Objective. The purpose of this study was to compare the sonographic features as well as the results of fine-needle aspiration biopsy (FNAB) of follicular variant papillary thyroid carcinoma (FVPTCs) and conventional papillary thyroid carcinoma (PTCs). Methods. Forty patients with 44 FVPTCs and 59 patients with 74 conventional PTCs were enrolled in this study. The sonographic features, sonographic gradings, and FNAB results were compared between the two groups. Results. The mean nodule size of FVPTCs was larger than that of conventional PTCs (17.70 versus 10.53 mm; P < .001). Sonographic features of an ovoid-to-round shape (95% versus 73%), isoechogenicity (52% versus 8%), and a hypoechoic halo (25% versus 3%) were more frequent in FVPTCs than conventional PTCs (P < .001). Sonographic features of a taller-than-wide shape (5% versus 22%), a spiculated margin (7% versus 32%), marked hypoechogenicity (5% versus 38%), and microcalcification (7% versus 24%) were rarer in FVPTCs than conventional PTCs (P < .05). The incidence of a sonographically malignant grade was also lower in FVPTCs (48%) than conventional PTCs (81%; P < .001). A diagnosis of PTC on FNAB of FVPTCs was less common than that of conventional PTCs (28% versus 56%; P = .0393); however, a diagnosis of an indeterminate cytologic type such as atypical cells or follicular lesions in FVPTCs was higher than that in conventional PTCs (46% versus 19%; P = .0418). Conclusions. Follicular variant papillary thyroid carcinomas show a relatively larger size, more benign sonographic features, a lower incidence of a sonographically malignant grade, and a lower diagnostic rate of PTC on FNAB compared with conventional PTCs.
After conventional papillary thyroid carcinoma (PTC), which is the most common type of thyroid carcinoma, follicular variant papillary thyroid carcinoma (FVPTC) is the second most common variant of PTC and variably accounts for 9% to 50% of all PTCs.1–6 Follicular variant papillary thyroid carcinomas have been reported to have a lower incidence of lymph node metastasis and capsular invasion as well as higher rates of lung and bone metastasis compared with conventional PTCs. Recent studies have reported that the same treatment strategies should be applied to both types because the prognostic factors and survival rates were similar between the two groups.1,2,6,7
Since Lindsay8 described FVPTCs as neoplasms having nuclear features of the usual PTCs with a predominantly follicular architectural pattern in 1960, preoperative diagnosis of FVPTCs with fine-needle aspiration biopsy (FNAB) has been attempted. However, this has proved difficult compared with conventional PTCs because of their lack of cytologic clarity.5,8,9
Recently, many studies have reported relatively consistent sonographic features of thyroid malignancy, especially for PTCs10–14; however for FVPTCs, only a few radiologic reports have been published, and they have suggested that FVPTCs had relatively benign sonographic features.15,16 Until now, it has not been determined whether and how an FVPTC is different from a conventional PTC on the basis of sonographic features. In this study, we compared the sonographic features and FNAB results between FVPTCs and conventional PTCs.
Materials and Methods
Institutional Review Board approval was obtained for this retrospective study, and informed consent was waived.
A consecutive series of 200 patients at a single hospital, including 110 with surgically proven FVPTCs between April 2001 and July 2007 and 90 with conventional PTCs between January 2007 and July 2007, was established. Of these patients, 40 with 44 FVPTCs (31 female and 9 male; mean age, 48.3 years) and 59 with 74 conventional PTCs (52 female and 7 male; mean age, 48.6 years) who underwent thyroid sonography were enrolled in this retrospective study. In 2 patients with FVPTCs, 2 conventional PTCs coexisted, so they were included only in the evaluation of multiplicity on the sonographic analysis. All of the 59 patients with conventional PTCs underwent total thyroidectomy, whereas, the 40 patients with FVPTCs underwent total thyroidectomy (n = 28), lobectomy (n = 11), or partial thyroidectomy (n = 1).
