Olsen KD, Moore EJ, Lewis JE. Frozen Section Pathology for Decision Making in Parotid Surgery. JAMA Otolaryngol Head Neck Surg. 2013;139(12):1275-1278. doi:10.1001/jamaoto.2013.5217
For parotid lesions, the high accuracy and utility of intraoperative frozen section (FS) pathology, compared with permanent section pathology, facilitates intraoperative decision making about the extent of surgery required.
To demonstrate the accuracy and utility of FS pathology of parotid lesions as one factor in intraoperative decision making.
Design, Setting, and Participants
Retrospective review of patients undergoing parotidectomy at a tertiary care center.
Evaluation of the accuracy of FS pathology for parotid surgery by comparing FS pathology results with those of permanent section.
Main Outcomes and Measures
Documented changes from FS to permanent section in 1339 parotidectomy pathology reports conducted from January 1, 2000, through December 31, 2009, included 693 benign and 268 primary and metastatic malignant tumors.
Changes in diagnosis were found from benign to malignant (n = 11) and malignant to benign (n = 2). Sensitivity and specificity of a malignant diagnosis were 98.5% and 99.0%, respectively. Other changes were for lymphoma vs inflammation or lymphoma typing (n = 89) and for confirmation of or change in tumor type for benign (n = 36) or malignant (n = 69) tumors. No case changed from low- to high-grade malignant tumor. Only 4 cases that changed from FS to permanent section would have affected intraoperative decision making. Three patients underwent additional surgery 2 to 3 weeks later. Overall, only 1 patient was overtreated (lymphoma initially deemed carcinoma).
Conclusions and Relevance
Frozen section pathology for parotid lesions has high accuracy and utility in intraoperative decision making, facilitating timely complete procedures.
Parotid neoplasms are uncommon tumors that present unique challenges in diagnosis and effective treatment planning. The patient’s history, examination findings, and imaging all contribute to an overall clinical assessment. Operative decision making, including the amount of gland removed (partial, superficial, deep, or total), facial nerve resection (partial or total), and removal of parotid and regional cervical nodes, is generally based on clinical and surgical findings, imaging results, and the pathologic diagnosis from fine-needle aspiration (FNA) cytology, frozen section (FS) pathology, or permanent section pathology (standard hematoxylin-eosin–stained sections). Accurate pathologic diagnosis is challenging because of the rarity and variety of parotid tumors.
Many centers use preoperative FNA cytology to aid in diagnosis and management decision making. However, FNA cytology of salivary gland lesions has a high false-negative rate that limits its usefulness, so few surgeons are willing to commit to a comprehensive or extended surgical procedure on the result of this test alone. Physicians may use FNA cytology for an initial impression and delay any decisions on major surgery or other treatment until permanent section pathology is available.
Some medical centers also use FS pathology for parotid pathology and margin analysis. However, many pathologists do not because of the rarity of these neoplasms, the variety of tumors, and the challenges in making an accurate diagnosis from FS pathology. Prior authors evaluating the value of FS pathology have determined that it offers greater accuracy and more improved sensitivity than FNA cytology. Most studies, though, have examined only the accuracy of FS pathology in determining benign vs malignant disease and margin extent.1
At our institution, FNA cytology is rarely performed for parotid lesions. Instead, we rely on FS pathology to provide histopathologic intraoperative information to guide surgical management. The surgeon needs to know not only whether the tumor is benign or malignant but also whether it is lymphoma or carcinoma, whether it is low grade or high grade, the status of any intraparotid and extraparotid lymph nodes, and the margin analysis. The extent of parotid and neck surgery is influenced by pathologic information. Operative surgical decisions include whether a partial, superficial, deep, or total parotidectomy should be performed. Margin analysis of an involved nerve or adjacent structures also factors into the extent of an operation. Finally, a decision regarding the removal of regional neck nodes and deep lobe parotid nodes can be guided by the parotid tumor histology and/or involvement of the nodes in the superficial lobe. Frozen section pathology can help in all of these decision points, especially the last one.
If FS pathology provides reliable information, patients could benefit by receiving all surgical procedures in a single operation. There would be no need to return later to the operating room for an operation the surgeon would have performed if the correct diagnosis had been known at the time of the initial surgery or to rely instead on radiation therapy postoperatively. Accurate FS pathology can provide the surgeon with an important factor in planning a rational oncologic procedure that is based on knowledge of tumor behavior.
However, if the information from FS parotid pathology is not accurate, it may increase the potential for performing the wrong operation and possibly harming the patient. The patient may undergo unnecessary surgery, return for another operation, or accept a nonsurgical treatment because surgeons may be reluctant to operate again in a parotid bed after prior parotid surgery.
