I. INTRODUCTION. Intraoperative consultations fall into two general categories. Microscopic consultations, usually performed as frozen sections, are undertaken to establish a tissue diagnosis, determine the nature of a lesion that may require ancillary testing, establish that sufficient diagnostic tissue has been obtained, identify metastatic disease, and assess surgical margins or extent of disease. Microscopic consultation can also be performed using touch preparations, a practice that has the advantage of preserving valuable tissue. Nonmicroscopic consultations are gross examinations of a specimen that provide the surgeon with real-time information on tissue margins and the anatomic extent of disease processes. They also facilitate the triage of fresh tissue for ancillary diagnostic studies or research protocols.
II. FROZEN SECTIONS. High-quality frozen sections can be performed with remarkable speed if equipment is kept in optimum working condition and if the operator is well versed in the technique. In experienced hands, the entire consultation can often be performed in 10 to 15 minutes from the time of the arrival of the specimen in the frozen section room to the notification of the surgeon of the diagnosis. For larger tissue specimens, proper interpretation requires a thorough gross examination of the tissue before sectioning. Good communication with the surgeon regarding operative findings and knowledge of pertinent clinical history are also absolutely essential for optimization of the process.
A. Indications. Frozen sections are indicated to establish a tissue diagnosis (such as the presence of malignancy, which will guide intraoperative patient management and extent of surgery); for tissue identification (e.g., to confirm the presence of parathyroid tissue in a parathyroidectomy specimen); to determine the nature of a lesion that may require ancillary testing that requires special fixatives or media (e.g., RPMI for flow cytometry or glutaraldehyde for electron microscopy); to establish that sufficient diagnostic tissue has been obtained; to identify metastatic disease; and to assess surgical margins or extent of disease.
Frozen sections should not be used merely to satisfy a surgeon’s curiosity, to compensate for inadequate preoperative evaluation, or as a mechanism to communicate information more quickly to the patient or the patient’s family.
B. The frozen section procedure. Frozen sections are performed by freezing the tissue in a block of specialized embedding medium, followed by cutting thin (usually 5 µm) sections from the block using a cryostat (refrigerated microtome). The sections are adhered to glass slides, fixed in ethanol, and stained with hematoxylin
and eosin (H&E). Small specimens may be completely utilized for frozen section slide preparation, but if possible, a portion of the tissue should be preserved for routine handling to avoid freezing artifacts that can compromise interpretation of the permanent sections (e.g., in the case of brain biopsies). For larger tissue samples, judgment must be exercised in gross sampling so that the area(s) of highest diagnostic yield is selected for frozen section. Cytological imprints from tissues can be an important adjunct in diagnosis, especially for hematolymphoid abnormalities, lymph node biopsies, and thyroid lesions.
C. Interpretation. The interpretation of frozen sections requires integration of the histologic morphology in the H&E-stained sections; the gross features of the specimen; information from the surgeon regarding the origin of the tissue, the indication for the consultation (including the clinical history, radiological findings, and intraoperative observations); and the ways in which the frozen section diagnosis will affect the operative strategy. In some difficult cases, it may be necessary to request additional tissue for frozen section analysis. If additional tissue cannot be obtained, deferral of a definitive diagnosis pending examination of formalin-fixed paraffin-embedded sections is acceptable. In routine clinical practice, such deferrals are employed in <5% of frozen section diagnoses.
D. Communication of findings. Clear communication of a concise diagnosis to the surgeon is the last step of the intraoperative consultation. Usually, there is a narrowly defined clinical problem that frozen section is to solve; this should be specifically and unambiguously addressed in the diagnosis. The diagnosis is written and signed, with the date and time, by all interpreting pathologists and made part of the final pathology report. If the written diagnosis is verbally transmitted, the pathologist should first confirm with the recipient of the information the identity of the patient (using two identifiers) and the surgeon. The operating room staff member taking the diagnosis should repeat it back to the pathologist to confirm accurate communication.
