Papillary thyroid carcinoma (PTC) is the most common thyroid malignancy, whereas follicular variant PTC (FVPTC) is the second most common subtype after classic PTC (cPTC), accounting for 9% to 22.5% of all PTCs.¹ There are 2 main subtypes of FVPTC: encapsulated FVPTC (EFVPTC) and infiltrative FVPTC (IFVPTC), with different biological and clinical characteristics.^2,3 The former could be divided into 2 subtypes: invasive EFVPTC and noninvasive EFVPTC. In 2016, noninvasive EFVPTC was renamed as “noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP),”⁴ which had the characteristics such as a complete capsule, absence of invasion, and nuclear features of PTC. It was reported that NIFTP accounted for 9.1% of all PTCs.⁵ Because of the very low risk of adverse outcome, NIFTP has become a new research hotspot to reduce overtreatment of this tumor.^2,6

Conventional ultrasound (US) is the most common diagnostic tool for thyroid nodules. Previous studies have compared the ultrasonographic characteristics of NIFTP and non-NIFTP, however, there were few studies which compared NIFTP with cPTC or with follicular neoplasm (FN), and the sonographic features of NIFTP remain unclear.^7–9 NIFTP has a noninvasive capsule or a clear demarcation of the tumor from adjacent thyroid tissue, and the growth pattern is similar with FN. Regarding pathological features, follicular adenoma (FA) has a complete capsule, while follicular thyroid carcinoma (FTC) has infiltration to the capsule and blood vessels, which may be micro-infiltration to the capsule or extensive infiltration.¹⁰ Meanwhile, cPTC is the most common pathological type among all PTCs. An NIFTP has papillary-like nuclear characteristics similar with cPTC. Because of the complex sonographic features of thyroid nodules, several US risk stratification systems have been developed to standardize the execution process of US and the diagnostic criteria in recent years. In this study, we retrospectively analyzed the clinical and ultrasonographic characteristics of NIFTP so as to investigate the differences among FN (FA and FTC), cPTC, and NIFTP. We also investigated the capability of the US risk stratification systems including ATA and ACR-TIRADS of differentiation the NIFTP from cPTC and FN.

PATIENTS AND METHODS

Patients

This retrospective study was approved by the Institutional Review Board, with waiver of informed consent. One hundred seventy-eight consecutive patients with a single nodule who underwent ultrasonography and surgery in our hospital from January 2018 to October 2019 were collected. There were 29 cases of NIFTP, 95 cases of FN (41 FTC and 54 FA) and 54 cases of cPTC. All nodules were confirmed by postoperative pathology. The hematoxylin and eosin (H&E) staining slides were evaluated by 2 pathologists with more than 10 years of pathological diagnosis experience who were blinded to the clinical and imaging data. The following histological criteria must be met to diagnose the NIFTP: (1) encapsulation or clear demarcation; (2) follicular growth pattern with less than 1% papillae, no psammoma bodies, and less than 30% solid/trabecular/insular growth pattern; (3) nuclear score 2–3; (4) no vascular or capsular invasion; (5) no tumor necrosis; and (6) no high mitotic activity.¹¹

The inclusion criteria were as follows: (1) solitary tumor confirmed by histopathological examination; (2) patients with complete medical information, including personal information, sonographic images and reports, blood tests, medical case records, surgical records and postoperative pathological diagnosis; (3) US examination was completed within 2 weeks before surgery; and (4) patients had at least undergone hemithyroidectomy and ipsilateral central compartment node dissection. Exclusion criteria were as follows: (1) multiple lesions or coexisting with other types of thyroid malignancies; (2) patients with history of neck surgery or neck irradiation; (3) US examination was not performed in our hospital before operation; and (4) nodules could not be classified by ATA.

Ultrasound Examination

During the sonographic examination, the patient was in the supine position, with the neck fully exposed. Both gray scale and color Doppler examinations were performed with a 4- to 13-MHz linear probe (MyLab 90; EsaoteSpA, Genoa, Italy; iU22 system, Philips, Seattle, WA) by 2 radiologists with more than 5 years of experience in thyroid imaging. The focus, gain, and depth were adjusted to obtain the best image. Doppler parameters were optimized to maximize Doppler sensitivity in the absence of any color noise in normal thyroid tissue. Images of each suspicious nodule were obtained both in transverse and longitudinal orientations. The patients who had nodules with no malignant features needed a regular follow up every 6 months. If the size of nodule remained stable, further follow-up was performed. All images were recorded and uploaded to a picture archiving and communication system for further analysis.

