Thyroid Bethesda Atypia of Undetermined Significance or Follicular Lesion of Undetermined Significance (AUS/FLUS): A Heterogenous Group.

BACKGROUND: The Bethesda system of reporting thyroid cytopathology (BSRTC) was introduced in 2007. The third category of atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS) poses difficulties for the pathologist, and different papers have been published varying the risks of malignancy. AIMS: (1) Evaluation of the cytological features of thyroid lesions according to BSRTC. (2) After resection, correlation with histopathological report to evaluate the risk of malignancy (ROM) and the risk of neoplasm (RON). (3) Division of category III into six subgroups based on cytological findings and assessment of ROM and RON.
MATERIALS AND METHODS: A total of 282 patients with diagnosed thyroid lesions underwent fine-needle sampling under ultrasound guidance. Smears were prepared and stained with May-Grunwald-Giemsa stain and Papanicolaou stain.
RESULTS: Of 282 cases, there were 9 cases (3.1%) of category I, 157 cases (55.8%) of category II, 24 cases (8.5%) of category III, 20 cases (7.1%) of category IV, 14 cases (4.8%) of category V, and 58 cases (20.7%) of category VI. The RON was 60, 17.1, 63.1, 77.7, 91.7, and 98.2% and the ROM was 60, 14.3, 26.3, 38.9, 91.7, and 96.3% in categories I, II, III, IV, V, and VI, respectively. The RON was 0, 75, 50, 100, 66.6, and 100% and the ROM was 0, 25, 50, 100, 16.6, and 0% in subgroups 1, 2, 3, 4, 5, and 6, respectively. We have proposed a system of subgrouping AUS/FLUS that may help to dispel the confusion generated by an AUS/FLUS report, and provide with a more exact and reproducible diagnostic and prognostic tool. Copyright:

INTRODUCTION

The Bethesda system of reporting thyroid cytopathology (BSRTC) was introduced in 2007.[1] It has six categories: category I – nondiagnostic/unsatisfactory (ND/US), category II – benign (B), category III – atypia of undetermined significance/follicular lesions of undetermined significance (AUS/FLUS), category IV – follicular neoplasm/suspicious of follicular neoplasm (FN/SFN), category V – suspicious of malignancy (SM), and category VI – malignancy (M). Each category has an implied risk of malignancy (ROM) and a usual scheme of management, though the actual management may depend upon factors other than fine-needle sampling (FNS) results. The purpose of BSRTC is to delineate patients who require surgical excision of thyroid lesions from patients who can be managed conservatively. There are several studies describing the usefulness of BSRTC.[23456]

Out of the six categories of BSRTC, the third category of AUS/FLUS poses difficulties for the pathologist.[78910111213] Different papers have been published showing varying risks of malignancy for category III lesions.[7] The AUS/FLUS category is a waste-basket of compromised morphology, either characterized by sparse cellularity with nuclear atypia only or in combination with microfollicles or Hürthle cell changes. This category is usually overreported above the 7% threshold level.

In our study we evaluated the cytological features of thyroid lesions and reported according to BSRTC. Later, after resection, we correlated the cytological report with histopathological report and evaluated the ROM and the risk of neoplasm (RON).

We divided category III into six subgroups based on cytological findings and assessed the ROM and RON in each subgroup.

MATERIALS AND METHODS

A prospective study was conducted from January 2017 to June 2018 in collaboration with the Departments of Otolaryngology and Endocrinology at our institute. Patients clinicoradiologically diagnosed with thyroid lesions were enrolled in the study. Detailed history was obtained and informed consent was taken. A total of 282 patients were enrolled in the study.

FNS was performed under ultrasound guidance. The procedure was done with the patient in supine position and a pillow under the upper back. 25G needles were used and aspiration was avoided unless the lesion was cystic. In case of multiple nodules, more than one aspiration was done from prominent or ultrasonographically suspicious (heterogeneous echo texture/high vascularity on Doppler) nodules. In cystic nodules, the cyst contents were aspirated, centrifuged and smears were made from the sediment for microscopic examination. Smears were prepared on four glass slides for each case. Two slides were air dried and stained with May–Grunwald–Giemsa (MGG) stain. Two slides were fixed in 95% ethyl alcohol and stained with Papanicolaou (Pap) stain.

