Frozen section and the surgical pathologist: a point of view.
Abstract: * Frozen section is a prominent point of intersection between surgeons and pathologists. It is regarded as the most definitive--but not the sole--form of intraoperative consultation. Its role in tissue triage, diagnosis, and intraoperative management should not be misconstrued as a shortcut to a definitive diagnosis. Although the pathologist remains in control of the tissue disposition, frozen sections are ideally requested and executed as a collaborative effort. Frivolous requests with no direct consequences for the conduct of a given procedure should not be honored. Frozen section plays a material role in resident education and may be the last vestige of general surgical pathology in an era of organ system specialization. Frozen section will retain its relevance only in the context of broad clinical knowledge by the pathologist and judicious utilization by the surgeon, both in the ultimate service of the patient.
Article Type: Report
Subject: Histology, Pathological (Practice)
Pathologists (Powers and duties)
Breast cancer (Diagnosis)
Immunohistochemistry (Usage)
Immunohistochemistry (Analysis)
Author: Taxy, Jerome B.
Pub Date: 07/01/2009
Publication: Name: Archives of Pathology & Laboratory Medicine Publisher: College of American Pathologists Audience: Academic; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2009 College of American Pathologists ISSN: 1543-2165
Issue: Date: July, 2009 Source Volume: 133 Source Issue: 7
Topic: Event Code: 200 Management dynamics
Accession Number: 230152046
Full Text: Surgical pathology as we currently recognize it was initially conceived and developed by surgeons. Concurrent with its development was the idea of a rapid intraoperative diagnosis, which was first attempted at The Johns Hopkins Hospital a little more than 100 years ago on a breast biopsy. (1) It took awhile for this new technique to gain popularity, but by the 1920s the use of frozen section as an intraoperative guide for the surgeon was deemed essential. (2) The technology has changed over the years: there are cryostats capable of rapid freezing, built-in microtomes sectioning at 5 [Am or less, better stains, and improved light microscope optics. The technical advances and the variations of implementation among present-day institutions notwithstanding, the essentials of the preparation and reporting of a frozen section would remain recognizable to our predecessors. This paper presents a personal statement on the contemporary use of frozen section based on my own experience in both university and community hospital settings. The examples cited are tumors, but the principles of examination and reporting are easily extrapolated to nonneoplastic and inflammatory conditions.

The receipt of specimens varies, including direct pickup by the pathologist/pathology resident, delivery by operating room personnel, or a tube system. The physical space for performing frozen sections also varies, ranging from a dedicated room within the pathology laboratory to a separate space proximate to the operating rooms. The potential questions would primarily still be: what is it, where is it, and is there enough tissue? The communication of results may be by telephone or intercom, or it may require the pathologist to return to the operating room and verbally deliver the result. The clarity of this communication is obviously essential. Direct transmission of the microscopic field via closed-circuit television or intranet access from pathology to surgery is available but not universal. In any event, some hard copy record of the results of the intraoperative examination (with or without an actual frozen section) will enter the patient's record. The frozen section interpretation is also incorporated into the final surgical pathology report, including an acknowledgment of any discrepancies with the final diagnosis and the resolution of those discrepant issues.

Although the general public may be unaware of the active role of the pathologist during a given operative procedure, intraoperative pathology consultations are a mainstay of clinical practice, mostly in the care of patients with tumors. A frozen section is one of several modalities, which also include gross specimen evaluation with appropriate triage, imprint preparation, or the actual appearance in the operating room by the pathologist to view the operative field and speak directly to the surgeon. The intimate interactions between surgeons and pathologists in the conduct of these activities should suggest that the construction of new operating rooms or surgical wings without appropriate provisions for the performance and reporting of intraoperative consultations, including frozen sections, reflects a lack of multidisciplinary understanding and is undertaken at a disservice to surgeons and patients.

