Fine-needle aspiration cytology (FNAC) and core needle biopsy

Transcription

1 146 CANCER CYTOPATHOLOGY A Comparison of Aspiration Cytology and Core Needle Biopsy in the Evaluation of Breast Lesions Pieter J. Westenend, M.D., Ph.D. 1 Ali R. Sever, M.D. 2 Hannie J. C. Beekman-de Volder 1 Sik J. Liem, M.D. 2 1 Laboratory for Pathology, Albert Schweitzer Hospital, Dordrecht, The Netherlands. 2 Department of Radiology, Albert Schweitzer Hospital, Dordrecht, The Netherlands. BACKGROUND. Several arguments are used for choosing either fine-needle aspiration cytology (FNAC) or core needle biopsy (CNB) in the evaluation of breast lesions. Comparison of published data on both methods is complicated by differences in study design, calculations, and operator experience. The objective of this study was to make a direct comparison of both methods. METHODS. In 286 breast lesions (cysts and microcalcifications without a soft tissue mass excluded), both ultrasound-guided FNAC and CNB were performed in the same session by the same operator. Histologic follow-up was collected, and for those lesions that were not excised the results of the most recent mammography was used. A combination of the findings of both FNAC and CNB were evaluated. RESULTS. Core needle biopsy and FNAC do equally well for sensitivity (88% vs. 92%), positive predictive value for malignancy (99% vs. 100%), and inadequate rate (7% vs. 7%). However, statistical differences are found for the specificity (CNB, 90%; FNAC, 82%). In addition, differences are found in the positive predictive value of both suspicious (CNB, 100%; FNAC, 78%) and atypia (CNB, 80%; FNAC, 18%) and for the suspicious rate (CNB, 5%; FNAC, 13%) reflecting difficulties in interpreting some FNACs. Combining the findings of both FNAC and CNB results in an increase in absolute sensitivity, a decrease in the positive predictive value of atypia compared with FNAC and CNB per se, and a decrease in the inadequate rate for cancers. CONCLUSIONS. For the lesions selected in this study, FNAC and CNB are comparable for most parameters, but CNB has a higher specificity and lower suspicious rate. Combining results of FNAC and CNB leads to an increase in absolute sensitivity without affecting specificity and a decrease in the inadequate rate for cancers. Cancer (Cancer Cytopathol) 2001;93: American Cancer Society. KEYWORDS: breast, pathology, diagnosis, aspiration cytology, core needle biopsy, ultrasound guidance, sensitivity, specificity. Presented in abstract form at the 5th Annual Multidisciplinary Symposium on Breast Disease, Rome, Italy, February 13 16, Address for reprints: Pieter J. Westenend, M.D., Ph.D., Laboratory for Pathology Dordrecht, Jkvr van den Santheuvelweg 2a, 3317 NL Dordrecht, The Netherlands; Fax: ; pwestenend@paldordt.com Received June 28, 2000; revision received November 30, 2000; accepted December 6, Fine-needle aspiration cytology (FNAC) and core needle biopsy (CNB) are both used in the evaluation of breast lesions and play an important role in their management. Fine-needle aspiration cytology is a fast and simple procedure but highly operator-dependent requiring special training on the part of the pathologist. Consequently, a wide range of different results have been reported. 1 3 Core needle biopsy takes more time, often requires anesthesia, and, as with cytology, pathologists, radiologists, and surgeons need to be familiar with core biopsy technique to avoid certain pitfalls. 4,5 Core biopsy results also demonstrate considerable variation depending on needle size, number of cores, and operator experience. 6,7 Because of this, it is difficult to decide from published studies which method is preferred. Results may be influenced by patient selection, incidence of breast 2001 American Cancer Society

