| Abstract|| |
Background: Frozen Sections (FS) are used to assess margins, for staging, and primary diagnosis. FS guide intraoperative treatment decisions in oncological gastro-intestinal tract surgeries and further management of the patients. Aim: To analyze the distribution, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of frozen sections in gastrointestinal pathology in our institution during the period of 3 years (2016–2018). Material and Methods: This study was an audit to determine the accuracy of FS reports by comparing them with the paraffin section (PS) reports. The FS diagnoses and their PS diagnoses were noted in 1704 gastrointestinal surgeries during the period from 2016 to 2018. Discrepancies were noted and slides of discrepant cases were reviewed to determine the cause. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated using the standard formulae. Results: Out of 1704 cases, correct diagnosis on frozen section was made in 1649 cases (96.77%), 20 (1.17%) were deferred cases, and 35 (2.05%) were discrepant cases. The commonest discrepancies were seen in the primary diagnosis of the gall bladder and gastrectomy margins. The commonest causes for discrepancies were interpretation errors and technical errors. Sensitivity was 91.71%, specificity was 99.69%, positive predictive value was 98.84%, negative predictive value was 97.68%, and accuracy was 97.92%. Conclusion: FS diagnosis is a reliable guide to surgeons for intraoperative management. Studying deep cuts and careful sampling at frozen sections will help reduce discrepancies.
Keywords: Audit, frozen section, gastrointestinal pathology
|How to cite this article:|
Jadhav AS, Deodhar KK, Ramadwar M, Bal M, Kumar R, Goel M, Saklani A, Shrikhande SV. An audit of frozen sections for suspected gastrointestinal malignancies in a tertiary referral hospital in India. Indian J Pathol Microbiol 2022;65:796-801
|How to cite this URL:|
Jadhav AS, Deodhar KK, Ramadwar M, Bal M, Kumar R, Goel M, Saklani A, Shrikhande SV. An audit of frozen sections for suspected gastrointestinal malignancies in a tertiary referral hospital in India. Indian J Pathol Microbiol [serial online] 2022 [cited 2022 Nov 30];65:796-801. Available from: https://www.ijpmonline.org/text.asp?2022/65/4/796/359360
| Introduction|| |
Frozen sections (FS) are commonly used in gastrointestinal oncology surgeries to assess margins (adequacy of excision), for staging, and sometimes for primary diagnosis. The FS diagnosis plays a pivotal role in guiding the intraoperative management of cancer patients. Hence, determining the accuracy of FS reports by comparing them with the paraffin section (PS) reports helps to review the efficiency of our FS workstation at a technical level, grossing/sampling level, and interpretation level. An audit will help us do the same and find ways to improve the accuracy of our FS diagnosis, thereby, reducing the discrepancies between FS diagnosis and final diagnosis and understand the utility of FS diagnosis specifically in gastrointestinal pathology.
The objectives of our audit were to analyze the distribution, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of frozen sections in gastrointestinal pathology in our institution during the period of 3 years (2016–2018).
| Materials and Methods|| |
Data retrieval and analysis
The data of frozen reports of patients who had undergone surgeries in the gastrointestinal surgery unit and had FS examination during the period from 2016 to 2018 were obtained from in-house search engines and electronic medical records. These cases were pertaining to surgeries for gastrointestinal and hepatobiliary malignancies. Surgeries for nonmalignant lesions such as Hirschsprung disease, biliary atresia are not undertaken in our center. These cases were reported by 18 pathologists in rotation. For 3 years, the total workload at the FS workstation in our institute was 14,660 cases (oncological surgeries) out of which gastrointestinal specimen constituted 1704 cases (11.62%). The FS diagnoses and their PS diagnoses were noted in these 1704 surgeries. The discrepancies between the two were noted considering PS diagnoses as the gold standard. The FS slides and PS slides of the discrepant cases were reviewed to determine the cause of the discrepancy. The number of deferred cases was noted. We calculated the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy using the standard formulae.
Workflow of FS reporting in our institute
We receive the fresh specimen from the operation theatre at the FS workstation in plasticsealed bags with proper labeling, requisition with an indication for FS analysis, and the clinical chart. The pathologist at the FS workstation carries out grossing of the specimen and documents the specimen dimensions, gross appearance, gross distances from margins, the size of the sample that is processed. Small bits of tissue are sampled entirely. A cryostat is used for processing the sample, and two slides are prepared with two different stains which are toluidine blue and hematoxylin and eosin stain. The pathologist then reports the microscopy, informs the FS diagnosis to the surgical team on the telephone, and is documented.
