Role of imprint cytology and frozen section in the diagnosis of tumours

Introduction: Imprint cytology is an economical, simple and quick method with excellent cytological clarity but cannot provide information on the depth of invasion. Frozen section is the standard method of intraoperative diagnosis but is limited by its cost, need of qualified technical staff and the freezing arte facts. Thus, both the techniques have their own limitations and combining these techniques helps in achieving high diagnostic accuracy rates. Aims: 1. To assess the combined role of imprint cytology and frozen section in the diagnosis of tumours of various organs. 2. To compare the diagnostic accuracy of imprint cytology and frozen section individually with that of the histopathological diagnosis. Materials and Methods: The present study was done in the pathology department of Bangalore medical college and research institute for a period of one year. A total of 50 surgical specimens from various organs were studied using imprint cytology and frozen section. The results of both the techniques were compared with the gold standard histopathological diagnosis. Results: Out of 50 cases, majority were ovarian neoplasms comprising of 21 (42%) cases followed by thyroid (14%), breast (12%) and salivary gland (10%) neoplasms. The majority of the cases were malignant (50%) followed by beingn (40%), borderline (6%) and metastatic (4%). The overall diagnostic accuracy of imprint cytology and frozen section in various organs were 94% and 98%respectively. The combined diagnostic accuracy was 98%. Conclusion: Imprint cytology and frozen sections are interdependable intraoperative diagnostic modalities. Imprint cytology alone may not be able to provide a correct diagnosis. Thus, a combination of imprint cytology and frozen section helps in achieving high diagnostic accuracy rates and in distinguishing benign and malignant lesions.


Introduction
History of scrape cytology can be tracked back to 1927 when Leonard S. Dudgeon and Vincent Patrick at the University of London raised the horizons of the rapid cytological diagnosis of freshly cut specimens with reliable accuracy rates [1,2]. Imprints prepared from fresh surgical specimens give excellent cytological clarity. It provides valuable information when frozen section interpretation is equivocal [3]. The method is simple, rapid, inexpensive and does not require special techniques or instruments [4,5]. It does not alter the utility of the specimen for subsequent histopathological examination [6]. The limitation of imprint cytology is its inability to distinguish in-situ from infiltrating carcinomas and to evaluate the depth the invasion.
Also, it gives poor cellular yield in cases with extensive fibrosis and poses difficulty in the diagnosis. The frozen section method was standardized in 1905 by Dr. Louis B Wilson. It is the standard method for intraoperative diagnosis of tumours and provides information on the depth of infiltration but is limited by its cost, the need for qualified technical staff and the unavoidable freezing arte facts [7,8].
Thus, a combination of imprint cytology and frozen section helps in achieving high diagnostic accuracy rates and in distinguishing benign and malignant lesions.

Materials and Methods
Place of study: Pathology department of Bangalore Medical College and Research Institute. 1. All the surgically resected specimens with suspected neoplastic lesion from various organs. 2. Fresh specimens submitted in clean, airtight and properly labelled containers. 3. Cases with prior information/appointment.

Exclusion criteria
1. Cases with known inflammatory and/or infectious conditions. 2. Specimens sent in formalin, water or saline. 3. Specimens sent in unlabelled containers.
Sample collection method: Ethical committee clearance was obtained from the institution. The fresh unfixed specimens were sent in clean, airtight and properly labelled containers along with the requisition forms. The clinical and radiological details were obtained by good communication with the operating surgeon.
Prior appointment for frozen sections was usually taken at least one day before by various departments.
The specimens were grossly examined by inspection and palpation. The specimen was then dissected with a sharp scalpel into two halves. The cut surface was examined for the presence of any tumour. The representative areas were selected and clean glass slides were placed on them for taking the imprint smears. Depending upon the type and consistency of the tissue, touch imprints, scrape or crush techniques were used to prepare the smears. Slides were immediately fixed in 95% alcohol and stained with haematoxylin and eosin stain. Air dried smears were stained with Leishman's stain.
Simultaneously, bits from the representative areas were taken for frozen section examination. The tissue was embedded in OCT (Optimal Cutting Temperature) compound and sectioning was done using the Leica CM1850UV cryostat followed by rapid haematoxylin and eosin staining. The remaining tissue was fixed in 10% neutral buffered formalin for routine histopathological processing. The results were interpreted on microscopic examination and the data was entered in the excel sheet. The results of imprint cytology and frozen section were compared with that of the gold standard histopathological diagnoses.