2 FNAC versus CNB in Breast Lesions/Westenend et al. 147 carcinoma in the population, availability of follow-up data, duration of follow-up, and differences in biopsy technique. Many published series are from academic centers and not from a general hospital setting. Calculations for sensitivity and specificity are made in different ways sometimes reflecting the clinical effectiveness of the procedure, sometimes only evaluating the laboratory performance. 2,8 Comparisons between series can only be made after recalculation. 2 This study compares the results of FNAC and CNB taken from the same breast lesion by the same operator in one session under ultrasound guidance. MATERIALS AND METHODS From the beginning of 1994 until the end of 1997, both ultrasound-guided FNAC and CNB were performed on 286 breast lesions in the same session by one person (S.J.L.). Of these lesions, 232 were palpable. The mean patient age was 46 years (standard deviation, 16 years), and both screen-detected and symptomatic lesions were included. Cysts and microcalcifications without a soft tissue mass were excluded from this study. The mean size of the lesions as measured by ultrasound was 16 mm (standard deviation, 10 mm). In all procedures, ultrasound was used to verify the position of the needle tip. The first procedure was always FNAC performed with a 21-gauge hypodermic needle on a syringe with application of a vacuum. Usually one pass was made. Subsequently, CNB was performed with a 18-gauge needle with a length of the sample notch of 15 mm on a spring-loaded biopsy gun by using the same angle and depth as for the FNAC. For each CNB procedure, the number of biopsies taken was recorded. Eight different pathologists read FNAC and CNB slides, and these results were used in this evaluation without revision because these reports were used in patient management. Results of FNAC were reported as malignant, suspicious for malignancy, atypia of uncertain significance, no malignancy, or insufficient material. For this study, results of CNB were reported in the same way. Follow-up data were collected up to the end of Histologic follow-up was available in 115 cases including 79 cancers. For the remaining 171 patients, radiologic follow-up was available in 92 cases, and in all these cases the breast lesion had remained unchanged. In the remaining 79 cases, no histologic or radiologic follow-up was available. Calculations of sensitivity, specificity, positive predictive value of malignant, suspicious or atypia, false-negative rate, false-positive rate, inadequate rate, inadequate rate for cancers, and suspicious rate were performed for both biopsy methods. 8 These parameters were calculated for the combined results of FNAC and CNB. A combined result TABLE 1 Diagnostic Accuracy of Fine-Needle Aspiration Cytology Parameter This study (95% CI) Acceptable values (%) 8 Absolute sensitivity 72 (67 78) 60 Complete sensitivity 92 (89 95) 80 Specificity (biopsy cases only) 58 (53 64) Specificity (full) 82 (78 87) 60 Ppv malignant 100 ( ) 95 Ppv suspicious 78 (73 83) Ppv atypia 18 (13 22) False-negative rate 6 (3 8) 5 False-positive rate 0 (0 0) 1 Inadequate rate 7 (4 10) 25 Inadequate rate from cancers 2 (1 4) Suspicious rate 13 (9 17) 20 CI: confidence interval; Ppv: positive predictive value. was obtained by taking the highest score, e.g., FNAC atypia and CNB malignant as a combined result malignant, FNAC suspicious, and CNB benign as a combined result suspicious. The method of calculating was chosen to reflect the results of the whole procedure, not only the laboratory performance, as is customary in the National Health Service Breast Screening Programme (NHSBSP) in Great Britain. 8 This means that inadequate samples are included in the calculation of sensitivity and specificity. Sensitivity is calculated in two ways, one including only samples with a diagnosis malignant called absolute sensitivity, and one including samples with a diagnosis malignant, suspicious, or atypia called complete sensitivity. Specificity also is calculated in two ways, one including only samples with a diagnosis benign for which histologic follow-up was available called specificity (biopsy cases only). The other way, called specificity (full), includes all samples with a diagnosis benign with or without histologic follow-up and those samples with diagnosis atypia without histologic follow-up assuming a benign lesion. RESULTS The results of the FNAC are summarized in Table 1. Included in this table are the acceptable values as used in the national breast carcinoma screening program in the U.K. 8 As can be seen in this table, FNAC in this study meets these standards and is an effective procedure. The results of CNB are summarized in Table 2. Core needle biopsy and FNAC perform equally well on most reported values such as sensitivity, positive predictive value for malignancy, and inadequate rate. However, statistical differences are found for the specificity (both biopsy cases only and full), for the positive

