Fine Needle Aspiration
Application of the Probabilistic Approach to Reporting Breast Fine Needle Aspiration in Males
A
TE
Rebecca F. MacIntosh, M.D., F.R.C.P.(C), Jennifer L. Merrimen, M.D., F.R.C.P.(C), and Penny J. Barnes, M.D., F.R.C.P.(C)
Objective
Conclusion
C
sus review. The κ statistic for benign and atypical/suspicious/malignant categories was 0.90.
P L M a I
To apply the probabilistic approach to a series of fine needle aspiration (FNA) samples of male breast lesions and determine the accuracy and reproducibility of this method of reporting in men.
Based on this series, the probabilistic approach can be applied to the reporting of FNAs of male breast lesions. Gynecomastia may result in an atypical cytologic diagnosis. (Acta Cytol 2008;52:530–534) We demonstrate that the
Study Design
ht
rig
©
C
op y
T
F
D
O
N O
ed
D
U
te
ria
l
All male breast surgical specimens with a preoperative breast FNA at our institution probabilistic approach is both Keywords: aspiration cytolfrom 1994 to 2005 were identified. The FNAs were blindly accurate and reproducible in the ogy, fine-needle; gynecomale breast; probreviewed by 2 groups of obassessment of male breast masses. mastia; abilistic. servers and classified in 1 of 5 categories using published reine needle aspiration porting guidelines: positive, (FNA) of palpable suspicious, atypical, proliferative without atypia and unrebreast lesions is a widely used method to achieve a premarkable. The histologic and cytologic diagnoses were correoperative diagnosis of breast cancer. The reported lated. The interobserver variation was determined. sensitivity and specificity of breast FNA ranges from Results 94% to 100% and 98% to 100%, respectively.1-3 It is known that there are some benign proliferative breast A total of 138 FNAs were performed for 123 male patients. lesions that can exhibit cytologic features similar to Histologic correlation was available for 23 satisfactory those of low-grade or sparsely cellular carcinomas.4 FNAs. A total of 11 of 11 carcinomas (100%) were classiMany authorities have recognized this limitation of fied as positive, suspicious or atypical. Of 12 benign masses, breast cytopathology and have advocated the use of a 11 (91.6%) were classified as proliferative without atypia categorical or probabilistic approach for reporting or unremarkable. One case of gynecomastia was classified as breast FNA to limit false positive and false negative diatypical by 1 observer but deemed not atypical with consenFrom the Division of Anatomical Pathology, Department of Pathology, Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia, and Department of Pathology, Credit Valley Hospital, Mississauga, Ontario, Canada. Dr. MacIntosh is Associate Professor, Department of Pathology, Dalhousie University. Dr. Merrimen is Fellow in Anatomical Pathology, Credit Valley Hospital. Dr. Barnes is Associate Professor, Department of Pathology, Dalhousie University. Address correspondence to: Rebecca F. MacIntosh, M.D., F.R.C.P.(C), Anatomical Pathology and Laboratory Medicine, 7th Floor, Mackenzie Building, Queen Elizabeth II Health Sciences Centre (VG site), 5788 University Avenue, Halifax, Nova Scotia, Canada B3H 1V8 (
[email protected]). Financial Disclosure: The authors have no connection to any companies or products mentioned in this article. Received for publication July 3, 2007. Accepted for publication August 6, 2007.
