- Email: [email protected]
Feasibility of breast duct lavage performed by a physician extender
Feasibility of breast duct lavage performed by a physician extender Philip N. Redlich, MD, PhD, Anna C. Purdy, RN, MSN, C-ANP, Vinod B. Shidham, MD, MRCPath, FIAC, Hyun J. Yun, PhD, Alonzo Walker, MD, and David Ota, MD, Milwaukee, Wis
Background. The safety and feasibility of ductal lavage (DL), a risk-assessment tool utilizing a minimally invasive technique that permits sampling of breast duct epithelium, performed primarily by a nurse practitioner (NP), was studied prospectively. Methods. Women at high risk for breast cancer with a normal clinical breast exam and mammogram were enrolled. Nipple aspirate fluid (NAF)-yielding ducts were identified, cannulated, and lavaged primarily by an NP in collaboration with a breast surgeon. Samples with sufficient cellularity were categorized as benign, mild atypia, marked atypia, or malignant. Pain and adverse events were recorded. Results. Thirty-seven women, with a mean age of 51.7 years, were enrolled. Thirty-one (83.8%) women yielded NAF and, of those, 28 (90.3%) had one or more ducts successfully cannulated. Of 65 lavaged ducts in these 28 women, cellularity was adequate for diagnosis in 44 (67.7%) samples. Cytologic findings were as follows: 24 benign, 15 mild atypia, 4 marked atypia, and 1 malignant. The procedure was well tolerated with a mean pain score of 3.2 (SD ± 1.81). The most frequent adverse event was breast fullness, reported by 44.8% of the women. Two women with marked atypia were evaluated further and found to have intraductal papillomata. The woman with malignant cytology had ductal carcinoma in situ. Conclusion. DL is a safe, generally well-tolerated procedure that can be performed successfully by a trained NP. (Surgery 2004;136:1077-82.) From the Departments of Surgery, Pathology, and Biostatistics, Medical College of Wisconsin, Milwaukee, Wis
THE ASSESSMENT OF RISK FOR BREAST CANCER development may guide a woman’s choice of risk-reducing interventions.1 Although determination of risk as assessed by the Gail model is useful as a general tool, individualized information can be gained by analysis of breast duct epithelium.2 The detection of cellular atypia was found to correlate with an increased risk of breast cancer in a prospective study of 6,904 women that compared nipple aspirate fluid (NAF) cytology to breast cancer incidence, with a median follow-up of 21 years.3 Similar findings were observed from analysis of breast cytology obtained by random periareolar fine-needle aspiration (FNA).4 Ductal lavage (DL) is a procedure that allows the collection of breast epithelial cells more Presented at the 61st Annual Meeting of the Central Surgical Association, Chicago, Illinois, March 4-6, 2004. Supported by a grant from the Medical College of Wisconsin Cancer Center. Reprint requests: Philip N. Redlich, MD, PhD, Department of Surgery, Froedtert Memorial Lutheran Hospital, 9200 West Wisconsin Avenue, Milwaukee, WI 53226. 0039-6060/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.surg.2004.08.004
effectively and to a greater degree than by NAF and less invasively than by FNA by means of microcatheter cannulation of individual ducts, lavage with normal saline, and collection of ductal effluent for cytologic analysis.5 Furthermore, DL enables localization of abnormal cytology to an individual ductal tree. A multiinstitutional trial by Dooley et al demonstrated the safety and efficacy of DL for collection of ductal epithelial cells, but most of the procedures were performed by surgeons or physicians involved in breast care.5 DL can be a time-intensive procedure, depending on the number of identified NAF-yielding ducts and the ease of the cannulation step. Acceptance of DL by physicians as a risk-assessment tool may be enhanced by successful performance by a physician extender. We sought to corroborate the results of the multiinstitutional trial in our patient population with the DL procedure performed primarily by a trained nurse practitioner (NP) in collaboration with a breast surgeon. METHODS Women were enrolled into an institutional review board--approved protocol after giving their SURGERY 1077
