SEARCH

SEARCH BY CITATION

Keywords:

  • estrogen receptor;
  • cytology;
  • immunocytochemistry;
  • fixative;
  • antigen retrieval;
  • Abbott method;
  • formalin;
  • Papanicolaou (Pap);
  • Carnoy's fixative

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND

The techniques for immunostaining estrogen receptor (ER) in cytologic specimens have varied, as have the detection rates. The authors compared various fixation methods for their effect on ER detection in cytologic smears of breast carcinoma.

METHODS

Smears were prepared by gently scraping the cut surfaces of 47 resected breast carcinoma specimens and placing immediately in 1of the following conditions: 1) a sequence of 10% formalin-methanol-acetone fixatives at −20 °C (Abbott method); 2) 10% formalin at room temperature; and 3) Carnoy's fixative at room temperature and then Papanicolaou stained (Carnoy's-Pap). Destaining of Carnoy's-Pap smears (Carnoy's-dPap) was initially attempted before ER staining. One set of smears was also air-dried for 3 minutes before using the Abbott method. Smears and corresponding tissue sections were immunostained with anti-ER antibody 6F11 using a similar protocol except for antigen retrieval, which was not initially applied on cytologic slides. All the ER-negative smears that had been fixed with 10% formalin or Carnoy's-Pap were restained after antigen retrieval. Agreement between cytologic and histologic findings was expressed by both concordance and the kappa coefficient.

RESULTS

ER detection in smears processed with the Abbott method correlated best with findings from tissue samples, with an overall correlation of 91.5% (κ = 0.80). Findings from air-dried smears were less optimal (concordance, 84.4%; κ = 0.65), followed by Carnoy's-Pap (concordance, 71.4%; κ = 0.45), formalin (concordance, 31%; κ = 0.05), and Carnoy's-dPap (concordance, 29.4%; κ = 0.04). Antigen retrieval converted most of the ER-negative smears to positive (18 of 20 smears in formalin and 6 of 8 smears in Carnoy's-Pap), leading to a final concordance of 93% and κ = 0.83 for both conditions. Antigen retrieval also led to stronger staining intensity without causing false positivity.

CONCLUSIONS

Antigen retrieval was found to greatly improve ER immunodetectability and staining intensity in formalin-fixed and Carnoy's-Pap smears. The former may offer an alternative to the Abbott method because of its easiness and the latter can be reliably used in archival Pap-stained smears for retrospective analysis of ER. Air-drying, destaining Pap smear, and fixation in formalin or Carnoy's-Pap without antigen retrieval are not recommended. Cancer (Cancer Cytopathol) 2004;102:34–40. © 2003 American Cancer Society.

Determination of estrogen receptor (ER) status by immunostaining is now standard practice in the evaluation of patients with breast carcinoma. This information is principally used to determine treatment strategy, predict response to hormonal therapy, and determine prognosis.1–4 Although ER status is usually evaluated on surgically removed tissue specimens, evaluation of ER in cytologic material would be advantageous for some patients, particularly those undergoing preoperative chemotherapy or irradiation5, 6 or those for whom surgery would otherwise not be indicated because of advanced inoperable primary tumor, metastases, local disease recurrence, or malignant effusions.7–9 Assessment of ER status in cytologic specimens may also be a useful diagnostic adjunct in the evaluation of metastatic tumors of unknown origin.10

Although immunohistochemical (IHC) methods for determining ER status have been standardized, immunocytochemical methods (ICC) for determining ER status in cytologic smears have varied. In addition to the Abbott method,8 which was originally introduced with a first-generation ER antibody (i.e., H222) and required fixing samples in a formalin-methanol-acetone sequence at −20 °C, a variety of other fixatives have been used along with subsequently developed antibodies. These included the use of periodatelysine-paraformaldehyde at room temperature (RT), a formalin-acetone sequence at −10 °C,11 95% alcohol,12–14 10% buffered formalin at RT,15 acetone/formalin fixative,16 and cytospin collection fluid, a mixture of ethanol, isopropanol, and polyoxyethylene.17 These methods, however, have produced conflicting results. The purpose of the current study was to evaluate the effect of the most commonly used fixation methods on ER detection in cytologic smears of breast carcinoma. Results from ICC were compared with those obtained by IHC of the corresponding tissue sections.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patient Population

