1%) compared with patients with PTC (97.4%, P = 0.008). The authors found that a diagnosis of DSV was an independent factor associated with mortality, with a hazard ratio of 1.8.
DSV has been considered to be a variant which typically occurs in young patients (mean age 30 years). Most studies with this finding have had small samples sizes. For example, in a single institutional study of 68 patients aged less than 20 years who underwent thyroidectomy for PTC, histologic types of thyroid cancer included classic PTC in 83.8% of patients, follicular carcinoma in 11.8%, and aggressive variants of PTC in 4.4% of patients . The authors found that DSV tumors were more often multifocal (60.7 vs. 17.2% PTC, P = 0.003) and had more extrathyroidal extension (89.3 vs. 59.6% PTC, P = 0.009). Patients were followed for an average of 55 months. Disease recurrence was more common in patients with DSV than those with PTC (39.3 and 17.2%, P = 0.032). The authors recommended aggressive surgical treatment for these patients with total thyroidectomy, possible lymph node resection, and postoperative RAI treatment.
Regalbuto et al.  analyzed a single institutional series of 34 patients with DSV and compared them to 245 patients with PTC. The authors divided the PTC patients into low-risk and high-risk PTC groups. High-risk patients were defined as those harboring PTC with at least one of the following characteristics: tumor size greater than 4 cm, advanced tumor grade (G2–G4), bilaterality, extrathyroidal extension, or locoregional and distant metastases. Patients with DSV were then compared to both low-risk and high-risk PTC patients in regard to pathologic findings, disease recurrence, and development of distant metastases. The authors found that patients with DSV had similar outcomes to those with high-risk PTC, although DSV was independently associated with the development of recurrent disease. These patients required additional treatment with surgery and repeated RAI administration. The authors concluded that DSV behaves clinically similarly to high-risk PTC and that it should be treated aggressively at the time of diagnosis.
The diagnosis of DSV is made usually on postoperative pathologic examination; however, experienced surgeons may suspect this diagnosis intraoperatively, because of the often firm and fibrosed appearance of the tumor. Several studies have focused on the potential of identifying DSV preoperatively, thus allowing better operative planning. In a study of eight patients with pathologically documented DSV, Zhang et al.  retrospectively analyzed the thyroid ultrasound examinations of these patients. All eight cases displayed diffuse involvement of the ipsilateral thyroid lobe; 63% of cases showed hyperechogenicity; 88% had diffuse scattered microcalcifications; and all cases showed cervical metastatic lymph nodes. They concluded that the sonographic findings of DSV are characteristic to the tumor type, and that ultrasound may therefore be a useful preoperative diagnostic modality.
Diagnosing DSV on preoperative fine-needle aspiration (FNA) biopsy has remained a challenge. Bongiovanni et al.  reported a case and performed a thorough review of the literature regarding FNA diagnosis of DSV. The pathologic features which point to a diagnosis of DSV include cells with nuclear features of PTC, innumerable psammoma bodies in the tumor and parenchyma, metaplastic squamous epithelium, extensive lymphocytic thyroiditis, and broad bands of collagenous tissue. These changes also can be seen in classic PTC, but they are more pronounced in DSV tumors. The amount of lymphocytic thyroiditis found in DSV may lead to cytologic pitfalls of entertaining alternative diagnoses such as possible thyroid lymphoma or severe Hashimoto's thyroiditis. To avoid this, the authors recommended extensive and diffuse FNA sampling of the patients’ thyroid gland, in order to capture other cytologic features, such as the nuclear changes of classical PTC, abundance of psammoma bodies, and metaplastic squamous epithelium. These findings, along with the lymphocytic thyroiditis, would support the diagnosis of DSV on preoperative FNA, and therefore allow better treatment planning.
TALL CELL VARIANT OF PAPILLARY THYROID CANCER
TCV tumors differ from classic PTC in that they are comprised of distinctly rectangular cells, with less colloid, but similar nuclear features as PTC. The tall or pink cell variant of PTC was originally described by Hawk and Hazard  in 1976. The authors described a group of PTC with a ‘distinctive cell type in a columnar shape with the height of the cell being at least twice the breadth’. The most common definition of TCV is when at least 30% of all tumor cells are twice as long as they are wide; however, the World Health Organization defines TCV as a tumor that is composed ‘predominantly of cells whose heights are at least 3 times their width’ .
Silver et al.  wrote a review of the existing literature on TCV, among other thyroid cancers, and included seven of the most encompassing studies on TCV published up to 2010. The studies included in the review were institutional series with relatively small number of patients. The reported incidence of TCV ranged between 5 and 11% of PTC. All studies found that TCV tumors had more aggressive features, such as lymph node metastases, and patients had higher overall mortality at 10 years (up to 22% for patients with TCV, compared to 8.6% for PTC).
