Cyto‐morphological features of parathyroid lesions: Fine‐needle aspiration cytology series from an endocrine tumor referral center

Descriptions of parathyroid cell cyto‐morphology are limited. Fine‐needle aspiration cytology (FNAC) with immunocytochemistry (ICC) or biochemical PTH measurements may help verify the parathyroid origin in extraordinary cases, although these methods are nowadays largely replaced by imaging techniques.


| INTRODUCTION
Primary hyperparathyroidism (pHPT) is recognized as symptomatic or asymptomatic hypercalcemia and high to borderline high levels parathyroid hormone (PTH) in the blood. If secondary reasons for hyperparathyroidism have been ruled out (such as renal failure), the cause is one or several parathyroid neoplasms, such as an adenoma (single or multiple) or carcinoma (almost always single nodules). 1,2 Traditionally, patients were surgically treated using a bilateral neck exploration approach. However, since most cases are caused by single adenomas, less invasive approaches have replaced bilateral neck exploration, facilitated by preoperative localization of the abnormal gland using high resolution ultrasound and/or technetium sestamibi scintigraphy. 3,4 Even so, some of these lesions are still difficult to find by radiological methods. Sestamibi scans exhibit limitations when it comes to small tumors, as well as tumors with water clear cell appearance, 5 and thyroid nodules may sometimes create false positive scan results. High-resolution ultrasound may be optimal for the identification of cysts and smaller lesions, but is generally sub-par in distinguishing parathyroid lesions from thyroid nodules and adjacent lymph nodes. 6 When the patient exhibits concomitant thyroid disease, radiological modalities may not be entirely sufficient, and ultrasonography guided fine-needle aspiration cytology (FNAC) could in certain cases be used to pinpoint the location. 3 FNACs of parathyroid tumors are not recommended in general, and are usually utilized as part of a work-up of a thyroid nodule.
Indeed, it is estimated that 0.12% of all thyroid FNACs report intrathyroidal parathyroid glands. 7 Normal parathyroid glands as well as adenomas usually contain conventional chief cells, with or without smaller populations of clear-type cell chief cells and oxyphilic cells.
Thyroid lesions on the other hand contain follicular epithelial cellsbut they are still traditionally difficult to distinguish from parathyroid cells in cytology preparations. 2,4,[7][8][9][10][11][12][13] The literature regarding cytomorphological hallmarks for parathyroid cells is rather limited, and specifically meager in terms of studies that offer clues how to differ parathyroid cells from thyroid epithelium. If the cytopathologist, already while investigating a quick-stained smear, could raise the suspicion of a parathyroid lesion, part of the aspirate can be submitted for biochemical estimation of PTH to establish the diagnosis. Absher et al and Dimashkieh et al made the first attempts to find common hallmarks for parathyroid cells to differ them from thyroid cells using fine-needle aspiration biopsies from 12 to 20 parathyroid lesions, respectively. 8,10 Up until these studies were launched, contemporary scientific literature on the subject was restricted to occasional case reports, or were based on small studies performing aspirations or imprints from surgically resected specimens. Since then, several studies have been performed to define cyto-morphological criteria for parathyroid lesions, 4,7,[11][12][13] but no clear-cut definitions or recommendations in terms of morphological classification have been proposed.
Generally, when listing subjective parameters based on pattern recognition, the interobserver variability is usually high. Moreover, different terminology and definitions for cytological patterns have been used, adding to the heterogeneous results obtained. In this study, we correlated cytological characteristics of parathyroid nodules with postoperative histology in a large series of patients in order to see whether or not the FNAC report had an impact on the clinical handling, and if the morphological aspects observed on FNACs were correlated to specific histological attributes. Of course, if parathyroid lesions could be recognized cyto-morphologically, it would be of direct clinical value, to avoid falsely diagnosing a parathyroid lesion as thyroid-derived.