Image Analysis

The sonographic appearances of thyroid nodules were evaluated carefully, including composition, echogenicity, shape, margin, echogenic foci, and halo. According to the proportion of solid components, the composition was classified as cystic, solid, and mix. The echogenicity was determined as anechoic, hyperechoic or isoechoic, hypoechoic, and very hypoechoic, and it referred to the nodule's reflectivity relative to peripheral thyroid gland, except for very hypoechoic nodules, in which the strap muscles were used for comparison.¹² Margin was categorized into smooth, lobulated or irregular, and extrathyroidal extension. Shape was divided into wider than tall (T ≤ W) and taller than wide (T > W). Echogenic foci was assessed as macrocalcifications (with diameter ≥2 mm or coexisted with microcalcifications), peripheral calcifications, punctate echogenic foci (with diameter <2 mm), and none or large or comet-tail artifacts. Halo was defined as absent or present. Color Doppler sonographic features of thyroid nodules included vascularity degree and vascularity pattern. Compared with the surrounding thyroid tissue, vascularity degree was classified as high, medium, or low. Vascularity pattern was classified perinodular, mild intranodular, and mix. The sonographic features of the target nodules on grayscale and color Doppler were assessed by 2 radiologists in consensus.

After evaluating the ultrasound indicators, the nodules were classified according to ACR-TIRADS¹² and ATA guidelines,¹³ respectively. The ATA assessments of the thyroid nodules were as follows: high suspicion: a solid hypoechoic nodule with 1 or more malignant features such as irregular margins, microcalcifications, taller than wide shape, and evidence of extrathyroidal extension; intermediate suspicion: a hypoechoic solid nodule with a smooth regular margin, but without malignant features; low suspicion: an isoechoic or hyperechoic solid nodule, or partially cystic nodule with eccentric uniformly solid areas without suspicion for malignancy. The nodules that could not be classified by ATA such as solid or mixed nodules, with hyperechoic or isoechoic echogenicity, accompanied by the following malignant features including punctate echogenic foci, T > W, irregular margin, or extrathyroidal extension were excluded. According to the ACR-TI-RADS, scores were assigned to the nodules, and the classifications of nodules were as follows: TR 1: 0 points; TR 2: 2 points; TR 3: 3 points; TR 4: 4–6 points; and TR 5: 7 points or more.

Statistical Analyses

Continuous variables were expressed as the mean with standard deviation. Measurement data were compared using t test. Categorical variables were compared using χ² or Fisher exact tests. A P value less than 0.05 was considered to indicate statistical significance. The statistical analyses were performed using SPSS version 25 (IBM Corporation, Armonk, NY).

RESULTS

1. Clinicopathological Characteristics

The clinicopathological characteristics of the patients were listed in Table 1. A total of 29 NIFTPs, 54 cPTCs, 41 FTCs, and 54 FAs were confirmed by postoperative pathology. There were no statistically significant differences in age and sex between NIFTP group and the other groups; however, the average maximum diameter of NIFTP was larger than that of cPTC (P = 0.000). No cervical lymph node metastases were observed in NIFTP, FTC, and FA groups. However, there were 17 cases with lymph node metastases in cPTC group. Significant difference was found in lymph node metastases between NIFTP and cPTC groups (0% vs 31.5%, P = 0.000).

• 2. Ultrasonographic Characteristics

The ultrasonographic characteristics of the nodules were listed in Table 2. Most NIFTPs had the following sonographic features: solid composition, hypoechoic echogenicity, smooth margin, T ≤ W shape, none echogenic foci, absence of halo, and perinodular vascularity. No significant differences were detected in the sonographic features between NIFTP (Fig. 1) and FN groups (FA [Fig. 2] and FTC [Fig. 3], P > 0.05 for both). However, there were significant differences in echogenicity, shape, margin, echogenic foci, absence of halo, and vascularity degree between NIFTP and cPTC groups (P < 0.05, for all). Solid composition, hypoechoic, irregular margin, punctate echogenic foci, and low vascularity were the most common appearances in cPTC group (Fig. 4).

• 3. The ACR-TIRADS and ATA Category

The US risk stratifications of the nodules were listed in Figure 5. The sonographic features of NIFTPs in ATA risk classification were showed in Table 3. There were statistical differences both in ACR-TI-RADS and in ATA classification between NIFTP and cPTC groups (P = 0.000, for both); however, no differences were observed between NIFTP and FA or FTC groups (P > 0.05, for both).

