Abstract
Keywords
Introduction
The incidence of thyroid nodules has risen to 67% of general population in recent years [4]. Nevertheless, only 5% of the nodules were confirmed as malignant [5].
Thus it is important to distinguish a malignant one from all. As a non-invasive and available tool worldwide, high-resolution ultrasound has been the preferred imaging modality of thyroid nodules. It provides information regarding sonographic features of thyroid nodules, such as border and margin characteristics, echogenicity, and presence or absence of microcalcifications, etc. However, none of the above is sensitive or specific enough for diagnosing malignancy individually as considerable overlap has occurred in some studies [6, 10].
Contrast-enhanced ultrasound (CEUS), using a microbubble contrast material to investigate dynamic enhancing pattern and alteration of nodal perfusion, has been widely used for assessment of thyroid-related diseases. Researchers are divided over the clinical value of CEUS. Some of them suggested that CEUS could be a potential tool in differentiating benign from malignant nodules [7, 14], while others have stated that it did not improve the characterization of thyroid nodules [1, 2].
At present, the contrast-enhancement pattern of CEUS has been classified as “absent” with no change in enhancement of the lesion before and after contrast agent injection; “dotted” with tiny separate spots of enhancement distributed throughout the lesion; or “diffuse” with homogeneous or heterogeneous enhancement of the entire lesion. Others included echogenicity (with respect to thyroid parenchyma), classified as hyperechoic, isoechoic, or hypoechoic. In Bartolotta’s study, malignant nodules showed absent, faint dotted contrast enhancement while both homogeneous and heterogeneous pattern were found in benign and malignant nodules [1]. However, the conclusion was not quite the same from Ma’s observation: benign nodules appeared homogeneous, whereas malignant nodules appeared heterogeneous; malignant nodules showed absent or incomplete ring enhancement as a result of the infiltrative growth [7]. Therefore, there is no consistent comprehension of the evaluation criteria of contrast enhancement to the best of our knowledge.
To further extend one of the contrast-enhancement pattern, our current study sought to investigate the unique diagnostic value of “absent” pattern of contrast-enhancement in the differentiation of suspicious thyroid nodules, and analyze corresponding characteristics of conventional B-scan US and real-time elastography (RTE).
Materials and methods
Patients
This retrospective study was approved by local institutional review board and written informed consent was obtained from all patients before RTE and CEUS. Medical records of 8476 consecutive patients were reviewed for patient demographics, clinical and radiological evaluation results and pathology examination from November 2012 to June 2014. Among them, 96 patients who underwent further RTE and CEUS with suspicious thyroid nodules (difficult to identify with conventional US) were eventually enrolled. An additional recruitment of the study individuals was the presence of “absent” pattern of contrast-enhancement.
There were 56 female and 40 male patients, whose median age was 53 years (30 to 84 years), with thyroid nodules (size ranged from 3.8 to 25 mm, mean: 9.5 mm) at diagnosis.
Fine-needle aspiration (FNA) cytology/histology
All included patients received either cytology using FNA and/or histology from thyroid surgery to verify the diagnosis. FNA under US guidance was performed by a skilled radiologist using a 25-gauge needle attached to a 10-ml syringe.
Conventional B-scan US
All patients underwent an ultrasound examination of the thyroid gland by ACUSON S2000 ultrasound system (Siemens, Erlangen, Germany) with a linear probe (4–9 MHz) by one radiologist with 20 years of experience in thyroid sonography. The patients were positioned in a supine position with dorsal flexion of the head, and thyroid nodules were evaluated in size, border, margin, echogenicity (hyperechoic, isoechoic, hypoechoic, echocomplex with respect to normal thyroid parenchyma), echotexture (homogeneous, inhomogeneous), presence/absence of microcalcification (calcification less than 2.0 mm in diameter).
