Frequency. Powdery chromatin. Nuclear grooves. Nuclear overlapping. Intranuclear cytoplasmic inclusions*. Papillary structures. Giant cells. Psammoma bodies.
Anatomic Pathology / ORIGINAL ARTICLE
Comparison of Intraoperative Cytology With Frozen Sections in the Diagnosis of Thyroid Lesions Joseph A. Tworek, MD, Thomas]. Giordano, MD, PhD, and Claire W. Michael, MD Key Words: Intraoperative cytology; Thyroid; Neoplasms
Abstract
Frozen section (FS) analysis of thyroid masses is a common procedure and often is used to guide definitive surgical management. To our knowledge, there have been few studies of the usefulness of intraoperative cytology (IOC) as an adjunct to FS in the rapid diagnosis of thyroid lesions. One study was in abstract form at the time of writing this study.1 Another study solely addresses the usefulness of IOC in distinguishing colloid nodules from follicular adenomas and does not fully specify the cytologic preparations that were used.2 It is our goal to determine the usefulness of IOC as an adjunct to FS in the diagnosis of thyroid lesions.
Materials and Methods Case Selection Sixty-eight thyroidectomies performed between January 1, 1994 and September 1, 1996 were obtained from our departmental files. Cases were selected if both FS analysis and IOC had been performed. By using permanent sections as the "gold standard," the accuracy of diagnoses given by IOC was compared with the accuracy of diagnoses made by FS. Fisher's exact test was used for statistical comparisons. Papillary thyroid carcinomas that were diagnosed on permanent sections and not sampled by IOC and FS were not considered to be a misdiagnosis by IOC or FS. Cytology
IOC consisted of touch imprints or directed scrape preparations on lesions that were subsequently submitted for FS analysis. The cytologic preparations were alcohol fixed and stained by H&E. In each case, the IOC was reviewed retrospectively without knowledge of the gross appearance of the specimen of the FS and permanent section diagnoses.
456
AmJCIinPathol 1998;110:456-461
© American Society of Clinical Pathologists
Downloaded from http://ajcp.oxfordjournals.org/ by guest on April 10, 2016
We retrospectively studied the usefulness of intraoperative cytology (IOC) and frozen section (FS) in the rapid diagnosis of 68 thyroid lesions. In 14 cases of papillary thyroid carcinoma, IOC correctly diagnosed 13 cases, while FS correctly diagnosed 11 cases. There was no significant difference in sensitivities, and both methods had similar specificities. In 21 cases of colloid nodule, IOC was slightly more sensitive than FS; IOC correctly diagnosed 16 cases, while FS correctly diagnosed 15 cases. However, the specificity of IOC was only 71%, but was 98% for FS. Of 17 follicular adenomas, FS diagnosed 16 as follicular neoplasms and misdiagnosed only 1 as a colloid nodule. By contrast, IOC misdiagnosed 9 follicular adenomas as colloid nodules, most of which were macrofollicular variants with abundant colloid. Of 11 follicular carcinomas, FS diagnosed all as follicular neoplasms, while IOC misdiagnosed 3 as colloid nodules. While IOC is not as accurate as FS in the diagnosis of colloid nodules and follicular neoplasms, it is highly sensitive and specific in the diagnoses of papillary carcinoma and performance of the technique is rapid and easy. In an intraoperative setting, IOC is a useful adjunct to FS in screening thyroid nodules for the presence of papillary carcinoma.
Anatomic Pathology / ORIGINAL ARTICLE •Table 11 Diagnoses Made on Permanent Sections Compared With Diagnoses Made on IOC and FS* Permanent Section Diagnosis Colloid nodule (n = 21) Follicular adenoma (n = 17) Follicular carcinoma (n = 11) Papillary carcinoma (n = 14)
Colloid Nodule 16 9 3 0
Follicular Neoplasm
vs15 vsl vsO vsO
4 8 8 1
vs6 vs16 vs 11 * vs3
Papillary Carcinoma 1 0 0 13
vsO vsO vsO vs11
IOC = intraoperative cytology; FS = frozen section. *lncluding oncocytic lesions. Data are given as intraoperative cytology vs frozen section. 7 In 2 cases, the diagnosis of follicular carcinoma was correctly made on frozen section.
•Table 21 Frequency of Cytologic Features in Papillary Thyroid Carcinoma Frequency
Cytologic Features Powdery chromatin Nuclear grooves Nuclear overlapping Intranuclear cytoplasmic inclusions* Papillary structures Giant cells Psammoma bodies
Sticky colloid
14/14 14/14 13/14 13/14 9/14 5/14 2/14 1/14
Cytoplasmic inclusions were sometimes scarce.
extracapsular vascular invasion, or both. Papillary carcinomas were defined primarily on their nuclei, which were enlarged and overlapping and contained clear chromatin, abundant nuclear grooves, and cytoplasmic inclusions. Oncocytic lesions were diagnosed when a majority of the lesion was composed of cells with abundant eosinophilic granular cytoplasm. These lesions were classified according to the previously defined categories.
Histology In each case, the FS also was interpreted without knowledge of the gross appearance of the specimen and the diagnosis given for the permanent section. Our FS diagnostic categories were similar to those used for IOC with the exception of the rare cases in which capsular invasion was apparent in a follicular neoplasm, enabling a diagnosis of follicular carcinoma. On permanent section, our diagnostic categories included colloid nodule, follicular adenoma, follicular carcinoma, papillary carcinoma, and other. We followed generally accepted criteria in establishing diagnoses by FS and permanent sections. Colloid nodules were unencapsulated, contained follicles of varying sizes, which were similar to follicles elsewhere in the thyroid, and often had regressive changes, such as hemorrhage. By contrast, follicular adenomas were well encapsulated and were architecturally distinct from the adjacent uninvolved thyroid gland. Follicular carcinomas were similar histologically to follicular adenomas but had established capsular invasion,
© American Society of Clinical Pathologists
Results Papillary Carcinoma The diagnosis of papillary carcinoma made by IOC was more sensitive, correctly diagnosing 13 of 14 cases, than FS, correctly diagnosing 11 of 14 cases ITable II. However, the difference was not statistically significant (f
