Article in Press
The contemporary utility of intraoperative frozen sections in thyroid surgery☆
, , , ,
Abstract
View Table in HTML
View Table in HTML
View Table in HTML
Full Text
ReferencesArticle Outline
Abstract
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
Purpose
To determine the accuracy of intraoperative frozen section analysis on thyroidectomy specimens stratified by the Bethesda classification scheme and its utility for intraoperative decision-making.
Study design
Retrospective chart review.
Methods
A retrospective review was performed on all patients who underwent thyroidectomy or thyroid lobectomy with intraoperative frozen sections at a tertiary care academic center from 2009 to 2015.
Results
There were 74 total patients who underwent partial or total thyroidectomy with intraoperative frozen section analysis of a thyroid nodule whom had previously undergone a thyroid fine needle aspiration of the nodule. The sensitivity, specificity, positive predictive value, and negative predictive value for a thyroid frozen section with respect to its prediction for malignancy was 81%, 95%, 98%, and 66%, respectively, with a diagnostic accuracy of 85%. For 37 patients with an indeterminate cytologic diagnosis on fine needle aspiration (Bethesda categories III–V), the sensitivity, specificity, positive predictive value, and negative predictive value for a thyroid frozen section was 81%, 91%, 95%, and 67%, respectively, with a diagnostic accuracy of 84%. False positives and false negatives resulted in 1 completion thyroidectomy for benign pathology and 3 reoperations for malignancy not discovered on frozen section.
Conclusion
While intraoperative frozen sections on thyroid specimens may be helpful if positive, the false negative rate remains high. There appears to be limited value in routine frozen sections to guide clinical management and decision-making in the era of the Bethesda system.
Keywords:
Thyroidectomy, Frozen section, Fine needle aspiration biopsy, Bethesda classification, Indeterminate cytology1. Introduction
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
The increasing incidence and detection of thyroid nodules over the past decade [1] has led to an increase in fine needle aspiration biopsies. While there was initially much variability with regards to the classification and subsequent management of these results, this has improved greatly with the standardization brought on by the Bethesda classification system, introduced in 2007 and used widely since 2009 [2]. Within the classification scheme, studies have shown a high level of accuracy for the Bethesda II (benign) and Bethesda VI (malignant) categories [3]; however, the indeterminate Bethesda categories III (atypia of undetermined significance or follicular lesion of undetermined significance), IV (follicular neoplasm or suspicious for follicular neoplasm), and V (suspicious for malignancy) have highly variable reported rates of malignancy and greater discordance amongst cytopathologists, resulting in a lack of a clear management strategy for the clinician.
One proposed solution for the ambiguity regarding the indeterminate Bethesda categories is the use of intraoperative frozen section analysis on thyroidectomy specimens to determine the need for completion thyroidectomy. While frozen sections are used frequently and with great accuracy for head and neck aerodigestive tract malignancies, prior studies on thyroid specimens have shown mixed results [[4], [5], [6],[7]], leading some to assert that cytologically indeterminate nodules cannot be accurately diagnosed intraoperatively [8]. Most of the aforementioned reviews, however, were performed prior to the adoption of the Bethesda system; therefore, there is a paucity of data regarding the accuracy and utility of frozen section analysis on nodules classified within this system. This lack of data is reflected in the 2015 American Thyroid Association (ATA) management guidelines, which neither argues for nor against the use of intraoperative frozen sections [9].
The purpose of our study was to determine the accuracy of intraoperative frozen section analysis on thyroidectomy specimens stratified by the Bethesda classification scheme and its utility for intraoperative decision-making, with a focus on indeterminate (Bethesda categories III–V) nodules.
2. Materials and methods
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
The study was approved by the Cleveland Clinic institutional review board. A retrospective chart review was performed on all patients who underwent thyroid lobectomy or total thyroidectomy with intraoperative frozen section analysis at a tertiary care academic center from 2009 to 2015, after the adoption of the Bethesda classification system. Frozen sections were performed at the clinician's discretion in order to determine the need for completion thyroidectomy, for immediate feedback regarding the disease process, and/or for determining the need for central neck lymph node sampling. All patients included had a prior fine needle aspiration biopsy (FNAB) of a thyroid nodule performed with available cytopathologic results interpreted within the Bethesda classification, as well as subsequent intraoperative frozen section analysis of the same nodule. Patients with FNABs and/or frozen section analyses performed on adjacent lymph node specimens only were excluded, as were patients for whom final pathologic results were unavailable for comparison.