All sonographic examinations were performed with either an Accuvix XQ system (Medison Co, Ltd, Seoul, Korea) or a Philips HDI 5000 system (Philips Healthcare, Andover, MA) equipped with a 7.5- to 12- or 8- to 15-MHz linear array transducer. The scanning protocol in all cases included both transverse and longitudinal real-time imaging of the thyroid nodules, with representative Digital Imaging Communications in Medicine images. The mean duration between preoperative sonographic examination and thyroidectomy was 21.8 days (range, 0–32 days).
Two head and neck radiologists (J.-H.K. and D.G.N.) with 5 and 10 years of experience in performing and evaluating thyroid sonography, respectively, reviewed all of the sonograms with the use of representative Digital Imaging Communications in Medicine images. The reviewers were blinded to the clinical history and final pathologic findings of the patients. Final decisions were reached by consensus. The maximal diameter of the nodules was measured with an electronic caliper on picture archiving and communications system monitors. For each thyroid nodule, more than 2 gray scale images were acquired in transverse and longitudinal planes (mean, 3.57 images; range, 2–6 images).
The following sonographic features were evaluated: nodule size, internal content, shape, margin, echogenicity, calcification, and presence of a hypoechoic halo as described previously by Moon et al.10 Nodule size was described by the longest and shortest dimensions. The internal content of a nodule was described as pure solid, predominantly solid, predominantly cystic, or pure cystic. The shape of a nodule was categorized as ovoid to round, taller than wide, or irregular. The character of the margin was categorized as smooth, spiculated, or poorly defined. The echogenicity of the solid portion of a nodule was assessed with respect to the thyroid parenchyma and strap muscles and was categorized as markedly hypoechoic, hypoechoic, isoechoic, or hyperechoic. Calcification was assessed according to its size and was categorized as microcalcification, macrocalcification, or rim calcification. The presence of a hypoechoic halo was determined if there was a thin hypoechoic rim at more than 75% of the peripheral margin of a nodule.
We classified the nodules as malignant if even 1 of the following suspicious sonographic features was present, as suggested by Moon et al10: taller-than-wide shape, spiculated margin, marked hypoechogenicity, and presence of microcalcification or macrocalcification.
Criteria for sonographically guided FNAB included nodules 1 cm or larger and those smaller than 1 cm with at least 1 suspicious sonographic feature. However, in the cases in which the patient or the corresponding clinician requested FNAB for a nodule regardless of whether it met our criteria for the procedure, FNAB was performed according to the hospital policy. Cytologic material obtained by blinded or sonographically guided FNAB was smeared on a slide immediately after aspiration (solid lesions) or after cytologic centrifugation (in the case of cystic fluid) and processed with May-Grunwald Giemsa and Papanicolaou stains. Cytologic specimens were then evaluated by experienced cytopathologists. Surgical samples were sectioned before fixation for macroscopic examination and then fixed in 10% buffered formalin, embedded in paraffin, and stained with hematoxylin-eosin for histologic examination. Cytologic diagnostic criteria for FVPTCs were small follicles containing dense colloid surrounded by cells having nuclei characteristic of papillary carcinoma.9,17,18
Continuous variables were compared by unpaired Student t tests, and categorical variables were compared by χ2 analysis or the Fisher exact test as appropriate. P < .05 was considered to indicate statistical significance. All analyses were performed with SPSS version 13 software for Windows (SPSS Inc, Chicago, IL).
Sonographic Features of FVPTC and Conventional PTC
Table 1 summarizes the sonographic features of FVPTCs and conventional PTCs. The mean nodule size of the FVPTCs was larger than that of the conventional PTCs (longest diameter, 17.70 versus 10.53 mm; P < .001; shortest diameter, 10.98 versus 6.85 mm; P = .005). The most common sonographic features of FVPTCs were a pure solid internal content (93%), an ovoid-to-round shape (95%), the presence of a smooth margin (73%), isoechogenicity (52%) or hypoechogenicity (43%), and the absence of calcification (68%). In contrast, the most common sonographic features of conventional PTCs were a pure solid internal content (99%), an ovoid-to-round shape (73%), the presence of a smooth margin (55%), hypoechogenicity (54%), and the absence of calcification (53%; Figure 1).