530
ACTA CYTOLOGICA
0001-5547/08/5205-0530/$21.00/0 © The International Academy of Cytology
Probabilistic Approach to Male Breast FNA
tolyte for ThinPrep (Cytyc Corp, Boxborough, Massachusetts, U.S.A.) processing. All smears, cytospins and liquid-based preparations were stained with Papanicolaou stain. The slides were reviewed retrospectively by 2 groups of pathologists (P.B. and R.M./J.M.) in a blinded manner. The cytopathologic diagnoses were classified into 5 categories as described by Wang and Ducatman9: unremarkable, proliferative without atypia, atypical, suspicious and positive. In those cases in which there was a discrepancy in the cytologic interpretation, a consensus review was performed. The cytologic diagnoses were correlated with the histologic findings. The positive predictive value (PPV), negative predictive value (NPV) and κ statistic were calculated using standard statistical methods. Results
te
ria
P L M a I
Over a period of 10.5 years (June 1994 to December 2005), 4,183 breast FNAs were reported at our institution. Of this group, 138 FNAs were performed on 123 male patients. Of the male breast FNAs, 46 (33.3%) were reported as unsatisfactory. Breast FNA specimens at our institution are considered inadequate or unsatisfactory if they are acellular or sparsely cellular if mass lesions are sampled. Histologic correlation was available for 23 satisfactory FNAs of unilateral breast masses performed on 23 male patients (age range, 27–79 years). These 23 cases formed our study set.
ed
ht rig
D
U
The computerized files of the Queen Elizabeth II Health Sciences Centre were searched for all FNAs of breast masses performed between June 1994 and December 2005. FNAs from male patients with histologic follow-up were selected for review. The histopathologic data on these patients was obtained from the computerized surgical pathology files. We obtained the required institutional review board approval for this study. Clinicians performed all of the FNAs. Those performed between 1994 and 2003 were submitted to the laboratory in 50% ethanol for smear and/or cytospin preparation. After 2003, they were submitted in Cy-
l
Materials and Methods
C
A
TE
agnoses.5-9 Wang and Ducatman9 and Ayata et al10 published well-defined criteria for each category and demonstrated the accuracy and consistency of this reporting scheme. Male breast carcinoma is a rare disease, forming < 1% of all breast cancers.11 Our experience in assessing male breast FNA is therefore limited compared to that in female breast FNA. The spectrum of benign proliferative breast lesions in men also varies from that in women. The majority of unilateral male breast masses are due to gynecomastia, a lesion that may lead to cytologic misdiagnosis of malignancy.2,12,13 We aimed to determine whether the probabilistic approach can be effectively applied to a series of male breast FNAs with tissue correlation.
D
O
op y
Prep
C
74 66 61 52 38 79 54 72 77 72 72 73 33 75 33 69 30 63 74 57 46 34 27
Cytospin ThinPrep Cytospin ThinPrep ThinPrep Cytospin Conventional smear Conventional smear Cytospin/conventional smear ThinPrep ThinPrep ThinPrep Cytospin Conventional smear Cytospin ThinPrep Conventional smear Cytospin Cytospin ThinPrep Conventional smear Conventional smear ThinPrep
©
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Patient age (yr)
N O
Case no.
T
Table I Cytologic Diagnoses of FNAs of Male Breast Lesions with Corresponding Histologic Findings Cytologic interpretation
Tissue diagnosis
Unremarkable Proliferative Proliferative Unremarkable Unremarkable Positive Positive Positive Proliferative Suspicious Positive Unremarkable Unremarkable Positive Suspicious Positive Suspicious Positive Unremarkable Suspicious Unremarkable Unremarkable Proliferative
Gynecomastia Gynecomastia Gynecomastia Gynecomastia Gynecomastia Invasive ductal carcinoma Invasive ductal carcinoma Invasive ductal carcinoma Negative Invasive ductal carcinoma Invasive ductal carcinoma Gynecomastia Gynecomastia Invasive papillary carcinoma Invasive ductal carcinoma Invasive ductal carcinoma Secretory carcinoma Invasive ductal carcinoma Gynecomastia Invasive ductal carcinoma Gynecomastia Gynecomastia Fibroadenoma