3 148 CANCER (CANCER CYTOPATHOLOGY) April 25, 2001 / Volume 93 / Number 2 TABLE 2 Diagnostic Accuracy of CNB Parameter CNB (95% CI) Compared with FNAC (P value) Absolute sensitivity 75 (70 80) NS Complete sensitivity 88 (84 91) NS Specificity (biopsy cases only) 78 (73 83) 0.05 Specificity (full) 90 (87 94) 0.05 Ppv malignant 99 (97 100) NS Ppv suspicious 100 ( ) 0.05 Ppv atypia 80 (75 85) 0.05 False-negative rate 9 (6 12) NS False-positive rate 1 (0 2) NS Inadequate rate 7 (4 10) NS Inadequate rate from cancers 3 (1 6) NS Suspicious rate 5 (2 7) 0.05 CNB: core needle biopsy; FNAC: fine-needle aspiration cytology; CI: confidence interval; NS: not significant; Ppv: positive predictive value. predictive value of both suspicious and atypia, and for the suspicious rate. In all of these cases, CNB performs better. The diagnostic accuracy of CNB was further studied in relation to the number of cores taken (Table 3). As shown in this table, accuracy of CNB biopsies increases with the number of cores taken. However, the inadequate rate is highest for lesions for which greater than or equal to five cores were taken, probably reflecting the limited accessibility of some lesions. The results of the combined score for FNAC and CNB are shown in Table 4. Importantly, combination of the score of FNAC and CNB results in an increase in absolute sensitivity and a decrease in the positive predictive value of atypia compared with the results of FNAC and CNB evaluated separately. In addition, there is a decrease in the inadequate rate for cancers. DISCUSSION The results of our study show that FNAC is a reliable procedure compared with the targets set by the NHS- BSP (included in Table 1) used nationwide in the evaluation of breast screening units in Great Britain 8 and with several reviews. 1 3 Although to our knowledge similar targets have not been set for CNB, our results show that CNB is also a reliable procedure compared with the NHSBSP targets for FNAC and with several reviews and a large multi-institutional study of CNB Our results with CNB are similar to those of Britton et al. 13 who used the same calculations. The main difference is the lower absolute sensitivity that we calculated (75% vs. 89,3%), whereas complete sensitivity, full specificity, positive predictive value, falsenegative rate, and inadequate rate are the same. 13 This difference can be explained by the smaller biopsy needle that we used (18 vs. 16 gauge), a factor that has been shown to affect the results of CNB. 7 The effect of the number of cores on the diagnostic accuracy was similar as reported by others. 6 The design of our study makes it possible to compare FNAC and CNB directly because it excludes several confounding variables. Most importantly, FNAC and CNB were performed on the same lesion, so there are no differences in patient and lesion selection, and the operator dependence has been standardized. In addition, the follow-up and the way the results are calculated is the same. Our data have shown that FNAC and CNB do equally well in both absolute and complete sensitivity, positive predictive value of malignant, false-positive, and false-negative rate, inadequate rate, and inadequate rate for cancers. The higher specificity for CNB compared with FNAC calculated for cases with histologic follow-up only and for all cases reflects the limitations of FNAC as a procedure for making a specific benign diagnosis. Although the categories suspicious and atypia were rarely used in CNB, the higher predictive value for malignancy for both compared with those of FNAC reflects the easier interpretation of CNB. This also shown by the lower suspicious rate. Apparently suspicious for malignancy was not needed as often in CNB as in FNAC. The use of both FNAC and CNB on the same lesion in the same visit has been reported previously In most of these series, all the lesions were palpable, and needle positioning was by palpation without the use of ultrasound guidance In addition, there was a high malignancy rate of 75% or greater. 