For the patient with a parotid mass, we use FS evaluation of the lesion as a key factor to determine whether to proceed with deep lobe removal or neck dissection during the initial operation. Frozen section pathology has been used for several decades in all parotid operations at Mayo Clinic. This intraoperative diagnostic approach has potential value for our patients. If accurate, it could be cost-effective if it eliminated the need for FNA, additional surgery, or radiotherapy.
This report is the largest series to date evaluating the accuracy of FS pathology for parotid surgery. We sought to determine whether the FS diagnosis of the pathologist, as conveyed to the surgeon, provided information that led to overtreatment or undertreatment. We wanted to determine whether FS pathology results had a positive or negative effect overall on surgical treatment decision making, beyond just determining whether a tumor was malignant or benign. An FS pathology report of a high-grade malignant tumor with known potential for lymphatic spread or the pathologic finding of positive superficial parotid nodes often leads to the decision at surgery to proceed with a deep lobe parotidectomy and neck dissection to remove potential metastatic cancer. Clinical, surgical, and pathologic findings are all used to guide intraoperative and postoperative management.
Frozen section analysis of nearly all surgical specimens is standard practice at our institution. Approximately 20 pathologists with general expertise participate in the FS laboratory. In most cases, a final report is generated on the basis of the FS pathology result. However, in more complex cases, such as those requiring immunostains and/or intradepartmental consultation, a preliminary FS report is issued, and the final report is deferred to permanent section. For every FS procedure performed, a corresponding hematoxylin-eosin section is prepared and reviewed the following day.
Approval from the Mayo Clinic Institutional Review Board was obtained to review all FS and permanent section pathology reports for all cases of parotid gland surgery in patients who had given written permission for the use of their medical records for research purposes. Any change in diagnosis on the permanent section pathology report was noted, even if a preliminary FS pathology report stated that the diagnosis was suspect or suspicious, and it was then confirmed on permanent section. When there was any change in the final pathology report, we reviewed it to determine whether the change led to overtreatment or undertreatment surgically. All applicable surgical records were then reviewed.
During a 10-year period from January 1, 2000, through December 31, 2009, a total of 1339 parotid surgical procedures were conducted at Mayo Clinic, Rochester, Minnesota, with FS analysis performed on each case. Subsequent pathologic review led to a permanent pathologic diagnosis for each case. There were 693 benign tumors and 268 primary and metastatic malignant lesions. The remaining 378 cases included infections, sialadenitis, and inflammatory nodes.
No change was found in the pathology report from FS to permanent section in 1119 of the 1339 cases of parotid surgery that underwent the FS procedure. The remaining 220 cases were studied in detail. The change from FS to permanent section included both those cases for which the diagnosis was preliminary and confirmed on permanent section and those cases that were indeterminate or that had any substantive change made in diagnosis from FS to permanent section (Table).
Of the 220 cases, 89 were suggestive of lymphoma or an inflammatory process. The final diagnosis was confirmed or determined with additional pathologic study. There was only 1 case of presumed lymphoma, which proved to be a carcinoma (ie, Merkel cell carcinoma). In addition, only a single case of lymphoma was incorrectly identified on FS as high-grade carcinoma.
Our review identified 69 cases of suspected malignant tumor on FS that had some change in the final pathology report. A specific suspected malignant tumor was confirmed in 27 cases; in 40 of the other 42 cases, permanent section pathology revealed a different final pathologic malignant tumor. An example of the latter was adenocarcinoma on FS evaluation identified as salivary duct–type carcinoma on the final pathology report. There were no cases of high-grade cancer changed to low-grade cancer on the final pathology report or vice versa. Only 2 cases were thought to be malignant and instead proved to be benign on permanent section pathology. Both were initially reported to be low-grade malignant tumors.
Our review identified 53 cases of suspected benign tumor on FS pathology that had some change in the final pathology report. Thirty-six benign tumors had a pathologic classification on permanent section pathology that basically confirmed the presumed FS diagnosis. In the other 17 cases, the diagnosis was changed on permanent section pathology to another type of benign tumor, such as from monomorphic adenoma to basal cell adenoma tumor (n = 6). A benign tumor suspected on FS proved to be a malignant tumor on permanent pathology in only 11 cases. Ten of these 11 cases were initially reported to the surgeon as benign tumors or as possible low-grade malignant tumors. One case proved to be an intermediate-grade mucoepidermoid cancer. No cases were changed to a high-grade cancer.
Finally, in 8 cases the number of involved nodes noted on FS was different from the number identified on permanent section pathology, and in 1 case several margins initially reported as negative proved instead to be positive on permanent section pathology.