The largest population-level study to date on TCV is by Kazaure et al. [8▪▪] which included 576 cases. The authors found that the incidence of TCV increased by 158% (0.05 per 100 000 to 0.13 per 100 000) between 2001 and 2008. Compared with patients with PTC, those with TCV were older, had higher rates of total thyroidectomy, lymphadenectomy, and postoperative RAI treatment. TCV tumors were most commonly greater than 4 cm. Rates of distant metastases in patients with TCV were 2.6% higher than for those with PTC. Disease-specific mortality rates at 5 years were 6.4% for patients with TCV and 2.3% for patients with PTC (P < 0.001). Patients with TCV who underwent treatment with total thyroidectomy and postoperative RAI had improved survival compared with those who had less surgical resection and no RAI.
Like DSV, a diagnosis of TCV is most commonly achieved on postoperative pathology. Accurate preoperative diagnosis has been difficult to secure. Choi et al.  studied in a retrospective fashion the clinical and sonographic features of eight patients with 10 TCV tumors. The inclusion criteria for a diagnosis of TCV included a tumor ‘composed of more than 50% tall cells, a tall cell height at least twice its width, eosinophilic tall cell cytoplasm, and nuclear features characteristic of papillary thyroid carcinoma such as nuclear irregularities, clearing and overlapping, grooves, and pseudoinclusions’. They examined the preoperative ultrasound findings for very specific criteria, such as nodule size, shape, echogenicity, texture, internal composition, margins, pattern of calcifications, extrathyroidal extension, and associated lymph node metastases. They found that the majority of the tumors had malignant sonographic features that included the presence of microcalcifications or macrocalcifications, a spiculated margin, marked heteroechogenicity, solid composition, and a taller-than-wide shape. Most had evidence of pathologic lymphadenopathy. Eight of ten cases had sonographic evidence of extrathyroidal extension confirmed intraoperatively with clear invasion into surrounding tissues. The authors concluded that TCV has significant overlap sonographically with PTC, but often harbors more aggressive features, such as variable hypoechogenicity, evidence of capsular invasion, microcalcifications and macrocalcifications, and lymphadenopathy. Identification of such sonographic findings preoperatively should raise suspicion for the possibility of TCV.
INSULAR THYROID CANCER
The insular variant of PTC was described first by Carcangiu et al.  in 1984. They delineated the pathologic criteria for ITC, including the presence of nests or ‘insulae’ of tumor cells with round, dark nuclei, and scant cytoplasm. Cytologic uniformity, hypercellularity, and scant colloid are typical; however, there are no psammoma bodies in these tumors. The nuclear-to-cytoplasmic ratio is high, but nuclear grooves and intranuclear inclusions are absent. ITCs are often large tumors, with frequent extracapsular extension. They are mixed tumors, often containing areas of classic components of PTC or follicular thyroid cancer . There appears to be a relationship with longstanding goiters .
The most encompassing recent population-level study in the USA was by Kazaure et al. [18▪▪]. The authors analyzed 114 patients with ITC, 497 patients with anaplastic thyroid cancer, and 34 021 patients with well differentiated thyroid cancer culled from the SEER database, 1999–2007. ITC represented 0.3% of the cohort. Compared with PTC patients, those with ITC were older (mean age 48.1 vs. 62.1 years, P < 0.001); ITC was common in men, with a nearly 1 : 1 sex ratio. ITC tumors were large, with a mean size of 5.9 cm. Extrathyroidal extension (47.3%) and lymph node involvement (61.9%) were common. Over 30% of patients with ITC presented with distant metastases, compared with 4.5% of PTC and 69% of anaplastic thyroid cancer patients. ITC patients often underwent total thyroidectomy (81.6%), postoperative RAI treatment (57%), and external-beam radiation therapy (15.8%).
ITC patients had worse disease-specific survival compared with PTC patients (5-year survival 72.6 vs. 97.2%, P < 0.001), but undergoing total thyroidectomy was protective, after adjustment for other variables (hazard ratio 0.2, P < 0.04). Age at diagnosis and sex did not appear to play a role in survival for patients with ITC. The authors concluded that ITC is an aggressive tumor and that current Tumor Nodes Metastasis staging for well differentiated thyroid cancer is not applicable to ITC. Given the demonstrated survival benefit of tumor debulking, patients with ITC (including those with distant metastasis) should be considered candidates for aggressive treatment involving total thyroidectomy, RAI, and appropriate lymph node dissections for those with lymph node metastases.