Volume 52 Number 5
September–October 2008
ACTA CYTOLOGICA
531
A
TE
MacIntosh, Merrimen and Barnes
B
C
A
ria
te
O
©
C
op y
rig
ht
ed
D
T
N O
l
of invasive papillary carcinoma and 1 invasive secretory carcinoma. Four of the histologically proven cancers were called suspicious on cytology, including the secretory carcinoma, which cytologically is a low grade cancer (Figure 2). The other 6 cases were interpreted as malignant, including 1 invasive papillary carcinoma (Figure 3). In current practice at our institution, a diagnosis of suspicious or malignant on FNA would result in a recommendation for excision. An FNA diagnosis of atypical would prompt a needle core biopsy or an excision. Because 11 of 11 cases of carcinoma were classified as either malignant or suspicious on FNA, the PPV value of applying the probabilistic approach to this subset of male breast FNAs is 100%. Likewise, the NPV, following consensus review of 1 case, is 100%. With omission of consensus review, the NPV value is 92%. The κ statistic for the combined unremarkable and proliferative without atypia (benign) categories
U
On cytologic review, 12 (52%) cases were classified as unremarkable or proliferative without atypia. The remaining 11 cases (48%) were classified as atypical, suspicious or malignant (Table I). In case no. 2, there was a discrepancy with interpretation as a proliferative lesion without atypia by 1 group and as atypical by the other. This prompted a consensus review, which resulted in a final interpretation of proliferative without atypia. There was agreement on all other cases reviewed. Histologic correlation revealed 12 benign lesions, including 10 cases of gynecomastia (43% of cases) (Figure 1). The 1 discrepant cytologic interpretation was a case of gynecomastia on excisional biopsy. This was correctly classified as a proliferative lesion without atypia on consensus review (Table II). The remaining 11 cases (48%) were malignant on histologic follow-up. These included 9 cases of invasive ductal carcinoma of no special type (NST), 1 case
P L M a I
Figure 1 FNA of gynecomastia. (A) Cohesive epithelial group. Note myoepithelial cell nucleus in the background (arrow). (B) Epithelial group in a stromal fragment (Papanicolaou stain, × 400).
D
Table II Histologic Correlation with Consensus Preoperative FNA Diagnoses Histologic diagnosis
Cytologic diagnostic category Unremarkable/proliferative without atypia Atypical Suspicious Positive Total
Benign
Malignant
Other
Gynecomastia
2
10
2
10
Ductal NST
Papillary carcinoma
1 3 5
1
9
1
NST = no special type.
532
ACTA CYTOLOGICA
Volume 52 Number 5
September–October 2008
Secretory carcinoma
1 1
Probabilistic Approach to Male Breast FNA
A
TE
negative diagnoses.14 Male breast cancer is a rare disease, accounting for < 1% of all breast cancers.11,13 These generally present as unilateral, firm subareolar masses. Except for the rarity of lobular variants of invasive breast cancer, the distribution of tumor type and grade between males and postmenopausal females appears similar.15 Gynecomastia is the most common lesion of the male breast encountered on FNA and generally affects the same age-group as male breast cancer. Gynecomastia is recognized as a “gray zone” lesion on FNA and may be a source of false positive diagnoses. Interpretive errors are attributed to the florid hyperplasia that often occurs in association with this lesion.2,5,12,16-18 Male breast FNA may also result in high unsatisfactory results reported on the order of 15%.2 Possible factors for low-yield specimens include dense fibrosis associated with the late or inactive phase of gynecomastia and desmoplastic carcinomas.2 In spite of this, FNA is recognized as a useful preoperative tool for the diagnosis of male breast cancer, having a sensitivity of 87–95.5% and a specificity of 78–100%.2,12 To our knowledge, our study is the first to apply the probabilistic approach to a series of male breast FNA cases with tissue correlation. In this study of 23 cases, we demonstrate that the probabilistic approach is both accurate and reproducible in the assessment of male breast masses. The accuracy of this approach is suggested by the excellent PPV and consensus NPV of 100%. There were no false negative diagnoses. One case of gynecomastia was deemed atypical by 1 investigator but, on consensus review, was interpreted as proliferative without atypia. In the event that this discrepancy had not been changed to proliferative without atypia, the PPV would have been 92%. In practice at our institution, an FNA diagnosis of “atypical”