1078 Redlich et al
written informed consent. To be eligible for participation, women must have had an increased risk of invasive cancer as determined by either a Gail index $1.7% at 5 years, known breast cancer gene 1 (BRCA1) or breast cancer gene 2 (BRCA2) mutation, previous diagnosis of lobular carcinoma in situ, or previous diagnosis of invasive or in situ breast cancer in the contralateral breast. All women had a normal clinical breast exam and mammogram with low or no suspicion of carcinoma (breast imaging reporting and data system, or BI-RADS, 1, 2, or 3) before enrollment. Women were excluded if recently pregnant or had lactated within 12 months of enrollment, if they had received chemotherapy or radiation therapy 6 months before enrollment, if they had prior periareolar surgery, or if they had a breast implant. The lavage procedure has been described previously.5 Briefly, topical anesthetic is applied to the nipple and the breast is warmed and massaged. After cleansing with an exfoliant (Nuprep; D.O. Weaver & Co, Aurora, Colo) and alcohol, a nipple aspirator is applied (Cytyc Health Corp, Boxborough, Mass) and NAF-yielding ducts identified. A discharging duct is progressively dilated and cannulated with a microcatheter (Cytyc). Up to 5 mL of 1% lidocaine (Abbott Labs, Chicago, Ill) is infused, followed by up to 20 mL sterile saline in 2 to 5 mL increments. The breast is intermittently massaged and effluent collected, placed in preservative (SurePath Red; TriPath Imaging Inc, Burlington, NC), and processed with standard Papanicolaou staining for cytologic examination. Slides were examined by one cytopathologist (V.B.S.) for all samples and classified as either having inadequate cellular material for diagnosis (ICMD) or having cells that were interpreted as benign or demonstrating mild atypia, marked atypia, or malignant features. Results were provided to the women and, if they desired, their referring physicians. General recommendations regarding risk counseling were provided at the time of result disclosure. Specific recommendations and follow-up occurred under the direction of their health care providers. Training in the use of the microcatheter was provided by a surgeon sponsored by the Cytyc Health Corporation. The cytopathologist completed an Internet-based online training course entitled ‘‘Breast Ductal Lavage Fluid Cytology Training Program,’’ sponsored by the Cytyc Health Corporation. Either the surgeon alone or the NP with the surgeon in attendance performed the first 4 DL procedures, with the remaining 33 lavages performed primarily by the NP with physician oversight.
Surgery November 2004
Tolerability of the procedure was assessed on a 10-point pain scale, with 10 representing the highest pain rating. Adverse events were recorded on the day of the procedure. Follow-up information was obtained by phone the following day and 2 weeks later. The most significant adverse event and highest pain scale rating per subject in either breast was used in the analysis. Descriptive statistics were used to summarize data. Comparison of data to the multi-institutional trial (Dooley et al) was performed by two-sided chisquare or the Fisher exact test.6 A P value of < .05 was considered statistically significant. Analysis was performed with SAS statistical software version 8 (SAS Institute, Cary, NC).7 RESULTS Thirty-seven white women, with a mean age of 51.7 years (SD ± 7.9), were enrolled. Two women were eligible based on a positive test for a mutation in either the BRCA1 or BRCA2 gene, 5 based on a previous diagnosis of invasive breast cancer in the contralateral breast, 1 based on a diagnosis of ductal carcinoma in situ, 1 based on a diagnosis of lobular carcinoma in situ, and 28 based on an elevated 5-year Gail risk, with a mean value of 3.4% (SD ± 1.5%). One patient who underwent cannulation did not have NAF-yielding ducts identified and was excluded from analysis except for calculations of adverse events. Thirty-one (83.8%) women yielded NAF, and among them, 28 (90.3%) women underwent successful microcatheter cannulation in one or more ducts, which is comparable to a 91.8% success