In a review of thyroid cancers with intermediate differentiation, Sywak et al. included 213 patients with ITC. This revealed a female-to-male ratio of 2.7; mean tumor size was 5.5 cm; 44% had extracapsular extension; and 51% cervical lymph node metastases. Their mean follow-up was 72 months. Recurrent disease or development of distant metastases occurred in 64% of patients. Calculated disease-specific mortality was 32%. Their results were similar to prior single-institutional series and to the most recent population-level study by Kazaure .
MOLECULAR CHANGES IN AGGRESSIVE VARIANTS OF THYROID CANCER
There has been a growing interest in molecular analysis of thyroid cancers for both diagnostic and prognostic reasons. BRAF(V600E) mutation has emerged as a marker of aggressive behavior in thyroid cancer. BRAF mutation has been associated with clinical progression and recurrence in PTC . The reported prevalence of BRAF mutations in DSV and TCV has varied. Several studies have shown a high prevalence of BRAF(V600) mutations in TCV, while others have not [14,20]. On the basis of these discordant findings, the aggressiveness of the PTC variants has been ascribed to their phenotype rather than well defined molecular mutations.
Addressing the genotype and phenotype question, Virk et al. [21▪] examined 129 papillary thyroid microcarcinomas (<1 cm) by testing their BRAF(V600E) mutational status by single-strand conformation polymorphism. The authors then correlated the BRAF status with tumor size, focality, extrathyroidal extent, architecture and histologic subtype, and nodal metastases. Over 70% of the tumors were found to be positive for the BRAF(V600E) mutation. This was observed more frequently in TCV (91%), SCV (>70%), and classic PTC (>75%) than follicular variant of PTC (21%). Tumors without the mutation were significantly more likely to be solid, well circumscribed, and lacked fibrosis or sclerosis. Nodal metastases and extrathyroidal extension were associated with BRAF mutations. The authors concluded that BRAF(V600E) mutations occur early in tumorigenesis and are associated with specific morphology and aggressive features, even in microcarcinomas.
Soares et al.  examined genetic and epigenetic alterations in poorly differentiated and undifferentiated thyroid cancers. The authors reviewed insular or insular-like carcinomas displaying predominantly trabecular or solid growth pattern. Using immunohistochemistry, lectin histochemistry, electron microscopy, and molecular genetics, they determined that ITC resemble widely invasive follicular carcinomas, while other trabecular cancers were closer to classic PTC. As BRAF mutations have been observed more commonly in PTC rather than follicular thyroid cancers, no ITC cancers reviewed in their study harbored BRAF mutations. The cells which were BRAF positive were found to have PTC-like nuclei and were located in foci of PTC adjacent to the less differentiated portions of the tumors. This supported the relationship of ITC to widely invasive follicular cancers, rather than PTC.
Other genes, such as RAS (including N-RAS, H-RAS, and K-RAS), have been found to be mutated in up to 60% of aggressive variants . Novel genes, such as frameshift mutations and genetic loss of the LRP1B gene, have been described recently [24▪]. The LRP1B gene encodes a low-density lipoprotein receptor family located at the 2q21 chromosomal locus. This area has been implicated in familial nonmedullary thyroid cancers. Up to 80% of poorly differentiated thyroid cancers have been shown to have genomic loss of expression of the LRP1B gene. Clinical implications of this mutation have not been examined.
Better molecular and genetic markers have raised the suspicion of a thyroid nodule harboring an aggressive variant of PTC; this, in turn, allows surgeons and endocrinologists to formulate a more complete treatment plan, including total thyroidectomy, possible lymphadenectomy, and postoperative RAI administration. Given the rarity of these tumors, population-level data are necessary to understand the impact of appropriate treatments on patient survival.
Conflicts of interest
None of the authors have any conflicts of interest.
REFERENCES AND RECOMMENDED READING
Papers of particular interest, published within the annual period of review, have been highlighted as:
- ▪ of special interest
- ▪▪ of outstanding interest
Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 100).
1. Clark OH. Thyroid cancer and lymph node metastases. J Surg Oncol 2011; 1003:615–618.
2▪▪. Volante M, Collini P, Nikiforov Y, et al. Poorly differentiated thyroid carcinoma: the Turin proposal for the use of uniform diagnostic criteria and an algorithmic diagnostic approach. Am J Surg Pathol 2007; 31:1256–1264.
This study laid the groundwork for unified pathologic diagnostic criteria.
