Introduction

Fine-needle aspiration cytology (FNAC) holds a main role in assessing thyroid nodules [1–4]. The large majority of thyroid lesions undergone FNAC has benign (70–80 %), suspicious for malignancy or malignant (5–10 %) diagnosis, being the negative predictive value very low (i.e., 1–2 % of benignancies). On the other hand, a not negligible rate of thyroid cytologies is read as uncertain and indeterminate for diagnosis (10–20 %) or as inadequate sample (10–15 %) [1–3]. Even if the rate of malignancy in the latter cytologic categories (Thy 3/Class 3/Category III–IV and Thy 1/Class 1/Category I according to the main International guidelines [1–3]) is low, to exclude a cancer is one of the major challenges in thyroidology. In the last decade, several single-center original papers described the use of core needle biopsy (CNB) as a second-line approach to diagnose those thyroid nodules with Thy 3/Class 3/Category III–IV or Thy 1/Class 1/Category I at FNAC. Data from these papers are interesting, and more recently, this biopsy option was included in AACE/AME/ETA guidelines [1] that should encourage a more large diffusion of CNB. Here, we reviewed the literature on this topic and outlined several aspects:

  1. 1.

    The “large” needle biopsy in 90’s.

  2. 2.

    Complications by and patient’s comfort with thyroid CNB.

  3. 3.

    Advantages provided by examination of a microhistologic sample of thyroid nodule.

  4. 4.

    Use of CNB in thyroid nodules with previous not adequate (Thy 1/Class 1/Category I) cytology.

  5. 5.

    Use of CNB in thyroid neoplasms (Thy 3/Class 3/Category III–IV) cytology.

  6. 6.

    Use of CNB in specific ultrasonographic presentations of thyroid nodules or in patients with peculiar clinical contexts.

  7. 7.

    First-line approach by CNB in thyroid nodules.

  8. 8.

    Immunohistochemistry and molecular tests on CNB samples.

  9. 9.

    Future perspective.

Materials and methods

Search strategy, study selection and data extraction

We searched studies evaluating the use of CNB in thyroid nodules. A comprehensive computer literature search of the PubMed/MEDLINE, Embase, and Scopus databases was conducted to find published articles on this topic. The search algorithm was based on the combinations of the terms (“thyroid” AND “core” AND “needle” AND “biopsy”). Initially, we did not use a beginning date limit, the search was updated until May 31, 2014, and no language restriction was used. To identify additional studies and expand our search, references of the retrieved articles were also screened.

Original articles that investigated CNB in thyroid nodules were eligible for inclusion. The two authors of the present study independently reviewed the retrieved articles. The main exclusion criterion was article that did not specifically report CNB in thyroid lesions, and review papers were also excluded. For each included study, information was abstracted concerning study data (authors, year of publication, and country).

The comprehensive computer literature search revealed 125 articles. Titles and abstracts of these articles were reviewed, and 31 original papers were selected by retrieving their full-text version and screening their references list [5–35] (Table 1). All these papers was published after 2003, but one [11] which was excluded from the review because it is too old. Almost all of these papers were retrospective, and the majority was from USA, Korea, or Italy. Subtle differences were recorded regarding the type of the needle used for CNB (Ace-cut, Menghini, Quick-core, Temno evolution, or other); the gauge of these needles ranged from 18 to 22. Overall, low rate of complications from CNB was described.

Table 1 Overview of the thirty studies included in the present review
Full size table

The large needle biopsy in 90’s

A large needle biopsy (LNB) was first proposed in thyroid nodule in the 90’s. The main goal was the possibility to evaluate the nodular architecture and the potential application of ancillary techniques. Despite the advantages in result obtained from the thick core examination, LNB has not become a standard procedure because of its relatively high risk of complications, especially hemorrhages, and its consequent technical difficulties [36, 37]. The recently described CNB has roots in the LNB of the 90’s but has been diffused due to the more accurate sampling and rare complications recorded (see below). In fact, in the last decade more refined cutting systems with smaller gauge have been commercialized (such as ace-cut, temno, and quick-core). These devices, with automatic or semiautomatic operation, are carefully inserted into the lesion under ultrasound guide, and the needle can be fired with high precision to obtain a nodule’s core specimen.