14,16,17 In contrast, our series contains both palpable and nonpalpable lesions, needle positioning was always ultrasound-guided, and the malignancy rate was 28%. The reported series differ also from ours in several other aspects. Elston et al. 14 in a study in 1978 selected lesions of large size and used Tru-cut biopsies and FNAC. They concluded that Tru-cut biopsies are reliable in the diagnosis of carcinoma but that the false-positive rate of FNAC was unacceptably high. This may reflect a lack of experience in the interpretation of FNAC because the same diagnostic unit later reported in 1996 that 80% of the patients had a diagnosis of carcinoma confirmed with FNAC and 88% with CNB. 17 The problem of an unacceptably high false-positive rate for FNAC is no longer mentioned in this later study. In this series, patients were selected for age older than 70 years or locally advanced disease, and CNB was performed with a biopsy gun. The combined results of FNAC and CNB confirmed a diagnosis of carcinoma in 97% of the patients. Cheung et al. 15 obtained almost identical values for sensitivity and specificity of FNAC and CNB with

4 FNAC versus CNB in Breast Lesions/Westenend et al. 149 TABLE 3 Accuracy Related to the Number of Cores Taken Core number >5 No. of patients a Accuracy (%) b Inadequate (%) a For 262 patients, the number of cores was recorded; for the remaining these data were missing. b Calculated for the adequate biopsies only. TABLE 4 Diagnostic Accuracy of Combined FNAC and CNB Parameter FNAC and CNB (95% CI) Compared with FNAC (P value) Compared CNB (P value) Absolute sensitivity 88 (84 91) Complete sensitivity 98 (96 99) NS 0.05 Specificity (biopsy cases only) 64 (58 69) NS 0.05 Specificity (full) 86 (82 90) NS NS Ppv malignant 99 (97 100) NS NS Ppv suspicious 73 (68 78) NS 0.05 Ppv atypia 7 (4 10) False-negative rate 2 (1 4) NS 0.05 False-positive rate 1 (0 2) NS NS Inadequate rate 3 (1 5) NS NS Inadequate rate from cancers 0 (0 0) Suspicious rate 9 (6 13) NS NS FNAC: fine-needle aspiration cytology; CNB: core needle biopsy; CI: confidence interval; NS: not significant; Ppv: positive predictive value. the use of Tru-cut biopsies and concluded that one technique is not superior to the other. The malignancy rate of this series (16%) is somewhat lower than in our series, but the inadequate rate for both FNAC (27.4%) and CNB (33.3%) is much higher than our series, 7% for both CNB and FNAC. This could very well be the result of the difference in needle positioning by palpation as opposed to ultrasound. Consequently Cheung et al. 15 report a lower sensitivity than we do for FNAC (78.4%) and CNB (82.9%) and a lower specificity for FNAC (71.6%) and CNB (61.7%) calculated with the nondiagnostic samples included. Ballo and Sneige 16 selected only patients in whom a malignancy ultimately had been confirmed by an excision biopsy or mastectomy. They concluded that FNAC is more sensitive in detecting cancer than CNB (sensitivity, 97.5% vs. 90%). In another report, both FNAC and CNB samples were obtained on nonpaplable lesions with a stereotactic device. 18 Besides mass lesions, microcalcifications also were included in this series. No differences were found for sensitivity or specificity of FNAC and CNB. As in our study, some of these series report a favorable effect of combining the results of FNAC and CNB. 15,17 This supports a strategy of using CNB for patients with inadequate FNAC. Other studies have shown the same effect, and CNB is claimed to be more cost-effective. 