The final diagnosis is made on formalin-fixed and paraffin-embedded tissue with hematoxylin and eosin staining.
| Results|| |
The list of specimens received for FS analysis during the 3 year period (2016–2018) and their distribution is shown in [Table 1].
The commonest gastrointestinal oncological surgery specimen that we received for FS diagnosis was cholecystectomy for primary diagnosis and common bile duct cut margin (402 cases; 23.60%). It was followed by gastrectomy margins (318 cases; 18.66%), followed by margins of colorectal resection specimen (300 cases; 17.60%). Other specimens sampled in the frozen section included Whipple surgery for pancreatic neck margin, peritoneal nodule biopsy, liver biopsy, lymph node biopsy, etc., (constituting collectively approximately 684 cases; 40.14%). In the case of biopsies, the intent of the surgeons was also to know if they were representative. Apart from the gall bladder, the primary diagnosis of malignancy in the stomach, colon, and pancreas was generally not asked for in the Frozen section.
A correct diagnosis on the frozen section was made in 1649 cases (96.77%) out of 1704 cases. There were 20 (1.17%) deferred cases (cases in which FS report mentioned the comment to wait for paraffin section report for definitive diagnosis) out of 1704 cases. Most of the deferred cases were biopsies either for metastatic disease (10/20) or primary diagnosis (8/20). Only two cases were resection margins of gastrectomy and cholecystectomy (2/20). [Table 2] and [Table 3] summarize the deferred cases and causes for the same. These 20 deferred cases were excluded while calculating the accuracy, sensitivity, specificity, PPV, and NPV.
There were 35 discordant cases out of 1684 cases (excluding the 20 deferred cases out of 1704 cases).
True positive was defined as a case in which FS diagnosis and PS diagnosis were concordant and positive for malignancy. There were 343 out of 1684 cases (20.4%) that were true positive. Out of these, 49 were for margin involvement, 204 were for metastatic disease, and 90 were for primary diagnosis.
True negative was defined as a case in which FS diagnosis and PS diagnosis were concordant and negative for malignancy. There were 1306 out of 1684 cases (77.5%) that were true negative. Out of these, 772 were for margin involvement, 411 were for metastatic disease, and 123 were for primary diagnosis.
False-positive was defined as a case in which FS diagnosis was positive for malignancy while PS diagnosis was negative for malignancy. Four out of 1684 cases (0.24%) were false positive. Out of these, one case was for margin involvement, and three cases were for the presence/absence of metastatic disease.
False-negative was defined as a case in which FS diagnosis was negative for malignancy, whereas PS diagnosis was positive for malignancy. Thirty-one out of 1684 cases (1.8%) were false negative. Out of these, 16 cases were for the involvement of margins, 8 cases were for the presence or absence of metastasis, and 7 cases were for primary diagnosis.
Sensitivity was 91.71%, specificity was 99.69%, positive predictive value was 98.84%, negative predictive value was 97.68%, and accuracy was 97.92%. The above calculations are summarized in [Table 4]. The commonest discrepancies were seen in the primary diagnosis of gall bladder (8 out of 35 cases) and gastrectomy margins (6 out of 35 cases). The reasons for discrepancies included interpretation error (due to freezing and crushing artifact) and superficial cuts (technical error), followed by sampling errors. The discrepant cases, their FS diagnosis, PS diagnosis, and causes for the discrepancy are summarized in [Table 5].
| Discussion|| |
In the literature, there has been a good number of studies on frozen section utility in onco-pathology. However, the majority of them are concerned with breast and head neck pathology. Those that are concerned with FS utility in GI pathology have focused on
the parameter of single specimen type only (for e.g. Margin of gastrectomy or sentinel nodes in gastrectomy).,,, Our study is rare because of the large number and variety of GI specimens considered in this study.
Specimens such as gastrectomy, colorectal resections, pelvic exenteration, and small bowel resection are sent for FS diagnosis for margin involvement by tumor. If there is the presence of a tumor at the margin, then FS diagnosis gives a chance to the surgeons to intraoperatively convert the R1 resection to R0 resection. In our study, gastrectomy and colorectal resections which on biopsy showed signet ring cell morphology were more commonly sent for FS diagnosis. In these cases, although grossly the distance of margin from the tumor is far away, there is a possibility of submucosal spread of the tumor which makes it impossible to grossly determine the involvement of the margin.