| MATERIAL AND METHODS
The pathology database at the Karolinska University Hospital, Stockholm, Sweden was searched retrospectively for parathyroid lesions diagnosed by cytology. For this purpose, we used the SNOMED nomenclature for "parathyroid" (T97xxx) and investigated the cytology records between January 1, 2000 and May 20, 2020. In total, 37 cytology reports were found, which should be compared to the >1600 thyroid-related FNACs performed annually at our institution. The diagnoses, biochemical tests, ultrasonography-guided FNAC smears and available histological tissue sections (if surgical resection was performed) were retrieved. All FNA smears had been airdried and stained with May Grunwald-Giemsa (MGG). Unstained smears were also previously acquired for the purpose of PTH ICC which had been performed for the majority of cases. The PTH ICC was conducted in a clinically accredited laboratory setting using the    were assessed cyto-morphologically. Seven of these cases did not proceed to surgical intervention and were therefore not included in the comparative analyses with histology, and for one case the histological sections were not available for evaluation, leaving 20 cases in which comparison with histopathologic characteristics of the subsequently excised tumor could be made.
The clinical parameters of this cohort are summarized in Table 1.
In detail, out of the 28 patients investigated, the female: male ratio was 3:1, and the median age at the time of cytology was 55 years (range 16-76). Twenty-one patients (75%) had a pre-FNAC clinical diagnosis of pHPT. Of these, the FNAC investigation was initiated with the aim to verify the assumed location of a parathyroid lesion in 17 patients (81%), of which many exhibited a history of inconclusive tumor localization using radiology and/or negative findings using surgical exploration. Therefore, the reason for performing the FNAC was primarily not to establish a preoperative parathyroid tumor diagnosis, but rather to verify the location in cases in which standard clinical work-up was inconclusive. Cases with an established PHPT and initial negative findings using imaging and/or surgical exploration were followed biochemically and re-investigated using scintigraphy, and if the results were inconclusive, the patients were then assessed with ultrasonography-guided FNAC in order to try to establish the exact location in order to avoid excessive explorative surgery. The majority of patients (15/28; 54%) was somewhat unconventional in its clinical presentation, as patients either exhibited synchronous thyroid nodules upon ultrasonographic investigations (n = 4), presented with an intrathyroidal nodule (n = 4) or had no previous diagnosis of PHPT (n = 7). In these cases, the FNAC was therefore needed to pinpoint the parathyroid origin in order to plan the correct surgical procedure.
An additional 12 patients (12/28; 43%) had previous neck surgery (either parathyroid surgery or surgery for other causes), and the FNAC was ordered to verify the parathyroid origin in order to plan the reoperation carefully. When looking into these latter cases more care- fully, it appears that the preoperative ultrasonographic investigation in many cases was uncertain, and the surgeons specifically ordered another round of ultrasound with FNAC verification before attempting another round of surgery. In these instances, the risk of  T A B L E 2 Histological and cyto-morphological attributes of the study cohort

| Cytological hallmarks of parathyroid lesions
The cyto-morphological aspects of the parathyroid FNACs are listed in Table 2, and illustrated in Figures 1-3. A few common denominators were found among parathyroid FNACs. In terms of cellular characteristics, the smears were generally monomorphic to slightly pleomorphic. Cells were commonly small, ranging from sizes slightly larger than an erythrocyte up to twice the size of an erythrocyte in the majority of cases (24/28 cases, 86%). In most smears, the cells showed a low to moderate nuclear-to-cytoplasm ratio (seen in 23/28 cases, 82%), a round to oval (seen in 25/28 cases, 89%) decentralized nucleus (seen in 21/28 cases, 75%) with stippled chromatin (seen in 25/28 cases, 89%). Cytoplasmic granularity was seen in 14 cases (50%) and intracytoplasmic vacuolization was noted in seven cases (25%; Figure 3). The background characteristics were generally without macrophages (only seen in four cases, 14%) and colloid (only seen in four cases, 14%). No mitotic activity was identified in any smear.
Outliers in many parameters were two cases (7%) in which the

