Accuracy of definitive cytopathological diagnoses with rapid on-site assessment: assessment of potentially critical diagnoses as a means of quality assurance (2023)

Quote excerpt:

Sensitivity and specificity decrease if the atypicality is found to be false-negative or false-positive, respectively. Geisler et al20 reported a 99.1% concordance rate with a definitive diagnosis, selecting cases with a definitive diagnosis of ROSE in all anatomic sites and found concordance similar to an intraoperative frozen section. Nakajima et al21 showed that a smear-based ROSE could achieve a 94.3% concordance rate with a definitive diagnosis of lung cancer staging in EBUS-TBNA nodules.

Rapid site assessment (ROSE) typically uses swabs taken at the treatment site ("smear-based ROSE"). This requires considerable time, usually 2 people, technical knowledge and it can be difficult to estimate the material available for auxiliary research. We developed an alternative ROSE method using ThinPrep liquid cytology with hematoxylin and eosin (H&E) staining ("liquid-based ROSE") and evaluated its advantages.

Physicians wash the sample(s) in CytoRich Red and send them to Pathology. A defined part of the washed needle is removed for ThinPrep stained with Quick H&E. Suitability and diagnosis were compared with the final result. Total time recorded.

Of the 52 liquid-based ROSE readings, 28 (53.8%) were interpreted as "adequate" and the final as appropriate; 17 (32.7%) were interpreted as "insufficient" and the last as inadequate; 7 (13.5%) were interpreted as "insufficient" and the last as adequate. Of the 23 readings obtained during the on-site diagnosis, 15 (65.2%) were interpreted as definitive positive or negative diagnoses; 6 (26%) were interpreted as non-diagnostic; and 2 (8.7%) were interpreted as atypical. All definitive diagnoses were consistent with the final diagnoses. Liquid ROSE time ranges from 6 to 22 minutes (average 13 minutes) and only requires 1 person.

The liquid ROSE method allows for accurate determination and fitting, takes about 15 minutes of the cytologist's time, and can be performed by only 1 person. This technique produces well-preserved and stained slides, may allow better estimation of the total amount of material in a sample vial, and may provide a better platform for telecytology.

Malignant peritoneal mesothelioma is an extremely rare cancer and is a difficult diagnosis to make based on cytology alone. We report 3 cases in which cytological features were misdiagnosed as carcinoma/lymphoma, but histopathology and immunohistochemistry (IHC) established the diagnosis of malignant mesothelioma.

Case 1 was a 60-year-old man with multifocal ascites and omental agglutination. The peritoneal fluid was reported to be malignant on cytology but was misclassified as an adenocarcinoma. Case 2, a 45-year-old man with ascites and peritoneal nodules radiographically simulating peritoneal carcinoma, was also reported as positive malignant on ascitic fluid cytology. Fine needle aspiration (FNA) of the omentum fat revealed signet ring cells, thus confounding the cytological diagnosis of adenocarcinoma. Case 3 was a 63-year-old man with a parasplenic mass with extensive omental crowding and peritoneal nodules, who also had suspected peritoneal carcinoma on radiological examination. FNA smears from the peripheral mass showed layers of plasmacytoid cells. In the case of cytology, the differential diagnoses offered were neuroendocrine tumor or non-Hodgkin's lymphoma. The diagnosis of malignant mesothelioma was established only after IHC in histopathological sections in all these cases. None of our patients had a history of previous exposure to asbestos.

In these clinical settings, when radiology suggests peritoneal carcinoma, cytological features require confirmation by IHC/fluorescence in situ hybridization on a cell block or biopsy to correctly identify malignant mesothelioma and distinguish it from metastatic carcinoma deposits and benign mesothelial proliferation.

About 15% to 30% of fine needle aspiration (FNAB) thyroid nodules have indeterminate cytology. The Afirma (Veracyte Inc, South San Francisco, CA) Gene Expression Classifier (GEC)/Gene Sequencing Classifier (GSC) assays are designed to improve risk stratification of unspecified thyroid nodules. The aim of this study was to evaluate and compare the effectiveness of the Afirm GEC and GSC tests in unspecified thyroid lesions.

Thyroid FNAC cases with indeterminate cytology were searched in the pathology database and only those with Afirma results available were selected for this study. Demographic, ultrasound, cytological, molecular, and subsequent surgical procedure results for each patient were collected and analyzed.

There were 100 cases with indeterminate thyroid FNAB results, including 49 cases investigated by the GEC and 51 cases by the GSC. In the GEC group, the mild calling rate (BCR) was 53% (26 of 49), and the calculated negative predictive value (NPV) and positive predictive value (PPV) were 88% and 47%, respectively. In the GSC group, the BCR was 63% (32 of 51), and the calculated NPV and PPV were 100% and 64%, respectively. While only 17% (1 in 6) of benign oncocytic lesions were classified as benign by the GEC, 60% (3 in 5) of benign oncocytic nodules were classified as benign by the GSC.

