Intraoperative Decision-Making: A Comparative Analysis of Scrape Cytology vs. Frozen Section in Eighty Consecutive Samples

Abstract

Introduction:

Intraoperative tumor evaluation is essential for optimizing surgical decision-making and can prevent the need for unnecessary radical surgeries. Although frozen section (FS) is the gold standard for such evaluations in advanced centers, scrape cytology presents a simpler, cost-effective alternative that could be beneficial in resource-limited settings. However, this technique is often overlooked in favor of FS due to greater pathologist confidence in FS.

Materials and Methods:

This study included 80 consecutive intraoperative samples from the Department of Pathology. Both scrape cytology and FS were independently evaluated, and results were compared with those of the final histopathology, taken as the gold standard. Statistical analysis assessed the concordance, sensitivity, specificity, and diagnostic accuracy for each method.

Results:

Scrape cytology and FS demonstrated a concordance rate of 87.5%, with sensitivity and specificity of 89.3% and 90.4%, respectively, for scrape cytology. Positive predictive value was 83.3%, and negative predictive value was 94%. Scrape cytology significantly reduced diagnostic time, averaging 10 min compared to 20 min for FS. Tissue-specific discrepancies were noted, particularly in lymph nodes and parathyroid cases.

Conclusion:

Scrape cytology proved to be a reliable, cost-effective, and time-efficient alternative to FS, especially in settings where FS is unavailable. Although FS remains optimal for architectural detail, scrape cytology provides high diagnostic accuracy and utility for intraoperative decision-making in resource-limited environments.

INTRODUCTION

Intraoperative consultations provide critical, real-time information to guide surgical decisions, often determining the nature of the lesion identified accidentally while operating, the extent of resection needed, confirmation of tissue types, or evaluation of the margins of excised tumors.[1,2] This process can prevent unnecessary radical surgeries or additional interventions. Traditionally, frozen section (FS) has been the gold standard for intraoperative evaluation, providing rapid histopathological results through cryostat-prepared tissue sections. However, FS requires specialized equipment, such as a cryostat, and trained personnel, resources typically available only in advanced medical centers.[3,4,5] Scrape cytology is an alternative technique where cells are scraped from the freshly cut surfaces of surgical specimens and smeared onto slides for cytological evaluation. This method is simpler, faster, and more cost-effective, making it accessible even in resource-limited settings. However, pathologists are hesitant to adopt scrape cytology due to the perceived superiority of FS in terms of accuracy. Most studies have compared the diagnostic utility of scrape cytology and FS in ovarian neoplasms.[6,7,8,9,10] This study aims to compare the performance of scrape cytology and FS in a large cohort of diverse tissue types, identifying the tissue-specific challenges faced.

MATERIALS AND METHODS

Study design

This prospective study was conducted over 2 years in the Department of Pathology and Laboratory Medicine. A total of 80 consecutive intraoperative samples were evaluated using FS and scrape cytology. After receiving the fresh operational sample, a prompt gross examination was performed. The fresh-cut surface was assessed to determine whether the specimen had a homogenous or variegated appearance. Depending on the nature of the specimen, an ideal technique was decided. A scalpel blade or edge of a glass slide was taken for homogenous surfaces of the lesion to scrape the surface and prepare smears. If the lesion had a variegated appearance, the scrapping was done from the most suspicious area. A minimum of three smears were prepared [Figure 1a and b]. Two smears were immediately fixed in absolute alcohol and stained with the rapid hematoxylin and eosin (H and E) technique, and one smear was air-dried and stained with Giemsa. For FS, the most suspicious area was sampled. The sections were cut at 5 mm thickness using a Leica cryotome and stained with rapid H and E for evaluation. Time was noted for each technique from the start of the procedure to the final diagnosis. The remaining sample was sent to the Histopathology Lab for routine tissue processing and subsequent paraffin sectioning. The scrape cytology smears were evaluated by a single cytopathologist who was blinded to the results of FS. All cases were categorized into three diagnostic groups: Group 1: Benign, Group 2: Borderline and Group 3: Malignant. The results for both scrape cytology and FS were recorded separately. The final paraffin section histopathological diagnosis was considered the gold standard.

Figure 1.

Technique of smear preparation for scrape cytology. (a) A scalpel blade used for a homogeneous surface. (b) Spreading of the obtained material over the slide.

Statistical analysis

Statistical analyses were performed to compare the performances of FS and scrape cytology. Concordance between FS and scrape cytology was evaluated using Cohen’s Kappa. The association between tissue type and concordance/discordance was analyzed using the Chi-square Test of Independence. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of scrape cytology and FS were calculated independently compared to the gold standard histopathology. Paired t tests and the Wilcoxon signed-rank test were used to assess time differences between the two methods.

