Journal of the American Society of Cytopathology
Volume 11, Issue 3,
May to June 2022
, pages 133-141
Author links open an overlay panel, , , , , , ,
https://doi.org/10.1016/j.jasc.2022.02.002Get rights and content
Rapid In-Situ Intraoperative Assessment (ROSE).cytologysamples increases the practice of cytopathology. More recently, ROSE diagnoses such as frozen section (SF) diagnoses have guided immediate clinical decisions. In this study, we evaluated the diagnostic accuracy of definitive ROSE diagnoses in our quality assurance system over a period of 52 months.
Materials and methods
ROSE cytology cases from January 2017 to April 2021 were retrieved from ourlaboratory information system. After excluding deferred or non-diagnostic/unsatisfactory cases, each final ROSE diagnosis (ie, negative malignant cells or positive malignant cells) was classified as concordant or discordant with the final diagnosis. For comparison, concordance of FS diagnoses from the same period was compiled and compared with ROSE diagnoses using χ2test with P < 0.05 considered statistically significant.
Of the 1649 ROSE diagnoses, there were 15 discrepancies (0.9%) with 1 moderate final interpretation disagreement (0.06%). In comparison, of 17,469 FS diagnoses, there were 141 discrepancies (0.8%) with 49 final moderate or severe discrepancies in interpretation (0.3%). Other divergences were minor. There were no statistically significant differences in moderate and severe final misinterpretation rates.
Final misinterpretation rates for final diagnoses of ROSE and FS were similar in this study. Given the growing role of ROSE and its application in some cases to immediate clinical decisions, monitoring the accuracy of the final diagnosis can serve as an initial quality assurance measure.
Rapid intraprocedural in-situ evaluation (ROSE) of cytology specimens has been used for several decades, primarily for fine needle aspiration (FNA) smears and core biopsy specimens for palpation preparation.1,2,3 Selected procedures performed by radiologists, physicians, surgeons and pathologists benefit from the ROSE exam performed by cytopathologists to ensure the intended lesion is sampled, reduce mismatching rates, ensure proper screening of specimens for ancillary examinations, minimize the number of repeat procedures and make a initial diagnosis in selected cases.4However, ROSE practices are complex and vary widely from institution to institution. Differences include technical details (e.g., slide fixation, types of stains used), type of services offered (e.g., fitting and/or diagnosis), staff involved (e.g., cytotechnicians, pathology residents, cytopathology staff, and /or cytopathologists) and use of technology such as telecytology.4The clinical considerations and purpose of ROSE may vary depending on the patient's situation and the clinician's treatment plans. For example, ROSE for thyroid ANF can be managed by assessment of suitability by a cytotechnician, without on-site diagnosis (which would require the involvement of a cytopathologist). In contrast, a proceduralist who performs an FNA of a lung mass may want an on-the-spot diagnosis to assess the possibility of recurrence of the adenocarcinoma. If a final diagnosis of malignant ROSE is made, the treating physician may proceed to place baseline data on the patient. Such practices are increasingly found in tertiary care settings.
While ROSE practices vary, the ability to make critical clinical decisions raises questions about the diagnostic accuracy of the final ROSE diagnoses, which are used by the procedure operator to potentially take the next step. In a survey published by the American Society of Cytopathology, the 3 groups surveyed (cytotechnologists, cytopathologists, pathology interns) were asked "why do physicians order ROSE" and "why does cytology do ROSE".5Regarding the question "why do physicians order ROSE", one response was "providing immediate interpretation to facilitate patient management", with affirmative responses in the three groups ranging from 44.3% to 57.9%. When asked "why cytology performs ROSE", the affirmative answer "to provide immediate interpretation to facilitate patient management" ranged from 48.3% to 63.2%. In this regard, ROSE diagnoses are used similarly to Frozen Section (SF) diagnoses in selected cases. Individual studies of organ-specific ROSE procedures evaluated statistical parameters (eg, sensitivity, specificity, predictive values) for accuracy.6,7,8,9,10,11 However, an ongoing systematic quality assurance (QA) process as this required for FS diagnoses, it is not routinely performed.2,12Given some similarities between intraoperative ROSE and intraoperative SF assessments, we performed a QA study to assess the accuracy of our institution's ROSE diagnoses and how they compare with our intraoperative FS diagnoses over a 52-month period.
Materials and methods
This study was conducted with permission from the University of Pittsburgh Medical Center, Division of Pathology Quality Improvement Program. Cases of cytopathology and surgical pathology diagnosed with ROSE and FS from January 2017 to April 2021 (52 months) were retrieved from our laboratory information system. The operator of the procedure decided to order a ROSE with the original diagnosis, a ROSE for adequacy assessment only, or no ROSE. The cytopathologist made a diagnosis of ROSE
A total of 5900 ROSE events were identified during the study period. Of these, 1649 ROSE events were identified with definitive diagnoses (malignant cells negative, malignant cells positive) in 1141 cases. The remaining ROSE events were either deferred (3080 events) or non-diagnostic (1171 events). Most ROSE cases were performed with Diff-Quik stains (73.5% of ROSE cases), while alcohol-fixed smears were occasionally used in our practice (26.5% of ROSE cases).