Statistical tool:
The data was analysed using the Microsoft Excel 2007 and IBM SPSS statistical analysis.

Results
A total of 50 surgically resected specimens from various organs were included in our study. The tumours were broadly classified into benign, borderline, malignant and metastatic. The majority of the cases were malignant (50%) followed by beingn (40%), borderline (6%) and metastatic (4%) ( Table 1).
The overall diagnostic accuracy rate of imprint cytology in various organs was 94% and that of frozen section was 98% (Table 3 and 4). The combined diagnostic accuracy rate was 98%. The false-negative and false-positive rates of imprint cytology were 4% and 2% respectively. The false negative rate of frozen section was 2%. Out of three discordant cases in imprint cytology, two were due to misinterpretation error and one was due to gross sampling error.
In two of the misinterpreted cases, frozen section was helpful to arrive at the final diagnosis and in distinguishing between the benign, borderline and malignant cases. The diagnostic accuracy of imprint cytology and frozen section for malignant tumours were 96% and 98% respectively.  Out of 50 cases, the majority were ovarian neoplasms comprising of 21 (42%) cases followed by thyroid (14%), breast (12%) and salivary gland (10%) neoplasms (Table 2). One case of ovarian neoplasm on imprint cytology was false positively reported as granulosa cell tumour as the imprint smears were cellular showing clusters of small, bland neoplastic cells occasionally arranged in small follicle structures and showing rare nuclear grooves ( Figure 2A, 2B and 2C). But the frozen section revealed alternate hypercellular and hypocellular areas composed of dual population ( Figure 2D and 2E). No nuclear grooves were noted. Therefore, a differential diagnosis of sclerosing stromal tumour and fibrothecoma was given. Finally, the histopathological examination showed cellular ps fibrous bands ( Figure 2F) and composed of two types of cells (spindle shaped cells and vacuolated cells resembling signet-ring cells). Also seen were many prominent thin diagnosis of sclerosing stromal tumour of ovary was given ( Figure 1D).
In another ovarian neoplasm the imprint smears were cellular with many papillary structures showing stratification of the lining epithelium and nuclear atypia. Therefore, a dia frozen section revealed the foci of stromal invasion and so, a diagnosis of serous papillary cystadenocarcinoma was finally made which was concordant with the histopathological diagnosis. was not diagnosed both by imprint cytology and frozen section because of gross sampling error in a case of adenomatoid goiter. Finally, on histopathological examination the focus of micropapillary carcinoma of th One case of ovarian neoplasm on imprint cytology was false positively reported as granulosa cell tumour as the imprint rs of small, bland neoplastic cells occasionally arranged in small follicle structures and showing rare nuclear grooves (Figure 2A, 2B and 2C). But the frozen section revealed alternate hypercellular and hypocellular areas composed of dual population of cells (spindle shaped cells & vacuolated cells) ( Figure 2D and 2E). No nuclear grooves were noted. Therefore, a differential diagnosis of sclerosing stromal tumour and fibrothecoma was given. Finally, the histopathological examination showed cellular pseudolobules separated by thick fibrous bands ( Figure 2F) and composed of two types of cells (spindle shaped cells and vacuolated cells resembling ring cells). Also seen were many prominent thin-walled blood vessels and marked interlobular fibrosis. diagnosis of sclerosing stromal tumour of ovary was given ( Figure 1D).
In another ovarian neoplasm the imprint smears were cellular with many papillary structures showing stratification of the lining epithelium and nuclear atypia. Therefore, a diagnosis of borderline serous tumour was given false negatively. The frozen section revealed the foci of stromal invasion and so, a diagnosis of serous papillary cystadenocarcinoma was finally made which was concordant with the histopathological diagnosis. One case of micropapillary thyroid carcinoma was not diagnosed both by imprint cytology and frozen section because of gross sampling error in a case of adenomatoid goiter. Finally, on histopathological examination the focus of micropapillary carcinoma of thyroid was revealed. One case of ovarian neoplasm on imprint cytology was false positively reported as granulosa cell tumour as the imprint rs of small, bland neoplastic cells occasionally arranged in small follicle-like structures and showing rare nuclear grooves (Figure 2A, 2B and 2C). But the frozen section revealed alternate of cells (spindle shaped cells & vacuolated cells) ( Figure 2D and 2E). No nuclear grooves were noted. Therefore, a differential diagnosis of sclerosing stromal tumour and eudolobules separated by thick fibrous bands ( Figure 2F) and composed of two types of cells (spindle shaped cells and vacuolated cells resembling walled blood vessels and marked interlobular fibrosis. Thus, a

4: Diagnostic accuracy of frozen section in various organs
In another ovarian neoplasm the imprint smears were cellular with many papillary structures showing stratification of the gnosis of borderline serous tumour was given false negatively. The frozen section revealed the foci of stromal invasion and so, a diagnosis of serous papillary cystadenocarcinoma was ne case of micropapillary thyroid carcinoma was not diagnosed both by imprint cytology and frozen section because of gross sampling error in a case of adenomatoid yroid was revealed.

Discussion
There are several studies in the literature which have confirmed the utility and efficacy of imprint cytology and frozen section as intraoperative consultation diagnostic modalities. These studieshave concluded that cytology has the advantage of being much less time consuming, easy to adopt, reliable and does not r special instruments [4,5]. Hence, imprintcytology can be employed routinely in the intraoperative diagnosis in conjunction with frozen section [9,10]. The use of either frozen section or cytological examination alone has an acceptable rate (93-97%) of correct diagnosis, with regard to interpretation of benign versus malignant [11,12,13].
The useof cytology smears during intraoperative consultation has often been neglected in favor of traditional examination of frozen sections. This appearsto be due to the surgical pathologist higher level of confidence in frozen sections, though many studies have demonstrated that the diagnostic efficacy of intraoperative cytology is comparable to that of frozen section [12]. There are several studies in the literature which have rmed the utility and efficacy of imprint cytology and frozen section as intraoperative consultation These studieshave concluded that cytology has the advantage of being much less time consuming, easy to adopt, reliable and does not require special instruments [4,5]. Hence, imprintcytology can be employed routinely in the intraoperative diagnosis in conjunction with frozen section [9,10]. The use of either frozen section or cytological examination alone has an of correct diagnosis, with regard to interpretation of benign versus malignant The useof cytology smears during intraoperative consultation has often been neglected in favor of traditional examination of frozen sections. This to the surgical pathologist′s relatively higher level of confidence in frozen sections, though many studies have demonstrated that the diagnostic efficacy of intraoperative cytology is comparable to that In our study, a total of 50 surgically resected specimens were studied using both imprint cytology and frozen section. The diagnostic accuracy of imprint cytology and frozen section were evaluated individually by comparing with the gold standard histopathological examination. The diagnostic accuracy rates of imprint cytology is 80-97 % in different studies [9,14 diagnostic accuracy rates of imprint cytology and frozen section of our study was comparable with that of other studies and we obtained very good results.
Out of 50 cases, the ovarian neoplasms constituted the major group comprising of 21(42%) cases followed thyroid (14%), breast (12%) and salivary gland (10%) neoplasms. The diagnostic accuracy rates of imprint cytology and frozen section in ovarian neoplasms w 90.5% (19/21) and 100% respectively. The false negative and false positive rates of imprint cytology in ovarian neoplasms were 2% each. In a study conducted by Stewart et al., [8] 402 ovarian tumours were studied and all the benign ovarian lesions were categorized using both imprint cytology and frozen 50 surgically resected specimens were studied using both imprint cytology and frozen section. The diagnostic accuracy of imprint cytology and frozen section were evaluated individually by comparing with the gold standard histopathological diagnostic accuracy rates of imprint 97 % in different studies [9,[14][15][16][17][18][19][20]. The diagnostic accuracy rates of imprint cytology and frozen section of our study was comparable with that of other studies and we obtained very good results. 50 cases, the ovarian neoplasms constituted the major group comprising of 21(42%) cases followed by thyroid (14%), breast (12%) and salivary gland (10%) neoplasms. The diagnostic accuracy rates of imprint cytology and frozen section in ovarian neoplasms were 90.5% (19/21) and 100% respectively. The false negative and false positive rates of imprint cytology in ovarian neoplasms were 2% each. In a study conducted by Stewart et al., [8] 402 ovarian tumours were studied and all the benign ovarian lesions were accurately categorized using both imprint cytology and frozen

