Abstract
Introduction
When indicated, intraoperative use of frozen sections may assist in determining the surgical course or appropriate processing of surgical specimens. Knowing the accuracy of a preliminary frozen section diagnosis is important. The purpose of this study is to determine the rate of correlation between frozen and permanent histopathologic diagnoses of adult orbital lesions, analyze characteristics of discordant cases, and examine the effects of discordance on surgical decision-making.
Methods
A 15-year retrospective chart review was conducted at a tertiary care center of all adult patients with orbital lesions for which frozen section and corresponding permanent section tissue diagnoses were obtained.
Results
Sixty-five orbital surgeries were performed with a total of 89 frozen sections sampled. In 63 surgeries (96.9%), at least 1 frozen section diagnosis matched the final permanent section diagnosis. Overall, frozen section diagnosis corresponded with permanent section diagnosis in 81 of 89 (91.0%) specimens. Of the 8 (9.0%) specimens from 5 unique patients that did not correlate, the final diagnoses on permanent sections were amyloidosis (5), margin-positive infiltrating breast carcinoma (2), and lymphoid hyperplasia (1). The discrepancy between frozen and permanent sections did not alter care in any patient.
Conclusion
Frozen section diagnoses correlate with permanent histopathologic tissue diagnosis in adult orbital biopsies in greater than 90% of cases. Among non-correlated specimens, amyloidosis was the most common diagnosis. Although rare, orbital amyloid disorders may be considered in the differential diagnosis of cases of orbital biopsies with nonspecific findings on a frozen section.
Keywords: Orbital tumors, Histopathology, Ocular pathology, Frozen section, Orbital amyloidosis
Introduction
Histopathologic examination of permanent sections remains the gold standard for the definitive diagnosis of suspected malignancies in orbital lesions [1, 2]. Frozen sections may be used to quickly assess the general cellular architecture of a sample when the care plan could potentially be altered by a preliminary intraoperative diagnosis [1, 2, 3]. Indications for frozen section analysis for orbital lesions include identifying lesions to allow for appropriate specimen handling for additional testing (i.e., determining if a lesion is lymphoproliferative in origin and needs flow cytometry, assuring correct processing and adjunctive staining for sebaceous cell carcinoma, or obtaining cytogenetics for certain sarcomas), obtaining a preliminary intraoperative diagnosis that could alter the surgical plan, and assessing surgical margins [1, 2, 3]. The use of frozen sections for ocular specimens [1, 2, 3] as well as the accuracy of frozen section diagnosis in other specialties has been studied [4, 5, 6, 7, 8, 9, 10, 11]. This paper examines the rate of correlation between frozen and permanent section histopathologic diagnoses specifically for adult orbital lesions.
Purpose
The purpose of this study is to assess the rate of correlation between frozen section and permanent section diagnoses of adult orbital lesions, analyze the histopathologic characteristics of discordant cases, and examine any effects of discordance on surgical decision-making or plan of care.
Materials and Methods
A chart review was conducted to analyze all adult patients who had orbital biopsies with frozen section diagnoses and corresponding permanent section pathology results between January 1, 2000 and October 31, 2014. The reports of frozen section and permanent section diagnoses were compared and discrepancies were recorded. For each non-correlated case, available frozen and permanent slides, as well as clinical management, were reviewed to assess potential effects of discordance on patient care. This study was reviewed and approved by the Northwestern University Institutional Review Board.
Results
There were 65 patients who underwent orbital surgery with submission of tissue for frozen section and subsequent permanent tissue diagnoses between January 1, 2000 and October 31, 2014. In these 65 surgeries, a total of 89 frozen sections were obtained. An average of 1.4 frozen section samples were taken per patient (range 1–6). Fifty-three (81.5%) patients had one frozen section sample taken, 4 (6.2%) had two samples, 6 (9.2%) had three samples, and 2 (3.1%) patients had more than three samples taken.
Of the 65 surgeries performed with frozen sections during this timeframe, five of these surgeries had some instance of discrepancy between frozen and permanent section diagnoses. However, in 3 of these 5 surgeries with discrepancy, multiple frozen sections were done, and there was concordance between at least one pair of frozen and permanent section diagnoses. Therefore, the objective of obtaining a correct intraoperative diagnosis from a frozen section during a procedure was achieved in 63 of 65 surgeries (96.9%), as shown in Figure 1.
Fig. 1.
Correlation between frozen and permanent section diagnoses per surgery (left) and per individual frozen section sampled (right). In 63 of 65 surgeries performed, at least 1 frozen section sampled matched the final permanent section diagnosis. Eighty-one of 89 frozen sections were correlated to the final permanent section diagnosis. Eight non-correlated specimens were found in 5 patients.