Of all 220 cases, only 4 could be identified in which the FS diagnosis would have altered intraoperative decision making if the final pathology report had been known. One case involved a rapidly growing destructive lesion that was replacing the parotid gland. Initial FS biopsy during surgery identified it as a high-grade carcinoma. A total parotidectomy was performed. On permanent pathology, the case proved to be a high-grade lymphoma. This was the single case of overtreatment in the series. Three cases with FS pathology had the diagnosis changed on permanent section pathology such that the patient later returned to the operating room for additional surgery. If the correct diagnosis had been made initially, the additional operation would not have been necessary. Of these 3 cases, the first was initially believed to be a lymphoma, whereas permanent section pathology proved it to be a Merkel cell carcinoma. The second case was initially believed to be a benign tumor, whereas permanent section pathology identified it as an intermediate-grade mucoepidermoid cancer. The third case that required additional surgery was initially identified as node-negative metastatic melanoma, whereas permanent section pathology found a parotid node that had been removed to be positive for metastasis. A neck dissection was performed later.
Most prior studies examining the accuracy and use of FS pathology have compared it with FNA cytology and have documented improved accuracy of FS pathology over FNA cytology.1- 3 Fine-needle aspiration cytology has been found to have a high false-negative rate, up to 20%, that limits its usefulness.1
Reports to date of FS pathology as a diagnostic tool for parotid surgery have largely focused on the ability to differentiate a benign from a malignant tumor. One analysis of this factor found the efficacy of FS pathology to be good, with specificity of 99% and sensitivity of 90%.2 Others have reported 100% accuracy in both sensitivity and specificity.2 Our study found 98.5% sensitivity and 99.0% specificity for FS pathology in detecting malignant tumors.
However, rather than just being a useful adjunct to intraoperative decision making, FS pathology is most helpful if it can reliably determine more than whether a lesion is benign or malignant. It is important to know whether the tumor is lymphoma or carcinoma, whether it is a low-grade or a high-grade malignant tumor, and what the status is of the intraparotid nodes and peripheral margins.
Our findings indicate that with experienced pathologists, FS evaluation of the tumor and superficial parotid nodes can be helpful in intraoperative decision making. Our patients undergoing parotid surgery are told that any tumor found with a high likelihood of metastatic spread to the deep portion of the gland or neck nodes will be managed with deep lobe parotidectomy and select neck dissection during the same operation. In addition, FS pathology is helpful in evaluating margin status and aiding in determining the extent of surgery. It is one factor in decision making that is pertinent to sacrifice or preservation of the facial nerve. However, most decisions concerning the facial nerve rest equally on intraoperative findings, and a preoperative functioning facial nerve is almost always preserved. Most low-grade malignant tumors and benign tumors are generally treated the same.
Cost is an important consideration with the routine use of FS pathology. Because FNA cytology is rarely performed in our practice and FS pathology helps some patients avoid returning at a later date to the operating room, the cost of FS pathology may be justified. Its use in guiding appropriate surgery also helps avoid postoperative radiotherapy in some cases.
Intraoperative parotid surgical decision making as a benefit to the use of FS pathology has been described by other authors.4 Reports have documented its value in making a diagnosis, delineating margins, and determining whether the facial nerve or neck nodes are involved. Its main value has been discussed in margin analysis and in differentiation of benign from malignant tumors.2
We believe that the histopathologic information provided by the pathologist is valuable in guiding the extent of treatment and the management of metastatic nodal disease. Decisions made at the time of surgery include proceeding with deep lobe removal, performing neck dissection, or extending the initial resection for management of adjacent structures and the facial nerve.5
Frozen section pathology for parotid lesions can be quite accurate. In parotid surgery, the pathologist remains one of the most important members of the surgical team. The routine use of FS pathology for parotid pathology is valuable in that it can provide information that may alter surgical management and add to patient safety, efficacy of treatment, and improved outcomes. Frozen section pathology for parotid tumors can inform the surgeon of more than just whether the tumors are benign or malignant. It is accurate in distinguishing carcinoma from lymphoma, high-grade from low-grade tumors, and the involvement of parotid nodes. This information can be valuable in surgical and nonsurgical management.
Submitted for Publication: April 1, 2013; final revision received July 3, 2013; accepted August 5, 2013.
Corresponding Author: Kerry D. Olsen, MD, Department of Otorhinolaryngology–Head and Neck Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (email@example.com).
Published Online: October 17, 2013. doi:10.1001/jamaoto.2013.5217.
Author Contributions: Drs Olsen and Lewis had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition of data: Olsen, Moore.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Olsen, Moore.
Critical revision of the manuscript for important intellectual content: All authors.
Administrative, technical, and material support: Olsen, Lewis.
Study supervision: Olsen.
Conflict of Interest Disclosures: None reported.
Previous Presentation: Presented at the Eighth International Conference on Head and Neck Cancer; July 23, 2012; Toronto, Ontario, Canada.