Breast carcinoma specimens were obtained from 47 women (age range, 25–85 years) who underwent surgery at The University of Texas M. D. Anderson Cancer Center (MDACC; Houston, TX) between June 2002 and January 2003. The patients had infiltrating ductal carcinoma (n = 41), invasive lobular carcinoma (n = 3), and mixed invasive ductal and lobular carcinoma (n = 3). Forty-four specimens were from primary lesions and 3 from recurrent carcinomas. The study was conducted with the approval of the MDACC surveillance committee (protocol number LAB02-558).

Collection and Processing of Cytologic and Histologic Materials and Immunostaining

Biopsy, lumpectomy, and mastectomy specimens were brought to the frozen section suite immediately after surgical excision. To avoid dilution of tumor cells by nonneoplastic components, only specimens with macroscopically identifiable tumors were included in the current study. As soon as a specimen was received, the tumor was sectioned and cytologic smears were prepared by gently scraping the cut surface of the tumor using a scalpel blade. For each specimen, one smear was stained with Diff-Quik (Statlab, Lewisville, TX) to confirm the presence of adequate numbers of malignant cells. The other smears were processed and kept in one of the following conditions until ER immunostaining (Table 1): 1) Abbott method, 2) air-Abbott method, 3) 10% formalin, and 4) Carnoy's fixative.

Table 1. Fixation Conditions for Estrogen Receptor Immunoanalysis in Cytologic Smears
StepsAbbottAir-Abbott10% formalinCarnoy's
  1. Abbott: Abbott method; Air-Abbott: air-drying before fixation with the Abbott method; min: minutes; RT: room temperature; Pap: Papanicolaou; PBS; phosphate-buffered saline.

Time before fixationNone (immediate)3 min air-dryNone (immediate)None (immediate)
10% phosphate-buffered formalinYes, 10 minYes, 10 minYes, 10 minNo
6:1 70% ethanol-to-acetic acidNoNoNoYes, 10 min
Fixative temperature4 °C4 °CRTRT
Postfixation treatmentMethanol, −20 °C, 4 min; acetone, −20 °C, 2 min; PBS, 5 min × 2Methanol, −20 °C, 4 min; acetone, −20 °C, 2 min; PBS, 5 min × 2NonePap staining
Storage conditionGlycerol-sucroseGlycerol-sucroseAir-dryNone
Storage temperature−20 °C−20 °C−20 °CRT

Using the Abbott method, slides were placed immediately in 10% phosphate-buffered formalin at 4 °C for 10 minutes, followed by methanol and acetone at −20 °C, rinsed in phosphate-buffered saline, and stored in a cold glycerol/sucrose storage medium at −20 °C. The method was initially recommended by Abbott Laboratories (Abbott Park, IL).8 In the air-Abbott method, smears were air-dried for 3 minutes and subjected to the process described above. In the third method, slides were placed immediately in 10% phosphate-buffered formalin at RT for 10 minutes and stored air-dried at −20 °C. Using the final method, slides were placed immediately in Carnoy's fixative (a 6:1 ratio of 70% ethanol to glacial acetic acid) at RT for 10 minutes and subjected to routine Papanicolaou (Pap) staining. The Pap-stained slides (Carnoy's-Pap) were kept in cardboard slide folders at RT for up to 16 weeks. For ER immunostaining, the coverslips of Carnoy's-Pap smears were removed by immersion in xylene for 18–24 hours and the slides were then processed through xylene and grades of alcohol. In the first 17 cases, Pap-stained slides were destained (Carnoy's-dPap) with acid alcohol (1% HCl in alcohol) before immunostaining for ER. However, problems with low detectability and poor staining quality led us to abandon the destaining procedure for the subsequent 28 cases. Instead, ER staining was performed directly on the Carnoy's-Pap smears without destaining.

The IHC results of the tissue sections from which the scrape preparations were obtained were considered the gold standard for comparison of ER status. Ten percent formalin-fixed and paraffin-embedded tumor tissue specimens were sectioned at 4 μm and deparaffinized. Antigen retrieval was performed by steaming the tissue sections in 10 mM citrate buffer (pH 6.0) for 45 minutes. The remaining steps were identical for the smears and the tissue sections. For example, the samples were incubated with primary mouse monoclonal antibody 6F11 against the ER (Novacastra/Vector Laboratories, Burlingame, CA) at a dilution of 1:50 for 30 minutes at RT. The EnVision+ method was employed on Dako Autostainer instrument for the rest of the procedure according to the manufacturer's instructions (Dako Corporation, Carpinteria, CA). The slides were then counterstained with hematoxylin, cleared, and mounted. Appropriate negative and positive controls were included.

In an attempt to retrieve the ER antigen after fixation-induced epitope loss in cytologic smears, an antigen retrieval technique similar to that used for the tissue sections was applied on all negatively-staining smears that had been fixed in 10% formalin or subjected to the Carnoy's-Pap procedure, followed by repeat ER staining.

Scoring of Estrogen Receptor Positivity and Statistical Analysis

At least 100 tumor cells were evaluated per slide by two pathologists. Immunoreactivity was scored as the percentage of positive nuclei. Specimens were interpreted as positive for ER if ≥ 10% of the cells demonstrated nuclear staining.18, 19 Staining intensity was graded semiquantitatively as negative, weak, or strong. A mean value was assigned for smears in which the staining intensity varied from area to area.

Results of the ICC of the smears fixed by each technique were compared with those obtained by IHC of the tissue sections and the sensitivity and specificity for each fixation method were calculated. Two-by-two contingency tables were constructed and agreement between smears and tissue sections was expressed by both concordance and the Cohen kappa coefficient. The relation between the kappa value and the level of agreement was suggested by Landis and Koch20: a κ value of 0.00–0.20 indicates slight agreement; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1.00, almost perfect.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

ER expression was detected in 36 (76.6%) of the 47 tissue sections. ER immunoreactivity in the cytologic smears without antigen retrieval and in the corresponding tissue sections is shown in Table 2. Staining results for the smears that had been processed immediately by the Abbott method agreed best with those of the tissue sections, with an overall correlation of 91.5% (43 of 47) and a κ value of 0.80 (P < 0.0001), indicating substantial agreement.20 Air-drying the smear before applying the Abbott method (Air-Abbott) yielded a lower concordance rate (84.4% [38 of 45]) because of the lower detection rate. The other fixation conditions led to much lower concordance (71.4% for Carnoy's-Pap, 31.0% for formalin, and 29.4% for Carnoy's-dPap) and much less staining intensity (Fig. 1 and Table 2). No false-positivity was associated with any method. Some variability of staining intensity was observed in different areas of individual smears, most often in specimens with lower ER scores. In contrast, samples with higher percentages of ER-positive cells stained more intensely and less heterogeneously.

Table 2. Estrogen Receptor Immunoreactivity in Cytologic Smears Processed by Different Methods (without Antigen Retrieval) versus That in Corresponding Histologic Sections
CharacteristicsCytologic smears (ICC)
Abbott (n = 47)Air-Abbott (n = 45)a10% formalin (n = 29)bCarnoy's-dPap (n = 17)cCarnoy's-Pap (n = 28)c
+++++
  • ICC: immunocytochemistry; Abbott: Abbott method; Air-Abbott: air-drying before the Abbott method; Carnoy's-dPap: Papanicolaou-stained smears with destaining; Carnoy's-Pap: Papanicolaou-stained smears without destaining; IHC: immunohistochemistry; ER: estrogen receptor.

  • a

    Two smears were excluded because of poor cellularity.

  • b

    Formalin fixation was not available in the first 18 cases.

  • c

    Because of poor estrogen receptor detection in the first 17 Carnoy's Papanicolaou-stained smears with destaining, destaining was omitted from the remaining 28 cases.

Histologic sections          
 IHC+324277220112138
 IHC−011011070407
ER detection rate in smears (%)68.160.06.95.946.4
Sensitivity (%)88.979.49.17.761.9
Specificity (%)100100100100100
Concordance (%)91.584.431.029.471.4
Kappa0.800.650.050.040.45
P value0.0000.0000.5690.7060.0054
thumbnail image

Figure 1. (A) Estrogen receptor immunostaining of a tissue section. (B) Corresponding cytologic smears processed by the Abbott method. (C) Air-drying preceding the Abbott method. (D) Ten percent formalin staining. (E) Carnoy's Papanicolaou-stained slides (Carnoy's Pap). (F) Destaining of Carnoy's Pap smears. No antigen retrieval was applied to all the smears.

Download figure to PowerPoint

Use of antigen retrieval markedly improved immunoreactivity in the formalin-fixed and Carnoy's-Pap smears (Table 3). Specifically, 18 of 20 previously negative formalin-fixed smears and 6 of 8 previously negative Carnoy's-Pap smears were converted to positive after antigen retrieval (Figs. 2, 3). Consequently, the sensitivities of the 2 methods improved substantially, from 9.1% to 90.9% for formalin and from 61.9% to 90.5% for Carnoy's-Pap. The final concordance rate was 93.1% and κ = 0.83 (P < 0.0001) for formalin, and 92.8% and κ = 0.83 (P < 0.0001) for Carnoy's-Pap, both indicating near-perfect agreement. The staining intensity was also stronger after antigen retrieval, particularly for the formalin-fixed smears. The two smears that were not converted in the formalin-fixed preparation were the same two smears that were not converted in Carnoy's-Pap preparation. The ER scores for the concurrent tissue sections by IHC were 90% and 50%, respectively. All of the previously ER-negative smears that had ER-negative tissue counterparts remained ER negative after repeat staining. Therefore, no false-positive results were introduced by antigen retrieval. Some degree of cytoplasmic staining was observed in a few cases, but was generally faint and did not interfere with interpretation. Background staining was unremarkable.

Table 3. Estrogen Receptor Immunoreactivity in Cytologic Smears after Antigen Retrieval Versus that in Corresponding Histologic Sections
CharacteristicsCytologic smears (ICC)
10% formalin after antigen retrieval (n = 29)Carnoy's Pap after antigen retrieval (n = 28)
++
  1. ICC: immunocytochemistry; Carnoy's-Pap: Papanicolaou-stained smears without destaining; IHC: immunohistochemistry; ER: estrogen receptor.

Histologic sections    
 IHC+202192
 IHC−0707
ER detection rate in smears (%)69.067.9
Sensitivity (%)90.990.5
Specificity (%)100100
Concordance (%)93.192.9
Kappa0.830.83
P value0.0000.000
thumbnail image

Figure 2. (A) Estrogen receptor immunocytochemical staining of formalin-fixed smears without antigen retrieval. (B) Repeat staining after antigen retrieval on the same slide.

Download figure to PowerPoint

thumbnail image

Figure 3. (A) Estrogen receptor immunocytochemical staining of smears processed with the Carnoy-Papapanicolaou procedure without antigen retrieval. (B) Repeat staining after antigen retrieval on the same slide.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Determination of ER status in cytologic material is frequently required to direct therapy or identify the primary site of metastatic disease. Studies evaluating ER status in cytologic material using the Abbott technique have demonstrated good correlation between IHC and ICC results, with a concordance rate ranging from 81% to 94%. Our results with the Abbott method showed a concordance rate of 91.5%, which is in keeping with those reported in the literature.8, 19, 21–23

Although the Abbott method yields good staining results in ICC analysis, it requires a sequence of 3-reagent fixation, several rinse steps, and storage of smears in storage medium at −20 °C to −80 °C before immunostaining—a fairly cumbersome and time-consuming procedure to be applied, especially in clinicians' office or clinics. It would also be difficult to transport the various reagents to the patient's bedside for proper fixation of the FNA smears when ER study is indicated. Simpler alternatives have been sought. As a result, a variety of fixatives have been used by different investigators, but with no validation. In addition to the Abbott method, the reliability of which has been documented by several studies, we included 10% formalin and Carnoy's-Pap because formalin is an ideal ER fixative for tissue sections and is readily available and Carnoy's-Pap is used routinely for preparing cytologic specimens.

Initially, immunostaining was performed without antigen retrieval because antigen retrieval is used mainly to improve the immunoreactivity of ER in formalin-fixed tissue sections24, 25 and the putative usefulness of antigen retrieval in cytologic specimens remains controversial. Unsatisfactory experience with antigen retrieval has been reported in ER immunostaining of smears processed with various fixatives.11, 26 In contrast, other reports showed that microwave-stimulated epitope retrieval can be applied successfully on smears fixed in alcohol, formalin, or processed by ThinPrep.15, 16, 27, 28 The low ER immunoreactivity in our initial smears, however, prompted us to try heat-induced epitope retrieval on all negatively staining smears that had been fixed in 10% formalin, followed by repeat ER staining. A negative-to-positive conversion was observed in 18 of 20 smears with an increased nuclear staining intensity. The final sensitivity and concordance values for formalin-fixed smears were slightly better than those of the Abbott method. No false-positive results were found in association with antigen retrieval. Therefore, formalin fixation combined with antigen retrieval is a potential alternative to the more laborious Abbott method because of its simplicity, accuracy, and superior staining quality.

A single fixative that could be used to study both morphology and ER status would be ideal. Carnoy's solution, a fixative used in Pap staining, has been tested for this purpose because archival Pap-stained smears are often the only material available in routine practice for determining ER, especially when the need for ER study is not anticipated during the initial specimen collection. Both Pap-stained smears and destained Pap smears have been used for retrospective immunostaining. The immunostaining on Carnoy's-dPap showed low detectability in the initial smears, presumably due to a loss of antigenicity during the harsh destaining procedure. In the subsequent cases, Pap-stained smears without destaining (Carnoy's-Pap) were used and better staining results were obtained although they were less optimal compared with the Abbott method.29 Applying antigen retrieval on Carnoy's-Pap smears, however, substantially improved the ER detection rate. Koyatsu et al.30 used cell transfer and a cutting technique with archival Pap-stained smears and obtained 96.9% concordance rate with IHC analysis of ER status. In another study, however, Schmitt et al.27 achieved similar agreement only by using microwave treatment. In the current study, the use of antigen retrieval converted six of eight previously ER-negative Pap-stained smears to ER positivity, raising the final sensitivity and concordance value to the level of formalin-fixed smears with antigen retrieval. Moreover, no antigen retrieval-associated false-positivity was observed. These results suggest that Pap-stained smears, without destaining but with antigen retrieval, can be reliably used for retrospective analyses of ER status.

We also evaluated the effect of air-drying versus immediate fixation in the Abbott method. Air-drying has been advocated by some for better retention of cells in the cytologic specimen,12, 15, 31 but others found air-drying to be a major cause of loss of ER antigenicity in addition to poor cellular morphology.11, 16, 26, 32 In the current study, the ER detection rate was better when smears were fixed immediately rather than air-dried, thereby attesting to the necessity of prompt fixation to reduce the risk of false-negative results.

The disconcordant results between IHC and ICC analyses observed in two cases even after antigen retrieval may have resulted from intratumoral heterogeneity and sampling variations.16, 19, 33 However, in view of the ER scores of the 2 cases by IHC (90% and 50%, respectively), tissue heterogeneity appears to be an unlikely factor contributing to the negative results in smears. The exact cause of the false-negative staining is difficult to determine and may be due to unknown preanalytic factors rather than tissue heterogeneity.

Identifying suitable fixative and processing conditions for the preservation of the antigen of interest is of fundamental importance in determining ER immunoreactivity. Fixation of smears with formalin or using Carnoy's-Pap smears in conjunction with antigen retrieval provides an overall accuracy similar to or slightly better than that of the Abbott method. The simplicity and the wide availability of formalin make the former technique a reliable alternative to the Abbott method. Whereas Pap-stained smears facilitate retrospective analysis of ER, and allow previsualizion of representative tumor cells in the smear. Laboratories must adhere strictly to previously well established processing conditions or validate their own technique before relying on the test results.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES