Differential Diagnostic Ultrasound Criteria of Papillary and Follicular Carcinomas: A Multivariate Analysis

Key words

thyroid carcinoma - papillary thyroid carcinoma - follicular thyroid carcinoma - ultrasound

Introduction

For some years an increasing incidence of differentiated thyroid cancer has been noted in Germany [1]. This increased incidence is mainly due to a rise in observed papillary thyroid cancer (PTC). For differentiated follicular carcinomas of the thyroid (FTC), the incidence has remained constant. An investigation in the USA taking into account the time period between 1973 and 2002 showed that in particular small PTCs influenced the increase in incidence [2]. Of the increase observed in this time frame, 49 % was due to PTCs smaller than 1 cm in size, and 87 % involved papillary tumors smaller than 2 cm.

Ultrasonography is playing an increasingly important role in the detection of nodules in the thyroid [3] [4]. Using currently available technology, the tiniest lesions down to a size of approx. 1 mm can be detected [5]. Characteristics discovered during sonography that indicate a potential malignant thyroid tumor include calcifications, irregular contours, hypoechogenicity, a so-called “taller than wide” (TTW) sign and increased blood flow disclosed in power Doppler imaging [6].

However, on the whole these criteria do not have a high test accuracy, resulting in many malignant thyroid tumors being classified as benign [7].

Differentiated papillary thyroid carcinomas can be reliably detected using fine needle biopsy due to their cytological characteristics [8]. In contrast, for reasons of principle, differentiation between FTC and follicular adenomas cannot be provided by cytology[9]. Such cases are classified as “follicular neoplasia” requiring histological clarification after a thyroid operation.

To date numerous published ultrasound investigations have described the findings of differentiated thyroid carcinomas and related test accuracy. The purpose of this article, however, is an investigation of the hypothesis that papillary thyroid cancers differ from follicular thyroid cancers with respect to their characteristics as revealed by ultrasound.

Patients and Methods

Patients

The study included 90 examined patients (70 female, 20 male), ranging in age from 16 to 87 and a standard deviation of ± 14.8 years.

The cohort was made up of patients suspected of thyroid disease examined at RNZ outpatient thyroid clinic between 07/01/2009 and 06/30/2012, who subsequently underwent thyroidectomy in various hospitals due to thyroid nodules which consequently could be determined to be papillary or follicular thyroid cancer.

The control group (CON) contained 33 examined patients (25 female, 8 male), ranging in age from 25 to 73 and a standard deviation of ± 12 years, all of whom underwent thyroidectomy during the same time period, and for whom a follicular adenoma (micro- or macrofollicular) could be histologically ascertained.

In the event of cytologically atypical results, a thyroid operation was always indicated. Cytologically atypical results included ambiguous follicular changes, follicular neoplasia or findings that could be attributed to a PTC. Further, in the event an FNAB was not performed or was not usable, a thyroid operation was also indicated if there were factors that indicated the necessity of histological evaluation [10].

Preoperative examinations included a physical examination, an ultrasound examination of the neck region as well determination of in vitro thyroid gland parameters. Patients with thyroid nodules > 10 mm diameter additionally underwent a thyroid scintigram with [^99 mTc]0₄ to exclude focal thyroid gland autonomy.

Upon receipt of the histological results, the ultrasound examinations stored in the PACS were reevaluated retrospectively according to the sonographic criteria. Excluded from the assessment were patients for whom an unambiguous assignment of differentiated thyroid carcinoma to a corresponding sonographic correlate was not possible (n = 4). Also excluded were carcinomas ≤ 1 mm, since due to their small size, sonographic criteria could not be reliably evaluated (n = 4). In all cases, these were papillary thyroid cancers.

Ultrasound Examinations

Ultrasound examinations of the neck region were performed using a SonoAce 8000 SE unit with a 7.5 MHz linear probe. These examinations were performed by physicians specialized in radiology and/or nuclear medicine. Representative images in two planes were stored in a PACS unit.

Tumor size was measured on three planes using an implemented cursor; this provided the diameter in the lateral (dx), anterior-posterior (dy) and craniocaudal (dz) axes. The largest maximum diameter was used for statistical processing.

Upon receipt of the histological results, a physician specializing in radiology and nuclear medicine performed a retrospective image analysis on a PACS console. Tumors were assessed according to six sonographic criteria and the following characteristics:

• Size

  • Maximum diameter on one plane (dx, dy or dz)

• Shape

  • Round: Diameter dx = dy = dz (± 10 %)

  • Oval: one axis surpasses the other axes by > 10 % (Exception: Taller-Than-Wide, see below)

  • Irregular: undulating or complex shape

  • Taller-Than-Wide (TTW): anterior-posterior diameter > lateral diameter, craniocaudal diameter not considered.