Another scenario, where FS diagnosis played an important role is post-chemotherapy resections, especially in gastric and colon cancers. The tumor in these specimens is shrunken or grossly there is no tumor visible and in such cases the surgeons need a pathologist to see microscopically the margin involvement by residual tumor. This task is made more difficult in tumors of signet ring cell morphology with neoadjuvant chemotherapy as these cells are singly scattered amidst chemotherapy response of fibrosis and foamy macrophages. This challenge increases the rate of false-negative discrepancies as observed in our study.
Gall bladder specimen (bile duct cut margin and sometimes for primary diagnosis) was the commonest type of specimen in our study. FS diagnosis of bile duct cut margin guided the extent for resection in order to achieve R0 resection. Its utility has also been studied by Shiraki et al. and found it useful.
Clinically and radiologically (thickened gall bladder wall), at many instances xanthogranulomatous cholecystitis mimics gall bladder carcinoma. To prevent overtreatment in such situations, FS analysis is useful. However, in our study, it is observed that out of the 8 discrepancies in cholecystectomy specimens, 4 were due to sampling error. During FS processing of gall bladder sent for primary diagnosis, limited sections are examined. Thus, it becomes very crucial to sample accurate areas which greatly depends on the skill and experience of the grossing pathologist. In patients with gall bladder carcinoma, interaortocaval lymph node tissue for metastatic involvement was another common sample that we received for FS diagnosis. This guides the surgeons to curtail the procedure in case of metastasis as these patients are then offered chemotherapy options. A study by Agarwal et al. supports the use of FS analysis of interaortocaval nodes before radical surgery in gall bladder carcinoma patients.
Whipple resection specimen was sent most commonly for pancreatic neck margin. Four discrepancies were noted in the pancreatic neck cut margin out of 159 Whipple specimens received for FS analysis in our study. Multiple studies have shown that even though intraoperative analysis of pancreatic neck cut margin helps to obtain margin negative resection, it does not improve the survival of patients with aggressive pancreatic tumors. Thus, the utility of FS analysis in pancreatic carcinomas is still questionable.,,
Other tissues such as retro-pancreatic node, mesenteric node, liver nodule, and peritoneal nodule were sampled in FS to determine the presence or absence of metastasis which would upstage the disease and guide the surgeon to decide whether to go ahead with the surgery or abandon surgery and use palliative treatment.
There were 35 discordant cases out of which 31 were false negative (FN) cases, and 4 were false positive (FP) cases.
Interpretation errors and superficial cutting of sections were the commonest cause of the discrepancy. Interpretation errors accounted for 15 out of 35 discrepant cases (42.86%). The majority of these interpretation errors were due to crushing artifacts, freezing artifacts, and florid inflammation obscuring the malignant cells. All four FP cases were due to interpretation errors. As mentioned before, FN due to interpretation errors is commonest in post-chemotherapy signet ring cell tumors of the stomach, colon, and rectum. A pathologist's diagnostic skills and experience in reporting FS slides also play a major role in reducing interpretation errors. [Figure 1] and [Figure 2] show examples of two false-negative discrepancies.
|Figure 1: Example of a false negative discrepancy - Tissue over bile duct a) Frozen section (FS) shows dense inflammation and crushing. H and E x 400 b) Paraffin section shows atypical glands amidst inflammation in same area as FS. H and E x 400 c) and d) These atypical glands highlighted by CK7 and AE1/AE3 immunohistochemistry DAB x 400|
Click here to view
|Figure 2: Example of false negative discrepancy- Gall bladder for primary diagnosis a) Frozen section shows dense inflammation and freezing artefacts obscuring the morphology of singly scattered large tumor cells. H and E x 400 b) Paraffin section shows the tumor cell morphology clearly. H and E x 400 c) Singly scattered tumor cells are highlighted by AE1/AE3 immunohistochemistry DAB x 400|
Click here to view
Superficial cutting of sections accounted for 15 out of 35 discrepant cases (42.86%). Thus, studying deeper sections would help to reduce the discrepancies, especially when the surgical margin is very close grossly. Miyashiro et al. suggested that studying sections in different planes would help to reduce the false-negative cases, especially in gastric sentinel node biopsies.
As mentioned above, sampling error was the common cause of the discrepancy in the primary diagnosis of the gall bladder.
