INTRODUCTION
Pathology intraoperative consultation is a high-risk procedure performed by the pathologist to obtain a tissue diagnosis at the request of the surgeons with important consequences (1). It requires not only experience but also knowledge of clinical medicine and pathology, as well as the ability to make judgements and be aware of the limita-tions (2). The frozen section technique was developed for immediate intraoperative pathologic diagnosis by Wilson and MacCarty (3) in 1905 at the Mayo Clinic. Frozen section practice has been developed and increased in number since the invention of cryostat in 1959 (4,5). Numerous studies have confirmed the high rate of diagnostic accuracy routinely achieved with the frozen section method (6). The indications and limitations of frozen section diagnosis depend on the individual organ and may show variations. The principal purposes of a frozen section are to determine the presence of a lesion, to define surgical margins and to establish whether the sample contains diagnosable material (2). Therefore, it is important, and sometimes critical to determine the efficiency of frozen section performance periodically.
MATERIALS AND METHODS
This study included frozen sections of 552 specimens from 401 cases performed between June 2001 and January ABSTRACT
Intraoperative consultation is a high-risk procedure of pathology departments with important consequences. There-fore, it is critical to determine institutional efficiency of frozen section performance periodically. Frozen section diagnoses of 552 specimens from 401 cases between June 2001 and January 2005 were rewieved and correlated with subsequent histopathological examination, to assess concordant, discor-dant and deferred diagnoses as well as false negative and false positive rates and to determine reasons for discordance. Each individual specimen from the same case was considered and evaluated as one case. Excluding deferred diagnosis constituting 4.53% of the cases, accuracy, false negative and false positive rates were consecutively found 97.47%, 2.08%, and 0.57%. Interestingly, a great proportion of false negative cases were found to be associated with papillary carcinoma of thyroid. Gross sampling error and misinterpretation appeared as the leading reasons for frozen section discordance. Considering discordant frozen section diagnoses have a great impact on patient’s care, intra and interinstitutional monitoring of frozen section performance may serve as a quality control programme.[Turk J Cancer 2006;36(4):157-161].
KEY WORDS:
Frozen section, pathology, diagnosis
Frozen section experience with
emphasis on reasons for discordance
fiÜKRÜ O⁄UZ ÖZDAMAR1, BURAK BAHADIR1, TULU EMRE EKEM1, GÜRKAN KERT‹fi1, BANU DO⁄AN GÜN1, GAMZE NUMANO⁄LU1, ZAFER YÜNTEN2, GAMZE MOCAN KUZEY1
2005. The total biopsy count was 11,964. All frozen sections were reevaluated by three pathologists. Cases with more than one sample were considered as one. The definitions previously described and used in this study are as follows (7):
Case: A single per-patient episode (accession case number) for which tissue was submitted for examination and reporting. The tissue may or may not have originated from the operating room. A case may have contained more than one specimen. More than one specimen may have been evaluated by frozen section in a single case, and an individual specimen may have been evaluated by one or more frozen sections; a frozen section diagnosis may have required one or more frozen section blocks.
Number of Frozen Sections: Total number of all frozen section blocks examined.
Frozen Section Diagnosis: Diagnosis rendered on a frozen section consultation, which may require one or more frozen section blocks.
Frozen Section Block: Tissue frozen on one chuck for which a corresponding slide was prepared for histologic evaluation.
Frozen Section Specimens: Number of specimens for which at least one frozen section was performed.
Frozen Section Cases: Number of cases for which at least one frozen section was performed.
Adequate Frozen Section: Frozen section study done, and frozen section diagnosis rendered with a diagnostic agreement or disagreement.
Concordance: An adequate frozen section study and diagnostic agreement with the permanent sections.
Discordance: An adequate frozen section study and diagnosis disagreement with the permanent sections.
Deferred Diagnosis: Diagnoses that were indeterminate at the time of frozen section examination. These were not considered discordant diagnoses.