The frozen section is regarded as the most definitive form of consultation because it employs the microscopic examination of tissue designated by the surgeon as important to the conduct of an operation already underway. The distribution of requests for frozen sections related to specific organ systems is institution driven based on the staffing of surgical specialists and patient referral patterns--in other words, the number and the types of specialists have a direct effect on the requests for frozen sections. For example, plastic surgeons would be expected to generate requests for margins on basal cell carcinoma, and head and neck surgeons would be expected to generate requests for margins on the resection or diagnosis of squamous carcinoma. The active and evolving demand for frozen section not only underscores it as an established parameter of medical practice but also reflects the changes in the clinical questions related to advances in surgical treatment, mostly for tumors in various organ systems. The clinical relevance of specific examinations, especially for cancers, is and will continue to be debated, but requests for frozen sections are a fact of life and embedded into all surgical training programs. Requests need not always be honored, because the responsibility for the nature and extent of the tissue examination belongs to the pathologist. Nonetheless, the role of the pathologist in this process is inherently collaborative, so that relevance issues as a function of practice patterns are best resolved between surgeons and pathologists on an individual institutional basis. This is perhaps best illustrated by the historic evolution of the use of frozen section in breast cancer.

Intraoperative consultations are a major responsibility in our role as physicians; the frozen section is essential to and part of the practice of every pathologist. As surgical pathologists, we expect the requests for frozen section to be selective, not frivolous; we expect that the consequences of the diagnoses will help to manage the operation; and we expect the surgeon to be prepared to make these adjustments when he or she initiates the request. The pathologist's awareness of the treatment algorithm is clearly essential. If there is no intent to modify the procedure, the frozen section should not be requested or, if requested, not done. The frozen section is a management tool, not a diagnostic shortcut. Therefore, the comprehensive analysis of a specimen, biopsy or major resection, by frozen section is not part of mainstream medical practice.

The requesting, performance, and reporting of frozen sections present a range of considerations, only some of which are discussed here. In addition to the above, there are 3 points deserving special emphasis. First, there is an obvious impact on the education of pathologists. Next, the technical artifacts of frozen section must be incorporated into our interpretative database. Finally, as a utilization issue, frozen sections affect the interactions between pathologists and clinicians and, ultimately, patient care.


Doing frozen sections is an inseparable part of any pathology training program. Currently, resident training in most university hospitals uses, at least in part, the approach of organ system specialization. However, many if not most residents end up as community hospital pathologists, where they are expected to have broad organ system expertise. How effective is this specialty training in emphasizing breadth of knowledge and experience? Has general surgical pathology teaching been abandoned? Do we expect residents to absorb specialty knowledge and integrate the general base on their own?

The need to function as a generalist may present a conflict for faculty members who, despite acknowledged areas of expertise, may be regularly required to do frozen sections involving organs outside their comfort zone. In addition, there may be issues of trust with clinical physicians outside the pathologist/specialist's accepted area(s) of expertise. There are no simple resolutions other than a very understanding clinical staff and, perhaps, readily available intradepartmental consultant colleagues when ever a frozen section is done. The latter might not be a problem during daytime hours, but is potentially problematic on nights and weekends. Perhaps we need to look more closely at the diagnostic and educational expectations for faculty in this context.

The frozen section may be the final vestige, perhaps uniquely, of general surgical pathology that has been retained by implicit general agreement as an obligatory, clinically relevant task for all department members. Obviously, any surgical procedure may present an unanticipated finding that needs to be resolved immediately by frozen section, or a known complex procedure may require ongoing intraoperative assistance for its successful completion. Admittedly, the diagnostic expectations in this setting are different and may lessen the pressure, because a specific diagnosis or tumor cell type may be less relevant than the adequacy of the sample, the status of the margin, or simply an assessment of the lesion as benign or malignant. A specific lesion classification may be less important. Therefore, it would be safe to conclude that in both university and community settings there is a prescient and continuing requirement for pathologists with broadly based clinical medical knowledge, a capacity for decision making, good judgment, ease in communication, and the possession of expert analytic morphologic skills. In this role, faculty members function as teachers of morphology, modelers of good communication, and builders of the respect and trust with clinical colleagues so essential for good patient care. The evolving and ongoing debate about whether pathology residencies should be training specialists, generalists, or some form of hybrid is perhaps high-lighted in daily practice by the frozen section.