As for nodule shape, an ovoid-to-round shape was more frequent (95% versus 73%; P = .03) and a taller-than-wide shape was rarer (5% versus 22%; P = .016) in FVPTCs than conventional PTCs (Figures 1 and 2). The presence of a spiculated margin was also rarer in FVPTCs than conventional PTCs (7% versus 32%; P = .001). As for echogenicity, marked hypoechogenicity was rarer (5% versus 38%; P < .001), microcalcification was rarer (7% versus 24%; P = .0234), and isoechogenicity was more frequent (52% versus 8%; P < .001) in FVPTCs than conventional PTCs. The presence of a hypoechoic halo was also more frequent in FVPTCs than conventional PTCs (25% versus 3%; P < .001; Figures 1 and 2).
The incidence of a solid internal content (93% versus 99%; P = .145), a smooth margin (73% versus 55%; P = .078), hypoechogenicity (43% versus 54%; P > .999), and macrocalcification (20% versus 22%; P > .999) were not significantly different in FVPTCs and conventional PTCs (Figures 3 and 4). The incidence of multiplicity was not significantly lower in FVPTCs than conventional PTCs (15% versus 20%; P = .9107).
The incidence rates of sonographically malignant grades for FVPTCs and conventional PTCs are shown in Table 2. A total of 23 of 44 FVPTCs (52%) did not meet the criteria to predict the presence of thyroid malignancy, whereas only 14 of 74 conventional PTCs (19%) did not meet the criteria (P = .0004).
Results of FNAB
Table 3 shows FNAB results in the 2 groups. Fine-needle aspiration biopsy was performed in 35 of 44 FVPTC nodules (80%) and 36 of 74 conventional PTC nodules (49%) before thyroid surgery in our hospital. A diagnosis of PTC on FNAB of FVPTCs was less common than that of conventional PTCs (28% versus 56%; P = .0393); however, a diagnosis of an indeterminate cytologic type such as atypical cells or follicular lesions in FVPTCs was higher than that in conventional PTCs (46% versus 19%; P = .0418). Only 1 of the 35 FVPTCs was diagnosed as FVPTC on FNAB.
According to our study, the common sonographic features of FVPTCs are different from those of conventional PTCs. The relative frequency of an ovoid-to-round shape and relative paucity of a taller-than-wide shape and spiculated margin that we observed in FVPTCs may be explained by the tendency of these lesions to grow parallel to the normal tissue plane rather than infiltratively across normal tissue.10,19 The frequent isoechogenicity or hypoechogenicity rather than marked hypoechogenicity in FVPTCs might be related to abundant follicles and a lesser degree of nodule compaction, fibrosis, or cellularity compared with conventional PTCs.19,20 We also found a relatively more frequent hypoechoic rim in FVPTCs, and this may have been due to the tendency of compression of the normal thyroid tissue or capsule.19
Table 1. Comparison of Sonographic Features Between FVPTCs and Conventional PTCs
FVPTCs (n = 44)
Conventional PTCs (n = 74)
aFisher exact test or χ analysis.
bData are mean ± SD.
cTotal numbers of patients (FVPTCs, 40; conventional PTCs, 59) were standardized for this category. Two conventional PTCs coexisting in 2 patients with FVPTCs were included only in the evaluation of multiplicity for the sonographic analysis.