Original Research Article
Pathology Update: Tropical Journal of Pathology & Microbiology Available online at: www.pathologyreview.in 584 | P a g e section. 30(86%) of borderline tumours and 137(97%) of malignant tumours were accurately diagnosed on frozen section whereas the corresponding results for cytology were 23(66%) and 131(93%) respectively. In a study by Abdel-Baset et al., [21] out of 60 ovarian masses, the overall accuracy of frozen section was 95.5%. Out of 7 (14%) thyroid cases in our study, one case of micropapillary thyroid carcinoma was not diagnosed both by imprint cytology and frozen section because of gross sampling error.
The diagnostic accuracy of both imprint cytology and frozen section in thyroid lesions was 85.7% (6/7). Sukumar Shaha et al., [22] studied a total of 57 thyroid lesions. The diagnostic accuracy of frozen and imprint cytology were 88.46% and 81.25% respectively. This was almost similar to our findings.
In the present study, the diagnostic accuracy of both imprint cytology and frozen section in breast tumours was 100%. In a study conducted by Amar Ranjan et al [7] out of 31 cases of breast tumours all the cases were accurately categorized into benign, locally infiltrative and malignant tumours on frozen section.
The diagnostic accuracy of both the techniques in other organs was 100% in the present study. Lee et al., [23] retrospectively studied 522 histologically proven tumours. The total diagnostic accuracy of imprint cytology was 92.9% for both benign and malignant tumours. Individually for benign tumours, accuracy rate was 97.2% and for malignant tumours it was 89.9%.
In our study, all the three CNS cases were accurately diagnosed by frozen section and imprint cytology. SAM Ahmed et al., [24] studied 33 clinically diagnosed patients of CNS tumours. The sensitivity of frozen section was 100% and that of imprint was 75%.
The accuracy of both the techniques were comparable in our study. The experience of the pathologist in interpretation of imprints is the mainstay in improving the diagnostic accuracy when both the techniques are combined [25].
Thus, a combination of these techniques plays a significant role in arriving at an accurate diagnosis rapidly and in guiding the surgeon intraoperatively in the management of patients.

Conclusion
Imprint cytology and frozen sections are interdependable intraoperative consultation diagnostic modalities. Imprint cytology alone may not be able to provide a correct diagnosis. Thus, a combination of imprint cytology and frozen section helps in achieving high diagnostic accuracy rates and in distinguishing benign and malignant lesions. They together form a very important aspect of surgical pathology that guide the surgeon's hand and play a significant role in rapid intraoperative pathological evaluation.

What does this study adds to existing knowledge:
Many of the studies done in the past have either used imprint cytology or frozen section as an intraoperative diagnostic tool in arriving at a diagnosis. Only few of the studies have showed the combined role of imprint cytology and frozen section in intraoperative diagnosis. In our study, we have assessed the diagnostic accuracy rates of both the methods in various organs and also have evaluated their combined accuracy.