Because it is intuitive that a frozen section taken from a nonrepresentative area will be clinically meaningless, the gross features of the disease process must be known, either by direct visualization of the specimen or operative field, a direct physician communication, or an informative requisition. In some cases, an intraoperative consultation could or even should be limited to an examination of the gross specimen, with triage for special studies as needed. Thyroid lobectomies are very suitable for gross examination without frozen section. Surgeons should be discouraged from opening specimens in the operating room without the pathologist being present and then asking for a frozen section.

In addition, imprints stained either by hematoxylin-eosin or Diff-Quik are greatly underutilized. Imprints are perhaps "tainted" by being a cytopathologic technique. In many university hospitals, cytopathology is a separate division often staffed by specialists, and the cytology laboratory may be physically remote from the frozen section room. Surgical pathology faculty members, who often do not do cytopathology, may not use imprints as diagnostic aids. Residents are undoubtedly influenced by this practice. Imprints are extremely useful in the sharp representation of individual nuclear and cytoplasmic detail that may not be achieved on a frozen section. Imprints and frozen sections have positive redundant value and should be used as complements, not as competing modalities.

Because the tissue is hardened by freezing and not by routine fixation, infiltration, and embedding, the frozen section histopathology may be distorted by folds, tears, and repeated knife marks (ie, "chatter"). These common artifacts are abetted by the microtome blade and/or an unsteady hand. These may be ameliorated by having well-sharpened instruments available and the acquisition of experience over time, respectively. Mature fat will not freeze, and attempts to obtain good diagnostic sections are often fruitless. Requests that involve freezing fat should not be honored. Tissues with substantial water content, such as the brain, often yield ice crystals during the actual freezing in the cryostat and result in nonrepresentative architecture of tumor growth or inflammatory infiltrate. Imprints, smears, or snap freezing with liquid nitrogen are often employed to mitigate these artifacts. Drying artifacts are many (eg, enlargement of nuclei, pallor of the chromatin, and nonuniform eosinophilia of the cytoplasm, to name just a few) and may be encountered on hematoxylineosin imprints or tissue sections not immediately immersed in fixative. Artifactually crushed cells often obscure the distinction between a small cell neoplasm and inflammation, or even lymphoma. Additional artifacts are encountered on the permanently embedded blocks of the frozen tissue, such as various architectural aberrations, distortions of cell size, abnormalities of chromatin clumping ("crush" artifact), difficulty in appreciating mitotic activity, and apoptotic cells. In the end, artifacts are unavoidable but also surmountable. Pathologists learn to recognize and incorporate these artifacts into the interpretation of the frozen samples, as well as anticipate their potential to obscure the definitive pathology. It is also imperative that, insofar as clinically feasible, additional non-frozen tissue be available for standard histopathologic study.


Artifacts notwithstanding, the accuracy of frozen section diagnosis across the board has been exceptionally good. This excellent historic and contemporary track record may have led some clinical physicians to regard the frozen section as a convenience to satisfy their curiosity or to shortcut the usual fixation and processing to obtain a definitive diagnosis. Although this approach could be regarded as a complement to the talents of the pathologist, such requests dismiss the technical hazards and the possibility of a revised interpretation the next day based on adequately fixed and embedded tissue. Paradoxically, the pathologist's credibility may suffer as a consequence, not to mention a potential deleterious effect on the patient. However, although clinical physicians should be disabused of the notion that a frozen section is equivalent to permanent embedding and thorough microscopic study, it would be foolish for pathologists to ignore that in some circumstances, the frozen section is regarded as definitive or virtually definitive, such as in margins on a basal cell carcinoma, ureter margins on a cystectomy, or sentinel lymph nodes for breast cancer.