Intraoperative pathology assessments (ie, ROSE, FS) have become an integral part of interventional and surgical practices. FS assessments, required by the surgeon and performed by the surgical pathologist, are often used to guide the surgical procedure (eg, whether to perform a resection). In contrast, the practice of ROSE has multiple objectives (improving sampling adequacy, ensuring proper segregation of samples for ancillary testing, minimizing repeat procedures, and making an initial diagnosis) and is
In conclusion, we evaluated the diagnostic accuracy of our final ROSE diagnoses in our quality control system. While direct comparisons with the FS process are not perfect, the bottom line is that the final diagnosis, whether in practice ROSE or FS, can have a significant clinical impact because it gives the clinician receiving the information the freedom to take immediate and meaningful action. , based on this information. We believe that monitoring these diagnoses can be the first step in ROSE
Disclosure of a Conflict of Interest
All authors have no conflict of interest. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Daniel L. Geisler, MD: Formal Review, Validation, Investigation, Resources, Original Draft Writing, Review Writing, and Editing. Rick J. Nestler, BS: Research, Resources. Beth L. Mosley, SCT: Research, Resources, Data Collection. Adrianna L. Walko, BS: Research, Resources. Jacqueline M. Cuda, SCT: Research, Resources. Karen E. Schoedel, MD: Formal review, review writing, and editing. Jon M. Davison, MD: Formal Review, Review, and Editing. N. Paul Ohori: conceptualization,
ROSE diagnoses may warrant immediate clinical intervention in some cases. Assessment of its diagnostic accuracy, as with frozen sections, can be implemented as a quality assurance measure.
- j.Wojtenme inside
Lessons learned from small-set biopsy clinical trial surveys: the importance of rapid on-site assessment
J Am Soc Cytopatol
- T.Nakajimame inside
Rapid assessment of on-site cytology during ultrasound-guided transbronchial needle aspiration for lymph node staging in patients with lung cancer
Ann Thorac Surg
- NCJhalame inside
Providing on-site diagnosis of malignancy in fine-needle aspirates under endoscopic and ultrasound guidance: should it be done?
Ann Diagn Pathol
- JL.Sauterme inside
Results of the American Society of Cytopathology Rapid On-Site Search (ROSE) 2019 - Part 2: Subjective Views of the Cytopathology Community
J Am Soc Cytopatol
- annuallyVanderLaanme inside
American Cytopathology Society 2019 Onsite Rapid Assessment Survey Results - Part 1: Objective Best Practices
J Am Soc Cytopatol
- S.Satturwarme inside
Update on tactile preparations for small biopsies
J Am Soc Cytopatol
- J.F.Silvermanme inside
Diagnostic accuracy and role of immediate interpretation of fine needle aspiration biopsy from different sites
- right sidetamborme inside
Cytopathology plus FNA cytology: rapid on-site assessment – differences in practice. cap today
- MOTHER.Eloubeidime inside
Agreement between rapid cytological interpretation of EUS-guided ANF samples: implications for the endosonographer and patient management
Am J Gastroenterol
- MJCubicme inside
The diagnostic value and accuracy of impression cytology in image-guided biopsy: a review of our experience at a large academic center
Agreement between rapid on-site assessment and final cytological diagnosis of salivary gland samples
Fast and gentle on-site evaluation of endobronchial ultrasound cytology
2022, Journal of the American Society of Cytopathology
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: slippery as an eel to be diagnosed in a series of 3 cytology cases
Journal of the American Society of Cytopatology, tom 11, wydanie 1, 2022, s. 40-45
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.
Performance of the Afirma gene sequencing classifier compared to the gene expression classifier in thyroid nodules with indeterminate cytology
Journal of the American Society of Cytopatology, tom 11, wydanie 2, 2022, s. 74-78
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.
A prospective randomized study to compare the safety, diagnostic yield, and utility of 22- and 19-gauge transbronchial ultrasound aspirates and processing techniques for cytology and histopathology
Journal of the American Society of Cytopatology, tom 11, wydanie 2, 2022, s. 114-121
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.
A significant remaining challenge in cervical adenocarcinoma screening
Journal of the American Society of Cytopatology, tom 11, wydanie 1, 2022, s. 1-2
Evaluation of cytologically indeterminate thyroid nodules and ACR TI-RADS molecular profiles: a single institution experience
Journal of the American Society of Cytopatology, tom 11, wydanie 3, 2022, s. 165-172
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.
The use of molecular tests to improve the treatment of thyroid nodules with indeterminate cytology: an institutional experience with review of molecular alterations
Journal of the American Society of Cytopatology, tom 11, wydanie 2, 2022, s. 79-86
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.
This study was presented at the 69th Annual Scientific Meeting of the American Cytopathology Society in Las Vegas, Nevada, November 11-14, 2021.
© 2022 American Society of Cytopathology. Published by Elsevier Inc. All rights reserved.