Overall, the frozen section diagnosis corresponded with the permanent section diagnosis in 81 of the 89 unique frozen sections sampled (91.0%). A total of 8 discrepancies were noted in 5 patients. Of the 53 patients with only one frozen section sampled, there was discordance between frozen and permanent section diagnoses in one case (permanent section diagnosis of amyloidosis).
The final permanent section diagnoses were divided into four categories by type. Of the 89 orbital lesions sampled, 24 (27.0%) were non-lymphomatous neoplasia, 21 (23.6%) were lymphoproliferative neoplasia, 24 (27.0%) were nonneoplastic inflammatory, and 20 (22.5%) were nonneoplastic noninflammatory, as shown in Figure 2. The list of final histopathologic diagnoses in each of these categories is shown in Table 1.
Fig. 2.
Permanent section diagnosis categories: non-lymphomatous neoplasia, lymphoproliferative neoplasia, inflammatory, noninflammatory (n = 89). Two discordant non-lymphomatous specimens had a permanent section diagnosis of positive margins for infiltrating breast carcinoma. One discordant inflammatory specimen had a permanent section diagnosis of lymphoid hyperplasia. Five discordant noninflammatory specimens were found to contain amyloidosis on permanent section analysis.
Table 1.
Final diagnoses in each of the neoplastic and nonneoplastic categories
| Neoplastic: nonlymphoid (14) | Neoplastic: lymphoid (20) | Nonneoplastic: inflammatory (22) | Nonneoplastic: noninflammatory (9) |
|---|---|---|---|
| Basal cell carcinoma (3) | Extranodal marginal zone lymphoma (15) | Orbital inflammation (18) | Amyloid (5) |
| Squamous cell carcinoma (2) | Diffuse large B-cell lymphoma (3) | IgG4-related disease (2) | Normal orbital tissue (3) |
| Schwannoma (2) | Follicular lymphoma (2) | Sarcoidosis (2) | Fibro-osseous lesion (1) |
| Metastatic carcinoma (2) | − | − | − |
| Adenoid cystic carcinoma (1) | − | − | − |
| Cavernous hemangioma (1) | − | − | − |
| Fibroblastic carcinoma (1) | − | − | − |
| Inverted Schneiderian papilloma (1) | − | − | − |
| Meningioma (1) | − | − | − |
The 8 specimens where frozen and permanent section diagnoses did not correlate were further assessed. There were 5 (62.5%) instances of discordant nonneoplastic noninflammatory sections that were read on frozen sections as acellular hyalinized and calcified tissue, connective tissue with degenerative changes, or lymphoid/necrotic tissue. All of these had a permanent section diagnosis of amyloidosis. Two (25.0%) non-lymphomatous specimens were read as normal muscle intraoperatively but displayed positive margins for infiltrating breast carcinoma on deeper permanent sections. One (12.5%) nonneoplastic inflammatory case was read as necrotic tissue on a frozen section, but permanent section analysis yielded a diagnosis of lymphoid hyperplasia, as shown in Figure 2.
A chart review was conducted for each non-correlated specimen to identify the effect of discordance on subsequent patient care. Additionally, all available frozen and permanent section glass slides were reexamined to analyze the cause of the discrepancy, as shown in Table 2.
Table 2.
Pathology reports for each patient with one or more discordant frozen/permanent section diagnoses.
| Patient | Frozen section diagnosis | Permanent section diagnosis | Discordant? |
|---|---|---|---|
| 1 | Acellular hyalinized and calcified tissue | Amyloidosis | Yes |
| 2 | Connective tissue with degenerative changes | Amyloidosis | Yes |
| 2 | Amyloidosis | Amyloidosis | No |
| 3 | Lymphoid/necrotic tissue, evaluated for lymphoma | Amyloidosis/plasma cell dyscrasia | Yes |
| 3 | Lymphoid/necrotic tissue, evaluated for lymphoma | Amyloidosis/plasma cell dyscrasia | Yes |
| 3 | Lymphoid/necrotic tissue, evaluated for lymphoma | Amyloidosis/plasma cell dyscrasia | Yes |
| 4 | Normal skeletal muscle* | Infiltrating breast carcinoma | Yes |
| 4 | Normal skeletal muscle* | Infiltrating breast carcinoma | Yes |
| 4 | Infiltrating breast carcinoma | Infiltrating breast carcinoma | No |
| 5 | Necrotic tissue with occasional atypical cells | Lymphoid hyperplasia | Yes |
| 5 | Glandular tissue with lymphocytic infiltrate | Focal chronic inflammation | No |
Multiple frozen sections were done in 4 of the 5 patients with discordant diagnoses.