DISCUSSION

Because of the benign behaviors of noninvasive EFVPTC, it was redefined as NIFTP by the Global Endocrine Pathology, and the term carcinoma was avoided.⁴ An NIFTP presents as an encapsulated and well-demarcated noninvasive follicular neoplasm, with the papillary-like nuclear features on pathology.¹⁴ It was reported that a correct diagnosis of NIFTP could reduce the incidence of thyroid carcinoma by 10% to 20%.¹⁵ It is also of great significance to reduce overtreatment. Most studies suggested that NIFTP should get lobectomy without radioactive iodine.¹⁶ Some guidelines even recommended more conservative treatment.^13,17,18 Because it was no longer called carcinoma, there was no need to perform tumor staging, which could greatly reduce the psychological burden of patients, as well as the risk of conservative treatment.¹⁷

The recognition of the clinical and ultrasonographic characteristics of NIFTP might be helpful for accurate differential diagnosis, which was particularly important before surgery. In the present study, there were no differences in age and sex among the 4 groups; however, significant differences were observed in nodule size and cervical lymph node metastases. The results in our study indicated that the nodule size of NIFTP was similar to those of FTC and FA but significantly larger than cPTC. Previous research¹⁹ also showed that FA was significantly larger than cPTC; however, NIFTP could not be differentiated based on nodule size alone. The increase of nodule size affects cancer risk in a nonlinear way. A threshold size was observed at 2.0 cm, above which cancer risk remained unchanged. The risk of FTC and other thyroid malignancies was raised with the increasing nodule size.²⁰ Moreover, a large size (>4.0 cm) was a risk factor for a more advanced disease in patients with differentiated thyroid carcinoma, and the size of the nodule was related to the prognosis and the risk of extrathyroidal extension tissue infiltration outside the thyroid.

Papillary thyroid carcinoma is frequently associated with cervical lymph node metastases, whereas distant metastases, such as the lung and bone metastases, are more common in patients with FTC,^21,22 and regional lymph node metastases are rare. With regard to NIFTP, it was considered to have a good prognosis. Sowder et al²³ reported that no lymph node metastases were detected in patients with NIFTP either at the time of surgery or at follow-up. In our study, the incidence of lymph node metastases was 31.5% in cPTC group, which was statistically different from those in the other groups. Although the biological behavior of NIFTP was similar to that of benign tumor, metastases and recurrences had also been reported in some cases.²⁴ It was probably because that the pathological diagnostic criteria were not strict, and the invasion of capsules was not detected thoroughly in the retrospective analysis.²⁵ In our study, there were no lymph node metastases in NIFTP group; however, a long-term follow-up study with a larger sample size should be performed to observe if there would be any cervical lymph node metastases, distant metastases, or recurrences in the future.

The sonographic features of thyroid nodules have been compared in NIFTP and non-NIFTP. You et al⁸ found that macrocalcification was less common in NIFTP than in non-NIFTP, and all of the NIFTPs were classified as low or moderate suspicion. The non-NIFTP in the study referred to IFVPTC and invasive EFVPTC; however, no other pathologic types were included. In addition, there have been several reports on the comparison of FVPTC and cPTC; however, in most of these studies, follicular type of PTC was not subdivided into NIFTP and non-NIFTP. We compared the differences of sonographic findings between NIFTP and cPTC/FN, to understand the ultrasonographic characteristics of NIFTPs and make a correct diagnosis.

Microcalcification, irregular border, and higher-than-wide shape were usually considered to be the characteristics of cPTCs.²⁶ The results in our study showed that NIFTPs had the sonographic appearances similar with benign nodules, including hypoechoic echogenicity, T ≤ W, absence of halo, smooth margin, and non–echogenic foci. However, these features were statistically different from those of cPTCs, although NIFTP also showed papillary nuclear on pathology. These results were consistent with the study by Brandler.¹⁹ The difference between these 2 tumors was probably because of the complete tumor capsule with no invasion. On color Doppler, high vascularity degree was commonly observed in the NIFTP group. Yang et al²⁷ also found that NIFTP mostly showed high blood supply. The results showed that cPTCs frequently showed low vascularity degree in our study, which was similar to the results found in a previous study. Yang and Fried²⁸ reported that most PTCs showed a lack of vascularity on US. The pathological vessels of thyroid cancer were significantly different from normal vessels. The lack of muscle and nerve support, vascular tortuosity, and dysfunction could result in low blood supply in cPTCs.²⁹ As for vascularity pattern, Brandler et al¹⁹ found that cPTC was mostly manifested as intranodular vascularity, whereas NIFTP was manifested as mixed vascularity pattern (including perinodular and intranodular). However, there was no significant difference in vascularity pattern between NIFTP and cPTC in our study.