| Histological correlations
Out of the 20 cases with available histological sections, the dominating growth pattern was solid formations, which was present in 18 cases (90%; Figure 4). Of these, seven cases exhibited this growth pattern exclusively. The corresponding smears to the seven cases with sections containing solid growth only included two to four cytological patterns and thus did not pinpoint any architectural correlation. The remaining two histological cases without solid growth contained solely microacinar growth and sinusoidal patterns, respectively. The from cystic lesions (not included in this cohort), it may be invaluable since morphological and immunocytochemical methods may be of little use. 6,8 In this study, we have collected one of the largest series of parathyroid lesions investigated by FNAC and describe the cytomorphological aspects with the aim to correlate various aspects to the final histopathology report. The cytological parameters in our study include many of the parameters previously described by Kumari  difficult. 14 At our department, we use ICC using antibodies against TTF1 and PTH, but previous work indicate that also GATA3 may facilitate the discrimination between these two tissue types. 15  Similarly, intracellular fat tissue content has been proven to inversely correlate with the endocrine activity in normal and neoplastic parathyroid glands. 16  Moreover, some cytological hallmarks may be less pronounced in histological preparations and vice versa, which may be due to the different methods employed in terms of fixation.
FNAC is generally a good modality in differentiating benign from malignant tumors when the diagnosis is based on cellular morphology.
However, the histological criteria for parathyroid carcinoma do not rely on cellular characteristics, but rather on tumor behavior (such as capsular and vascular invasion). 16,17 As of this, one cannot rule-in a parathyroid carcinoma diagnosis from FNAC of a primary parathyroid tumor. Even so, certain cellular attributes are overrepresented in parathyroid carcinoma compared with adenoma, including macronucleoli and brisk pleomorphism. 16,17 Although our study only included two carcinomas, we still saw cytological characteristics (larger nuclei with more prominent atypia) that clearly set them apart from the adenomas, indicating possible features to be recognized using FNAC that potentially could alert the surgeons preoperatively.
In our series, pHPT patients with inconclusive imaging and/or negative surgical exploration were followed biochemically, and assessed using ultrasonography-guided FNAC at a later stage in order to verify locations of the culprit lesions and to avoid excessive explorative surgery. In all, the result of the FNAC investigation affected clinical decisions concerning treatment options in 27 patients (96%), most notably correct localization leading to a more focused surgical approach rather than bilateral exploration. Therefore, parathyroid FNAC may be an aiding tool for subsets of cases when imaging analyses or previous surgery fail to localize the culprit lesion. However, it must be stressed that the cohort described in this manuscript is a highly selected group of extraordinary cases in which the FNAC was deemed necessary during the clinical work-up to guide the surgical approach. This is by no means standard care in our department, and given the fact that we have performed >4000 parathyroid surgeries spanning the study period, the use of FNAC in localizing parathyroid tissue at the Karolinska University Hospital is exceedingly rare. Moreover, the usage of FNAC in the clinical work-up of parathyroid tumors is not without disadvantages. FNAs can complicate parathyroid surgeries due to fibrosis developing after a biopsy procedure, which potentially could lead to increased patient morbidity (for example, damage to adjacent nerves). 18 However, surgical difficulties or complications attributable to the FNAC procedure were not noticed in this series. Additionally, the appearance of FNA-induced fibrosis on subsequent histology may confuse the responsible pathologist, as this feature might be worrisome for an atypical parathyroid tumor or parathyroid carcinoma. 16 In our series, biopsy artifacts were noted in 40% of cases, including six adenomas and two atypical tumors. In the two atypical cases, the fibrosis was accompanied by several other risk features (such as trabecular growth, mitoses and nuclear pleomorphism), thereby arguing against misdiagnosis due to previous manipulation. Even so, the high proportion of cases with biopsy artifacts in our series advocate careful histological assessment of excised parathyroid tumors if a previous FNA procedure was performed, and the clinical team must be aware of the potential risks involved if planning for a parathyroid FNAC.
To conclude, parathyroid FNAC may be considered in a highly selected group of patients when localization of the diseased gland is imperative for the surgical procedure and not easily established using conventional imaging analysis. Although several morphological features seem recurrent on cytology, diagnosis should be confirmed with PTH measurements or PTH ICC. In this cohort, the FNAC investigation contributed with valuable information and affected clinical decisions for the majority of patients, but also induced biopsy artifacts in nearly half of the cases, which may cause diagnostic predicaments on the histological level. Also, although based on few cases, we also found cyto-morphological differences between parathyroid adenomas and carcinomas using FNAC worthy of follow-up studies.