Frozen sections reviewed. No carcinoma found on the frozen section slide. Carcinoma was found on deeper permanent sections.
Patients 1–3 all had at least one frozen section that did not correspond with the permanent section diagnosis of amyloidosis. Clinical records show that Patient 1 had a calcific lesion in the left superior orbit, and MRI showed a soft tissue enhancing infiltrative lesion in the extraconal space. The single frozen section sampled revealed acellular hyalinized and calcified tissue, as shown in Figure 3a. This was later found to be amyloidosis upon analysis of the permanent sections (Figure 3b). This discordance did not alter the surgical care plan.
Fig. 3.
Non-correlated frozen and permanent section slides, diagnosis: amyloidosis. a Patient 1 − (frozen section, ×10) read as acellular hyalinized and calcified tissue. b Patient 1 − permanent section (Congo red stain, ×10) read as amyloidosis. c Patient 2 − (frozen section, ×10) read as connective tissue with degenerative changes. d Patient 2 − permanent section (H&E, ×10) read as amyloidosis.
Patient 2 had two frozen sections analyzed intraoperatively; one was discordant with the final permanent section, as shown in Figure 3c, d. In the other frozen section, amyloid was diagnosed on both the frozen and permanent sections, as shown in Figure 4c, d. Intraoperative care was not altered because of this discrepancy.
Fig. 4.
Correlated frozen and permanent section slides, diagnosis: amyloidosis. a, b from Patient 6 who had only 1 frozen section sampled with a diagnosis of amyloidosis. a Patient 6 − (frozen section, ×10) read as amyloidosis. b Patient 6 − permanent section (H&E, ×10) read as amyloidosis. c, d from Patient 2. c Patient 2 − (frozen section, ×10) read as amyloidosis. d Patient 2 − permanent section (H&E, ×10) read as amyloidosis.
Patient 3 had three frozen sections that were all read as lymphoid aggregates with necrotic tissue and evaluated for lymphoma. Upon analysis of the corresponding permanent sections, the patient had a final diagnosis of a plasma cell dyscrasia with secondary amyloid. In this patient, the lymphoproliferative disorder was correctly identified on frozen section, but the secondary amyloidosis was not mentioned. There was no alteration of the operative plan because of this discordance.
For reference, a second example of a concordant frozen and permanent section diagnosis of amyloidosis for a 6th patient (Patient 6) was included, as shown in Figure 4a, b. Patient 6 had one frozen section sampled, and this sample correlated with the permanent section diagnosis of amyloidosis.
Patient 4 had three frozen section specimens analyzed, and two of these specimens were discordant with the final permanent section diagnosis. Two frozen sections were read as skeletal muscle with no tumor present, but the third frozen section revealed metastatic carcinoma. The permanent section diagnosis revealed that all three samples contained infiltrating metastatic breast carcinoma. Upon subsequent review of frozen section slides, one discordant frozen section revealed a small cluster of cells with round, inconspicuous nuclei; crush artifact; and no definitive tumor. A review of the second discordant frozen section slide did not reveal any tumor. The discordance in these two biopsies was due to the fact that there was no tumor in the part of the specimen cut for frozen section. Tumor was present on deeper sectioning of the tissue block for permanent slides for all three sections. This discordance did not alter the surgical care plan.
Patient 5 had two frozen sections analyzed, and one was found to be discordant with the permanent section diagnosis. For this patient, there was a known history of metastatic melanoma, and a mass was incidentally found in the right lacrimal fossa. The discordant frozen section was read as mostly necrotic tissue with occasional atypical cells. The final permanent section diagnosis was read as lymphoid hyperplasia. This discordance did not alter the surgical care plan.
Discussion
Intraoperative frozen sections may be indicated when the diagnosis could alter surgical management, tumor resection margins need definition, the specimen requires assessment to determine if it is representative for diagnosis, or the determination of subsequent workup and/or special studies are required [3, 12]. The American Academy of Ophthalmology states that intraoperative frozen sections are “time intensive and costly processes and should be used with discretion,” stressing the importance of communication between the surgeon and pathologist to determine if a particular case warrants frozen sectioning [12].
Prior studies assessing the accuracy of frozen sections demonstrate a high rate of concordance between frozen and permanent section diagnoses ranging from 83.6 to 96.8% [4, 5, 6, 7, 8, 9, 10, 11]. Further analysis of the correlation between the two has been carried out for individual subspecialties revealing concordance rates of 83.6% (n = 1,339) for parotid surgery, 90% (n = 1,023) for neoplasms of the thyroid, and between 92.4% (n = 331) and 94.7% (n = 460) for endometrial tumors [4, 5, 6, 7, 8, 9].