RESULTS
Distribution of specimens for frozen section according to clinical departments is shown in Graphic 1. The overall frozen section consultation rate was 3.35%. The concordant frozen section diagnosis, discordant frozen section diagnosis, and deferred diagnosis are displayed in Table 1. Excluding deferred diagnosis accuracy, false-negative and false-positive rates were consecutively found 97.47%, 2.08%, and 0.57%. Reasons for discordances
Table 2 shows the reasons for discordant diagnoses. The majority of frozen section discordances were associated with gross sampling error followed by misinterpretation of the original frozen section. Frozen section of one case with discordance had no diagnostic features at the time of frozen section examination, but permanent sections from the frozen block revealed evident malignancy.
False-negative and false-positive diagnoses Table 3 summarizes the overall negative and false-positive frozen section diagnoses of malignancy and the correlation with discordance. False-positive diagnoses of neoplasm accounts for 21.43% of the discordant frozen section diagnosis, and 78.57% of discordances were due to false-negative diagnoses of neoplasm.
Anatomic sites
Of the 14 discordant frozen section diagnoses, 9 were from the thyroid. The remaining 5 belonged to ovary, peritoneum, skin, brain, and bone, each as single cases.
Corrected diagnoses of discordances Table 4 lists the correlation between discordant frozen section diagnoses and corrected diagnoses. The leading corrected diagnosis was papillary carcinoma of thyroid, which constitutes 9 cases of 14 discordant diagnoses. Deferred Diagnosis
Table 5 illustrates anatomic sites and histopathological diagnoses of the deferred frozen section diagnoses. The most common site for deferred diagnosis was thyroid with four cases.
Table 1
Concordant, discordant, and deferred frozen section diagnoses
n %
Concordant Frozen Section Diagnoses 513 97.47*
Discordant Frozen Section Diagnoses 14 2.53*
Deferred Frozen Section Diagnoses 25 4.53
TOTAL 552 100.00*
* Excluding deferred diagnoses
Table 2
Reasons for discordant diagnoses
Gross sampling error 50.00%
Misinterpretation of the original frozen section 42.86%
Frozen section negative, permanent sections from frozen block positive 7.14%
Table 3
False-positive and false-negative diagnoses
Type of discordance n % of All Discordant Diagnoses % of All Frozen Section Diagnoses*
False-positive 3 21.43 0.57
False-negative 11 78.57 2.08
probably patients’ outcome. Frozen sections are performed on a wide variety of sites and organ systems.
The discordance rate appears to vary between sites and organ systems and depending on the method chosen for calculation; i.e. frozen section cases, specimens or blocks DISCUSSION
Frozen section is an intraoperative consultation carrying a high risk of misdiagnosis, and thus responsibility. Although sometimes misused by some surgeons, it supplies important information about the nature of the suspected lesions and
Table 4
Corrected diagnoses of discordant frozen section diagnoses
Anatomic Site n Frozen Section Diagnosis Corrected Diagnosis
Thyroid 4 Benign Nodular Goiter Follicular Variant of Papillary Carcinoma Thyroid 2 Benign Nodular Goiter Papillary Microcarcinoma
Thyroid 1 Benign Nodular Goiter Papillary Carcinoma
Skin 1 Foreign Body Reaction Malignant Fibrous Histiocytoma Ovary 1 Negative for Malignancy Metastatic Signet Ring Cell Carcinoma Peritoneum 1 Negative for Malignancy Undifferentiated Carcinoma
Brain 1 Negative for Malignancy Dysembryoplastic Neuroepithelial Tumor Thyroid 2 Suspected for Malignancy Benign Nodular Goiter
Bone 1 Suspected for Malignancy Osteomyelitis
Table 5
Anatomic sites and histopathologic diagnosis of the deferred frozen section
Anatomic Site n Histopathologic Diagnosis
Thyroid 1 Papillary Carcinoma
Thyroid 4 Benign Nodular Goiter
Uterus 2 Epithelioid Leiomyoma
Lymph Node 6 Reactive Hyperplasia
Mediastinum 1 Thymic Lymphoid Hyperplasia
Brain 1 Fibrillary Astrocytoma
Small Bowel 2* High Grade B Cell Lymphoma
Brain 1 Medulloblastoma
Ureter 1 Urothelial Carcinoma
Rectum 1 Adenocarcinoma
Ovary 1 Brenner Tumor
Middle Ear 1 Paraganglioma
Pleura 2* Scar Tissue
Ovary 1 Borderline Serous Cystadenofibroma
(7). Dankwa and Davies (8) suggested an irreducible minimum of 2% for discordances based on frozen section diagnoses. In this study, the discordant diagnosis rate of 2.53% was slightly higher, and concordant diagnosis rate of 97.47% was slightly lower than some previous reports (7,8). 21.4% of discordances were due to false-positive frozen section diagnoses in the current study and represented a higher rate compared to the previous studies. For speci-mens of the current study, the positive and false-negative rates were also somewhat higher but within the reported limits (7). Examination of frozen section deferral rates represents another parameter for quality assurance monitoring. Besides deferral rates may vary according to diagnostic expertise, rates ranging from 2.9% to 4.2% have been described (7). Moreover frozen section deferral may also be associated with the types of the specimen and resections as we encountered.