The most common presentation of FN is a hypoechoic or isoechoic, round tumor without calcification.³⁰ The accurate assessment of FN is challenging because of the similar appearances of FC and FA on US. As previously reported, the sonographic features of NIFTPs were similar to those of FAs,¹⁹ which was consistent with our results. To the best of our knowledge, there was no study that had been performed to compare the differences of sonographic features between FTC and NIFTP. Yang et al³¹ had conducted a research on NIFTP, which focused on the pathologic basis of the sonographic features that distinguished NIFTP from thyroid cancers. However, the differences of sonographic features were not discussed in detail. Our results showed that the sonographic appearances of NIFTPs overlapped with those of FTCs, and no differences were observed in the ultrasonographic characteristics between them. Pathologically, NIFTP is characterized by no capsular or vascular invasion.⁴ On the contrary, FTC has capsular and/or vascular invasion. Ultrasound could detect significant gross infiltration. Tumor protrusion, which was regarded as large forms of capsular invasion of FTC, could be observed by preoperative ultrasonography.³² However, histopathological examination is required to determine capsule microinfiltration or vascular infiltration. In our study, all FTCs presented with capsule microinfiltration on histological examination.

Although sonographic features are helpful in differentiating benign and malignant thyroid nodules, it may be affected by many factors such as intra-assessment and interassessment variability and suboptimal sensitivity and specificity. Ultrasound risk stratification systems could identify significant malignancies and reduce over examinations and treatments of benign nodules.^12,33–37 The ATA incorporated malignant indicators to distinguish low, intermediate, high suspicion, without counting, which was more convenient and fast, whereas ACR-TIRADS assigned scores to the indicators, which was more objective and practical. According to the ACR-TIRADS, most of the NIFTPs in our study were categorized as TR 4, which was similar to FTC and FA. TR 4 was defined as moderately suspicious, and fine needle aspiration (FNA) was not recommended if the nodule was less than 1.5 cm. However, most of the cPTCs in our study were categorized as TR 5, which was statistically different from NIFTP. A similar result was found in the study by Yang et al,³⁸ most NIFTPs showed TR 3–4, whereas none was TR 5. TR 5 was classified as highly suspicious, and FNA should be performed according to the recommendation when the size of nodule was more than 1.0 cm.

The same results were obtained by the ATA guidelines risk stratification of the 4 groups. Noninvasive follicular thyroid neoplasm with papillary-like nuclear features were more likely classified as low and intermediate suspicion, whereas 90.7% of cPTCs were classified as highly suspicion. There was a significant difference between NIFTPs and cPTCs (P < 0.05). Fine needle aspiration was recommended for nodules larger than 1.0 cm, regardless of whether the nodules were intermediate or highly suspicious. However, the estimated risk of malignancy was more than 70% to 90% for highly suspicion, whereas they were 10% to 20% for intermediate suspicion and 5% to 10% for low suspicion. Valderrabano et al³⁹ also found that most NIFTPs had a low or intermediate suspicion. If NIFTPs were considered as benign, the prevalence of malignancy in the low and intermediate suspicion groups would dropped by nearly half.³⁹ Our results showed that FNs were more inclined to low and intermediate suspicious, which were similar with NIFTPs.

All results in our study indicated that both ACR-TIRADS and ATA could discriminate NIFTPs and cPTCs with well agreement. We suggest that NIFTP can basically be ruled out if the nodule is highly suspicious on US and cytopathology shows classical features of PTC, and these nodules should be treated aggressively. However, because of the similarity of NIFTP and FN in ultrasonographic findings, the value of US alone was limited. As previously reported, the use of ultrasound elastography may improve the ability to discriminate benign from malignant thyroid nodules.⁴⁰ However, to the best of our knowledge, there was no comparative study on ultrasound elastography between NIFTP and FN. For those low to intermediate suspicious nodules with appropriate size, FNA should still be performed according to the guidelines.^12,13 If cytopathology shows follicular proliferative pattern with papillary-like nuclear features,⁴¹ and the nodule appears as a follicular nodule on US, these findings may indicate the possibility of NIFTP. However, these cannot distinguish NIFTP from FVPTC, molecular testing in FNA washout may be necessary for future research.

There are some limitations in this study. First, all cases were selected from patients undergoing surgery rather than the general population, and selection bias was inevitable because of the retrospective nature of this study. The patients who had nodules with no malignant features were followed up every 6 months; however, these nodules still had the potential for malignancy. These patients were not included in our study, which might also lead to selection bias. Second, the sample size was small, and it might be unable to show clinically meaningful differences. Therefore, a study with a larger sample size should be performed in the future. Finally, the mean size of cPTCs was small (9.6 ± 5.4 mm) in our study. Some patients with papillary thyroid microcarcinoma, which was less than 10 mm in diameter might not require surgery.

CONCLUSIONS

Noninvasive follicular thyroid neoplasm with papillary-like nuclear features had different clinical and sonographic features from cPTC, and there were some overlaps with FTC and FA. Before surgery, we should be aware that these may be NIFTP nodules, which show papillary-like nuclear characteristics on cytology, but had a benign appearance on ultrasound. Thus, it might avoid extensive surgery and additional therapies such as radioactive iodine ablation for NIFTP.