Faculty Bibliography

Background: Fine needle aspiration (FNA) of salivary gland lesions is a fast, minimally invasive and cost-effective procedure that aids in early patient management decisions. Recently, the Milan System for reporting Salivary Gland cytopathology (MSRSGC) was published in order to establish diagnostic categories with implied malignancy risks and recommended clinical follow-up. Our study aims to assess the interobserver reproducibility of salivary gland cytology diagnoses using the MSRSGC.
Design(s): Salivary gland cytology slides from 101 cases with surgical pathology follow-up from 11/2016-06/2019 were blindly and independently reviewed and classified according to the MSRSGC by four cytopathologists. Unweighted and linearly weighted percent agreement and Gwet's AC1 coefficients were calculated in AgreeStat 2015.6/Windows (AgreeStat Analytics).
Result(s): Unweighted percent agreement was 0.69 (substantial agreement) and weighted percent agreement was 0.92 (almost perfect agreement). Unweighted Gwet's AC1 was 0.64 (substantial agreement), and weighted Gwet's AC1 was 0.84 (almost perfect agreement) (Table 1). 50 of 101 (49%) cases had complete agreement among all 4 observers, 77 (76%) had at least 3 observers agreeing on the same diagnosis, and 99 (98%) had at least 2 observers agreeing on the same diagnosis. Category IVA (benign neoplasm) was the most likely to show interobserver agreement: among the 51 cases in which at least 2 cytopathologists agreed on a diagnosis of category IVA, 34 (67%) showed complete agreement among all 4 cytopathologists. Two cases showed no agreement among any observers. One low-grade mucoepidermoid carcinoma had MSRSGC diagnoses ranging from I to IVB, and one secretory carcinoma had MSRSGC diagnoses ranging from III to VI. Low-grade mucoepidermoid carcinoma is reportedly the most common malignant salivary gland tumor associated with false-negative diagnoses on cytology and is often misdiagnosed as a pleomorphic adenoma, due to the presence of bland-appearing intermediate cells as well as confusion between mucin and chondromyxoid stroma (Figure 1). The case of secretory carcinoma showed scant cellularity on cytology, confounding an accurate diagnosis (Figure 2).
Conclusion(s): Interobserver reliability analyses using the MSRSGC showed substantial to almost perfect agreement among the four observers in our study. Only two cases showed no agreement. Category IVA (benign neoplasm) is the most likely to show complete agreement among all observers

Background: Thyroseq next-generation sequencing assay and Afirma gene expression classifier (GEC) are used to risk-stratify thyroid nodules with indeterminate cytology: Bethesda III (atypia of undetermined significance, AUS/FLUS) and IV (suspicious for follicular neoplasm, SFN). In this study, we performed a paired comparison of both tests on the same group of indeterminate thyroid nodules with surgical followup.
Design(s): Of 645 AUS/FLUS/SFN cases with both molecular testing and surgical resection in 2014-2017, 40 cases had both Thyroseq (v2) and Afirma GEC performed on the same specimen. Cross-tabulations and ROC curves were created. McNemar tests were done to compare the performance of Thyroseq versus Afirma. The diagnostic performance of combined results were also examined: the combined result was called positive only if both Thyroseq and Afirma were positive/suspicious. Non-invasive follicular thyroid with papillary like nuclear features (NIFTP) on surgical resections was defined as ?positive.? Results: 20/40 (50%) cases were ?positive? on surgical pathology: 8 papillary thyroid carcinoma (PTC), 11 NIFTP, and 1 follicular carcinoma. Thyroseq and Afirma both showed high sensitivity and low specificity in diagnosing malignancy in indeterminate thyroid nodules. Next, the results of both tests were combined. The overall accuracy of combined testing was higher than either test alone (Figure 1). Compared to Afirma alone, the combined test had significantly higher specificity (30% vs 70%, p<0.05, Table 1), while the sensitivity declined from 90% to 75% (p=0.25, Table 1). Compared to Thyroseq alone, there was no significant difference in specificity (45% vs 70% p=0.06) or sensitivity (80% vs 75%, p=1.00, Table 1). Positive predictive value (PPV) improved compared to either test alone. Negative predictive value (NPV) improved compared to Thyroseq alone, and declined only slightly compared to Afirma alone.
Conclusion(s): Molecular testing of cytologically indeterminate thyroid nodules helps determine the extent of surgery. Low diagnostic performance metrics may limit the utility of molecular studies in distinguishing benign from malignant thyroid lesions. Our results show that the combined results of Thyroseq and Afirma improved the specificity and overall accuracy of molecular testing, and provided additional value in the surgical management of patients with indeterminate thyroid nodules. To the best of our knowledge, this is the first study that compares the performance of these two molecular tests on the same thyroid nodules

Introduction: In urine cytology, the atypical category suffers from interobserver variability and a wide range of risk of malignancy (ROM). This leads to confusion and inconsistent clinical management of patients with atypical urine diagnoses. Our study identified the rate of atypical diagnoses at our institution over a five-year period and examined the concurrent or follow-up surgical pathology diagnoses (SD) to understand the clinical significance of atypical urine specimens.
Material(s) and Method(s): A retrospective review of adult urine cytology specimens pre- 7/1/2014-6/30/2016 (TD) and post- 7/1/2017-6/30/2019 (TPS) implementation of the Paris System for Reporting Urinary Cytology was performed, which identified 10,132 total urine specimens. Of these, 2,457 specimens from 1,727 patients were categorized as "atypical." 177 cytology specimens were found to have corresponding urinary tract SD within 90 days. Pre- and post-Paris System (TD and TPS) urine cytology diagnoses were compared.
Result(s): The overall rate of "atypical" diagnoses was 2,457/10,132 (24.3%). 177 of 2,457 atypical specimens (7.2%) had corresponding urinary tract SD, of which the ROM was 37.9% (TD 37.8%, TPS 37.9%, p=0.992). 61 of 177 atypical urines (34.5%) were high grade (HGUC/CIS) on SD (TD 32.9%, TPS 35.8%). 66 of 177 atypical urines (37.3%) were low grade urothelial neoplasms (LGUNs), atypical, or dysplastic on SD (TD 39.0%, TPS 35.8%, p=0.660). 44 of 177 atypical urines (24.9%) were benign on SD (TD 23.2%, TPS 26.3%) (Table 1).
Conclusion(s): While the atypical diagnosis rate of nearly 1 in 4 urines is fairly high, over one-third of atypical results signified a high grade malignancy, and over an additional one-third of atypical urines represented LGUN, atypical, or dysplastic surgical pathology diagnoses. These findings emphasize the need for close follow-up in patients with atypical diagnoses on urine cytology. Our findings did not show a significant difference between pre and post TPS follow up data. [Formula presented]
Copyright

Introduction: Recent studies evaluating p16 immunohistochemistry (IHC) in cell blocks (CB) of fine needle aspirations (FNAs) in patients with oropharyngeal squamous cell carcinoma (OP-SCC) have shown good correlation between cytology and surgical pathology. Our study aimed to determine the reproducibility of p16 IHC scoring in CBs. Additionally, we evaluated whether interobserver variability would significantly affect the optimal threshold for p16 IHC positivity in CBs.
Material(s) and Method(s): 40 FNAs from 2014-2019 of head and neck squamous cell carcinoma with p16 IHC were obtained. Surgical pathology p16 IHC results were set as reference. p16 IHC stained CBs were scored independently by 5 cytopathologists and recorded as percentage of tumor cell positivity: 0%,0-1%,1-10%,10-50%,50-70%,70%. AgreeStat2015.6/Windows software was used to calculate the percent agreement (Pa) and Gwet's AC1 statistic to assess inter-rater reliability. ROC curves were examined to determine optimal cutoffs for each pathologist based on sensitivity and specificity values (IBM SPSS version 25).
Result(s): Overall performances of the raters were similar, with areas under the curve (AUCs) ranging from 0.88-0.95 (Figure 1). >10% appeared to be the optimal threshold for p16 positivity because this was the lowest threshold to reach 100% specificity with high sensitivity (55-84%) in all 5 raters. Using the >10% as threshold, the Pa was 86% (95% CI 0.78-0.94) and Gwet's AC1 coefficient was 0.72 (95% CI 0.56-0.89).
Conclusion(s): While the goal in developing guidelines for the interpretation of p16 IHC on cytology CBs is to provide generalizable standards for all cytopathologists, interobserver variability must be taken into account. Prior studies have shown optimal cutoffs ranging from >0% (any staining) to >70%, with sensitivity and specificity values ranging from 37%-100%. While our study did not show perfect agreement, all cytopathologists in our study displayed reproducible high sensitivity and specificity values at the >10% threshold with a percent agreement of 86%. [Formula presented]
Copyright

Introduction: Urothelial carcinomas (UC) of the upper urinary tract (UUT) are rare, and usually show higher grade and poorer prognosis than carcinomas of the bladder. The Paris System (TPS) has been integrated into our standard terminology for interpreting urine cytology. Here we compare TPS diagnoses to the traditional reporting system (TD) in interpreting UC of UUT and lower urinary tract (LUT) in correlation with surgical pathology diagnoses (SD).
Material(s) and Method(s): A search of the cytopathology database on urine specimens from 7/1/2014-6/30/2016 (TD) and 7/1/2017-6/30/2019 (TPS) with corresponding lesions in SD, yielded 14-TD cases and 19-TPS in UUT; and 178-TD and 243-TPS cases in LUT. The cytopathology diagnosis using TD and TPS were compared to their corresponding SD.
Result(s): In UUT, 51.5% (17/33) were UC on SD. High-grade UC (HGUC) was correctly identified in UUT in 100% (5/5) with TD, and 75%(3/4) with TPS. No low-grade (LG) diagnoses were rendered in TD or TPS though there were 8 low-grade urothelial neoplasms (LGUN). LGUN was classified as "atypical" (2/3-TD, 2/5-TPS) or "negative" (1/3-TD, 3/5-TPS). Rates of "atypical" diagnoses of UUT were 28.6% (TPS) and 26.3% (TD), with no HGUC on SD. The risk of LGUN with "atypical" diagnoses decreased using TPS from 75.0% (3/4) to 60.0% (3/5). In LUT, HGUC was correctly diagnosed in 55.0%(44/80) (TD) and 47.4%(46/97) (TPS). 4% of LGUNs were classified as LG (3/60-TPS, 1/40-TD). "Atypical" diagnoses for TD and TPS showed 29.5% (23/78) and 30.0% (27/90) risk of LGUN and 34.6%(27/78) and 37.8%(34/90) risk of HGUC respectively. (Table-1)
Conclusion(s): The TD and TPS systems showed high accuracy in reporting UUT-HGUC. However, urine cytology is not optimal to identify LGUN in both TD and TPS systems. Additionally, our results suggest that urine cytology may be more sensitive in detecting HGUC in UUT versus LUT samples.[Table presented]
Copyright

Introduction: Due to the diagnostic dilemma with indeterminate thyroid Bethesda categories III and IV (atypia of undetermined significance, AUS and Suspicious for follicular neoplasm, SFN), many laboratories utilize molecular testing to aid in risk stratification of these nodules. In this study, we evaluated the risk of malignancy (ROM) in AUS and SFN thyroid nodules with subsequent negative molecular (ThyroSeq) test results.
Material(s) and Method(s): This study was designed to evaluate the negative molecular thyroid fine needle aspiration (FNA) cases at a tertiary medical center in the metropolitan area. 109 cases of AUS and SFN thyroid FNAs over 3 years with surgical pathology follow up were included in the study.
Result(s): Of 109 AUS and SFN cases, 4 cases showed insufficient material for ThyroSeq testing (3.7%), 76 cases showed a molecular alteration (69.7%), and 29 cases were negative for an alteration on ThyroSeq (26.6%). Among the cases with negative ThyroSeq results, 26 cases were benign on surgical pathology (89.7%) (7/26 were follicular adenomas), and 3/29 cases were malignant on histopathology (papillary thyroid carcinoma) (ROM=10.3%, Table 1). AUS and SFN cases with molecular alterations showed a significantly higher ROM (ROM= 60.5%) compared to cases testing negative for molecular alterations (p<0.01, z = -4.61).
Conclusion(s): Our study found that indeterminate thyroid nodules that tested negative for a molecular alteration displayed an ROM of 10.3%. This ROM is comparable to the lower limit of ROM of FNA alone (without additional molecular testing data) in the AUS and SFN categories (10-30%), but is significantly lower than the ROM of indeterminate thyroid cases with known molecular mutations. Therefore, clinical follow-up is recommended for thyroid FNA indeterminate nodules, even those testing negative for a molecular alteration, due to the maintained, albeit lower, ROM. [Formula presented]
Copyright

OBJECTIVES/OBJECTIVE:Recent investigations have shown strong correlations between cytology and surgical non-small cell lung carcinoma (NSCLC) specimens in programmed death-ligand 1 (PD-L1) immunohistochemical (IHC) evaluations. Our study aims to evaluate the reproducibility of PD-L1 IHC scoring in NSCLC cytology cell blocks (CBs) and to assess the impact of CB cellularity, method of sample collection, and observer subspecialty on scoring agreement. METHODS:PD-L1 IHC was performed on 54 NSCLC cytology CBs and was scored independently by seven cytopathologists (three of seven with expertise in pulmonary pathology). Three-tier scoring of negative (<1%), low positive (1%-49%), and high positive (≥50%) and interrater agreement were assessed. RESULTS:Total and majority agreement among cytopathologists was achieved in 48% and 98% of cases, respectively, with κ = 0.608 (substantial agreement; 95% confidence interval, 0.50-0.72). Cytopathologists with pulmonary pathology expertise agreed in 67% of cases (κ = 0.633, substantial agreement), whereas the remaining cytopathologists agreed in 56% of cases (κ = 0.62, substantial agreement). CB cellularity (P = .36) and sample collection type (P = .59) had no statistically significant difference between raters. CONCLUSIONS:There is substantial agreement in PD-L1 IHC scoring in cytology CB specimens among cytopathologists. Additional expertise in pulmonary pathology, sample collection type, and CB cellularity have no statistically significant impact on interobserver agreement.

BACKGROUND:Ultrasound has become the initial approach to evaluating thyroid nodules, facilitating the distinction between benign and malignant nodules based on composition, echogenicity, nodule border or margin, shape, the presence of calcifications, and nodule dimensions. The American College of Radiology (ACR) recommended the Thyroid Imaging Reporting and Data System (TI-RADS) as a classification system to standardize thyroid ultrasound reports and to predict the probability of malignancy in thyroid nodules using a scoring system (TR1-TR5) based on multiple ultrasound characteristics and nodule size. Fine-needle aspiration (FNA) is recommended as the next step for nodules that warrant further workup. The authors assessed the accuracy of the ACR TI-RADS based on the corresponding FNA cytology results (Bethesda system diagnoses I-VI). METHODS:ACR TI-RADS ultrasound reports and corresponding FNA cytology diagnoses from January 1, 2018 to August 30, 2018 were evaluated. RESULTS:From January 1, 2018 to August 30, 2018, 2306 thyroid ultrasound-guided FNAs were performed at our institution. Of 2306 cases, 361 had ACR TI-RADS reports available. The majority of FNAs were TR4 (180; 49.9%) or TR3 (108; 29.9%). No TR2 or TR3 nodules were associated with Bethesda category V or VI diagnoses. The majority of TR4 nodules (142 of 180; 78.9%) and TR5 nodules (42 of 65; 64.6%) exhibited benign (Bethesda category II) cytology. Fourteen TR5 cases (21.5%) had malignant (Bethesda category VI) cytology. CONCLUSIONS:Although there were no TR2 or TR3 malignant (Bethesda category VI) diagnoses, and there were only a few malignancies in the TR4 and TR5 categories, the current results reassert the notion that the ACR TI-RADS scoring system shows at least some correlation between benign or malignant cytology diagnoses, as illustrated by the greater number of malignant cases in the higher ACR TI-RADS categories.

Extracardiac rhabdomyoma is an uncommon benign striated muscle tumor with a predilection for the head and neck region. However, it is extremely rare for extracardiac rhabdomyoma to present as a thyroid nodule. We report a case of rhabdomyoma diagnosed by thyroid fine-needle aspiration (FNA) in a patient with Birt-Hogg-Dubé (BHD) syndrome. A 60-year-old man with BHD syndrome presented for recurrent pneumothorax. Chest CT incidentally identified a thyroid nodule. Subsequent sonography confirmed a 4.44 × 2.28 × 2.82 cm solid, hypoechoic nodule with smooth margins in the right upper pole. Ultrasound-guided FNA revealed many clusters and scattered isolated large polygonal cells with abundant granular cytoplasm and small peripherally located nuclei. Vague striations in the cytoplasm were focally identified. No follicular cells or colloid was present. Immunocytochemistry on one direct smear slide demonstrated diffuse positivity for desmin, supporting muscular differentiation. Subsequent surgery identified an adult rhabdomyoma originating from the inferior constrictor muscle of the neck and anteriorly displacing the thyroid. Because the mass was intimately associated with the thyroid gland, it was initially mistaken for a thyroid nodule on ultrasound. Diagnosis of rhabdomyoma on FNA is challenging, especially when rhabdomyoma mimics a thyroid nodule on imaging. The differential diagnosis includes Hurthle cell neoplasm, granular cell tumor, colloid nodule, and normal striated skeletal muscle. Adequate radiologic data and familiarity with the cytologic features of rhabdomyoma are critical for an accurate diagnosis.