In this study, we showed that just over half of the unspecified thyroid nodules were Afirma GEC/GSC negative, and the BCR using the Afirma GSC test was greater than the GEC. GSC presented higher VPN and VPP than GEC. In addition, the Afirm GSC was superior in differentiating between benign and malignant oncocytic thyroid lesions.

Transbronchial needle aspiration (TBNA) guided by endobronchial ultrasound (EBUS) is a widely used method of minimally invasive lymph node sampling. The benefits of cytological processing of samples over "biopsies" are unclear. It is not known whether safety or diagnostic yield is affected by needle thickness.

Between June 2018 and July 2019, 40 patients (56 lesions) undergoing EBUS TBNA lymph node evaluation were enrolled in this prospective single-center study. Patients were selected by randomization of exchanged blocks to undergo EBUS TBNA starting with either 22-gauge (22g) or 19-gauge (19g) needles. Separate samples were sent for cytological and histopathological processing. Surgical pathologists and cytopathologists were unaware of needle size. The primary endpoint was diagnostic performance. Secondary endpoints compared sample suitability by rapid on-site assessment (ROSE), sample suitability for molecular testing, sample quality and safety.

The diagnostic yield of histopathology was 87.5% and 83.9% for 19 g and 22 g, respectively (P = 0.625). There was no significant difference in the diagnostic performance of cytology based on needle size. There was no significant difference in the quality of slides. Molecular suitability for core needle biopsy was 77% and 80% for 22g and 19g needles, respectively. Molecular suitability for cytology blocks was 77% and 80% for 22g and 19g needles, respectively. There were no significant procedural complications.

The 22 g and 19 g EBUS TBNA needles provided similar diagnostic performance and clinical utility in supportive studies. Cytology processing techniques or "core needle biopsy" methods did not show a significant impact on diagnostic yield or utility of molecular tests.

The American College of Radiology (ACR) Thyroid Imaging Reporting and Data Systems (TI-RADS) was developed to standardize thyroid ultrasound reporting and predict the likelihood of malignancy. In our study, we sought to correlate indeterminate cases of thyroid fine needle aspiration cytology with previous ACR TI-RADS ultrasound findings and concurrent molecular testing findings to investigate how well the use of ACR TI-RADS at our institution predicted which patients with indeterminate cytology may have molecular lesions.

We performed a retrospective review of thyroid nodules. Patients with US reports that included TI-RADS results, fine needle aspiration specimens with indeterminate cytology (Bethesda grade III-V), and molecular test results were included.

A total of 46 cases of indeterminate cytology had previous US reports with TI-RADS scores and molecular tests (Bethesda Class III, n=37; Bethesda Class IV, n=6; Bethesda Class V, n=3). Most unspecified cases had TI-RADS TR4 (31 of 46; 67.39%) or TR5 (9 of 46; 19.57%). RAS mutations were the most common alteration (n = 12). Of the 46 cases, 22 (47.85%) showed no changes. Ten cases were operated on, of which seven had malignant tumors.

Molecular studies in cytologically indeterminate thyroid nodules provided valuable information about changes in TR4 and TR5; however, TR2 and TR3 lesions often did not show molecular alterations. These findings highlight the potential value of considering ultrasound imaging capabilities when assessing the significance of indeterminate cytology findings.

Molecular testing helped clinicians and cytopathologists to further categorize unspecified thyroid fine needle aspiration (FNAB) specimens. The aim of this study was to evaluate the accuracy of commercially available molecular tests, review their impact on patient care, and correlate molecular changes with histological findings.

A search of the pathology laboratory information system identified thyroid FNABs performed at our facility between January 1, 2015 and June 30, 2020. Surgical observation and additional molecular test results were collected. We evaluated the accuracy of these tests and whether they could reduce the number of operations performed.

Our search of the laboratory information system identified 510 cases reported as atypia of undetermined significance, 94 as suspected follicular neoplasia, and 44 as suspected follicular neoplasia of Hurthle. Of the samples, 343 did not have further molecular testing, 146 were sent to ThyGenX/ThyraMIR and 136 were sent to ThyroSeq. Of patients without molecular testing, 50.4% underwent follow-up surgery compared to 30.8% after ThyGenX/ThyraMIR and 38.2% after ThyroSeq, resulting in 38.9% fewer surgeries and 24.2% fewer surgeries and odds ratios 0.04 (95% CI, 0.00-0.33) and 0.14 (95% CI, 0.01-0.95). For the ThyGenX/ThyraMIR assays, the risk of malignancy for high and moderate to high risk lesions was 80%, 28.6% for moderate and low to moderate lesions, and 23.1% for low risk lesions. For ThyroSeq, the cancer risk was 87.5% for high-risk lesions, 36.8% for intermediate or high-risk lesions, 27.3% for intermediate-risk lesions, and 0% for low-risk lesions. The areas under the curve for the ThyGenX/ThyraMIR and ThyroSeq assays were 0.65 and 0.85, respectively.

These findings suggest that both ThygenX/ThyraMIR and ThyroSeq can be used in our facility to stratify cytology samples based on malignancy risk and reduce the number of surgeries performed in our facility.