l
ria
ht
rig
op y
C
©
D
O
N O
T
D
Many authorities in the field of cytopathology have advocated for a uniform approach to reporting FNA specimens of breast lesions in an effort to improve diagnostic accuracy and avoid interpretive errors. In 1996 a classification of diagnostic terminology for breast FNAs was published in the National Cancer Institute–sponsored consensus conference document.6 At about the same time, other experts in cytopathology and breast pathology proposed similar approaches.5,7-9 Wang and Ducatman9 developed a classification system for reporting FNA of the breast based on the likelihood of identifying cancer. This scheme provides a basis for guidelines for the management of breast lesions.9 In our practice, we have found their criteria to be well defined and easily applied. The accuracy of this approach to breast FNA assessment was confirmed by Ayata et al10 in a study of 297 patients. This group reported a PPV of 100% and a NPV of 96% of FNA of breast lesions, independent of the level of experience of individual pathologists.10 Gornstein et al14 assessed the interobserver agreement of this method of reporting on a set of breast FNAs prepared on ThinPrep slides. The investigators reported a high level of interobserver agreement for the positive category (κ statistic 0.75), with lesser degrees of agreement for other categories. Of 120 cytologic diagnoses (20 cases assessed by 6 pathologists), there were no false positive diagnoses, although there were 10 false
ed
U
Discussion
te
and combined atypical, suspicious and malignant categories was 0.90, indicating excellent interobserver agreement for combined categories, which directs further patient management decisions.
P L M a I
C
Figure 2 FNA of secretory carcinoma. Cellular specimen shows discohesive epithelial cells with minimal nuclear atypia and secretory vacuoles (Papanicolaou stain, × 400).
Figure 3 FNA of invasive ductal carcinoma sample shows discohesive epithelial cells with marked nuclear atypia (Papanicolaou stain, × 400).
Volume 52 Number 5
September–October 2008
ACTA CYTOLOGICA
533
MacIntosh, Merrimen and Barnes
the management of male breast masses: Nineteen years of experience. Acta Cytol 1999;43:334–338 2. Siddiqui MT, Zakowski MF, Ashfaq R, Ali SZ: Breast masses in males: Multi-institutional experience on fine-needle aspiration. Diagn Cytopathol 2002;26:87–91 3. Collaço LM, de Lima RS, Werner B, Torres LF: Value of fine needle aspiration in the diagnosis of breast lesions. Acta Cytol 1999;43:587–592
TE
4. al-Kaisi N: The spectrum of the “gray zone” in breast cytology: A review of 186 cases of atypical and suspicious cytology. Acta Cytol 1994;38:898–908 5. Sneige N, Staerkel GA, Caraway NP, Fanning TV, Katz RL: A plea for uniform terminology and reporting of breast fine needle aspirates: M.D. Anderson Cancer Center proposal. Acta Cytol 1994;38:971–972
C
A
6. The uniform approach to breast fine-needle aspiration biopsy: National Cancer Institute Fine-Needle Aspiration of Breast Workshop Subcommittees. Diagn Cytopathol 1997;16:295– 311
P L M a I
7. Page DL, Johnson JE, Dupont WD: Probabilistic approach to the reporting of fine-needle aspiration cytology of the breast. Cancer (Cancer Cytopathol) 1997;81:6–9 8. Logrono R, Kurtycz DF, Inhorn SL: Criteria for reporting fine needle aspiration on palpable and nonpalpable masses of the breast. Acta Cytol 1997;41:623–627
Volume 52 Number 5
ria
te
ed
ht
11. Giordano SH, Buzdar AU, Hortobagyi GN: Breast cancer in men. Ann Intern Med 2002;137:678–687 12. Westenend PJ, Jobse C: Evaluation of fine-needle aspiration cytology of breast masses in males. Cancer (Cancer Cytopathol) 2002;96:101–104
op y
C
©
1. Joshi A, Kapila K, Verma K: Fine needle aspiration cytology in
ACTA CYTOLOGICA
10. Ayata G, Abu-Jawdeh GM, Fraser JL, Garcia LW, Upton MP, Wang HH: Accuracy and consistency in application of a probabilistic approach to reporting breast fine needle aspiration. Acta Cytol 2003;47:973–978
rig
D
T
N O
O
D
References
534
l
9. Wang HH, Ducatman BS: Fine needle aspiration of the breast: A probabilistic approach to diagnosis of carcinoma. Acta Cytol 1998;42:285–289
U
would lead to needle core biopsy for further evaluation. None of the benign lesions was regarded as positive or suspicious of malignancy by either group of reviewers. Of the histologically proven malignant lesions, all were identified as suspicious or malignant by both sets of reviewers. At our institution, either of these categories would prompt recommendation for an excision of the lesion. Of the FNAs of malignant tumors, the identification of myoepithelial cells was the most common reason for a suspicious interpretation. Presumably, these were aspirated from adjacent benign breast tissue. A case of invasive secretory carcinoma was also classified as suspicious because of its low-grade nuclear features. In addition to the high level of accuracy, the reproducibility of the probabilistic approach in males is shown by the interobserver agreement. We evaluated the interobserver agreement of combined benign categories (unremarkable, proliferative without atypia), which may not result in follow-up tissue biopsy, and the combined atypical, suspicious and malignant groups, which would mandate further investigation and biopsy. The κ statistic for these groups was 0.90, indicating excellent interobserver agreement with the application of the probabilistic approach to this set of male breast FNAs. The main limitation of our study was the relatively small number of cases of male breast FNAs with tissue correlation. This is due to the rarity of clinically worrying symptomatic male breast lesions, including male breast cancer. Many clinical decisions regarding diagnosis and management of male breast cancer are extrapolated for female breast cancer studies because of limited data and clinical experience. In summary, several authorities in the field of cytopathology have advocated for a uniform method of reporting of FNA specimens of breast masses in an attempt to improve diagnostic accuracy and avoid interpretive errors. The probabilistic approach, described by Wang and Ducatman,9 is one such approach. It has well-defined criteria that are easily applied. FNA of the male breast presents many of the same challenges as those with the female breast. In addition to those lesions encountered in women, gynecomastia is a common lesion unique to the male breast. This is recognized as a “gray zone” lesion, which may result in false positive diagnoses. We have applied the criteria for the probabilistic approach to reporting, as outlined by Wang and Ducatman,9 to a series of male breast FNAs. Using these criteria in our limited series, we demonstrated high PPVs and NPVs for identifying those cases requiring surgical management.
13. Sneige N, Holder PD, Katz RL, Fanning CV, Dekmezian RH, Shabb NS, Singletary SE: Fine-needle aspiration cytology of the male breast in a cancer center. Diagn Cytopathol 1993;9: 691–697 14. Gornstein B, Jacobs T, Bédard Y, Biscotti C, Ducatman B, Layfield L, McKee G, Sneige N, Wang H: Interobserver agreement of a probabilistic approach to reporting breast fine-needle aspirations on ThinPrep®. Diagn Cytopathol 2004;30:389– 395 15. Goodman MT, Tung KH, Wilkens LR: Comparative epidemiology of breast cancer among men and women in the US, 1996 to 2000. Cancer Causes Control 2006;17:127–136 16. Amrikachi M, Green LK, Rone R, Ramzy I: Gynecomastia: Cytologic features and diagnostic pitfalls in fine needle aspirates. Acta Cytol 2001;45:948–952 17. Das DK, Junaid TA, Mathews SB, Ajrawi TG, Ahmed MS, Madda JP, Mirza K: Fine needle aspiration cytology diagnosis of male breast lesions: A study of 185 cases. Acta Cytol 1995; 39:870–876 18. Gupta RK, Naran S, Dowle CS, Simpson JS: The diagnostic impact of needle aspiration cytology of the breast on clinical decision making with an emphasis on the aspiration cytodiagnosis of male breast masses. Diagn Cytopathol 1991;7:637–639
September–October 2008