RESULTS

Patient demographics

The study included 80 consecutive intraoperative consultations with patients aged 18 to 82 years. There were 52 benign cases and 28 malignant cases. The samples comprised various tissue types, including predominantly lymph nodes, parathyroid, ovary, salivary glands, bladder wall, inguinal masses, thyroid, and pelvic masses [Figures 2,3,4,5,6; Table 1].

Figure 2.

Parathyroid adenoma: (a: Scrape cytology and b: frozen section) (10×, hematoxylin and eosin stain)

Figure 3.

Pleomorphic adenoma: (a) Scrape cytology showing the basaloid cells intermingled in the chondromyxoid stroma. (b) Frozen section showing features of pleomorphic adenoma. (10×, hematoxylin and eosin stain)

Figure 4.

Papillary thyroid carcinoma: (a) Scrape cytology showing papillaroid architecture with cellular details. (10×, Giemsa), (b) frozen section showing papillary architecture with papillary-like nuclear details. (40×, hematoxylin and eosin stain)

Figure 5.

Reactive lymphoid hyperplasia (a: Scrape cytology and b: frozen section) (10×, hematoxylin and eosin stain)

Figure 6.

Metastatic breast carcinoma (a: Scrape cytology and b: frozen section) (40×, hematoxylin and eosin stain)

Table 1.

The concordance and discordance among frozen section and scrape cytology in different lesions

S No.	Case	Concordant Cases (Confirmed by Histopathology	Discordant Cases (missed on scrape cytology)	Final Histopathology Diagnosis
1	Lymph node	53 (42 benign, 14 malignant)	03 (2 benign, 1 malignant)	Benign: Chronic nonspecific reactive lymphoid hyperplasia (42) malignant: metastatic carcinoma deposits (14)
2	Parathyroid	01 (benign)	03 (benign)	Parathyroid adenoma
3	Ovary	06 (4 malignant, 2 benign)	00	Benign: fibrothecoma (1), luteoma (1) malignant: Serous carcinoma (2), mixed germ cell tumor (2)
4	Salivary Gland	04 (2 benign, 2 malignant)	01 (malignant)	Benign: Parathyroid adenoma (1), Warthin tumor (1) Malignant: Mucoepidermoid carcinoma (2), Lymphoma (1)
5	Bladder wall nodule	01 (malignant)	00	Malignant: Tumor deposits from serous carcinoma ovary
6	Inguinal mass	01 (malignant)	00	Malignant: Lymphoma
7	Thyroid	05 (2 benign, 3 malignant)	01 (malignant)	Benign: Colloid goiter (2) Malignant: Papillary thyroid carcinoma (2), Anaplastic carcinoma (1)
8	Pelvic mass	01 (malignant)	00	Malignant: High-grade serous Carcinoma

Overall, 70 out of 80 cases showed concordance between FS and scrape cytology, resulting in a concordance rate of 87.5%. Cohen’s Kappa for agreement between FS and scrape cytology was 1.0, indicating perfect agreement between the two methods. The cases with discordance were studied in detail, and efforts were taken to evaluate the cause of discrepancy in the results. Details of discordant cases are presented in Table 2.

Table 2.

Reason for discordance among frozen section and scrape cytology

Sl. No	Case	Scrape cytology	Frozen section	Reason for discrepancy
1.	Cervical lymph node (02 cases)	Poorly preserved cells	Positive for malignancy	Crushing artifact
2.	Right level II lymph node (01 case)	Negative for malignancy	Positive for malignancy	Focus of infiltration in the lymph node was not scraped
3.	Parathyroid tissue (03 cases)	Not present/ probably thyroid	Parathyroid identified	Singly scattered parathyroid cells. Cell morphology could not be discerned as parathyroid
4.	Salivary Gland (01 case)	Non-representative	Malignant	Necrotic tissue only
5.	Thyroid (01 case)	Poorly preserved cells	Poorly differentiated carcinoma	Crushing artifact

Diagnostic performance

Overall, the diagnostic accuracy of scrape cytology was 90% (95% CI, 0.834–0.966). The sensitivity was 89.3% (95% CI, 0.778–1.000), and specificity was 90.4% (95% CI, 0.824–0.984). The PPV was 83.3% (95% CI, 0.695–0.971), and the NPV was 94% (95% CI, 0.875–1.000). All tissues were accurately identified by FS, with a specificity of 100% and sensitivity of 100%. The receiver operating characteristic (ROC) curve for scrape cytology yielded an area under the curve (AUC) of 0.90, indicating excellent diagnostic performance. The high AUC suggested that scrape cytology had a strong ability to distinguish between benign and malignant lesions. The average time for reporting was significantly shorter for scrape cytology than for FS. Scrape cytology averaged 10 min, while FS averaged 20 min. Both the paired t test (P < 0.001) and the Wilcoxon signed-rank test (P < 0.001) confirmed that this time difference was statistically significant.