Nine of the 14 cases in this study with discordance were from thyroid gland, with 7 cases of false-negative
and 2 cases of false-positive. In addition, 4 of the 25 deferred diagnoses were from thyroid. Only one case of follicular variant of papillary carcinoma was correctly identified in our experience. As reported in the literature, frozen section evaluation of thyroid lesions may sometimes be challenging and has limited utility (9-13). Moreover, thyroid accounts for most of the deferred frozen section diagnoses as noted in the current study (9). Some authors also suggested that the majority of the frozen sections from follicular thyroid lesions should be deferred (9). If frozen sections from thyroid glands had been excluded from this study, the concordant and discordant rates would have been 98.6% and 1.4%, consecutively.
In summary, considering discordant frozen section diagnosis may cause important consequences, pathologists should be aware of the reasons for frozen section discordance and the limitations of frozen section practice. As a result, monitoring of frozen section performance may serve as a quality control programme.
References
1. Zarbo RJ, Schmidt WA, Bachner P, et al. Indications and immediate patient outcomes of pathology intraoperative consultations. Arch Pathol Lab Med 1996;120:19-25. 2. Rosai J. Introduction. In: Rosai J, editor. Rosai and Ackerman’s
Surgical Pathology, 9th ed. Philadelphia: Mosby, 2004:1-23. 3. Jennings ER, Landers JW. The use of frozen section in cancer
diagnosis. Surg Gynecol Obstet 1957;104:60-2.
4. Challis D. Frozen section and intra-operative diagnosis. Pathology 1997;29:165-74.
5. Sawady J, Berner JJ, Siegler EE. Accuracy of and reasons for frozen sections: A correlative, retrospective study. Hum Pathol 1988;19:1019-23.
6. Zarbo RJ, Hoffman GG, Howanitz PJ. Interinstitutional comparison of frozen-section consultation: a College of American Pathologists Q-Probes study of 79 647 consultations in 297 North American Institutions. Arch Pathol Lab Med 1991;115:1187-94.
7. Gephardt GN, Zarbo RJ. Interinstitutional comparison of frozen section consultations. A College of American Pathol-ogists Q-Probes study of 90,538 cases in 461 institutions. Arch Pathol Lab Med 1996;120:804-9.
8. Dankwa EK, Davies JD. Frozen section diagnosis: an audit. J Clin Pathol 1985;38:1235-40.
9. Montone KT, LiVolsi VA. Frozen section analysis of thy-roidectomy specimens: experience over a 12-year period. Pathol Case Rev 1997;2:241-5.
10. Lin HS, Komisar A, Opher E, et al. Follicular variant of papillary carcinoma: the diagnostic limitations of preoperative fine-needle aspiration and intraoperative frozen section evaluation. Laryngoscope 2000;110:1431-6.
11. Shen PUF, Kuhel WI, Yang GCH, et al. Intra-operative touch-imprint cytological diagnosis of follicular variant of papillary thyroid carcinoma. Diagn Cytopathol 1997;17:80-3. 12. Tielens ET, Sherman SI, Hruban RH, et al. Follicular variant
of papillary thyroid carcinoma: a clinicopathologic study. Cancer 1994;73:424-31.
13. Mandell DL, Genden EM, Mechanick JI, et al. Diagnostic accuracy of fine-needle aspiration and frozen section in nodular thyroid disease. Otolaryngol Head Neck Surg 2001;124:531-6.