Complications by and patient’s comfort with thyroid CNB

In general, thyroid CNB has been reported as a safe procedure, and the complications rate and comfort degree using CNB was not significantly higher than that recorded by FNAC [5–10, 12–35]. Specifically, minor complications were recorded by thyroid CNB; transient hoarseness, hematoma, parenchymal edema, bleeding, and bruising were rarely (i.e., 1–3 % of cases) present after these biopsies [5–10, 12–35]. Very recently, tolerability of CNB was investigated by one research group [33]. There, a questionnaire was administrated to patients undergone both CNB and FNAC, and patient’s comfort and biopsy tolerability were not significantly different. Another study reported similar results [34].

Advantages provided by examination of a microhistologic sample of thyroid nodule

Core biopsy represents an optimal complementary test for thyroid nodules with fibrous component that are difficult to be aspirated by FNAC. In these cases, CNB allows obtain adequate material from nodules [5–10, 12–35]. At microscopic level, the nodule’s core sample offers the possibility to evaluate the general architecture of the lesion, the alteration of follicular structures, and the relationship with adjacent tissues [9, 20, 21]. The latter tissue–tissue examination is highly important in all histologic assessments. Moreover, CNB is advantageous in the analysis of nuclear changes, being the tissue formalin fixed and paraffin embedded similarly to the final post-surgical histology [5–10].

Use of CNB in thyroid nodules with previous not adequate (Thy 1/Class 1/Category I) cytology

The inadequacy of sample from thyroid nodule FNAC occurs in about 10–15 % of cases. In this context, the international guidelines suggest to repeat FNAC [1–4], and two inadequate FNAC should prompt to diagnostic surgery [2]. Remarkable, the likelihood of malignancy in a thyroid nodule with Thy 1/Class 1/Category I is low [1–4], and other techniques or markers able to diagnose these lesions are strongly required to avoid surgery.

Very interesting papers have reported a main role of CNB as a second-line biopsy in these lesions with Thy 1/Class 1/Category I [12–16]. The more relevant contribution was published by Samir et al. [12]. There, the authors enrolled a series of 90 nodules with prior one or two inadequate FNAC, and all nodules underwent both CNB and repeat FNAC. The combination of repeat-FNAC and CNB classified as diagnostic 87 % of thyroid lesions with one prior inadequate FNAC and 86 % of those with two prior inadequate FNAC. CNB has been also used by Zhang et al. [13] in a large series of nodules with inadequate or suboptimal cytologic outcome. Subjects underwent FNAC or both FNAC–CNB, and the latter approach could reduce the sample’s inadequacy rate from 8.7 % obtained by FNAC alone to 3.4 %. Moreover, Na et al. [16] compared CNB and a second FNAC in 64 nodules previously reading of not diagnostic; a 28 % of inadequate specimens were found in FNAC, while 1.6 % was recorded by CNB.

In all, whether to repeat FNAC in nodules with previous not diagnostic cytologic report is a matter to debate [12–16]. The combined use of repeat FNAC and CNB should be strongly considered as an alternative to surgery in individuals with previous non-diagnostic cytology. This approach has been recently included in the AACE/AME/ETA guideline [1] and Italian consensus for thyroid cytology [38].