Tissue-specific variability

The chi-square test of independence (χ² = 21.97, P = 0.003) showed a statistically significant association between tissue type and concordance/discordance, indicating that certain tissue types, such as parathyroid, are more likely to show discordant results, which aligns with our observations about parathyroid cases.

DISCUSSION

Intraoperative diagnosis plays a crucial role in guiding surgical decision-making, enabling surgeons to assess tumor margins, confirm the nature of lesions, and determine the extent of tissue resection. This real-time evaluation can prevent the need for unnecessary radical surgeries, reduce the need for second surgeries, and minimize patient morbidity. Several techniques are available for intraoperative diagnosis, including FS, imprint cytology, fine-needle aspiration cytology (FNAC), and scrape cytology. Among these, FS is widely considered the gold standard as it provides rapid histopathological evaluation. However, FS requires specialized equipment (cryostats) and trained personnel, resources that are not available in all healthcare settings, particularly in resource-limited centers.

Scrape cytology, a simpler and more cost-effective alternative, involves scraping cells from freshly cut surfaces of surgical specimens for immediate cytological evaluation. Our study compared scrape cytology and FS across 80 consecutive intraoperative samples, evaluating their concordance, time efficiency, and diagnostic accuracy. The sensitivity (89.29%) and specificity (90.38%) of scrape cytology in our study were impressive, demonstrating its utility in most tissue types. The overall diagnostic accuracy was 90%, with scrape cytology offering a faster turnaround time, making it especially valuable in time-sensitive surgeries. Our findings are comparable to those reported in previous studies.[10,11,12,13,14] For instance, Tanya Sharma et al. (2009) reported the diagnostic utility of scrape cytology in ovarian neoplasms, highlighting its accuracy of 96%.[11] Similarly, Kolte and Satarkar evaluated scrape cytology in the intraoperative diagnosis of tumors and found a diagnostic accuracy of 93.5%, emphasizing that it can serve as an effective tool when FS is not available.[12] Malkiely et al. also found that the sensitivity of the cytological smears in their study was comparable to that of FS, and due to the shorter analysis time required, it was the preferred method at their institution.[15] Agarwal et al. performed a study on 61 cases of ovarian mass and compared the scrape cytology with the histopathology and found that compared with histopathology, the sensitivity and specificity of the scrape cytology technique were 93.55% and 96.67%, respectively, and the diagnostic accuracy was 95.08%.[16] Samaddar and Talukdar[17] also performed a comparative study on ovarian neoplasm and revealed that the accuracy of scrape cytology in the diagnosis of benign, borderline, and malignant ovarian tumors was 95.45%, 40%, and 89.74%, respectively, and the overall accuracy rate was 90.91%. Our study results align with these findings, further supporting the efficacy of scrape cytology in neoplastic lesion evaluation. The high AUC (0.90) of the ROC curve in our study further validates the strong diagnostic performance of scrape cytology.

In the present study, scrape cytology demonstrated several advantages, particularly in appreciating cellular and nuclear details. These finer details were more easily discernible on cytology smears compared to FSs. For instance, features such as extracellular mucin and basement membrane-like material were appreciated on cytology smears, aiding in diagnosis.

The lymph node scrape cytology was performed in 53 cases, among that three cases showed discordance with the frozen and histopathology. The majority of the cases were sentinel lymph nodes and diagnosed as reactive lymph node on both scrape cytology and FS. The reasons for discordance in case of lymph nodes were poor technique of spreading, and one case was having micro metastatic focus (cluster of 4–5 malignant cells), which was missed in the scrapping. In the case of papillary carcinoma of the thyroid, the nuclear features characteristic of this malignancy were better appreciated on cytology. Three instances of papillary carcinoma were diagnosed primarily through the cytological examination of nuclear grooves and inclusions, which are hallmark features [Figure 4a and b]. The authors Khuroo et al.[18] conducted a study on 50 cases of thyroid neoplasm by obtaining scrapings from the fresh cut surface of thyroid specimens before formalin fixation and compared the cytological features with the preoperative FNA cytology features and concluded that scrape cytology can act as a useful adjunct to FNAC and is a good procedure for intraoperative consultation. It is particularly beneficial in cytologically diagnosed suspicious cases of papillary carcinoma and may prevent the need for a second procedure for the completion thyroidectomy. Scrape cytology is a simple, rapid, and inexpensive adjunctive technique that should be used in intraoperative consultation for thyroid lesions. They also found that scrape cytology has limitations in the diagnosis of over-fixed and small tumors (micro-papillary carcinoma), which require extensive sampling and histologic architectural evaluation.[19] In the present scenario, scrape cytology and FS were planned to know the nature of the nodule due to the high clinical suspicion of the thyroid nodule, but FNA features and radiology were not concordant. In one case of PTC, scraping and frozen were performed to detect the extrathyroidal extension and assist the surgeon in determining the extent of thyroid surgery and central neck dissection.