Real-time elastography (RTE)
New technique RTE could evaluate tissue stiffness by measuring the degree of distortion under application of an external force [11]. Since malignant nodules tend to be much harder than the surrounding benign tissues, RTE could appropriately provide additional information of a malignant lesion with successful application in different fields [8, 16]. RTE of the thyroid gland was performed with a 6–13 MHz linear transducer (Vision Preirus; Hitachi Medical, Tokyo, Japan). The probe was placed on the neck and a light pressure on a 3-4 scale was applied for measurement. The region-of-interest (ROI) for the elastography examination was selected by the operator including the nodule and surrounding normal thyroid tissue. Elastography images were classified into a scale of 1–5 following the criteria of Rago et al (Score 1: the nodule is displayed in green homogeneously; Score 2: the nodule is displayed predominantly in green with a few blue spots; Score 3: the nodule is displayed predominantly in blue with few green spots; Score 4: the entire nodule is displayed in blue; Score 5: both the nodule and its surrounding area are blue) [13]. In this study, a lesion showing Rago scores of 3–5 was considered as a “hard” malignancy and scores of 1-2 as a “soft” one.
Contrast-enhanced ultrasound (CEUS)
CEUS imaging was performed by one experienced radiologist with a high-resolution ultrasound system (ESAOTE MyLab 90 Xvision, Italy) equipped with a linear probe of 12 MHz at low MI (0.05–0.07). As bubble disruption was strictly related to depth and focalization of a US beam, focus was always placed deeper than the nodule being examined so as not to destroy the microbubbles. Sulphur hexafluoride [3] (SonoVue®, Bracco International, Milan, Italy) was injected with an intravenous bolus of 1.2 ml via a 20-gauge cannula followed by a 10 ml saline flush. Meantime, the timer on the US machine was started, and each contrast imaging acquisition lasted at least 3 minutes was digitally stored as raw data on a personal computer-based workstation.
Statistical analysis
The statistical analyses were performed using a statistical package (SPSS 16.0, Chicago, IL, USA). Fisher’s exact test was used to evaluate the criteria for diagnosis of malignant thyroid nodules in conventional B-scan US and RTE, and a p-value less than 0.05 was considered statistically significant. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy of conventional B-scan US and RTE for thyroid nodules were calculated comparing the findings with cytological and/or histological reports.
Results
Cytology/histology
FNA was performed in 84 patients. Hereby, cytology revealed a benign nodule in 66 cases, a follicular lesion in 13 cases, a suspicious nodule in 3 cases, a papillary carcinoma in 2 cases, and there was no nondiagnostic aspiration during the study period. All indeterminate nodules (follicular and suspicious lesion) and malignant nodules were proposed to surgery. In addition, twelve additional patients who preferred direct surgery were operated without receiving prior FNA. Histology showed a papillary carcinoma in 4 nodules of 4 patients, and benign adenoma and/or regressive changes including hemorrhage, follicular epithelial hyperplasia and calcification in other 26 patients.
Conventional B-scan US
The values of five B-scan US-features in the differential diagnosis of benign and malignant thyroid nodules were shown in Table 1. The following suspicious features were found in thyroid nodules: obscure border (17/96), irregular margin (64/96), internal hypoecho (83/96), inhomogeneous echo (78/96), microcalcifications (58/96). Nodular border obscurity was the ultrasound pattern most predictive of malignancy (accuracy 84.4%, sensitivity 75.0%, specificity 84.8%), and showed significant difference between benign and malignant nodules (p < 0.05). While none of other sonographic features were significantly different (p > 0.05), and they all had a high sensitivity, low accuracy and specificity relatively (As all the lesions in our study showed no blood flow signals, which was consistent with the absent wash-in enhancement, features of vascular distribution would not be discussed in this article).
RTE
Real-time elastography score ES1 and ES5 wasn’t found in nodules; ES 2 in 4 nodules (all benign nodules); ES 3 in 52 nodules (50 benign nodules and 2 papillary carcinomas); and ES 4 in 40 nodules (38 benign nodules and 2 papillary carcinomas), respectively (Table 1).
Hence, 4 of 92 nodules with the final diagnosis of benign nodules showed ES 2, and 4 of 4 nodules with the final diagnosis of thyroid cancer showed ES 3-4 resulting in a sensitivity of 100%, but a very low specificity of 4.3%, respectively.
CEUS
The 92 benign thyroid nodules and 4 papillary carcinomas showed the same enhancement pattern: absent wash-in inside the lesion.
Discussion
Preoperative diagnosis of benign and malignant thyroid nodules has been shown clinical importance to the management of the nodules, and it still remains challenging due to increasing incidence of malignancy. The usefulness of both conventional B-scan US and elastosonography have made it more accurate to predict malignancy of thyroid nodules in a non-invasive way.