Use of CNB in thyroid neoplasms (Thy 3/Class 3/Category III–IV) cytology

Up to 20 % of thyroid nodules submitted to FNAC has an indeterminate diagnosis of follicular neoplasm [1–4]. These lesions represent one main hot topic in thyroidology [39]. In fact, only one in four of nodules with Thy 3/Class 3/Category III–IV cytology is a cancer [40, 41], but this possibility cannot be excluded before final histology. Then, surgery has been traditionally required to diagnose these lesions [1–4]. Because of the large majority of nodules with indeterminate FNAC is benign, thyroidectomy/lobectomy is often a posteriori not necessary, and several studies have attempted to analyze clinical, morphologic, instrumental, or immunocytochemical parameters to be used in the risk stratification of thyroid follicular neoplasms [41–45]; as a result, male gender, large nodule’s size, irregular margins, and microcalcifications at ultrasonography, MIBI uptake, cytologic atypias, and a positive Galectin-3 were reported as predictors of cancer, but they lack to be proved as reliable tests to avoid diagnostic surgery with 100 % specificity [41–45]. Thus, current guidelines do not recommend for the use of the above mentioned tools in the workup of patients with inconclusive thyroid cytology [1–4].

Several single-center papers described CNB in thyroid lesions with previous indeterminate FNAC, with remarkable findings [16–22]. Park et al. [20] addressed a series of patients with indeterminate neoplasms to CNB, or second FNAC, or diagnostic surgery; CNB was more effective with 77.8 % benign, 20.3 % cancer, and 1.8 % uncertain samples. Na et al. [16] evaluated 161 nodules with Category III–IV according to Bethesda system [3], and CNB correctly assessed 66.7 % of benign and 67.7 % of malignant tumors, with low rate of not diagnostic or indeterminate reports. These results were corroborated by other studies [17–19, 21, 22]. Finally, a paper by the authors of the present review described a new CNB approach which allows detect nodule’s capsule while present [21]; this sampling technique can distinguish among follicular lesions the adenomatous non-encapsulated nodules from truly neoplastic encapsulated lesions.

All in all, the above studies reported high accuracy of CNB to detect benign nodules with indeterminate cytology which can avoid unnecessary diagnostic surgery. Also, the rate of not adequate CNB sample is low [16–22]. At present, CNB is not recommended by guidelines in nodules with Thy 3/Class 3/Category III–IV cytology [1–4]. In fact, CNB cannot distinguish a follicular adenoma from a follicular carcinoma which should represent the majority of thyroid lesions with indeterminate cytologic outcome. The assessment of these neoplasias requires the evaluation of the whole nodule’s capsule at final postsurgical histology [1, 21].

Use of CNB in specific ultrasonographic presentations of thyroid nodules or in patients with peculiar clinical contexts

Few papers evaluated the diagnostic utility of CNB with ultrasound risk presentation [23–26]. As the first, Ha et al. [23] used CNB in thyroid nodules with previous benign FNAC outcome but having high risk at ultrasound examination. Interestingly, a very high cancer rate (32 %) was found. This finding was perfectly in agreement with that previous reported by papers analyzing the likelihood for malignancy in these specific nodules [46–48]. More recently, CNB or conventional FNAC was proposed to patients with newly discovered at risk thyroid lesions [24]; the diagnostic accuracy of CNB was significantly higher than that of FNAC [24]. These data should prompt to use CNB in those nodules suspected to be a cancer to avoid false negative FNAC reports, but more studies, ideally with prospective design and multicenter setting, are needed.

The approach by CNB in calcified thyroid nodules was reported by Ha et al. [25]. There, in a series of 272 nodules with micro- or macrocalcifications undergone CNB, only 2 (0.7 %) cases had inadequate sample, other 37 (13.6 %) were indeterminate, and there were 3 cancers with false negative microhistologic report [25]. Further data on CNB in calcified thyroid lesions are necessary.

Thyroid nodules occur with low frequency in pediatric population, and malignancy rate in these subjects ranges from 16 to 43 % [49]. A single institute experience of CNB in children was reported by Yunker et al. [27]; 36 % of nodules were malignant, and CNB had 13 % of inadequacy with sensitivity and specificity for cancer of 85 and 63 %, respectively. Of interest, no complications were recorded. The authors advocated this procedure to avoid diagnostic thyroidectomy in pediatric population. Because no significant difference was found between CNB and FNAC, the authors reject their initial hypothesis.