Similarly, scrape cytology proved advantageous in pleomorphic adenoma due to the well-preserved bland cellular details and the distinct presence of chondromyxoid matrix, which facilitated an accurate diagnosis in two cases [Figure 3a and b]. These findings highlight the ability of scrape cytology to reveal the subtle cytological features essential for identifying certain neoplasms. On the other hand, FS provided a significant advantage in appreciating the architectural patterns of tissues, which are often crucial for accurate diagnosis. In particular, identifying parathyroid glands was challenging on cytology but readily apparent on FS. The architectural arrangement of parathyroid acini interspersed with fat was easily identified on FS, which contributed to the correct diagnosis [Figure 2a and b]. This difficulty in identifying parathyroid tissue on cytology could also be attributed to the lack of prior experience of the cytopathologist in handling cytological smears of parathyroid glands. However, this highlights another important aspect of scrape cytology: its potential as a learning tool. By encountering such rare lesions during intraoperative consultations, cytopathologists can gain valuable experience and become more proficient in identifying less frequently seen tissue types. Rohaizak et al.[19] performed a prospective study on 29 cases of parathyroid scrape cytology, and it was well correlated with histopathology in 25 cases. The four cases were not correlated due to insufficient samples for the diagnosis and, hence, concluded that it is a rapid, economical test with acceptable sensitivity and high specificity. It can be used as an adjunct to FS and may be used as a tool in helping surgeons to identify parathyroid tissue.

One difficulty of scrape cytology observed in our study was the preparation technique. In some cases, excessive pressure applied during smearing resulted in crushing artifacts, which distorted the morphology of the cells and complicated the interpretation. This was particularly evident in a few lymph node cases, where the excessive pressure made it difficult to appreciate the normal or malignant cellular details. Therefore, we emphasize the importance of careful preparation techniques for cytology smears to avoid artifacts that could hinder diagnosis.

Moreover, the sampling site is critical for scrape cytology and FS. In cases of heterogeneous lesions, the most suspicious areas should be carefully selected for sampling to ensure that the tissue analyzed is representative of the pathology. Proper sampling site selection increases the likelihood of an accurate diagnosis, regardless of the technique used. While this study highlights the effectiveness of scrape cytology, it is essential to acknowledge certain limitations. The study’s sample size (80 cases) is relatively small, and a larger cohort would help better validate the findings across different tissue types.

Despite its advantages, scrape cytology has certain limitations that must be acknowledged. It is unsuitable for mucinous tumors due to the gelatinous nature affecting smear preparation. Additionally, its role in detecting micrometastases in lymph nodes is limited as small tumor clusters may not be adequately sampled. Similarly, in necrotic or hemorrhagic tumors, smear quality may be compromised, leading to non-representative findings. Furthermore, like FNA, scrape cytology does not provide insights into tumor depth or capsular invasion, making it less reliable for assessing infiltrative growth. These limitations should be considered when selecting the most appropriate intraoperative diagnostic technique. It is also not suitable for the detection of neuroepithelium in case of immature teratoma where extensive histopathological sections are required. It is also not helpful in the characterization of soft tissue tumors.

CONCLUSION

In conclusion, our study demonstrates that scrape cytology is a highly accurate and time-efficient alternative to FSs for intraoperative tumor diagnosis, especially in resource-limited settings. It performs well across most tissue types, with high sensitivity and specificity, and offers significant time savings compared to FS. However, tissue-specific challenges, such as those encountered in the parathyroid gland, underscore the importance of proper training and experience in cytological techniques. Improving smear preparation and increasing pathologists’ exposure to challenging tissue types can further enhance the diagnostic reliability of scrape cytology. As a cost-effective tool that requires minimal equipment, scrape cytology should be considered in centers where FS is unavailable or impractical, providing timely and accurate intraoperative diagnosis. Also, it helps in learning and teaching cytological features of tumors and can help improve the interpretation skills of FNAC.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.