The causes for deferred cases in our study included inadequate material due to scanty tissue or only necrotic tissue (14/20; 75%), especially in biopsies, followed by limitation of a number of sections while sampling during frozen processing (3/20; 15%) and cautery and crushing artifacts (3/20; 10%). Once the FS report was communicated to the surgeons in these cases, additional material was sent in all 14 cases with an FS report of scanty tissue or necrosis only. Additionally, immunohistochemistry helped to make a definitive diagnosis and confirm the presence of tumor cells in three of the deferred cases during the final report. Out of the 20 deferred cases, 18 were diagnosed as positive for malignancy, and two cases were negative for malignancy at the final diagnosis. Thus, communication with the surgeon regarding additional tissue for a definitive diagnosis after FS examination proved to be useful in the deferred cases.
The overall accuracy (97.92%), sensitivity (91.71%), specificity (99.69%), PPV (98.84%), and NPV (97.68%) in our study points to the fact that FS analysis in gastrointestinal oncological surgery is a very useful tool to obtain R0 resections and to guide the management in case of metastatic disease. Similar observations are seen in other studies.,
| Conclusion|| |
Frozen diagnosis is a reliable and helpful guide to surgeons for intraoperative management in gastrointestinal oncology. Accuracy of frozen section analysis can be increased by studying deeper tissues at the frozen workstation, communicating with surgeons regarding quantity and quality of the tissue, improving technical expertise to tackle freezing artifacts, and by improving skills as a pathologist to sample accurate areas and understanding the iartifacts of frozen section processing.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Spicer J, Benay C, Lee L, Rousseau M, Andalib A, Kushner Y, et al
. Diagnostic accuracy and utility of intraoperative microscopic margin analysis of gastric and esophageal adenocarcinoma. Ann Surg Oncol 2014;21:2580-6.
Keighley MR, Moore J, Lee JR, Malins D, Thompson H. Preoperative frozen section and cytology to assess proximal invasion in gastro-oesophageal carcinoma. Br J Surg 1981;68:73-4.
Kim SY, Hwang YS, Sohn TS, Oh SJ, Choi MG, Noh JH, et al
. The predictors and clinical impact of positive resection margins on frozen section in gastric cancer surgery. J Gastric Cancer 2012;12:113-9.
Miyashiro I, Hiratsuka M, Sasako M, Sano T, Mizusawa J, Nakamura K, et al
. High false-negative proportion of intraoperative histological examination as a serious problem for clinical application of sentinel node biopsy for early gastric cancer: Final results of the Japan Clinical Oncology Group multicenter trial JCOG0302. Gastric Cancer 2014;17:316-23.
Shiraki T, Kuroda H, Takada A, Nakazato Y, Kubota K, Imai Y. Intraoperative frozen section diagnosis of bile duct margin for extrahepatic cholangiocarcinoma. World J Gastroenterol 2018;24:1332-42.
Han SH, Chen YL. Diagnosis and treatment of xanthogranulomatous cholecystitis: A report of 39 cases. Cell Biochem Biophys 2012;64:131-5.
Agarwal AK, Kalayarasan R, Javed A, Sakhuja P. Role of routine 16b1 lymph node biopsy in the management of gallbladder cancer: An analysis. HPB (Oxford) 2014;16:229-34.
Kooby DA, Lad NL, Squires MH 3rd
, Maithel SK, Sarmiento JM, Staley CA, et al
. Value of intraoperative neck margin analysis during Whipple for pancreatic adenocarcinoma: A multicenter analysis of 1399 patients. Ann Surg 2014;260:494-503.
Barreto SG, Pandanaboyana S, Ironside N, Windsor JA. Does revision of resection margins based on frozen section improve overall survival following pancreatoduodenectomy for pancreatic ductal adenocarcinoma? A meta-analysis. HPB (Oxford) 2017;19:573-9.
Dillhoff M, Yates R, Wall K, Muscarella P, Melvin WS, Ellison ES, et al
. Intraoperative assessment of pancreatic neck margin at the time of pancreaticoduodenectomy increases likelihood of margin-negative resection in patients with pancreatic cancer. J Gastrointest Surg 2009;13:825-30.
Squires MH 3rd
, 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;21:4202-10.
Chatelain D, Shildknecht H, Trouillet N, Brasseur E, Darrac I, Regimbeau JM. Intraoperative consultation in digestive surgery. A consecutive series of 800 frozen sections. J Visc Surg 2012;149:e134-42. doi: 10.1016/j.jviscsurg. 2012.01.005.
Kedar K Deodhar
Department of Pathology, Annexe Building, 8th Floor, Tata Memorial Hospital, Dr. E. Borges Road, Parel, Mumbai - 400 012, Maharashtra
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]