In this study, 96.9% of orbital surgeries had at least one frozen section diagnosis that matched the permanent section diagnosis. This demonstrates that, when indicated, frozen sections can be utilized to make intraoperative management decisions for orbital lesions with a high degree of accuracy.
Despite a high concordance rate, however, most orbital biopsies do not require intraoperative frozen sectioning, and it should be used with discretion. Moreover, preparation of frozen section slides should not be performed if doing so reduces tissue availability for permanent, paraffin-embedded sections or decreases the amount of fresh tissue available for studies such as flow cytometry. In instances where a specimen is “sent for frozen section” to determine if the lesion is lymphoproliferative in nature, many centers use touch prep techniques to sample the fresh specimen and determine if there are adequate lymphocytes present. If so, then the frozen section process can be omitted, and the fresh tissue can be sent for flow cytometry. There is variability in surgeons' decisions to send for frozen sections and in pathologists' methods of processing it, so a clear communication between these individuals is necessary to determine how to optimize tissue processing for the highest yield diagnostic studies and correct diagnosis.
At our institution, all orbital frozen sections were interpreted by the general surgical pathology service. In all cases of discordance between frozen and permanent section diagnoses, the diseases were not unique to ocular pathology (e.g., amyloidosis, breast carcinoma, lymphoproliferative disease). This suggests that correlation rates are not affected by the reading of frozen sections by a general pathologist and permanent sections by an ocular pathologist.
There are a few potential reasons why a permanent section diagnosis might not match the initial frozen section diagnosis. First, pathology is present on the frozen section slide and is not recognized by the pathologist reading the frozen section. Second, the tissue sample processed for frozen section may not contain the pathology seen on deeper permanent sections (as was the case with Patient 4). Lastly, a pathologist may only be able to give a general description in diseases where ancillary testing, such as immunohistochemistry, is required to make the permanent section diagnosis.
In this study, the most common permanent section diagnosis missed on frozen section was amyloidosis, which accounted for 5 of the 8 discordant frozen and permanent section diagnoses (62.5%, Patients 1–3). This number may be considered to be artificially high because in 1 patient who had 3 frozen sections, the pathologist correctly identified the primary disorder (plasma cell dyscrasia) on each of the frozen sections but did not mention the secondary amyloidosis.
Frozen section diagnoses in discordant amyloidosis specimens included acellular hyalinized and calcified tissue, nerve and fibroconnective tissue with degenerative changes, and necrotic tissue. Analysis of available discordant slides with a permanent section diagnosis of amyloidosis and frozen section slides that correctly diagnosed amyloidosis reveals acellular, amorphous tissue. Slides from which amyloidosis was correctly diagnosed on frozen section had more focal, discrete amyloid deposits as compared to a more diffuse pattern of deposition on non-correlated cases (Fig. 3 and 4).
Amyloid in the orbit can be due to primary orbital amyloidosis, systemic amyloidosis, or secondary to diseases such as plasma cell dyscrasia. Increased awareness of orbital amyloid by the pathologist and communication from the surgeon to the pathologist regarding clinical suspicion for amyloid-related disorders may help decrease frozen section discordance rates.
In two other discordant specimens from one unique patient, metastatic carcinoma was not identified on two of the three frozen sections sampled. In the subsequent review of the slides, metastatic carcinoma could not be definitively identified in one frozen section slide, even with retrospective knowledge of the permanent section diagnosis. The second frozen section slide did not contain any tumor cells. Metastatic carcinoma was identified in the third frozen section sampled. These discordant cases, in particular, demonstrate that frozen section diagnoses might not correspond with final permanent section histopathologic diagnoses because the pathology is either absent from the frozen section specimen or cannot be conclusively identified until an analysis of deeper permanent section slides is performed. In cases of high clinical suspicion, the surgeon may consider multiple frozen section biopsies to help ensure that malignant tissue is sampled.
Conclusion
This study demonstrates that frozen section diagnoses for adult orbital biopsies are highly correlated to permanent section diagnoses. In 63 of 65 surgeries studied, one or more of the frozen sections obtained correlated with the final permanent section diagnosis for a concordance rate of 96.9%. Overall, 91.0% of all frozen sections sampled correlated to the final permanent section diagnosis.
The most common non-correlated diagnosis was amyloidosis in nearly 60% of the discordant specimens. When analyzing frozen sections, pathologists should consider amyloidosis in the pathological differential diagnosis they report to a surgeon for orbital specimens containing areas of acellular, eosinophilic tissue. As always, effective communication between a surgeon and a pathologist is important.
Statement of Ethics
The study was conducted in accordance with the Declaration of Helsinki, and this study and waiver of consent were reviewed and approved by the Northwestern University Institutional Review Board (STU00104442).
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This study was supported by an unrestricted departmental Grant from Research to Prevent Blindness Inc., New York, NY.
Author Contributions
Sarah Eichinger, Hans Barron Heymann, and Paul Bryar − data collection, data analysis, slide review, and manuscript preparation. Michael Mbagwu and Alexander Knezevic − data collection, data analysis, and manuscript preparation.
Data Availability Statement
All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.
References
- 1.Karcioglu ZA, Caldwell DR. Frozen section diagnosis in ophthalmic surgery. Surv Ophthalmol. 1984;28:323–32. doi: 10.1016/0039-6257(84)90095-x. [DOI] [PubMed] [Google Scholar]
- 2.Biswas J, Subramaniam N. Frozen section diagnosis in ophthalmic pathology. Indian J Ophthalmol. 1993;41((3)):114–6. [PubMed] [Google Scholar]
- 3.Chevez Barrios P. Frozen section diagnosis and indications in ophthalmic pathology. Arch Pathol Lab Med. 2005;129((12)):1626–34. doi: 10.5858/2005-129-1626-FSDAII. [DOI] [PubMed] [Google Scholar]
- 4.Olsen KD, Moore EJ, Lewis JE. Frozen section pathology for decision making in parotid surgery. JAMA Otolaryngol Head Neck Surg. 2013;139((12)):1275–8. doi: 10.1001/jamaoto.2013.5217. [DOI] [PubMed] [Google Scholar]
- 5.Paphavasit A, Thompson GB, Hay I, Grant CS, van Heerden JA, Ilstrup DM, et al. Follicular and hürthle cell thyroid neoplasms: is frozen-section evaluation worthwhile? JAMA Surg. 1997;132:674–9. doi: 10.1001/archsurg.1997.01430300116022. [DOI] [PubMed] [Google Scholar]
- 6.Quinlivan JA, Petersen RW, Nicklin JL. Accuracy of frozen section for the operative management of endometrial cancer. BJOG. 2003;108:798–803. doi: 10.1111/j.1471-0528.2001.00196.x. [DOI] [PubMed] [Google Scholar]
- 7.Squires MH, III, Kooby DA, Pawlik TM, Weber SM, Poultsides G, Schmidt C, et al. Utility of the proximal margin frozen section for resection of gastric adenocarcinoma: a 7- institution study of the US Gastric Cancer Collaborative. Ann Surg Oncol. 2014;21((13)):4202–10. doi: 10.1245/s10434-014-3834-z. [DOI] [PubMed] [Google Scholar]
- 8.Sala P, Morotti M, Menada MV, Cannavino E, Maffeo I, Abete L, et al. Intraoperative frozen section risk assessment accurately tailors the surgical staging in patients affected by early-stage endometrial cancer: the application of 2 different risk algorithms. Int J Gynecol Cancer. 2014;24((6)):1021–6. doi: 10.1097/IGC.0000000000000145. [DOI] [PubMed] [Google Scholar]
- 9.Amad Z, Barakzai MA, Idrees R, Bhurgri Y. Correlation of intra-operative frozen section consultation with the final diagnosis at referral center in Karachi, Pakistan. Indian J Pathol Microbiol. 2008;51((4)):469–73. doi: 10.4103/0377-4929.43733. [DOI] [PubMed] [Google Scholar]
- 10.Howanitz PJ, Hoffman GG, Zarbo RJ. The accuracy of frozen-section diagnoses in 34 hospitals. Arch Pathol Lab Med. 1990;114((4)):355–9. [PubMed] [Google Scholar]
- 11.White VA, Trotter MJ. Intraoperative consultation/final diagnosis correlation: relationship to tissue type and pathologic process. Arch Pathol Lab Med. 2008;132((1)):29–36. doi: 10.5858/2008-132-29-IFDCRT. [DOI] [PubMed] [Google Scholar]
- 12.AAO Frozen Section Indication (2020) [Internet] ophthalmic pathology and intraocular tumors part I: ophthalmic pathology, chapter 3: special testing and procedures in pathology. 2020. Available from: https://www.aao.org/bcscsnippetdetail.aspx?id=1216cf5e-8c26-4490-a29d-8d86f88dd8d6. Date Accessed 2021 May 27.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.