Should Liquid Based Cytology (LBC) be Applied to Thyroid Fine Needle Aspiration Cytology Samples?: Comparative Analysis of Conventional and LBC Smears.

CONTEXT: Liquid-based cytology (LBC) is a cyto-preparatory technique that may be applied to fine-needle aspiration (FNA) samples. However, its efficacy over conventional smears (CS) in thyroid is controversial. AIMS: The aim of this study was to compare CS versus LBC (SurePath, BD) preparations in thyroid FNA samples for diagnostic efficacy. SETTINGS AND
DESIGN: Prospective case-control study using split sample analysis in 200 non-consecutive cases of thyroid FNA reported as per the Bethesda system. METHODS AND MATERIAL: Detailed cytomorphological features were evaluated in CS and LBC preparations. Cellularity was scored as 0-3. Based on diagnostic efficacy, they were categorized into three groups:(i) CS and LBC equivalent for diagnosis, (ii) CS better than LBC, and (iii) LBC better than CS for providing the diagnosis. STATISTICAL ANALYSIS USED: Paired t test for cellularity scores and descriptive for diagnostic efficacy.
RESULTS: There were 7 unsatisfactory, 118 benign, 10 atypia or follicular lesion of undetermined significance, 25 follicular neoplasms, 6 suspicious for malignancy, and 34 malignant thyroid FNA cases based on routine conventional smears. Cellularity of conventional smears was significantly higher than paired LBC smears (paired t test, P < 0.005). Comparison of overall diagnostic efficacy showed that LBC and CS were equivalent in 59% of cases; CS was superior to LBC in 37% cases and LBC was better than CS in 4% cases only. LBC smears showed higher unsatisfactory rate as compared to CS (18% vs. 3.5%). In category 2, the two techniques were equivalent in 87% of cases. In categories 4 and 6, CS were superior to LBC.
CONCLUSIONS: In thyroid FNA cases, conventional smears are superior to LBC preparation in terms of cellularity and diagnostic efficacy and hence, LBC preparations should not replace conventional smears for making a routine cytomorphological diagnosis. Copyright:

INTRODUCTION

Fine-needle aspiration cytology (FNAC) is regarded as the best investigation for the preoperative evaluation and diagnoses of thyroid lesions because of its simplicity, safety, and cost-effectiveness.[1] Liquid-based cytology (LBC), introduced at the beginning of this century, revolutionized gynecological cytology. During the last several years, there has been a surge in liquid-based cytology slide preparation in non-gynecologic cytology specimens including thyroid FNAC.[2345] Essentially, these techniques use methanol or ethanol fixation followed by ultrafiltration or centrifugation steps which eliminates blood and extraneous material and produces a thin cellular monolayer spot. The two most common FDA-approved methods for processing the cytological samples are Thin- Prep2000TM (Thin Prep Hologic Co., Marlborough, Massachusetts), and Surepath (SurePathTM; TriPath Imaging, Burlington, North Carolina). The final result for both methods is 1 slide for each lesion where all cells are concentrated in a thin layer on the central area of the slide measuring 20 mm2 for ThinPrep and 13 mm2 for SurePath.[67] There are only a few published reports comparing the two preparation methods for the interpretation of thyroid lesions by FNAB, and these are mostly from the developed countries in the West.[8910] Although morphologic differences with the conventional smear were noted in these studies, the advantages of the LBC method included clear background, smaller area on the slide, lower number of slides and less time needed for evaluation. In addition, ancillary studies like immunohistochemistry were possible to perform on unstained LBC slides.[1112] There is no such study from India to the best of our knowledge. We hypothesized that LBC slide preparation was superior to conventional smears in thyroid FNAC samples. Hence, the aim of this study was to compare the two methods for cytomorphological differences and to determine their diagnostic efficacy.

PATIENTS AND METHODS

The study was approved by the Institutional Ethics committee. All subjects were recruited after informed consent. All cases of thyroid swelling were referred to the Cytopathology Department. All subjects underwent palpation-guided FNAC by an experienced operator who was a either a trained cytopathologist or a senior resident with experience in cytopathology, using 23- or 22-gauge needles attached to a 20-mL syringe in a holder (Cameco, Taby, AB, Sweden). A minimum of two passes were taken from each lesion. The material from the first pass was extruded onto the slides for preparation of 2–3 conventional smears and rest submitted for LBC. The material from the second pass was initially taken for LBC and from the residual material in the hub of the needle and the syringe, 2–4 conventional smear were made. Randomly, in half the cases, the reverse order was followed to minimize any bias. The material for LBC was submerged in CytorichTM solution for LBC [Becton Dickinson, India]. The air-dried and alcohol-fixed conventional smears (CS) were stained by May-Grunwald Giemsa and Hematoxylin-Eosin stains, respectively. A single LBC slide was prepared in each case and stained by the Papanicolaou stain using the BD [Becton Dickinson] SurePath TM system. Finally, the conventional smears and LBC preparation was evaluated by two independent cytopathologists. The following cytomorphologic features were evaluated, scored, and recorded in each case for both the LBC and CS slides upon microscopic examination: (1) cellularity; range from 0 to 3: 0––no cells, 1––mildly cellular, 2––moderately cellular and 3––highly cellular; the cellularity was scored on the most cellular slide in case of CS as 2–4 smears were made in each case. (2) amount and quality of colloid; (3) background blood; (4) cyst fluid and macrophages (5) inflammatory cells present or not and if so, their numbers and types. Finally, for each LBC and CS preparation, a remark of whether it was diagnostic of the entity or not was made.

Statistical methods

The cellularity scores for the paired CS and LBC smears were evaluated by the paired t test. For comparison of diagnostic efficacy, descriptive statistics was used.

RESULTS

A total of 200 cases of thyroid FNA with conventional smears and corresponding LBC smears were evaluated. Their ages ranged from 21 to 72 years. There were 170 women and 30 men. The Bethesda system of reporting thyroid cytopathology was used and the break-up of cases is shown in Table 1. There was a 3.5% (7/200) unsatisfactory rate among conventional smears. Benign lesions, which included colloid goiters and lymphocytic thyroiditis, comprised category 2, which constituted 59% (118/200) cases. There were 10 cases in the AUS/FLUS category and 25 in category 4 with 17 Follicular neoplasms and 8 Hürthle cell neoplasms. Overall there were 40 cases of malignant thyroid lesions, 6 in Category 5 (suspicious for malignancy), and 34 in category 6, malignant thyroid lesions, among which papillary thyroid carcinoma (PTC) was the most common malignancy.

Table 1

Thyroid FNA cases included in the study (n=200)

TBSRTC Category	No. of cases
Unsatisfactory	7
Benign	118
Colloid Goitre/cyst	87
Lymphocytic thyroiditis	31
AUS/FLUS	10
FN/SFN	25
Follicular Neoplasm	17
Hurthle Cell Neoplasm	8
Suspicious for malignancy	6
Malignancy	34
Papillary Thyroid Carcinoma	25
Medullary Carcinoma	03
Anaplastic carcinoma	03
Poorly differentiated carcinoma	01
Lymphoma	01
Langerhans Cell Histiocytosis	01

Comparison of morphology of conventional and LBC smears

Overall, the cellularity of conventional smears was better than LBC preparations. Out of 200 cases, the cellularity was equivalent in 56% (113) cases; conventional smears showed better cellularity in 37% (73) cases whereas in 7% (14) cases, the LBC preparation showed better cellularity. Out of the 113 cases with similar cellularity in both preparations, 4 cases showed no cells in either conventional or LBC smears. Both preparations showed mild (1+) cellularity in 23 cases, moderate (2+) cellularity in 70 cases and were highly cellular (3+) in 16 cases. Among the 73 cases where conventional smears showed better cellularity than LBC preparations, there were 17 cases with hardly any cells in LBC, but were adequate as per the Bethesda criteria in the conventional smears. In 10 of these cases, the conventional smears showed 2+ to 3+ cellularity. Conventional smears showed significantly higher cellularity as compared to paired LBC smears (paired t test, P < 0.005).

Although not all conventional smears made had uniform cellularity, we compared the most cellular slide with LBC. Thyroid follicular cells had a comparable morphology. Colloid was also better appreciated in conventional smears. In LBC preparations, colloid was appreciated to a far lesser degree; it was fragmented and seen only if there was thick colloid [1a and 1b]. When the aspirate was thin colloid or fluidy, the LBC smears showed mainly macrophages whereas the conventional smears showed the same cells in a thin fluidy colloid background. In lymphocytic thyroiditis, conventional smears were superior to LBC. However, scattered lymphocytes, fibro-inflammatory stromal fragments indicative of follicular destruction and Hürthle cells were all appreciated in LBC preparations [1c and 1d]. Among category 4, follicular neoplasms, the cellular elements were appreciated equally well in LBC smears and conventional smears. However, microfollicle formation was far more abundant in conventional smears and was less evident in LBC preparations [Figure 2a and 2b]. Hürthle cells were well appreciated in LBC preparations and the morphology was similar to conventional preparations [Figure 2c and 2d].