Finally, some studies indicate that microhistologic samples from CNB are more reliable than cytologic ones from FNAC in detecting thyroid lymphoma [28, 29]. Data on this topic are sparse. In this context, it has to be cited that current guidelines suggest CNB in patients with suspicious anaplastic thyroid carcinoma, thyroid lymphoma, pathologic lymph nodes, or other malignant neck disease [1].

First-line approach by CNB in thyroid nodules

Only few studies on the specific use of CNB as first-line approach to thyroid nodules exist. Sung et al. [10] submitted 555 thyroid nodules to both CNB and FNAC; the main result was the improved accuracy obtained by combining the two procedures. Also, Trimboli et al. [24] enrolled a selected series of 72 patients with nodule at ultrasound risk for malignancy, and higher accuracy of CNB with respect to FNAC was found. At present, the first-line use of CNB remains a matter to be evaluated.

Immunohistochemistry and molecular tests on CNB samples

The CNB sampling allows obtain a tissue fragment with size up to 500 µ and length up to 1.5 cm. These specimens should be the optimal material for extensive studies and ancillary techniques [9]. In fact, on one hand the microhistologic examination detects nuclear changes, architectural alterations of follicular structures, and pathologic relations between adjacent tissues; on the other hand, the paraffin core sections permit automated immunostaining with high reproducibility and very low cost [9, 30].

Of relevance, a very recent paper by the authors of this review reported 100 % concordance between immunohistochemical examination for the BRAF(V600E) antibody VE1 on paraffin section, and the mutational analysis performed by pyrosequencing on DNA obtained by extraction from the same core samples [30]. These results look to have very high interest and advises for future studies. In fact, lower accuracy of VE1 has been recorded in cell block and thin layer cytologic preparation [50, 51]. Few papers evaluated Galectin-3 expression in core samples from cytologically indeterminate thyroid neoplasms. The accuracy of the integrated morphologic–phenotypic diagnostic approach was high [21, 22], but data on this topic are very sparse to extend its use in routine workup of thyroid nodule. Very interesting findings were also described by Yousaf et al. [32] who applied thyroperoxidase (TPO) immunostaining in CNB samples to discriminate malignant and benign thyroid lesions with cold appearance at scintiscan.

All in all, the feasibility of immunohistochemistry is maintained with poor cellular or fibrous samples, as demonstrated in small cell aggregates from BRAF-mutated cases with inconclusive genetic results [52]. Although few studies reported immunohistochemistry applied to core specimens, CNB samples should be used for these extensive research.

Future perspective

In general, future prospective multicenter studies are required to better set the use of CNB in thyroid nodule. In fact, this study setting is lacking in thyroid CNB literature. Immunophenotyping thyroid nodules represent one major challenge in thyroidology. Different cytologic preparations (i.e., cell block, thin layer) failed to reach high reliability as samples for immunohistochemistry of thyroid cancer potential markers (i.e., Galectin-3), and their routinely use has not be recommended by current guidelines due to several reasons. On the contrary, microhistologic samples should be the ideal specimens to perform these ancillary studies. Future researches are needed to confirm that.

The cost of thyroid CNB is not negligible; not published data from the authors of the present review report that CNB may achieve 1,000 euros per patient, which is higher than the cost to perform a conventional FNAC (about 150 euros). The latter issue could represent a potential limit for the worldwide diffusion of thyroid CNB.

Conclusions

CNB has been reported as more accurate than FNAC in diagnosis of thyroid nodules. In particular, the second-line use of CNB can assess as diagnostic those thyroid lesions with previous inadequate or indeterminate cytology. At present, international guidelines include CNB only for thyroid nodules with prior not adequate FNAC. Information of first-line use of CNB is poor, and this approach should be considered only in specific conditions where conventional FNAC might be not accurate. The microhistologic samples from thyroid CNB should allow to obtain tissue fragments ideal for immunochemistry. In next future, the immunophenotype analysis of thyroid nodules should be improved by this approach.