Certain phraseologies tend to be resistant to misinterpretation. For instance, “negative for malignancy” or “positive for malignancy” are generally understood well. Since some diagnosis can easily be incompletely heard or misunderstood in a busy operating room, the diagnosis should always be repeated back as a safeguard.
E. Accuracy of frozen sections. The accuracy of frozen sections will vary from institution to institution on the basis of the types of surgical cases evaluated and the experience of the involved pathologists. Table 52.1 highlights the fact that the accuracy of frozen section diagnosis is dependent on the anatomic site. Regular self-audits of the frozen section service are desirable so that surgeons and pathologists are aware of the performance characteristics of the modality in their own hands. Such audits of single institutions, and pooled data across hundreds of institutions, show that accurate diagnoses are made overall in >95% of cases, while discordance with the final diagnosis occurs in 1% to 2% of cases. Deferral of the diagnosis until permanent section diagnosis occurs in 1% to 4% of cases.
F. Sources of error in frozen sections. Errors can be divided into errors of interpretation and errors of sampling; both usually result in false-negative diagnoses. False-positive diagnoses are rarer, likely because experienced pathologists tend to appropriately defer to permanent section rather than making an incorrect diagnosis on substandard material.
Misinterpretation accounts for about 40% of errors overall (Arch Pathol Lab Med. 1996;120:804; Arch Pathol Lab Med. 1996;120:19). Interpretations of frozen sections are more prone to this error than interpretation of permanent sections due to the presence of artifacts that are not encountered in routinely fixed, paraffin-embedded sections. Some tissues are more likely to show significant artifact, especially those with high fat content, such as most pelvic lymph nodes.
TABLE 52.1 Examples of Frozen Section Evaluation
Tissue
Concordance with permanent section diagnosis
False negative Comments on utility
Breast
Limited (see text)
Cervix
73% for evaluation of dysplasia
Poor for evaluation of dysplasia
Gallbladder
95% when used to evaluate a mural lesion
Useful in the rare instances in which it is required
Liver
Diagnostic dilemmas that generate deferrals:
1. Hamartoma vs. cholangiocarcinoma
2. Regenerative nodule vs. hepatocellular carcinoma
3. Adenoma vs. hepatocellular carcinoma
Lung
99%
Useful; deferral rate of only 3-4%
Axillary sentinel lymph nodes for metastatic breast cancer
Possibly useful (see text)
Lymph nodes for staging
˜100%c
20-40%c
Limited and dependent upon the lymph node location (see text)
Ovary
92%
5%
Useful; errors are disproportionately represented among mucinous tumors
Pancreas
98% when used for diagnosis of primary lesion; almost 100% for margins
1%
Atypical ductal structures, especially in the setting of pancreatitis, may mimic carcinoma; deferral rate of 6-7%
Parathyroid
99% (for parathyroid vs. nonparathyroid tissue)
Useful for distinction of parathyroid versus nonparathyroid tissue; inadequate for diagnosis of parathyroid carcinoma, or for differentiating adenoma from hyperplasia
Skin, melanoma
Up to 50%
Strongly discouraged (see text)
Skin, nonmelanoma
About 2%, based on recurrence rates following Mohs’ microsurgeryd
Useful for the evaluation of margins in the resection of tumors with infiltrating borders and tumors of the face, especially eyes, ears, and nose.
Thyroid
98%
10%
Inadequate for follicular lesions, as the sampling required is not practical (see text); better for lesions with papillary architecture; deferral rate is 6%
a When nodes with submicrometastases (<0.2 mm) are considered “positive.”
b When nodes with submicrometastases are considered “negative.”
c Staging pelvic lymph nodes at prostatectomy.
d Data on correlation with permanent sections is sparse, as follow-up permanent sections are not performed.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
Frozen Sections and Other Intraoperative Consultations
Frozen Sections and Other Intraoperative Consultations
Michael E. Hull
Peter A. Humphrey
John D. Pfeifer