22 This study has demonstrated that FNAC and CNB show only minor differences in specificity and suspicious rate. Therefore, the choice between FNAC and CNB in the evaluation of mass lesions whether palpable or not should be determined by the availability of experienced cytopathologists. The use of ultrasound guidance in needle positioning will reduce the number of inadequate biopsies. REFERENCES 1. Giard RWM, Hermans J. The value of aspiration cytologic examination of the breast. A statistical review of the medical literature. Cancer 1992;69: Wells CA. Quality assurance in breast cancer screening cytology: a review of the literature and a report on the U.K. national cytology scheme. Eur J Cancer 1995;31A: Collaço LM, de Lima RS, Werner B, Torres LFB. Value of fine needle aspiration in the diagnosis of breast lesions. Acta Cytol 1999;43: Liberman L, Cohen MA, Dershaw DD, Abramson AF, Hann LE, Rosen PP. Atypical ductal hyperplasia diagnosed at stereotactic core biopsy of breast lesions: an indication for surgical biopsy. AJR Am J Roentgenol 1995;164:

5 150 CANCER (CANCER CYTOPATHOLOGY) April 25, 2001 / Volume 93 / Number 2 5. Liberman L, Dershaw DD, Rosen PP, Giess CS, Cohen MA, Abramson AF, et al. Stereotactic core biopsy of breast carcinoma: accuracy at predicting invasion. Radiology 1995; 194: Brenner RJ, Fajardo L, Fisher PR, Dershaw DD, Evans WP, Bassett L, et al. Percutaneos core biopsy of the breast: effect of operator experience and number of samples on diagnostic accuracy. AJR Am J Roentgenol1996;166: Helbich TH, Rudas M, Haitel A, Kohlberger PD, Thurnher M, Wunderbaldinger P, et al. Evaluation of needle size for breast biopsy: comparison of 14-, 16-, and 18-gauge biopsy needles. AJR Am J Roentgenol 1998;171: Wells CA, Ellis IO, Zakhour HD, Wilson AR. Guidelines for cytology procedures and reporting on fine needle aspirates of the breast. Cytopathology 1994;5: Parker SH, Burbank F, Jackman RJ, Aucreman CJ, Gardenosa G, Cink TM, et al. Percutaneous large-core breast biopsy: a multi-institutional study. Radiology 1994;193: Nguyen M, McCombs MM, Ghandehari S, Kim A, Wang H, Barsky SH, et al. An update on core needle biopsy for radiologically detected breast lesions. Cancer 1996;78: Parker SH, Burbank F. A practical approach to minimally invasive breast biopsy. Radiology 1996;200: Bassett L, Winchester DP, Caplan RB, Dershaw DD, Dowlatshahi K, Evans WP III, et al. Stereotactic core-needle biopsy of the breast: a report of the joint task force of the American College of Radiology, American College of Surgeons, and College of American Pathologists. CA Cancer J Clin 1997;47: Britton PD, Flower CDR, Freeman AH, Sinnatamby R, Warren R, Goddard MJ, et al. Changing to core biopsy in an NHS screening unit. Clin Radiol 1997;52: Elston CW, Cotton RE, Davies CJ, Blamey RW. A comparison of the use of the Tru-Cut needle and fine needle aspiration cytology in the pre-operative diagnosis of carcinoma of the breast. Histopathology 1978;2: Cheung PS, Yan KW, Alagaratnam TT. The complementary role of fine needle aspiration cytology and Tru-cut needle biopsy in the management of breast masses. AustNZJSurg 1987;57: Ballo MS, Sneige N. Can core needle biopsy replace fineneedle aspiration cytology in the diagnosis of palpable breast carcinoma. A comparative study of 124 women. Cancer 1996;78: Poole GH, Willsher PC, Pinder PC, Robertson JF, Elston CW, Blamey RW. Diagnosis of breast cancer with core-biopsy and fine needle aspiration cytology. Aust N Z J Surg 1996;66: Lifrange E, Kridelka F, Colin C. Stereotaxic needle-core biopsy and fine-needle aspiration biopsy in the diagnosis of nonpalpable breast lesions: controversies and future prospects. Eur J Radiol 1997;24: Carty NJ, Ravichandran D, Carter C, Mudan S, Royle GT, Taylor I. Randomized comparison of fine-needle aspiration cytology and biopty-cut needle biopsy after unsatisfactory initial cytology of discrete breast lesions. Br J Surg 1994;81: Florentine BD, Cobb CJ, Frankel K, Greaves T, Martin SE. Core needle biopsy. A useful adjunct to fine-needle aspiration in select patients with palpable breast lesions. Cancer (Cancer Cytopathol) 1997;81: Litherland JC, Evans AJ, Wilson ARM, Kollias J, Pinder SE, Elston CW, et al. The impact of core-biopsy on pre-operative diagnosis rate of screen detected breast cancers. Clin Radiol 1996;51: Logan-Young W, Dawson AE, Wilbur DC, Avila EE, Tomkiewiecz ZM, Sheils LA, et al. The cost-effectiveness of fineneedle aspiration cytology and 14-gauge core needle biopsy compared with open surgical biopsy in the diagnosis of breast carcinoma. Cancer 1998;82: