Usefulness of Cytokeratin Fragment Antigen 21-1 Estimation in Fine-Needle Aspiration Washout Specimens in Diagnosis of Axillary Breast Cancer Metastasis

Abstract

Introduction:

Fine-needle aspiration biopsy (FNAB) is a valuable technique for diagnosing axillary breast cancer metastasis. However, its sensitivity can sometimes be affected by sampling errors and technical artifacts.

Aim:

We examined whether cytokeratin fragment antigen 21-1 (CYFRA) estimation by manual enzyme-linked immunosorbent assay (ELISA) on FNAB needle washouts could add diagnostic information to FNAB.

Materials and Methods:

This cross-sectional study involved FNAB from 138 lymph nodes in 108 suspected breast cancer patients. Papanicolaou and May–Grünwald–Giemsa smears were prepared for cytological examination. CYFRA was estimated by manual ELISA from fine-needle washouts after smear preparation. A cut-off of 1.93 ng/mL was considered to be a positive test for CYFRA.

Results:

The CYFRA estimates were significantly associated with the FNAB results (P < 0.001 by Kruskal–Wallis test). On FNAB, 84, 5, 36, and 13 cases were classified as malignant, suspicious, benign, and inadequate, respectively. CYFRA was reported as positive in 92 cases (75 in malignant FNAB, 3 in suspicious FNAB, 11 in benign FNAB, and 3 in inadequate FNAB). Six of 11 CYFRA positives in benign FNAB were confirmed by re-examination of cytology slides; the remaining 5 CYFRA-positive cases did not have further biopsy. Only two inadequate FNAB had histopathology follow-up; both revealed metastasis, of which one was CYFRA positive. Sensitivity and specificity with 95% confidence intervals of CYFRA for detecting axillary metastasis were 0.876 [0.794, 0.934] and 0.833 [0.653, 0.954], respectively.

Conclusion:

CYFRA estimation by manual ELISA is a simple and accurate test which supplements and enhances the diagnostic usefulness of axillary FNAB in suspected breast cancer metastasis.

INTRODUCTION

Fine-needle aspiration biopsy (FNAB) is a widely used technique for the diagnosis of axillary breast cancer metastasis.[1,2,3] However, it often shows low sensitivity due to sampling errors, dilution with blood, and technical artefacts.[4] As a result, the surgeon often relies on axillary sentinel lymph node biopsy and frozen sections. Any simple diagnostic test to increase the sensitivity of preoperative axillary lymph node staging will help manage the disease.

Cytokeratin fragment antigen 21-1 (CYFRA) is the soluble fraction of cytokeratin 19, which is present in nearly 90% of all breast cancers. Therefore, it is likely to be present in lymph node metastasis. CYFRA estimation by chemiluminescence-based immune assay on fine-needle washouts has been investigated as a simple diagnostic method for diagnosing lymph node metastasis.[5,6,7] It has demonstrated acceptable sensitivity and specificity in this setting for metastatic carcinoma of multiple sites, including the breast.[5,7,8,9,10] Validation of the cut-off defined in those studies remains to be carried out. Furthermore, estimation of CYFRA has been validated only using automated chemiluminescence-based assays. Therefore, we carried out a study to investigate the correlation of axillary fine-needle aspiration with CYFRA estimation by manual enzyme-linked immunosorbent assay (ELISA). Besides estimating the association between axillary FNAB and CYFRA, we also attempted to find out the additional cases of axillary metastasis diagnosed by CYFRA over FNAB, and whether the two techniques added diagnostic information to each other.

MATERIALS AND METHODS

This study was approved by the Institutional Ethics Committee. In this study138 axillary or supraclavicular lymph nodes from 108 suspected breast cancer patients were subjected to ultrasound-guided fine-needle aspiration after informed consent was obtained for the procedure. The age of the patients ranged from 19 to 79 years, with a median age of 47 years. In 73patients, breast ultrasound was also done, with 53 patients classified as Breast Imaging-Reporting and Data System (BI-RADS) 5, 16 as BIRADS 4c cases, and 4 as BIRADS 6. All the lymph nodes aspirated were suspicious on ultrasound.

Air dried and alcohol fixed smears were prepared for May–Grünwald–Giemsa (MGG) and Papanicolaou staining in all cases. After preparation of the smears, the needle hub was rinsed into a sterile 2 mL plastic tube using 0.5 mL of normal saline. The tubes were stored and then subsequently centrifuged at 1000 rpm for 20 min. The supernatant was used for estimation of CYFRA 21-1 using pre-coated ELISA kits (Elabscience, Houston, TX, USA, Catalog no. E-EL-H2077) as per the manufacturer’s instructions. Readouts were taken in the ERBA LisaScan EM ELISA reader (ERBA Mannheim, Mannheim, Germany). Standards were prepared by serial dilution of a 20 ng/mL reference standard, halving the concentration for each dilution, up to a concentration of 0.31 ng/mL (a total of seven serial dilutions). All standards and samples were run in duplicates.

A four-parameter logistic curve was fitted to the standard thus prepared using the online calculator available from mycurvefit.com. The average of the optical densities was used to fit the curves. The concentration of CYFRA for each unknown sample was calculated by back-calculation from the average optical densities of each unknown sample and the four-parameter logistic standard curve.

CYFRA 21-1 estimation by ELISA was the index test. A positive reference test was shown in one case where FNAB showed malignant cells during reporting, or after slide review, after CYFRA results, or after histopathology results. The area under the receiver operating characteristic (ROC) curve was estimated. A cut-off of 1.93 ng/mL, which was proposed in previous studies,[8,10] was used as the cut-off for differentiating benign from metastatic samples.

The relationship between the FNAB diagnosis and CYFRA estimates was examined using 2 × 2 cross tabulation, Fisher’s exact test, and the odds ratio. Cases where the CYFRA was negative, but FNAB was positive, were treated as a false negative CYFRA result. For cases where CYFRA was positive, but FNAB was negative, the FNAB smears were re-evaluated for any malignant cells on the MGG and Pap-stained smears. In these cases, immunocytochemistry for Pan-cytokeratin on Pap-stained slides was also attempted. The additional pick-up rate of CYFRA was estimated. Additionally, histopathology reports were examined, if available. The final diagnosis was considered positive if FNAB was positive for malignant cells either in the initial report or on re-examination, or if the histopathology report of any subsequent axillary dissection demonstrated metastasis. The final report was considered negative if axillary dissection on chemotherapy-naïve patients was negative, or if both CYFRA and FNAB were negative.

For the subsequent diagnostic accuracy estimates, all samples that were positive for CYFRA but negative on FNAB, even after re-examination and without histological follow-up, were considered negative. This prevented over-optimistic diagnostic accuracy results for CYFRA. For the diagnostic accuracy estimates, Inadequate FNAB cases not having histological follow-up were excluded, and lymph nodes having suspicious metastatic cells were considered positive.

All statistical analyses were conducted using R Statistical software, version 4.2, and the pROC package.[11]

RESULTS

The distribution of the CYFRA results is given in Table 1. There was a significant statistical difference in CYFRA levels between benign, suspicious, and malignant lymph node fine-needle aspirates (P < 0.001 by the Kruskal–Wallis test). At a cut-off of 1.93 ng/mL, CYFRA positivity was significantly associated with the cytological categories (P < 0.001 by Fisher’s exact test). The boxplot for visualizing the distribution of CYFRA across the various cytopathological categories is given in Figure 1.

Table 1.

The distribution of the CYFRA results in the different cytopathological diagnostic categories

Cytopathological categories	Mean	SD	Minimum	Median	Maximum	n
Inadequate	3.53	7.33	0.13	0.3	20	13
No metastasis	3.94	6.77	0.27	0.61	20	36
Suspicious of metastasis	9.57	9.92	0.3	7.16	20	5
Metastatic	13.79	7.86	0.18	20	20	84

All figures are in ng/mL.

CYFRA = cytokeratin fragment antigen, SD = standard deviation

Figure 1.

Boxplot with superimposed strip chart showing the distribution of the cytokeratin fragment antigen values across the different diagnostic categories

The distribution of the CYFRA results among the different cytopathological diagnostic categories is given in Table 2. There were 11 cases in which CYFRA was positive, but FNAB was negative. Six out of 11 cases were found to have malignant cells on microscopic review. Most such positive cases on review had only small clusters of malignant cells, which were missed [shown in Figure 2], or single tumor cells needing immunocytochemistry on Papanicolaou-stained smears. In addition, three cases with suspicious cytology were found to be positive on CYFRA, adding confidence to the results. Therefore, CYFRA provided additional information over and above FNAB in 9 cases out of 125 adequate reports. Additionally, one inadequate FNAB but showing CYFRA positivity was positive on histopathology subsequently.

Table 2.

The distribution of positive and negative CYFRA results

Cytopathological categories	CYFRA negative	CYFRA positive	Comments
Inadequate	10	3	One case of each was found to be a true positive and false negative on histopathology; the rest did not have histopathology.
No metastasis	25	11	Six out of 11 CYFRA-positive but reported as benign FNAB were found to have malignant cells on cytology re-examination. The other five patients have no cytologically detectable cells, but also do not have histological follow-up.
Suspicious of metastasis	2	3	Three out of five suspicious lymph nodes on cytology were found to be positive on re-examination.
Metastatic	9	75	Nine false negative cases

Diagnostic cut-off of 1.93 ng/mL

CYFRA = cytokeratin fragment antigen, FNAB = Fine-needle aspiration biopsy

Figure 2.

Photomicrograph of a case where cytokeratin fragment antigen was positive but fine-needle aspiration biopsy was initially reported negative and found positive on re-evaluation showing a small cluster of metastatic cells (May–Grünwald–Giemsa ×400)

The area under the ROC curve for the provisional diagnostic accuracy study was 0.888 (95% confidence interval = 0.814–0.962). The optimal threshold for classification from the ROC curve was reached at CYFRA values of >1.67 and ≥1.95/mL. The results of the provisional diagnostic accuracy studies are given in Table 3.

Table 3.

Diagnostic accuracy of CYFRA assuming that all cases of combined FNAB and CYFRA negative cases are benign

	Estimate [95% CI]
Diagnostic cut-off	1.93 ng/mL	1.6 ng/mL	1.98 ng/mL
Sensitivity	0.876 [0.794, 0.934]	0.876 [0.794, 0.934]	0.866 [0.782, 0.927]
Specificity	0.833 [0.653, 0.954]	0.800 [0.614, 0.923]	0.833 [0.653, 0.954]
Positive predictive value	0.944 [0.875, 0.982\	0.934 [0.862, 0.975]	0.944 [0.874, 0.982]
Negative predictive value	0.676 [0.502, 0.820]	0.667 [0.490, 0.814]	0.658 [0.486, 0.804]
Diagnostic accuracy	0.866 [0.794, 0.920]	0.858 [0.785, 0.914]	0.858 [0.785, 0.914]

Suspicious FNAB is considered positive (malignant) in this analysis.

CYFRA = cytokeratin fragment antigen, FNAB = fine-needle aspiration biopsy, CI = confidence interval

DISCUSSION

FNAB has proven utility in diagnosing lymph node metastasis in breast cancer patients. However, the sensitivity of this technique is lowered by sampling and interpretational errors. In such a setting, a rapid, simple supplementary test may be useful to avoid unnecessary surgery, frozen sections, and uncertainty in clinical decision making.

In this study, we have tried to show that CYFRA estimation by ELISA on FNAB needle washouts may serve such a purpose. This is a straightforward method requiring simple equipment and may be easily performed manually. In 9 cases out of 125 (7.2%), adequate cases, CYFRA estimation by ELISA could either detect a false negative FNAB report or lend support to an indeterminate (suspicious) report. At least 6 out of 36 cases (16.7%) reported as negative on FNAB could have been overturned using CYFRA results. The diagnostic accuracy of CYFRA was adequate and has a high sensitivity.

This study is also important in that it has validated a previously published cut-off for CYFRA. Previously published studies had used different cut-offs ranging from a cut-off of 1.6 [9] to 1.98 ng/mL.[5] We had a priori thought of using a cut-off of 1.93 ng/mL, but all three cut-offs, 1.6, 1.93, and 1.98 ng/mL, perform well. The cut-off of 1.93 ng/mL also corresponds to the optimum cut-off for the present study by examination of the ROC. It may be noted that the diagnostic accuracy study is biased against CYFRA since five cases of CYFRA-positive cases, which were FNAB negative, were considered negative, even though the negativity could not be confirmed by further studies. In practice, these results safeguard against a false negative FNAB result. Due to this bias in the diagnostic accuracy estimates, the counts of cases in which CYFRA added diagnostic usefulness to FNAB should be treated as the primary outcome measure of this study.

CYFRA represents a fragment of cytokeratin 19. The likely cause of false positive CYFRA levels in washouts are presence of other carcinomas.[5] Elevated serum CYFRA levels have been observed in benign liver disease and renal insufficiency, and have the potential to result in false positives in FNAB washouts if the serum level is too high.[12] Around 10% of breast carcinomas are either negative or only focally positive for CK19.[13] Hence, the expected sensitivity of CYFRA is 90%. These CK19-negative cases, along with a failure to aspirate even a few metastatic cells, are possible causes of false-negative CYFRA results.

The absence of histological follow-up in all cases is a significant limitation of this study. This is likely to bias our diagnostic accuracy estimates for CYFRA. A histological follow-up could not be possible in many cases due to loss of follow-up, as well as neoadjuvant chemotherapy. To counter reporting over-optimistic results for CYFRA estimation on FNAB washouts, we considered all CYFRA-positive but FNAB-negative results a false positive result for CYFRA unless malignant cells were seen in a re-examination of the cytological smear or there was histopathological evidence of metastasis. These cases may well be true positives for CYFRA and false negatives for FNAB, given that they have high CYFRA values. However, we had no means to confirm this since the patients were put on neo-adjuvant chemotherapy or did not follow up. Even treating all these cases as CYFRA false positives, we demonstrated additional utility for CYFRA in a small but significant subset of cases where, guided by the CYFRA results, we detected clusters of malignant cells in FNAB smears that were missed on initial reporting. Some false negatives are expected, since 10% of breast cancers may be CK19 negative. Our reported sensitivity of 87.6% matches the expected 90% sensitivity of CYFRA.

CYFRA estimation in FNAB needle washouts has been attempted earlier in at least four studies,[5,8,9,10] with results similar to ours. However, we have used much simpler equipment (ELISA performed manually) compared with tests using chemiluminescence performed in the above-mentioned studies. Therefore, this study validates the usefulness of CYFRA in more resource-limited settings compared with the earlier studies. The cost of the test presently using duplicate estimation for accuracy is approximately INR 750/- per test, with an additional requirement of a simple ELISA reader and washer, along with common laboratory consumables.

The primary alternative to axillary FNAB is core needle biopsy. Some studies have reported that core needle biopsy is more sensitive and accurate than FNAB.[4,14] However, FNAB has fewer complication making it an attractive alternative to Core-needle biopsy, and a systematic review has found similar accuracy between the two techniques.[15] The addition of supplementary CYFRA is likely to increase the sensitivity of FNAB. In experienced hands, FNAB is sensitive, especially if supplemented by rapid on-site evaluation (ROSE).[16] Pre-operative CYFRA estimation on FNAB washouts is similarly likely to improve the sensitivity of FNAB. FNAB is also dependent on the experience of the aspirator. Diagnostic trials incorporating ROSE with CYFRA estimation are needed to demonstrate the noninferiority of FNAB to core needle biopsy.

Absence of histological follow-up in benign reported cases is a significant limitation of this study, especially in cases that underwent neo-adjuvant chemotherapy. Additionally, this study could have been improved if immunocytochemistry had been attempted in all cases. Immunocytochemistry for Pan-cytokeratin is not routinely done on FNAB smears in our setting. We attempted immunocytochemistry on Pap-stained smears, but with mixed results; some smears floated off. In some cases, there were false positive strands, which limited our interpretation. We could attempt immunocytochemistry in one smear of all cases with a positive CYFRA but negative FNAB. Even then, not staining more slides due to technical problems remains a limitation. We plan to standardize our immunocytochemistry further. Routine cell-blocks with immunohistochemical staining for cytokeratin may have also enhanced our analysis and improved the detection of axillary metastasis.

CONCLUSION

In conclusion, CYFRA is suitable as a supplementary test enhancing the accuracy of FNAB, not its replacement. FNAB detects many cases that are negative on CYFRA testing and is more accurate. However, both tests add to one another. A positive CYFRA increases the probability of detecting a false-negative FNAB result, and a negative CYFRA may confirm a negative FNAB result. A study with complete follow-up of the patients is needed to accurately determine the extent to which CYFRA adds to axillary FNAB.

Author contributions

NC and MJ: Concepts, Design, Literature search, and Manuscript preparation; NC, FH, GRS, AG, and EA: Clinical Studies; MJ and GRS: Experimental studies and Data acquisition; All authors: Definition of intellectual content, Data analysis, and Manuscript review. NC: Statistical analysis; MJ, GRS, AG, EA, and NC: Manuscript editing. NC: Guarantor

Data availability statement

All result generated or analyzed during this study is included in this article. Further enquiries can be directed to the corresponding author.

Ethical policy and Institutional Review Board statement

This study protocol was reviewed and approved by the institutional ethics committee, AIIMS, Rishikesh (AIIMS/IEC/24/589 dated October 11, 2024). Written informed consent was taken for the procedure of USG-guided FNAB, and written informed consent was not required for this study, as the testing was conducted on leftover samples following cytology slide preparation. The use of such samples is in accordance with the joint Guidelines for Ethical Use of Leftover Clinical Samples issued by the Ministry of Health and Family Welfare and the Indian Council of Medical Research, INDIA.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.