The slides were examined and reported according to BSRTC. Category III of AUS/FLUS was further subdivided into six subgroups: 1 – cases with nuclear atypia, designated NA; 2 – cases with nuclear atypia and microfollicles, designated NA, MF; 3 – cases with nuclear atypia and lymphocytes, designated NA, L; 4 – cases with nuclear atypia and cyst macrophages, designated NA, CM; 5 – cases with microfollicles, designated MF; and 6 – cases with Hürthle cells, designated HC.

The patients were followed up, and after resection of their lesions histopathological examination was done. The cytological diagnosis was correlated with the histopathological diagnosis to estimate the ROM and RON of the total resected cases, the ROM and RON specific to each category of BSRTC (resected cases only), and in each subgroup of AUS/FLUS (resected cases only).

RESULTS

Of a total of 282 cases, there were 9 cases (3.1%) of ND/US, 157 cases (55.8%) of B, 24 cases (8.5%) of AUS/FLUS, 20 cases (7.1%) of FN/SFN, 14 cases (4.8%) of SM, and 58 cases (20.7%) of M. Out of 282 cases, follow-up thyroidectomies were done in 144 cases of which 44 (30.5%) were nonneoplastic lesions and 100 were neoplastic lesions, including both benign neoplasms (16 cases: 11.1%) and malignant neoplasms (84 cases: 58.3%). The overall RON was 69.4% and ROM was 58.3% in resected lesions. The RON for each category was ND/US 60%, B 17.1%, AUS/FLUS 63.1%, FN/SFN 77.7%, SM 91.7%, and M 98.2%. The ROM for each category was ND/US 60%, B 14.3%, AUS/FLUS 26.3%, FN/SFN 38.9%, SM 91.7%, and M 96.3% [Table 1].

Table 1

Details of distribution of FNA cases according to the Bethesda category and their histopathological correlation

BSRTC category	No. of cases	Cases that underwent surgery	Histopathological diagnosis		RON, %	ROM, %
			Benign lesion	Malignant lesion
I: Unsatisfactory	9 (3.1%)	5	Colloid nodule: 2	Papillary thyroid carcinoma: 3	60	60
II: Benign	157 (55.8%)	35	Colloid nodule: 21	Follicular carcinoma: 2	17.1	14.3
			Adenomatoid nodule: 5
			Hashimoto thyroiditis: 2
				Papillary thyroid carcinoma: 3
			Granulomatous thyroiditis: 1
			Follicular adenoma: 1
III: AUS/AFLUS	24 (8.5)	18	Follicular adenoma: 5	Follicular carcinoma: 1	63.1	26.3
			Adenomatoid goiter: 7	Follicular variant of papillary thyroid carcinoma: 1
			Hürthle cell adenoma: 2	Papillary thyroid carcinoma: 3
IV: FN/SFN	20 (7.1)	18	Follicular adenoma: 6	Follicular carcinoma: 3	77.7	38.9
			Adenomatoid goiter: 3	Hürthle cell carcinoma: 2
			Hürthle cell adenoma: 1	Follicular variant of papillary carcinoma: 2
			Hashimoto thyroiditis: 1
V: SM	14 (4.8%)	12	Hashimoto thyroiditis: 1	Papillary thyroid carcinoma: 9	91.7	91.7
				Medullary carcinoma: 1
				Lymphoma: 1
VI: Malignant	58 (20.7%)	55	Hyalinizing trabecular adenoma: 1	Papillary thyroid carcinoma: 46	98.2	96.3
				Medullary carcinoma: 3
			Hashimoto thyroiditis: 1	Lymphoma: 1
				Anaplastic carcinoma: 3

AUS/FLUS accounted for 8.5% (24 cases of 282) of the FNA diagnoses, and 19 out of 24 cases of AUS underwent partial or total thyroidectomies. Among these, 36.8% (7 cases of 19) were nonneoplastic lesions (adenomatoid goiter); 36.8% (7 cases of 19) were benign neoplasms, including Hürthle cell adenomas (2 of 7) and follicular adenomas (5 of 7); and 26.3% (5 cases of 19) were malignant neoplasms including follicular carcinoma (1 case), noninvasive follicular tumor with papillary-like nuclear features (NIFT-P) (1 case), and papillary thyroid carcinoma (PTC) (3 cases).

In the category of AUS/FLUS, most of the cases belonged to the subgroup MF (7 cases) followed by NA (5 cases); NA, MF (4 cases); NA, L (3 cases); HC (3 cases); and NA, CM (2 cases). The RON for each subgroup was NA 0%; NA, MF 75%; NA, L 50%; NA, CM 100%; MF 66.6%; and HC 100%. The ROM for each group was NA 0%; NA, MF 25%; NA, L 50%; NA, CM 100%; MF 16.6%; and HC 0% [Table 2, Figures 1 and 2].

Table 2

Subgroups of BSRTC third category of atypia of undetermined significance/follicular lesion of undetermined significance

Subgroup	No. of cases	Cases that underwent surgery	Histopathological diagnosis		RON, %	ROM, %
			Benign	Malignant
NA	5	3	Adenomatoid goiter=3	0	0	0
NA, MF	4	4	Adenomatoid goiter=1 Follicular adenoma=2	Follicular variants of papillary thyroid carcinoma=1	75	25
NA, L	3	2	Adenomatoid goiter=1	Papillary thyroid carcinoma=1	50	50
NA, CM	2	2	0	Papillary thyroid carcinoma=2	100	100
MF	7	6	Adenomatoid goiter=2 Follicular adenoma=3	Follicular carcinoma=1	66.6	16.67
HC	3	2	Follicular adenoma=2	0	100	0

Figure 1

(a) Bethesda category II: Multiple follicles with thin colloid in adenomatoid nodule (MGG stain, ×100). Inset shows high power view. (b) Bethesda category IV: Discrete cells and microfollicles in follicular neoplasm (MGG stain, ×100). Inset shows high power view.(c) Bethesda category V: Papilla-like structure (MGG stain, ×100). Inset shows cells with nuclear groove. (d) Bethesda category VI: spindle cells in a background of amorphous pinkish material in medullary carcinoma (MGG stain, ×100). Inset shows high power view

Figure 2

(a) Subgroup NA: Cells with nuclear enlargement and nuclear crowding (MGG stain, ×400). (b) Subgroup NA, MF: Microfollicles formed of enlarged cells (MGG stain, ×400). (c) Subgroup NA, L: Cells with nuclear enlargement and sprinkling of lymphocytes (MGG stain, ×400). (d) Subgroup NA, CM: Cells with nuclear atypia (MGG stain, ×400). Inset shows high power view of cyst macrophages. (e) Subgroup MF: Multiple microfollicles present (MGG stain, ×400). (f) Subgroup HC: Hürthle cells with large nucleus and abundant cytoplasm (MGG stain, ×400)

DISCUSSION

In thyroid cytopathology, the classification of “indeterminate lesions” was a source of confusion. Different subcategories of such lesions had different clinical outcomes. “Indeterminate lesions” were therefore classified into three categories in the BSRTC, namely AUS/FLUS, FN/SFN, and SM.

The category of AUS/FLUS consists of thyroid lesions in which the cytological and/or architectural atypia present is of insufficient degree to qualify for suspicious categories. On the contrary, the atypia is more marked than the benign lesions.

The inherent heterogeneity of the category, interobserver variability, and probable overuse of the category seem to be the reasons for varying malignancy rates in the AUS/FLUS category in the different studies included in the meta-analysis by Kholovaa et al.[7] In our study the ROM for resected AUS/FLUS nodules was 26.3%, while the standard proportion of patients with AUS/FLUS who prove to have cancer after surgery is 20–25%.

We divided the category of AUS/FLUS into six subgroups based on the initial cytological features. The first group consisted of cases with nuclear atypia that included nuclear enlargement, nuclear crowding, and intranuclear pseudoinclusions. This group had five cases out of which two were lost to follow-up. The other three cases underwent second FNS. Among these, two cases showed nuclear atypia again and the third showed nuclear atypia with microfollicle. All the three underwent thyroidectomy and each of the excised lesion was diagnosed as adenomatoid goiter histopathologically. The ROM and RON was 0 in this group. Therefore, in these three cases the presence of nuclear atypia pointed toward nonneoplastic lesion.

Cases with nuclear atypia and microfollicle were included in the second group. This group had four cases, all of which underwent surgery after first FNS. Out of these cases one was diagnosed as adenomatoid goiter, two were diagnosed as follicular adenoma and one was diagnosed as NIFT-P on consecutive formalin-fixed paraffin embedded examination. Thus, we can see that nuclear atypia with microfollicle can be present in both nonneoplastic and neoplastic lesions, including benign and malignant neoplasms.

The third group comprised cases with nuclear atypia and lymphocytes. One case was lost to follow-up while the other two underwent repeat FNS. Between these two cases one showed nuclear atypia only and the other showed nuclear atypia with microfollicles. Both the cases underwent thyroidectomy and were reported as adenomatoid goiter and papillary thyroid carcinoma, respectively.

The fourth group incorporated cases with nuclear atypia and cyst macrophages. There were two cases all of which underwent surgery after first FNS. Both were reported as papillary thyroid carcinoma in further examination. The ROM and RON were maximum in this group. Each had value of 100%. Thereby, the presence of nuclear atypia with cyst macrophage indicated toward malignant neoplasm.

Cases with microfollicles were included in the fifth group. They did not have nuclear atypia or lymphocytes or cyst macrophages. There were seven cases. One case was lost to follow-up. Two cases underwent thyroidectomy after first FNS, out of which one was reported as follicular adenoma and the other as follicular carcinoma histopathologically. Repeat FNS was performed on four cases. Out of these four cases, three cases showed microfollicles of which two were reported as adenomatoid goiter and one was reported as follicular adenoma histopathologically. The fourth case showed nuclear atypia with microfollicles and histopathologically it was diagnosed as follicular adenoma. Thus, we can see that mitotic figures can be present in both nonneoplastic and neoplastic lesions, including benign and malignant neoplasms.

Cases with Hürthle cells formed the sixth subgroup. One case was lost to follow-up after FNS, one case underwent surgery after first FNS and was histopathologically reported as Hürthle cell adenoma. The third case underwent repeat FNS, which showed Hürthle cells with microfollicles, and then underwent surgical excision when it was also histopathologically confirmed to be a Hürthle cell adenoma.

Kholovaa et al. pooled six studies and stratified AUS/FLUS lesions. AUS/FLUS with features suggestive of papillary carcinoma showed a higher ROM (28.9%) followed by architectural atypia (ROM 13.5%), oncocytic AUS/FLUS (8.5%), and technically compromised samples (7.9%).

It has been proposed that AUS/FLUS diagnosed nodules can be tested for BRAF, RET-PTC, PPARγ-PAX8, and RAS mutations.[1415] The ROM in mutation-positive cases is 88%, and in mutation-negative cases it is 5.9%. Mutation positive cases are therefore recommended for total thyroidectomy, while mutation negative cases may undergo lobectomy or observation, as clinically indicated.

The BSRTC category of AUS/FLUS requires further standardization and subgrouping in order to eliminate the diagnostic and prognostic variability associated with it. We have proposed one such system of subgrouping which may help dispel the confusion generated by an AUS/FLUS report, and provide both pathologists and surgeons with a more exact and reproducible diagnostic and prognostic tool. Further studies involving larger patient populations and correlating molecular and genetic parameters with cytology would help validate this subgrouping and its clinicopathological implications.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.