Data regarding patient demographics, nodule size and ultrasound characteristics, and perioperative complications and subsequent management were collected from the electronic medical records. Worrisome ultrasound characteristics are based on the 2015 ATA guidelines regarding sonographic risk patterns [9].
Performance statistics for sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were performed. Statistical analysis was performed using JMP (JMP®, Version <10>. SAS Institute Inc., Cary, NC, 1989–2007) software.
3. Results
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
There were 74 total patients, median age 44 years old (range 14–81), who underwent thyroid lobectomy or total thyroidectomy with intraoperative frozen section analysis whom also had a preoperative FNAB performed on a thyroid nodule that was classified by the Bethesda system. Demographics and clinical characteristics of all patients are shown in Table 1. Thirty-seven of the 74 patients (50%) had a cytologically indeterminate FNAB; characteristics of these patients are shown separately in Table 2. The total malignancy rate for all patients was 73%. The malignancy rate for the cytologically indeterminate group was 70%. Malignancy rates for the individual Bethesda categories on preoperative FNAB are shown in Table 3.
| Age (years) | <25 | 6 (8%) |
|---|---|---|
| 25–45 | 32 (43%) | |
| >45 | 36 (49%) | |
| Gender | Male | 23 (31%) |
| Female | 51 (69%) | |
| Nodule size on U/S (cm) | <1 | 2 (3%) |
| 1–4 | 48 (65%) | |
| >4 | 15 (20%) | |
| Unknown | 9 (12%) | |
| Suspicious U/S? | Yes | 32 (43%) |
| No | 36 (49%) | |
| Unknown | 6 (8%) | |
| History of XRT? | Yes | 2 (3%) |
| No | 72 (97%) | |
| Bethesda category | Non-diagnostic | 4 (5%) |
| Benign | 11 (15%) | |
| AUS | 13 (18%) | |
| Follicular neoplasm | 11 (15%) | |
| Suspicious | 13 (18%) | |
| Malignant | 22 (30%) |
U/S, Ultrasound; XRT, radiation therapy; AUS, atypia of undetermined significance.
| Age (years) | <25 | 4 (11%) |
|---|---|---|
| 25–45 | 17 (46%) | |
| >45 | 16 (43%) | |
| Gender | Male | 10 (27%) |
| Female | 27 (73%) | |
| Nodule size on U/S (cm) | <1 | 1 (3%) |
| 1–4 | 29 (78%) | |
| >4 | 7 (19%) | |
| Unknown | 0 (0%) | |
| Suspicious U/S? | Yes | 20 (54%) |
| No | 14 (38%) | |
| Unknown | 3 (8%) | |
| History of XRT? | Yes | 1 (3%) |
| No | 36 (97%) | |
| Bethesda category | AUS | 13 (35%) |
| Follicular neoplasm | 11 (30%) | |
| Suspicious | 13 (35%) |
U/S, Ultrasound; XRT, radiation therapy; AUS, atypia of undetermined significance.
| Bethesda category | Total patients | Malignant on frozen section | Malignant on final pathology | Malignancy rate |
|---|---|---|---|---|
| Non-diagnostic | 4 | 2 | 3 | 75% |
| Benign | 11 | 2 | 3 | 27% |
| AUS | 13 | 4 | 7 | 54% |
| Follicular neoplasm | 11 | 6 | 7 | 64% |
| Suspicious | 13 | 12 | 12 | 92% |
| Malignant | 22 | 19 | 22 | 100% |
AUS, atypia of undetermined significance.
Of the 74 total patients who underwent frozen section analysis, 45 (61%) were determined to be positive for malignancy intraoperatively. Only 1 of these 45 patients (2%) had a false positive; there was no evidence of malignancy on final pathology. In contrast, of the 29 patients who had no evidence of malignancy on initial frozen section, 10 patients (34%) had evidence of carcinoma on final histological analysis. The sensitivity, specificity, positive predictive value, and negative predictive value for a thyroid frozen section for all patients with respect to its prediction for malignancy was 81%, 95%, 98%, and 66%, respectively, with a diagnostic accuracy of 85%.