As our study focused on the diagnosis of suspicious thyroid nodules (difficult to identify with conventional US), results indicated considerable nodules should be in a high risk of malignancy. While there were only 4 malignant nodules conversely. It seemed that both conventional US and RTE had a poor diagnostic performance in the differentiation of thyroid nodules with absent wash-in pattern (Fig. 1). Comprehensive understanding of forementioned enhancement pattern is needed to analyse the contradictory result. CEUS has been a successful validated US method in the diagnosis of liver lesions, while its application in thyroid is still under investigation. CEUS allows to perform continuous imaging at low acoustic power, providing a detailed depiction of tumour vascularity and microcirculation.
As is known, the increase in vascularity is related to cellular proliferation in neoplasms. Since tumor grows with the complex neovascularization inside; once the growth outweighs neovascularization, necrosis and embolus formation happens within the tumor, finally leading to the lack of blood supply. On the other hand, benign thyroid nodules are often thought to be rich in blood supply [15, 17]. As a matter of fact, previous investigation have found that ring enhancement correlated highly with a benign diagnosis of thyroid nodule (sensitivity 83.0%, specificity 94.1%, and accuracy 88.5%), whereas heterogeneous enhancement correlated with a malignant lesion (sensitivity 88.2%, specificity 92.5%, and accuracy 90.4%) [19]. In addition, homogeneous enhancement, isoenhancement or hyperenhancement were also considered as benign [9]. However, to the best of our knowledge, there were few studies about the clinical value of absent wash-in pattern. Malignant nodules showed absent (4 out of 13) in Bartolotta’s study, and it seemed that CEUS didn’t change their precontrast diagnostic confidence [1]. While a literature once reported a case misclassified as malignant by FNA, showing an absent enhancement, was confirmed as nodular goiter with hemorrhage and cystis degeneration by surgery [18], our investigation was consistent with that actually. Since all the benign nodules with absent enhancement came up with benign adenoma and/or regressive changes including hemorrhage, follicular epithelial hyperplasia and calcification.
More encouragingly, while reviewing the medical history of each included patient, we found considerable benign nodules (68/92) had been in dynamic change with size and morphology. We speculated that it was the absorption of cystic degeneration or hemorrhage inside the lesion that led to a final shrinkage appearance. And the unique shrinkage appearance in conventional B-scan US and RTE could be similar to malignant features in thyroid nodules such as hypoechoic, inhomogeneous, irregular margin, and also a higher ES score since fibrosis caused the stiffness increase (Fig. 2). Nevertheless, it should be also noted that clear border was the only significant benign feature due to the shrinkage appearance. Besides, fibrosis could also decrease the density of the vascularity inside the nodules, which might explain no infusion in the target lesion as absent enhancement [9]. However, no imaging method is always reliable since a considerable overlap of characters occur in both benign and malignant thyroid nodules. Four malignant cases with suspicious features were inexplicable with the explanation above. After carefully reviewing the stored data of corresponding CEUS, we found there seemed to be faint infusion in the peripheral area of lesions, but it was difficult to differentiate from artifacts since they were all smaller than 10 mm (Fig. 3). A previous study also reported that nodules less than 1 cm showed mainly absent vascularization, while those larger than 1 cm revealed faint dotted or diffuse contrast enhancement. Hence the author assumed subjective findings of enhancement pattern could very well be related to the size of the lesion irrespective of the histology [1]. We supposed that the neovascular bed might not be completely developed to supply the uncontrolled cell proliferation of tumor in small malignant nodules.
Our present study has several limitations that merit consideration. First, the sample was constituted for nodules suspected of malignancy, hence they all underwent three examination modalities. Second, since we took the confirmation from cytology in 66 of 96 patients without histology, certain degree of false positive or negative results in FNA might be inevitable. However, in a different sense, as the absent enhancement of CEUS indicated benign nature of suspicious thyroid nodules, it could help significantly reduce the amount of unnecessary FNA and surgery particularly in benign nodules with shrinkage appearance similar to malignancy both in conventional B-scan US and RTE.
Conclusion
In conclusion, combined with the patient’s medical history, our preliminary study allows to identify the specific role for “absent” enhancement in the differentiation of solid thyroid nodules with suspicious features in conventional B-scan US and RTE.
Footnotes
Acknowledgments
This work was partially supported by Science Technology Department of Shanghai Grant N11ZR1427700.