Comparison of staining quality between rapid and routine hematoxylin and eosin staining of frozen breast tissue sections: an observational study

Abstract

Objective

To compare the staining quality between rapid hematoxylin and eosin (H&E) staining and routine H&E staining of frozen breast tissue sections.

Methods

In this cross-sectional observational study, 120 frozen breast tissue sections were randomly assigned to rapid or routine H&E staining (n = 60 per group). Rapid H&E staining used a 7:1 mixture of modified Gill’s hematoxylin and alcohol-soluble 1% eosin Y. The staining quality of each section was evaluated and scored. A score of >7 was considered excellent, a score of 6 to 7 good, and a score of ≤5 poor.

Results

The staining time for rapid staining was approximately 3 minutes, whereas that of routine staining was approximately 12 minutes. There were no significant differences in the staining quality scores or proportions of sections in each grade between the two staining methods. The proportions of sections that were classified as excellent or good were 96.7% and 98.3% for rapid and routine staining, respectively.

Conclusions

In frozen breast tissue sections, rapid H&E staining may provide staining quality that is comparable to that of routine staining, while markedly reducing the staining time.

Introduction

Breast cancer is the most common malignancy in women, with an estimated 2.3 million new cases worldwide in 2020. ¹ Surgery is the primary treatment for breast cancer. Histological staining of an intraoperative frozen section is usually required during lumpectomy to determine the type of breast tumor, assess the sentinel lymph node metastasis, and identify the resection margin.^2–4 Speed and reliability are the fundamental requirements for frozen section analysis. ²

Hematoxylin and eosin (H&E) staining is the most commonly used staining method for frozen sections; it demonstrates a wide range of cytoplasmic, nuclear, and extracellular matrix features.^4,5 However, routine H&E staining involves a two-step staining process (i.e., hematoxylin staining of the nuclei followed by eosin staining of the cytoplasm),^6,7 which results in a staining period that ranges from 10 to 15 minutes for frozen sections. Together with the sample processing and section preparation steps, it takes approximately 30 minutes to provide pathology results from frozen sections, which increases the total operating time for patients undergoing lumpectomy. ⁸ Thus, any modifications to the H&E staining procedure that can reduce staining time but maintain staining quality will be very valuable for clinical and diagnostic reasons—especially in cases where a rapid pathological diagnosis is required, such as for frozen sections and on-site evaluations.

At our institution, we have recently developed a rapid (one-step) H&E staining method using a combined mixture of H&E, which allows for simultaneous H&E staining and eliminates the intermediate steps of differentiation, bluing, and dehydration during routine staining. The staining time with this rapid H&E staining method is approximately 3 minutes. However, the main challenge of rapid H&E staining is to ensure consistent and repeatable staining quality. In the present study, we therefore aimed to compare the staining quality between rapid H&E staining and routine H&E staining of frozen sections of breast tissue, and to test the hypothesis that rapid H&E staining is a reasonable alternative to routine H&E staining in this setting.

Methods

Study design and patients

This was a cross-sectional observational study. The Ethical Committee of Jiaxing Hospital of Traditional Chinese Medicine, China, approved the study on 13 February 2023 (No. SL-2023-0010). The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. ⁹ Consecutive patients scheduled for lumpectomy requiring intraoperative frozen section evaluation between 15 February 2023 and 12 May 2023 were enrolled. All patients gave written consent after a written and verbal explanation of the study. We have de-identified all patient details so that they cannot be identified in any way in this paper.

Staining methods

All breast tissue specimens were submitted to the Pathology Department for standard frozen section processing. Two consecutive 5-μm frozen sections were made and coded: one for rapid H&E staining and one for routine H&E staining. Each section was randomly assigned a staining method.

Before proceeding with the present study, we performed a pilot study using 100 frozen sections to investigate the optimal H&E mixture and timing for rapid H&E staining. First, we compared the staining results of different hematoxylin solutions mixed with eosin solutions on 40 sections. A mixture of modified Gill’s hematoxylin and alcohol-soluble 1% eosin Y gave better staining results than Gill’s hematoxylin with water-soluble eosin Y. Second, we compared the effects of different ratios (6:1, 7:1, and 8:1) of modified Gill’s hematoxylin to alcohol-soluble 1% eosin Y with different staining times (1.5, 2, 2.5, 3, 3.5, and 4 minutes) on 60 sections. The optimal staining results were observed using a ratio of 7:1 with a staining time of 3 minutes. We therefore used a 7:1 mixture of modified Gill’s hematoxylin and alcohol-soluble 1% eosin Y for rapid H&E staining in the current study.

For the routine H&E stain, we used modified Gill’s hematoxylin and alcohol-soluble 1% eosin Y. The differences between the modified and standard Gill’s hematoxylin solutions are summarized in Table 1. The mixture of modified Gill’s hematoxylin and alcohol-soluble 1% eosin Y was freshly prepared prior to use. The routine H&E staining protocol comprised hematoxylin, water wash, differentiation, bluing, eosin Y, water wash, and dehydration steps. By contrast, the rapid H&E staining protocol only comprised a mixture of H&E and a water wash. The rapid and routine H&E staining protocols are shown in Figure 1. All sections were stained manually by one technician to avoid inconsistencies.

Table 1.

Comparison between modified Gill’s hematoxylin and Gill’s hematoxylin solutions.

Reagents	Modified Gill’s hematoxylin	Gill’s hematoxylin
Hematoxylin, g	4	2
100% ethanol, mL	250	250
Aluminum ammonium sulfate, g	18	17.6
Distilled water, mL	750	750
Sodium iodate, g	0.3	0.2
Glacial acetic acid, mL	10	20

Figure 1.

Staining protocols for the rapid and routine hematoxylin and eosin (H&E) staining methods.

Staining quality assessment

Two independent pathologists with more than 15 years of experience blindly assessed the staining quality to reduce subjective bias. The stained sections were viewed under a light microscope at 200× and 400× magnification in the order in which they were obtained. Each section was evaluated and scored based on nuclear staining, cytoplasmic staining, and clarity of staining (Table 2). Scores for each section were then summed. A score of >7 was graded as excellent, a score of 6 or 7 as good, and a score of ≤5 as poor. Sections classified as excellent or good were considered to be adequate for diagnosis, whereas those classified as poor were considered to be inadequate for diagnosis. In cases of disagreement between the two observers, a third senior pathologist was consulted for arbitration.

Table 2.

Scoring criteria for hematoxylin and eosin staining quality.

Score	Nuclear staining	Cytoplasmic staining	Clarity of staining
3	Blue-purple	Light pink	Distinct
2	Navy blue	Dark peach or red	Inconspicuous
1	Gray-blue	Deep red or purplish red	Ambiguous

Statistical analyses

Continuous data were tested for normal distribution using the Shapiro–Wilk test. Non-normally distributed continuous data are expressed as the median (interquartile range) and were analyzed using the Mann–Whitney U test. The correlation between rapid and routine H&E staining in staining quality grading was assessed with linear-by-linear association using the chi-squared test. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 20.0 (IBM Corp., Armonk, NY, USA), and P < 0.05 was considered significant.

Results

A total of 120 frozen sections of breast tissue were studied (two sections each from 60 patients). Sixty sections were stained with rapid H&E staining and 60 were stained with routine H&E staining, resulting in 60 matched pairs. The final pathological diagnoses of the 60 frozen sections were breast cancer (7), breast fibroma (41), and breast cystic hyperplasia (12).

For routine H&E staining, it took approximately 12 minutes to complete the staining process, whereas for rapid H&E staining, it took approximately 3 minutes (Figure 1).

The sections in both groups were similar in morphology and staining characteristics (Figure 2). The median (interquartile range) staining quality score was 8 (8–9) in the rapid H&E staining group and 8 (8–9) in the routine H&E staining group, with no significant difference between the two methods. Similarly, there was no significant difference in the proportion of stained sections in each grade between the two groups (Table 3). The proportion of sections graded as excellent or good was 96.7% in the rapid H&E staining group and 98.3% in the routine H&E staining group.

Figure 2.

Photomicrographs of frozen sections of breast cancer stained with rapid and routine hematoxylin and eosin (H&E) staining methods under 200× and 400× magnification.

Table 3.

Staining quality grading of rapid and routine hematoxylin and eosin (H&E) staining.

Grade	Rapid H&E stain (n = 60)	Routine H&E stain (n = 60)
Excellent	53 (88.3)	55 (91.7)
Good	5 (8.3)	4 (6.7)
Poor	2 (3.3)	1 (1.6)

Data are presented as the number (%).

Discussion

To reduce staining time, we developed a rapid, one-step H&E staining method that used a mixture of H&E instead of the respective staining of first hematoxylin and then eosin. Unlike the routine H&E staining method, the rapid H&E method did not require intermediate steps such as differentiation, bluing, or dehydration, thus further decreasing the staining time. However, it is essential to note that reducing the staining time did not affect the staining quality. In the present observational study, of 60 matched pairs of frozen breast tissue sections, 96.7% of the rapid H&E-stained sections were classified as diagnostic compared with 98.3% of the routine H&E-stained sections, suggesting that the rapid H&E staining method does not markedly affect the staining quality of frozen sections.

Intraoperative frozen section analysis plays a vital role in the surgical treatment of breast cancer, for the identification of tumor type, sentinel lymph node metastasis, and the resection margin.^2–4 However, the analysis of intraoperative frozen sections adds extra time to the operating period. ⁸ The most commonly used stain for frozen sections is H&E staining.^4,5 However, although H&E staining of frozen sections takes a shorter time than that of paraffin-embedded sections, it still requires approximately 10 to 15 minutes to complete the staining process. Reducing the staining steps without compromising staining quality is therefore clinically valuable.

H&E staining has been used for well over a century, and remains the cornerstone of the histopathological diagnosis of biopsy specimens. It is relatively simple to perform in many areas of the histology laboratory, and can be conducted using frozen sections, fine needle aspirates, or paraffin-embedded tissue. When performed correctly, H&E staining can demonstrate a wide range of normal and abnormal cellular and tissue components.^6,7 In addition to proper specimen processing, the result of H&E staining depends on the stains used and their components. Basophilic hematoxylin can accurately stain nuclear components, including heterochromatin and nucleoli, whereas acidophilic eosin can stain cytoplasmic components, including collagen, elastic fibers, muscle fibers, and red blood cells. The contrast between H&E allows for the easy differentiation of cellular components.

There are a number of different H&E formulations in common use, each with various advantages and disadvantages. Gill’s hematoxylin is an alum hematoxylin that is available at different concentrations. It can be used as a progressive or regressive stain. Eosin Y is the most commonly used form of eosin, and can be used in water and alcohol. A variety of combinations of both H&E stains can be used to achieve the desired staining results. Relative to Gill’s hematoxylin solution, modified Gill’s hematoxylin solution has a higher concentration of hematoxylin, aluminum ammonium sulfate, and sodium iodate, but a lower concentration of glacial acetic acid (Table 1), thereby shortening the nuclear staining time. The lower concentration of glacial acetic acid reduces the co-staining of nuclei and cytoplasm. In our pilot study using different mixtures of H&E, we found that modified Gill’s hematoxylin mixed with alcohol-soluble 1% eosin Y produced better staining (in terms of sufficient clarity, with crisp nuclear and cytoplasmic staining) than that mixed with water-soluble 1% eosin Y. This may be because alcohol-soluble eosin stains more sharply with hematoxylin than water-soluble eosin. In addition, the overall color of each section is the result of an appropriate balance between the intensity of the H&E and the length of time the section spends in the stains. We therefore investigated the staining of mixtures with different ratios of modified Gill’s hematoxylin to alcohol-soluble 1% eosin Y (6:1, 7:1, or 8:1) using different staining times in our pilot study; the 7:1 ratio gave the best staining result at a staining time of 3 minutes. Consequently, we adopted a mixture of modified Gill’s hematoxylin and alcohol-soluble eosin Y in a 7:1 ratio for rapid H&E staining.

In addition to the single (rather than double) staining step, our rapid H&E staining method eliminated the intermediate steps of differentiation, bluing, and dehydration during the staining process. The staining time was therefore reduced from approximately 12 minutes (with the routine H&E staining method) to approximately 3 minutes (with the rapid H&E staining method). Furthermore, there was no significant difference in staining quality between the two methods, and 96.7% of the rapidly stained sections were classified as diagnostic compared with 98.3% of the routinely stained sections. Our results may therefore provide a promising option for the rapid H&E staining of frozen breast tissue sections.

The current study has several limitations. First, given the observational nature of the study and the relatively small sample size, no definitive conclusions can be drawn, unlike with a randomized controlled trial. ¹⁰ Second, our study included only frozen breast tissue sections; the results may thus not be generalizable to other tissue sections or cytological smears. Third, the main disadvantage of rapid H&E staining is that the mixture of H&E is unstable and cannot be stored for long periods. This means that the mixture has to be prepared immediately before use, which increases the overall workload for technicians. Future research should therefore investigate ways to improve the long-term stability of the H&E mixture.

In conclusion, our study revealed that rapid H&E staining using a 7:1 mixture of modified Gill’s hematoxylin and alcohol-soluble eosin Y may provide staining quality that is comparable with that of routine H&E staining in frozen breast tissue sections, while markedly reducing the staining time. Rapid H&E staining may therefore be a promising alternative to routine H&E staining, especially in clinical scenarios such as for frozen sections and/or rapid on-site evaluation. However, further studies are required to validate the effectiveness of the rapid H&E staining method on other tissue sections and cytological smears, and to improve the long-term stability of the H&E mixture.