There were 37 patients with cytologically indeterminate FNABs who subsequently underwent thyroid lobectomy or total thyroidectomy with intraoperative frozen section. Twenty-two of the 37 specimens (59%) were read as positive for malignancy intraoperatively. One of the 22 patients (5%) had a false positive, while 5 of the 15 patients with no evidence of carcinoma on frozen section ultimately were diagnosed with thyroid cancer on final pathology (33%). The sensitivity, specificity, positive predictive value, and negative predictive value for a thyroid frozen section on a previously cytologically indeterminate nodule with respect to its prediction for malignancy was 81%, 91%, 95%, and 67%, respectively, with a diagnostic accuracy of 84%.
Thirty-one total patients underwent completion thyroidectomy after the results of the intraoperative frozen section analysis were obtained on an initial thyroid lobectomy specimen. While 9 of the 31 completion thyroid specimens (29%) showed evidence of carcinoma in the contralateral lobe, 8 of the 9 were microcarcinomas ≤5 mm in size; the only contralateral malignant nodule >1 cm in size was in a patient with a 7.7 cm thyroid carcinoma with extensive extrathyroidal invasion. The false positive and false negatives resulted in 1 completion thyroidectomy for benign pathology and 3 reoperations for malignancy not discovered on frozen section, respectively.
4. Discussion
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
The widespread use of a 6-tiered Bethesda classification system, along with the corresponding ATA guidelines, has greatly streamlined the cytopathologic diagnosis and subsequent management of thyroid masses and nodules. However, the cytologically indeterminate Bethesda categories still pose a challenge. One proposed solution to determine the extent of surgery is intraoperative frozen section analysis on thyroid lobectomy specimens. Proponents of frozen sections cite the high specificity and diagnostic accuracy when combined with preoperative FNAB [[10], [11]]. However, several large studies have found persistently low sensitivity and negative predictive value rates for frozen sections [[12], [13]]. This is in large part due to the inability to accurately diagnose follicular and Hurthle cell tumors, as many reviews have shown limited to no utility in performing frozen sections on follicular lesions [[4], [5], [12], [14]]. It should be noted that all of the above studies were performed prior to the adoption of the Bethesda system.
Recent reviews of frozen section analysis on thyroid specimen after Bethesda stratification have shown mixed results. Heterogeneous populations and highly variable, unpublished malignancy rates have led to a lack of consensus regarding frozen sections in this era. For example, while the positive predictive value of thyroid frozen sections has historically been high, one study showed a PPV of only 40%, likely due to the low rate of malignancy in their cohort [15]. Several of these reviews have shown low sensitivity for frozen sections, ranging from 22 to 51% [[15], [16], [17]]. Our study of patients undergoing both preoperative FNAB and intraoperative frozen section analysis on thyroid nodules since 2009 shows a high specificity and positive predictive value for frozen sections, but a relatively low sensitivity and a low negative predictive value for the test. Our results confirm prior longitudinal reviews that show prohibitively high false negative rates of intraoperative thyroid frozen sections. Interestingly, our results for the 37 patients with cytologically indeterminate nodules mirror those for the entire cohort. While frozen sections are less frequently performed for nodules with Bethesda II and VI classifications due to higher accuracy, this suggests that false negatives remain a concern, likely due to limited sampling and other technical and procedural concerns inherent to frozen section analysis [18]. It should be noted that the malignancy rate in our cohort was likely higher than the aforementioned reviews and higher than would be expected in a general population within each Bethesda category [3], owing to the fact that this selective cohort was recommended to undergo surgery largely due to a higher clinical degree of suspicion for malignancy. While a population with a lower prevalence of malignancy may have a lower false negative rate, the false positive rate would conversely increase, leading to an increasing number of unnecessary completion thyroidectomy procedures.