Figure 1

Paired Conventional smears (CS) versus Liquid based Cytology (LBC) in representative case of colloid goiter (a and b) and lymphocytic thyroiditis (c and d). [a and c, May-Grünwald Giemsa stain; b and d, Papanicolaou stain]

Figure 2

Paired Conventional smears (CS) versus Liquid based Cytology (LBC) in representative case of follicular neoplasm (a and b) showing fewer microfollicles in LBC and comparable morphology in Hurtle cell neoplasm (c and d). [a and c, May-Grünwald Giemsa stain; b and d, Papanicolaou stain]

Papillary carcinoma was the most frequent malignant lesion encountered. The LBC smears showed only some of the features of PTC that were well appreciated in conventional smears [Figure 3a–d]. These included papillary fragments, occasional giant cells and monolayered aggregates of cells. The most striking feature observed in LBC smears was the pale nuclear chromatin [3c and 3d] that leads to optically clear nuclei in the neoplastic follicular cells. Nuclear grooves were also less frequent. On the contrary, intranuclear inclusions were only occasionally observed in LBC smears and were more frequent in conventional smears [Figure 3a and 3b]. There were 2 cases of anaplastic carcinoma, which showed all features appreciated in conventional smears and in the LBC preparation including bizarre tumor cells in a necro-inflammatory background [Figure 3e and 3f]. One case of medullary carcinoma also showed similar features in CS and LBC.

Figure 3

Paired Conventional smears (CS) versus Liquid based Cytology (LBC) in representative case of Papillary thyroid carcinoma (a-d), showing intranuclear inclusions (a and b), and pale nuclear chromatin (c and d); anaplastic thyroid carcinoma (e and f) showing scattered tumour cells in a necroinflammatory background. [a,c and e, May-Grünwald Giemsa stain; b, d and f, Papanicolaou stain]

Diagnostic efficacy of conventional smears vs. LBC smears

In each case, the LBC preparation and the corresponding conventional smear was assessed independently by 2 pathologists as to whether a diagnosis could be effectively rendered by these two preparations. Comparison of the two techniques in each Bethesda category is shown in Table 2. Conventional smears had a 3.5% unsatisfactory rate as compared to 18% in LBC preparations. In category 2, there were 5 colloid goiters, 2 lymphocytic thyroiditis and 1 hurthle cell neoplasm where LBC was more cellular and diagnostic (4%) as compared to CS; however, in the majority (59%) of cases, the 2 were equivalent and CS was superior to LBC in 37% cases. Further, out of 31 cases of lymphocytic thyroiditis, CS was superior in 13 (42%) cases. Lymphocytic tangles were seen occasionally in LBC smears whereas they were frequent and of diagnostic value in CS. On the contrary, plasma cells and giant cells were more readily appreciated in LBC preparations. In category 3, LBC was either as good or CS was better for evaluation. In category 4, CS was superior to LBC preparations with an exception of 1 case of Hurthle cell neoplasm where LBC was more cellular than conventional smear. In category 5, CS was superior in majority of cases. In Category 6 comprising the entire spectrum of malignant lesions with PTC being the most common lesion, LBC was not superior to CS in any case and was either as good as CS or, CS was better than LBC in an equal number of cases. Thus, out of 200 cases, overall, the diagnostic efficacy of LBC and CS were equal in 119 or in 59% cases, CS was superior to LBC in 73 (37%) cases and LBC was superior to CS in 8 (4%) cases.