17.70 ± 11.76
10.53 ± 6.32
10.98 ± 6.49
6.85 ± 3.88
Internal content, n (%)
Shape, n (%)
Ovoid to round
Taller than wide
Margin, n (%)
Echogenicity, n (%)
Calcification, n (%)
Absence of calcification
Hypoechoic halo, n (%)
Multiplicity, n (%)c
The common sonographic features of solid internal content, an ovoid-to-round shape, a smooth margin, isoechoic echogenicity, and the relative frequency of a hypoechoic rim in FVPTCs are similar to the sonographic features of other follicular cell–derived lesions such as nodular hyperplasia, follicular adenoma, and follicular carcinoma.10,15,16,19,21,22 The sonographic similarities between FVPTCs and other follicular cell–derived lesions might be reflections of the similar gross features of encapsulated masses, which are composed entirely of variably sized follicles without virtually any papillary structures.7,17
In addition to the high incidence of indeterminate cytologic results such as atypical cell and follicular lesions in FVPTCs, the lower cytologic diagnostic results of PTC in FVPTCs in this study are concordant with many articles that have discussed the cytologic diagnostic difficulty of FVPTC.5,9,16–18,23,24 We surmise that the cytologic diagnostic difficulty of FVPTC might be attributable to a lack of thorough sampling of the nodule and a paucity of typical nuclear changes of papillary carcinoma in the underlying abundant colloid and follicular cells.4,9,18,20,23,24 Although it is not certain whether it is because FVPTCs usually do not manifest with a small size or FNAB of FVPTCs is difficult in small nodules, our study shows that FVPTCs manifest with a relatively larger size compared with conventional PTCs, which is in concordance with previous reports.2 If the relatively larger size of FVPTCs found in our study was influenced by the size criteria of 1.0 or 1.5 cm for FNAB, we can assume that it might also be difficult for small thyroid nodules to meet the cytologic diagnostic criteria for FVPTC because it requires detection of small nuclear features of PTC in the abundant follicular pattern.
In contrast to the high incidence of a sonographically malignant grade (81%) for conventional PTCs, the lower incidence of a malignant grade for FVPTCs (48%) in our study suggests that recently reported sonographically suspicious malignant features are applicable to conventional PTCs but not to FVPTCs. These results also suggest that the absence of sonographically suspicious malignant features in thyroid nodules cannot guarantee their benignancy; therefore, we cannot exclude follicular cell–derived malignancies such as FVPTC and follicular carcinoma.10,19 However, considering the relatively larger size of FVPTCs (maximal dimension ± SD, 17.70 ± 11.76 mm), strict application of size criteria is recommended even in benign-looking thyroid nodules.
Table 2. Comparison of Sonographic Grades Between FVPTCs Conventional PTCs
For these 118 nodules, the sonographic classification was statistically significant between FVPTCs and conventional PTCs (P < .05, χ test).
Malignant, n (%)
Not malignant, n (%)
Table 3. Comparison of FNAB Results Between FVPTCs and Conventional PTCs
aFisher exact test or χ analysis.
FNAB results, n (%)
Atypical cell or follicular lesion
The sonographic features of FVPTC in our study were comparable with the results reported by Yoon et al,15 although there were some differences in the fact that only our study included macrocalcification and rim calcification in the analysis of calcification and macrocalcification in the sonographically malignant criteria. Although there are some arguments that malignancy in cytologic follicular lesions has the tendency to show microcalcification and a larger size, the sonographic differentiation of FVPTCs from follicular neoplasms also seems as difficult as the sonographic differentiation between follicular adenoma and carcinoma.22,25 However, the characteristic sonographic feature of nodular hyperplasia with a spongiform appearance and isoechogenicity may exclude FVPTCs and follicular neoplasms.10
This study had several limitations in that it was a retrospective study with unavoidable selection bias, and there was a relatively smaller number of FVPTC cases. A larger-scale prospective study would overcome these limitations.
In conclusion, FVPTCs show a relatively larger size, more benign sonographic features, a lower incidence of a sonographically malignant grade, and a lower diagnostic rate of PTC on FNAB compared with conventional PTCs.