Intraoperative frozen sections have been used in the past in thyroid surgery to determine the need for completion thyroidectomy after initial lobectomy. While a positive frozen result would presumably lead the surgeon to remove the remaining thyroid lobe, the indications for doing so have become less clear than in prior years. The updated 2015 ATA guidelines in Recommendation 35 state that thyroid lobectomy can be used in properly selected patients with low-risk papillary thyroid carcinoma measuring 1–4 cm, in an effort to reduce recurrent laryngeal nerve damage, hypocalcemia, and thyroid hormone replacement [9]. The majority of pathologic factors that would lead one to perform total thyroidectomy such as aggressive histology and microscopic extrathyroidal extension are not reliably diagnosed on frozen sections [[19], [20]]. Patients with lymph node metastases for whom radioactive iodine would be indicated can often reliably be diagnosed on preoperative ultrasound and FNAB of the lymph node, thus thyroid frozen sections are rarely indicated. The multifocal nature of papillary thyroid carcinoma may lead one to consider completion thyroidectomy given the risk of contralateral disease; indeed, our review showed evidence of PTC in 29% of contralateral lobes. However, nearly all of these were microcarcinomas that would not have been detected otherwise, for which the ATA no longer recommends a FNAB even if there is evidence of sonographic suspicion [9]. The only patient who had a clinically significant contralateral nodule had an aggressive, T4a carcinoma for which total thyroidectomy was planned at the outset.
With a seemingly limited utility for intraoperative frozen section analysis in the era of the Bethesda classification system, there remains a need to attempt to risk stratify patients and avoid either subsequent or unnecessary operations. The 2015 ATA guidelines have identified and classified 5 basic sonographic risk patterns and their approximate risks of malignancy. Combined with a patient's FNAB result, universally standardized sonographic findings should assist clinicians in their management decisions and serve as a foundation for further research [21]. Another solution that may be used as an adjunct to FNAB is preoperative molecular testing. Molecular testing has been shown to improve the sensitivity of FNAB of a thyroid nodule under the Bethesda classification [[22], [23], [24]]. While there is a lack of consensus regarding the appropriate test and the cost of testing may be high, reviews have shown that molecular testing may eliminate the need for intraoperative frozen sections on cytologically indeterminate nodules [[7], [25]]. Molecular tests should be considered in clinical scenarios where the results would affect the management strategy.
Our study has limitations that deserve mention. This is a retrospective study with a relatively small sample size compared to past thyroid pathologic reviews due to the inclusion only of thyroid nodules stratified preoperatively by the six-tiered Bethesda classification, which is still in its relative infancy. In addition, the prevalence of differentiated thyroid cancer at our institution may not be equivalent to that at other institutions, which would undoubtedly affect the PPV and NPV of frozen section analysis. Despite this, our single-institution review of preoperative, intraoperative, and postoperative thyroid pathology demonstrates important limitations of frozen section analysis in the context of the Bethesda system, the 2015 ATA guidelines and recent advances in differentiated thyroid cancer management.
5. Conclusion
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
While intraoperative frozen sections on thyroid specimens with a classification of atypia, follicular neoplasm, or suspicious for malignancy on prior fine needle aspiration may be helpful if positive, the false negative rate remains high. There appears to be limited value in routine frozen sections to guide clinical management and decision-making in the era of the Bethesda system.
References
Jump to Section1. Introduction2. Materials and methods3. Results4. Discussion5. ConclusionReferences
- Davies, L. and Welch, H.G. Current thyroid cancer trends in the United States. JAMA Otolaryngol Head Neck Surg. 2014; 140: 317–322
- View in Article
☆Previous presentation: Preliminary data from this study were presented as a poster presentation at the Triological Society Combined Sections Meeting on January 21, 2017 in New Orleans, LA, USA.
☆Disclosures: The authors report no conflicts of interest or financial disclosures.
© 2017 Elsevier Inc. All rights reserved.
from ! Medicine by Alexandros G. Sfakianakis via Αλέξανδρος Σφακιανάκης on Inoreader http://ift.tt/2tKUjgh
via IFTTT
Δεν υπάρχουν σχόλια:
Δημοσίευση σχολίου
Medicine by Alexandros G. Sfakianakis,Anapafseos 5 Agios Nikolaos 72100 Crete Greece,00302841026182,00306932607174,alsfakia@gmail.com,