• Contour

  • Smooth

  • Spiculated or indistinctly delineated

• Structure

  • Homogeneous

  • Inhomogeneous

• Echogenicity

  • Hypoechogenic: nodule density between anechoic and less than the echogenicity of the perionodal tissue

  • Hyperechogenic: nodule density greater than that of the perinodal tissue

  • Cystic components: anechoic components within the nodule

• Calcification

  • Present

  • Not present

Statistical Evaluation

Statistical evaluation was performed using WinStat^R, version 2009. Statistical significance was presumed at values of p < 0.05.

Results

Distribution by Age and Sex

The average age was mean = 48.7 years with a standard deviation of ± 14.4 years for patients with papillary thyroid cancer and mean = 55.7 years with standard deviation of ± 14.3 years for patients with follicular thyroid cancer. The average age in the control group was mean = 48.5 years with a standard deviation of ± 12.3 years. The age difference among patients in the PTC, FTC and CON groups was not statistically significant (p = 0.09, ANOVA).

The patients were divided into age classes (Age class 1: (10 – 19 years, 2: 20 – 29 years, etc.). The percentage distribution of age classes did not vary significantly among PTC, FTC and CON (0.98, ANOVA) ([Fig. 1]).

Fig. 1 Age distribution (1: 10 – 19 years; 2: 20 – 29 years etc.), green bar: patients (PTCs), yellow bar: patients (FTCs), purple bar: control (CONs) (data in percent).

In the PTC group, 21 % of the patients were male, and 79 % female; in the FTC group, 25 % were male and 75 % female; the CON group was composed of 24 % male patients and 76 % female. Statistically this distribution between PTC patients and CON patients or between FTC patients and CON patients yielded no differences (p = 0.48 and p = 0.81, respectively, chi test).

Multifocality

PTC:: In the case of 53 patients there was a unifocal carcinoma, and 13 patients had a multifocal carcinoma.

FTC:: 23 patients had a unifocal carcinoma, and 1 patient had a multifocal carcinoma.

CON:: 28 patients had a unifocal adenoma, and 5 patients had a multifocal adenoma.

The greater proportion of multifocal tumors among PTC patients compared to the FTC group was statistically significant (p < 0.001, chi test). Regarding multifocality there was no difference between the PTC and CON groups (p = 0.16, chi test). However, difference between the FTC and CON groups was significant (p < 0.001, chi test).

Ultrasound Characteristics (including all PTC, FTC and CON patients)

Size:

For the PTC group, determination of maximum diameter was mean = 12.5 mm with a standard deviation of 8.1 mm; for the FTC group, mean = 35.4 mm and a standard deviation of 19.6 mm. For the CON group, mean = 22.7 mm with a standard deviation of 14.5 mm ([Fig. 2]).

Fig. 2 a Distribution of the ultrasound characteristic “size”, b Distribution of the ultrasound characteristic “shape”. c Distribution of the ultrasound characteristic “contour”, d Distribution of the ultrasound characteristic “structure”, e Distribution of the ultrasound characteristic “echogenicity”. f Distribution of the ultrasound characteristic "calcification".

Statistically, the following significances were found (t test):

• p < 0.001 for PTC vs. FTC,

• p < 0.001 for PTC vs. CON,

• p = 0.013 for FTC vs. CON.

The manifestations of the ultrasound characteristics “shape”, “contour”, “structure” “echogenicity” and “calcification” for PTC, FTC and CON patients are shown in [Table 1].

Table 1 Manifestation frequency of ultrasound characteristics of PTC, FTC und CON (in percent) as well as significance values p.

	PTC	FTC	CON	PTC vs FTC p	PTC vs CON p	FTC vs CON p
shape				< 0.001	< 0.001	= 0.001
round	39	24	34
oval	36	64	53
irregular	14	8	13
TTW	11	4	0
contour				< 0.001	< 0.001	< 0.001
smooth	47	24	74
irregular	53	76	26
internal structure				< 0.001	= 0.84	< 0.001
homogeneous	43	12	42
inhomogeneous	57	88	58
echogenicity				= 0.035	< 0.001	< 0.001
hypoechogenic	82	72	24
hyperechogenic	13	24	76
empty/cystic	5	4	0
calcifications				= 0.37	< 0.001	< 0.001
present	32	28	8
not present	68	72	92

[Fig. 3] show a PTC and an FTC; the respective ultrasound characteristics are described.

Fig. 3 a Ultrasound results for thyroid carcinoma, PTC with a size of 10 mm (maximum diameter). Shape: ttw, contour: irregular, structure: homogeneous, echogenicity: hypoechogenic, calcification: no. b Ultrasound results for thyroid carcinoma, FTC with a size of 59 mm (maximum diameter). Shape: oval, contour: irregular, structure: inhomogeneous, echogenicity: hypoechogenic, calcification: no.

Using an incremental multiple regression analysis, correlation to the dependent variable (PTC, FTC, CON) was calculated for the independent variables “size”, “shape”, “contour”, “structure”, “echogenicity” and “calcification”. It was demonstrated that the variables “size”, “echogenicity” and “calcification” correlated most strongly with the dependent variables with a correlation coefficient of r = 0.57. Ignoring the “size” variable, there was a correlation coefficient of r = 0.50 for the variables “contour”, “echogenicity” and “calcification”. ([Table 2]).

Table 2 Values of the incremental regression analysis of r and p taking into account size, contour, echogenicity and calcifications (column a) as well as the exclusion of size (column b).

	a		b
	r	p	r	p
size	0.57	< 0.001	–	–
contour	0.57	= 0.013	0.50	0.043
echogenicity	0.57	< 0.001	0.50	< 0.001
calcifications	0.57	= 0.002	0.50	= 0.002

Ultrasound Characteristics (excluding papillary microcarcinomas ≤ 10 mm)

Ultrasound characteristics for the patient group (n = 28) with papillary thyroid cancer without papillary microcarcinomas (PTCo) were identified separately compared to the FTC and CON groups.

Size:

For the PTCo group, determination of maximum diameter was mean = 19.3 mm with a standard deviation of 8.5 mm. Statistically, the following significances were found with respect to PTC and CON (t test):

• p < 0.001 for PTCo vs. FTC,

• p = 0.23 for PTCo vs. CON.

The manifestations of the ultrasound characteristics of the PTCo group compared to the FTC and CON groups are shown in [Table 3].

Table 3 Manifestation frequency of ultrasound characteristics of PTCo, FTC und CON (in percent) as well as significance values p.

	PTCo	FTC	CON	PTCo vs FTC p	PTCo vs CON p	FTC vs CON p
shape				< 0.001	< 0.001	= 0.002
round	21	24	34
oval	28	64	53
irregular	34	8	13
TTW	17	4	0
contour				= 0.086	< 0.001	< 0.001
smooth	28	24	74
irregular	72	76	26
internal structure				< 0.001	< 0.001	< 0.001
homogeneous	24	12	42
inhomogeneous	76	88	58
echogenicity				= 0.003	< 0.001	< 0.001
hypoechogenic	72	72	24
hyperechogenic	17	24	76
empty/cystic	10	4	0
calcifications				< 0.001	< 0.001	< 0.001
present	52	28	8
not present	48	72	92

Discussion

Epidemiological investigations indicate an increasing incidence of thyroid cancer. In Germany, the incidence of newly diagnosed cases has increased continuously, particularly for women under 60 years of age, but also for younger men [1]. The increasing rate of new diagnoses appears to be largely influenced by the rise in cases of PTC. One American survey documented an incidence rate increase of PTC diagnoses from 2.7 per 100,000 in 1973 to 7.7 per 100,000 in 2002 [2].

There are no uniform epidemiological data with respect to the incidence of FTC. A current study showed an increase in the USA of 30 % between 1980 and 2009 for occurrences of FTC [11]. In all age groups, FTC appeared more frequently among women. On the whole, men were less frequently affected. However, in the male population the increase in incidence was primarily in advanced age.

In our study, the average age of patients with FTC was greater compared to patients with PTC, the age classes were shifted upward and the proportion of women was lower; however none of these parameters reached a level of statistical significance. Higher age and lower percentage of women among patients with FTC compared to patients with PTC could be likewise be demonstrated. In a study originating in Würzburg, Germany, both parameters achieved a level of significance [12].

Among our patients, the size of the FTC tumors was larger on average than the PTC tumors. This finding has been confirmed by other investigations. Data from the National Cancer Institute´s Surveillance Epidemiology and End Results (SEER) program indicate a significantly higher proportion of larger tumors among FTC patients compared to PTC patients [13]. In their study, Verburg et al. could likewise demonstrate that at the time of diagnosis, patients with PTC exhibited larger tumors compared to patients with PTC. (FTC: average approx. 32 mm; PTC: approx. 19 mm) [12]. However, this study showed a greater percentage of multifocal carcinomas among FTC patients. We were not able to confirm these findings using our data. Among patients we examined, multifocality was indisputably statistically significant in the case of PTC.

A comparison of the size of PTCs, FTCs and CONs in our study demonstrates that PTC tumors are on average smaller than among the FTC and CON groups. There was likewise a statistical difference between the FTC and CON groups, even though the largest follicular adenoma was almost the size of the largest follicular carcinoma.

Papillary microcarcinomas were frequently found in operated thyroids. The publication by Davies et al. [2] notes that papillary carcinomas below 2 cm in size, including papillary microcarcinomas, essentially explain the increase in the incidence of differentiated thyroid cancers.

In our investigation, the portion of patients with papillary microcarcinomas (i. e. tumors ≥ 10 mm) was 57 %. This explains the smaller number of smaller PTC tumors on average compared to FTC tumors.

Excluding papillary microcarcinomas in the statistical size calculation, it can be demonstrated that FTC tumors are significantly larger than PTC tumors at the time of surgery. However, there would no longer be a difference between PTC tumors and follicular adenomas that functioned as controls in our study.

A number of factors are taken into consideration for the geographical and biographical differences in the different stages. In addition to genetic factors [14], nutritional reasons, among others, play a role [15]. In this context, there is an association of FTC with a lack of dietary iodine [15].

In a recently published study of 4955 patients and 7348 assessed thyroid nodules, it was shown that there is not a linear association of increasing nodule size and malignancy risk [16]. In addition, the authors found that with increased nodule size, the percentage of PTC decreased compared to FTC. In this study, the percentage of PTC with a nodule size up to 19 mm was 92 %; only 74 % had a nodule greater than 4 cm. In contrast to our study, other sonographic criteria were not included for the differentiated thyroid cancers.

Moleculargenetic analyses indicate that specific genetic changes are associated with both subtypes of differentiated thyroid cancer In cases of PTC, BRAF as well as modified RET/PTC and trc oncogenes could be verified, whereas in cases of FTC, PAX8/PPARγ rearrangements and RAS mutations could be found [14].

Whether FTC develops directly from follicular thyroid cells or from follicular adenomas is still a matter of scientific controversy [14] [17]. In addition to the described mutations, epigenetic factors leading to gene activation or deactivation may play a role in the development of differentiated thyroid cancer and are co-determinants of the phenotypical appearance of PTC and FTC tumors [18]. These phenotypical manifestations are reflected in the ultrasound characteristics of the thyroid carcinomas.

Popowicz et al. recently emphasized that hypoechogenicity, a TTW sign and verification of microcalcification represent independent risk factors for the presence of thyroid cancer [19]. Although this study distinguished between the sonographic manifestation of small and large nodules (cutoff 15 mm), it did not distinguish between PTC and FTC however. Our study showed that PTC and FTC exhibited hypoechogenicity in approx. 70 – 80 % of cases, but a TTW sign in only 4 – 11 % of cases and microcalcification in approx. 30 % of cases. With respect to manifestation differences between PTC and FTC, structure, in addition to size, was the best indicator: only 12 % of FTC cases exhibited a homogeneous internal structure, but 88 % showed a partially pronounced inhomogeneous internal structure. Among the PTC cases, this was 43 % and 57 % respectively.

However, regarding structure, it must be kept in mind that this does not represent a reliable criterion to distinguish between the PTC and CON groups, if papillary microcarcinomas must be taken into account. If we exclude papillary microcarcinomas from consideration, then structure – seen statistically – remains a differentiation criterion within the three investigated groups.

It is interesting to note that in the entire group of PTC and FTC cases, there was no distinction with respect to the presence of calcification. However, if papillary microcarcinomas are excluded, PTC more frequently exhibits calcification than FTC. In contract, calcification was less frequently detected in follicular adenomas.

This study contains several limitations. This concerns first of all the retrospective evaluation and the selection of patients involved. Only those patients for whom we had a histological finding could be assessed. Although we carefully correlated histologically defined carcinomas to corresponding nodules in the ultrasound examinations, a false correlation cannot be excluded in every instance.

Because the histological findings were made in different facilities, there was no standardization. Consequently there was, for example, no data regarding histological grading for the FTC cases.

Since in our study, assessment was performed retrospectively, and power Doppler or color-coded Doppler ultrasound results were not available for all malignant or benign nodules, we did not consider nodule perfusion as a criterion. Consequently, our study could not classify nodules detected during the ultrasound examination following a TIRADS breakdown, since perfusion behavior is a component of this classification method [20].

Since the study in which the TIRADS classification was initially suggested, no differentiation was made between perfusion behavior of PTC and FTC, it would be useful to examine this criterion in a future study.

Although the values of the sonographic characteristics of the PTCs and FTCs greatly overlap, the investigator should keep the differences in mind in order to control an adequate management process. A risk-adapted FNAP should be an integral part of the diagnostic evaluation of steadily increasing differentiated thyroid carcinomas [21] [22].

• References

• 1 Robert Koch Institut (ed.) Krebs in Deutschland 2005/2006. Häufigkeiten und Trends. Berlin: Westkreuz Druckerei; 2010
• 2 Davies L, Welch HG. Increasing incidence of thyroid cancer in the United States, 1973–2002. JAMA 2006; 295: 2164-2167
• 3 Hegedüs L. Thyroid ultrasound. Endocrinol Metab Clin North Am 2001; 30: 339-360
• 4 Solbiati L, Volterrani L, Rizzatto G et al. The thyroid gland with low uptake lesions: evaluation by ultrasound. Radiology 1985; 155: 187-191
• 5 Kim DW, Kim SH, Jung SJ. Successful sonography-guided fine-needle aspiration Biopsy of a 1-millimeter-diameter papillary thyroid microcarcinoma. AJNR 2010; 31: 1082-1084
• 6 Frates MC, Benson CB, Charboneau JW et al. Management of thyroid nodules detected at US: Society of Radiologists in US consensus conference statement. Radiology 2005; 237: 794-800
• 7 Moon WJ, Jung SL, Lee JH et al. Benign and malignant thyroid nodules: US differentiation-multicenter retrospective study. Radiology 2008; 247: 762-770
• 8 Hamburger JL. Diagnosis of thyroid nodules by fine needle biopsy: use and abuse. Journ Clin Endocrinology Metab 1994; 79: 335-339
• 9 Sugino K, Ito K, Nagahama M et al. Diagnostic accuracy of fine needle aspiration biopsy cytology and ultrasonography in patients with thyroid nodules diagnosed as benign or indeterminate before thyroidectomy. Endocr J 2013; 60: 375-382
• 10 Yeung MJ, Serpell JW. Management of the solitary thyroid nodule. The Oncologist 2008; 13: 105-112
• 11 Aschebrook-Kilfoy B, Grogan R, Ward M et al. Follicular thyroid cancer incidence patterns in the United States, 1980–2009. Thyroid 2013; 23: 1015-1021
• 12 Verburg FA, Mäder U, Luster M et al. Histology does not influence prognosis in differentiated thyroid carcinoma when accounting for age, tumour diameter, invasive growth and metastases. Europ Journ of Endocrinology 2009; 160: 619-624
• 13 Surveillance, Epidemiology and End Results (SEER) Program. Public – use database (1973–2009). Cancer Statistics Branch 2011; 1: 27
• 14 Greco A, Borrello MG, Miranda C et al. Molecular pathology of differentiated thyroid cancer. Q J Nucl Med Mol Imaging 2009; 53: 440-454
• 15 Zhu C, Zheng T, Kilfoy B et al. A birth cohort analysis of the incidence of papillary thyroid cancer in the United States 1973–2004. Thyroid 2009; 19: 1061-1066
• 16 Kamran SC, Marqusee E, Kim MI et al. Thyroid nodule size and prediction of cancer. J Clin Endrocrinol Metab 2013; 89: 564-570
• 17 Krause K, Prawitt S, Eszlinger M et al. Dissecting molecular events in thyroid neoplasia provides evidence for distinct evolution of follicular thyroid adenoma and carcinoma. Am Journ Path 2011; 179: 3066-3073
• 18 Eze O, Starker L, Carling T. The role of alterations in papillary thyroid carcinogenesis. Journ of Thyroid Res 2011; 2011: 895470
• 19 Popowicz B, Klencki M, Lewinski A et al. The usefulness of sonographic features in selection of thyroid nodules for biopsy in relation to the nodule’s size. Europ Journ of Endocrinology 2009; 161: 103-111
• 20 Horvath E, Majlis S, Rossi R et al. An ultrasonogram reporting system for thyroid nodules stratifying cancer risk for clinical management. J Clin Endocrinol Metab 2009; 90: 1748-1751
• 21 Watkinson JC. Fifteen years’ experience in thyroid surgery. Ann R Coll Surg Engl 2010; 92: 541-547
• 22 Paschke R. Abklärung des euthyreoten Schilddrüsenknotens: wann punktieren? Stellenwert der Sonographie. Dtsch Med Wochenschr 2009; 134: 2498-2503