Table 2

Comparison of paired conventional and Liquid based Cytology smears in 200 cases of thyroid fine needle aspirations

Bethesda Smear Category	Liquid Based Cytology Smear Diagnosis
Conventional Smear Diagnosis n (%) ↓	CAT 1	CAT 2	CAT3	CAT4	CAT5	CAT6
CAT 1 7 (3.5%)	7	-	-	-	-	-
CAT 2 118 (59%)	15	103	-	-	-	-
CAT 3 10 (5%)	1	-	8	-	1	-
CAT 4 25 (12.5%)	5	-	3	17	-	-
CAT 5 6 (3%)	-	-	-	-	6	-
CAT 6 34 (17%)	8	-	-	-	1	25
Total 200 (100%)	36 (18%)	103 (51.5%)	11 (5.5%)	17 (8.5%)	8 (4%)	25 (12.5%)

DISCUSSION

LBC technique revolutionized gynecological cytology or the cervical pap smears and has almost replaced conventional smears in the West. It has also been applied to non-gynecological specimen including thyroid FNA.[2345] However, there is no study from India regarding this. Furthermore, in this study, we compared liquid-based cytology (SurePath) smears of thyroid FNA with corresponding conventional made smears for cytomorphological features and diagnostic efficacy. Split sample approach was used and hence there was no or minimal sampling bias. The cases were also distributed across all the Bethesda categories. As per the Bethesda system, the adequacy criteria are met when at least six clusters of 10–20 well-preserved cells are observed in the smears. The overall inadequacy rate for LBC was 18% as compared to 3.5% inadequacy in conventional smears. The LBC inadequacy rate of 18% is lower than the 25% inadequacy rate reported previously with the SurePath system.[13]

Fadda and Rossi have reviewed the literature pertaining to the application of LBC to thyroid FNA and have detailed the cytomorphological differences between the CS and LBC ThinPrep preparations.[614] Afify et al.[15] performed a retrospective review and have also highlighted several differences between the two types of preparation. Our observations are somewhat similar in this respect. In the initial part of the study, we were on a learning curve to familiarize ourselves with the LBC morphology. In colloid goiters, thin colloid is not appreciated and thick colloid gets fragmented. In lymphocytic thyroiditis, with experience, one is able to observe most of the usual features although the numbers of inflammatory cells are lower in LBC preparations. In follicular and Hürthle cell neoplasms, there were lesser microfollicles and more naked nuclei. In PTC, nuclear overlapping and optically clear nuclei were striking in LBC preparations as compared to CS. Intranuclear inclusions were only occasionally observed in LBC smears as reported by Afify et al.[15] although others have frequently reported this to be present.[614]

Overall, conventional smears were definitely better in their diagnostic efficacy in any Bethesda category. LBC was only as good as CS in 58% cases and in only 5% cases was it better than CS. Recent reports from the College of American Pathologists survey by Fischer et al.[16] and another study by Nagarajan et al.[17] evaluating a large number of CS and LBC preparations have observed that LBC was associated with a significantly higher proportion of inadequate and a lower proportion of benign diagnosis. Cavaliere et al.[18] had a similar study design as our study and observed higher sensitivity in CS as compared to LBC smears (93.6% vs. 65.9%). The CAP survey,[16] as well as a study by Rossi et al.[19] have both concluded that a greater proportion of PTCs were missed in LBC. Tripathy et al.[20] used LiquiprepTM (Cytec corporation) in 18 cases of thyroid FNA and observed that LBCs were not very useful in goiter and infectious lesions, but they proved very useful in neoplastic lesions. In fact, we also observed great difficulty in making a primary diagnosis based solely on one LBC smear as compared to 3-6 well stained conventional preparations. Assessment of the background material and cellularity and microfollicles are so crucial to making a proper cytodiagnosis and are almost always better appreciated in conventional smears. LBC only adds to the cost with no added benefit. Hence, LBC should be viewed as a complementary technique to conventional direct smears as far as making a basic cytomorphological diagnosis is considered. It is pertinent to mention here that all aspirations in our setup are performed by trained operators (Cytopathologists/senior residents in cytopathology) and hence this resulted in superior conventional smears. It is possible that in institutions with alternate practices such as a physician performed FNA who may or may not have adequate experience in making smears, may affect the outcome.

However, there are some advantages offered by LBC as highlighted by Fadda and Rossi et al.[714] The residual cells in the LBC vial allow for the application of adjunct immunocytochemistry and molecular techniques, which are today essential to improve the sensitivity and diagnostic efficacy of thyroid fine-needle aspiration cytology.[1415] In our opinion, material is as easily collected for these ancillary techniques directly in appropriate fixatives rather than use residual LBC material.

Overall, conventional smears were definitely better in their diagnostic efficacy in any Bethesda category and are the preparation of choice in thyroid FNA. Hence, LBC should be viewed as a complementary technique if ancillary techniques like immunocytochemistry or molecular testing are required; conventional smears are the choice of preparation in thyroid fine-needle aspiration cytology as far as making a basic cytomorphological diagnosis is considered.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.