3. DeLellis RA, Loyd RV, Heitz PU, Eng C. Pathology and genetics of tumours of endocrine organs. Lyon: IARC Press; 2004.
4. Dettmer M, Schmitt A, Steinert H, et al. Poorly differentiated thyroid carcinomas: how much poorly differentiated is needed? Am J Surg Pathol 2011; 35:1866–1872.
5. Carling T, Ocal I, Udelsman R. Special variants of differentiated thyroid cancer: does it alter the extent of surgery versus well differentiated thyroid cancer? World J Surg 2007; 31:916–923.
6. Vickery AL Jr, Carcangiu ML, Johannessen JV, Sobrinho-Simoes M. Papillary carcinoma. Semin Diagn Pathol 1985; 2:90–100.
7. Carcangiu ML, Bianchi S. Diffuse sclerosing variant
of papillary thyroid carcinoma. Clinicopathologic study of 15 cases. Am J Surg Pathol 1989; 13:1041–1049.
8▪▪. Kazaure HS, Roman SA, Sosa JA. Aggressive variants of papillary thyroid cancer
: incidence, characteristics and predictors of survival among 43,738 patients. Ann Surg Oncol 2012; 19:1874–1880.
This study is the first population-level study of aggressive variants of PTC, with a large number of patients and long-term follow-up.
9. Koo JS, Hong S, Park CS. Diffuse sclerosing variant
is a major subtype of papillary thyroid carcinoma in the young. Thyroid 2009; 19:1225–1231.
10. Regalbuto C, Malandrino P, Tumminia A, et al. A diffuse sclerosing variant
of papillary thyroid carcinoma: clinical and pathologic features and outcomes of 34 consecutive cases. Thyroid 2011; 21:383–389.
11. Zhang Y, Xia D, Lin P, et al. Sonographic findings of the diffuse sclerosing variant
of papillary carcinoma of the thyroid. J Ultrasound Med 2010; 29:1223–1226.
12. Bongiovanni M, Triponez F, McKee T, et al.
Fine-needle aspiration of the diffuse sclerosing variant
of papillary thyroid carcinoma masked by florid lymphocytic thyroiditis: a potential pitfall. Diagn Cytopathol 2009; 37:671–675.
13. Hawk WA, Hazard JB. The many appearances of papillary carcinoma of the thyroid. Cleve Clin Q 1976; 43:207–215.
14. Silver C, Owen R, Rodrigo J, et al. Aggressive variants of papillary thyroid carcinoma. Head Neck 2011; 7:1052–1059.
15. Choi Y, Shin J, Kim J, et al. Tall cell variant
of papillary thyroid carcinoma: sonographic and clinical findings. J Ultrasound Med 2011; 30:853–858.
16. Carcangiu ML, Zampi G, Rosai J. Poorly differentiated (‘insular’) thyroid carcinoma. A reinterpretation of Langhans’ ‘wuchernde Struma’. Am J Surg Pathol 1984; 8:655–668.
17. Sywak M, Pasieka JL, Ogilvie T. A review of thyroid cancer with intermediate differentiation. J Surg Oncol 2004; 86:44–54.
18▪▪. Kazaure HS, Roman SA, Sosa JA. Insular thyroid cancer
: a population-level analysis of patient characteristics and predictors of survival. Cancer 2012; 118:3260–3267.
This study is the first population-level study of ITC, with a large number of patients and long-term follow-up.
19. Xing M. BRAF mutation in papillary thyroid cancer: pathogenic role, molecular bases, and clinical implications. Endocr Rev 2007; 28:742–762.
20. Lloyd R, Buehler D, Khanafshar E. Papillary thyroid carcinoma variants. Head Neck Pathol 2011; 5:51–56.
21▪. Virk RK, Van Dyke AL, Finkelstein A, et al.
BRAF (V600E) mutation in papillary thyroid microcarcinoma: a genotype–phenotype correlation. Mod Pathol 2012 [Epub ahead of print].
An interesting study of the genotype and phenotype correlation in rare aggressive microcarcinomas.
22. Soares P, Lima J, Preto A, et al. Genetic alterations in poorly differentiated and undifferentiated thyroid carcinomas. Curr Genomics 2011; 12:609–617.
23. Ricarte-Filho JC, Ryder M, Chitale DA, et al. Mutational profile of advanced primary and metastatic radioactive iodine-refractory thyroid cancers reveals distinct pathogenetic roles for BRAF, PIK3CA, and AKT1. Cancer Res 2009; 69:4885–4893.
24▪. Prazeres H, Torres J, Rodrigues F, et al. Chromosomal, epigenetic and microRNA-mediated inactivation of LRP1B, a modulator of the extracellular environment of thyroid cancer cells. Oncogene 2011; 30:1302–1317.