Pathologists regard frozen sections as challenging and do not approach their execution lightly. A frozen section is an emergent request requiring cessation of the activity of the moment (eg, being paged out of a meeting or halting a sign-out session or phone call with another physician). This is true not only for the pathologist on call for frozen sections but also for any colleague from whom he or she may seek an intradepartmental consultation. Because the occasional diagnostic difficulty cannot be overstated and because the clinical stakes may be high, seeking consultation or deferring a diagnosis pending additional tissue or permanent/nonfrozen embedding is quite appropriate.

Although most surgeons and pathologists do recognize that this is a cooperative effort of some gravity, the utilization of frozen section is dependent on the clinical physician, and thus varies among institutions. In the 1891 frozen section at Johns Hopkins, the surgeon (Halsted) had already left the operating room by the time the pathologist (Welch) had delivered the result.1 Allowing for the circumstance of the very first frozen section--that it took quite a while to prepare and deliver the result and that this may have been a genuine, if ultimately superfluous, exercise--it is somehow portentous that this would not be the last time such an event would occur. Today, this sequence of events would create friction in the perception by the pathologist that a frozen section had been frivolously requested. Specifically, working on a frozen section only to discover that the diagnosis has no potential to change what was done, that the surgeon is no longer in the room, or that the patient is already in recovery implies that the information was inherently irrelevant and the pathologist's effort inconsequential. The optimal use of the patient's diagnostic tissue may also have been compromised. Abuses of the system are potential anywhere and can only be resolved by ongoing direct and respectful communication between pathologists and clinical colleagues.


A successful intraoperative consultation with or without a frozen section has several components. There is nothing profound about this, just common sense and knowledge of the medical issues. The following is an admittedly incomplete suggested list:

1. Know the clinical history, preoperative diagnosis, radiographic images

2. Try to examine the gross--SAMPLING IS EVERYTHING

* Actual specimen receipt or

* Go into the operating room

* Don't freeze fat

* Freeze small specimens sparingly

* Triage, such as lymphoma (culture, cytogenetics, electron microscopy, flow cytometry)

3. Microscopic examination--KEEP IT SIMPLE

* Use imprints

* Sharp blades, working cryostat with backup, fresh stains

* Know the artifacts

4. Know the treatment algorithm

* Do not hesitate to DEFER

* Seek consultation from colleagues

* Ask for more tissue

5. Communicate effectively

Frozen section is a valuable tool in the management of clinical problems mostly involving tumors. The surgical pathologist brings to this procedure a morphologic expertise complemented by an awareness of the clinical setting and a familiarity with a given disease process. The standard evaluation of frozen sections is by hematoxylin-eosin staining--special procedures, such as immunohistochemistry, have no present role. An informed practice of frozen section is a collaboration between the clinical physician and the physician-pathologist for the benefit of the patients they both serve.


(1.) Carter D. Surgical pathology at Johns Hopkins. In: Rosai J, ed. Guiding the Surgeon's Hand: The History of American Surgical Pathology. Washington, DC: American Registry of Pathology; 1997:23-39.

(2.) Bloodgood JC. When cancer becomes a microscopic disease, there must be tissue diagnosis in the operating room. JAMA. 1927;88:1022-1023.

Jerome B. Taxy, MD

Accepted for publication December 2, 2008.

From the Section of Surgical Pathology, Department of Pathology, University of Chicago, Chicago, Illinois.

The author has no relevant financial interest in the products or companies described in this article.

Correspondence: Jerome B. Taxy, MD, Department of Pathology, Section of Surgical Pathology, MC6101, 5841 S Maryland Ave, Chicago, IL 60637 (e-mail:
Gale Copyright: Copyright 2009 Gale, Cengage Learning. All rights reserved.