rate calculated from data reported by Dooley et al (P = .73).5 DL yielded samples with adequate cellularity in 85.7% (24/28) of women, not statistically different when compared with the data from Dooley et al (78.1%, P = .48). Among adequate lavage samples, 54.2% (13/24) of women were interpreted as having cytologically abnormal cells in one or more ducts. The proportion of women with abnormal cytology in our study was higher than in Dooley et al (30.8%, P = .02). The distribution of cytologic diagnoses by patient is presented in the Table. In 31 women, 73 NAF-yielding ducts were identified, with a mean of 2.3 ducts per subject. Microcatheter cannulation was successful in 65 ducts (89%), yielding 44 samples (67.7%) of sufficient cellularity for cytologic diagnosis. Neither the success rate of cannulation by duct nor the adequacy of samples obtained was statistically different from data from Dooley et al, 84.2%
Redlich et al 1079
Surgery Volume 136, Number 5
Table. Distribution of cytologic diagnoses of ductal lavage samples Cytologic diagnosis ICMD* Benign Mild Atypia Marked Atypia Malignant
Distribution by subject (n = 28) 4 11 8 4 1
(14%) (39%) (29%) (14%) (4%)
Distribution by duct (n = 65) 21 24 15 4 1
(32%) (37%) (23%) (6%) (1.5%)
*ICMD: Inadequate cellular material for diagnosis.
(P = .38) and 70.7% (P = .67), respectively. Among ductal samples with adequate cellularity for cytologic diagnosis, 45.5% (20/44) were interpreted as demonstrating abnormal cytologic findings, compared with 25.1% in the study by Dooley et al (P = .004). The distribution of cytologic diagnoses by duct is presented in the Table. The primary difference in the distribution of cytologic diagnoses between our study and that of Dooley et al is the proportion of samples diagnosed as benign or mild atypia (P # .018) (Figure). Women found to have mild atypia were offered risk counseling as described8,9 and referred to their or other health care providers for specific treatment recommendations. Of the 4 women with marked atypia, 1 was lost to follow-up after moving to another state. One woman underwent ductoscopy and duct excision by her surgeon, revealing ductal papillomata. Two women underwent galactography, ultrasound, and magnetic resonance imaging (MRI) directed by their surgeons. In 1 of these women, imaging results identified an intraductal mass that proved to be an intraductal papilloma after excisional biopsy. The other woman had a negative imaging evaluation and is currently being closely monitored. The woman with malignant ductal cytology underwent galactography, ultrasound, and MRI. A wire-localized biopsy of narrowed ducts revealed ductal carcinoma in situ. Due, in part, to the extensive nature of her disease, she chose mastectomy as surgical treatment. The DL procedure was generally well tolerated, with a mean pain score of 3.2 (SD ± 1.81). Reported complaints included breast fullness (44.8%), nipple tenderness (22.9%), bruising (17.1%), and erythema (6.9%). Most symptoms resolved within 24 hours, and all were resolved by 2 weeks. DISCUSSION Cytologic analysis of breast duct epithelium in a population of women at high risk for breast
Figure. Comparison of the distribution of cytologic diagnoses between the current study (Redlich et al) and the multi-institutional study (Dooley et al).5
cancer who are otherwise asymptomatic serves as a tool for additional risk assessment at the current time.8,9 Toward this end, DL provides a minimally invasive method of collecting sufficient cells for cytologic analysis in the majority of subjects. In our study, >67% of NAF-yielding ducts sampled yielded sufficient cells for cytologic analysis. On a persubject basis, information from 1 or more ducts was available for risk assessment in >85% of women who had successfully cannulated ducts. The proportion of women who had cytologically evaluable samples, as achieved by a trained NP in all but 4 of the cases, is very similar and not statistically different from data from the multi-institutional study by Dooley et al, in which the majority of procedures were performed by surgeons.5 Our study differed from this large multi-institutional trial with respect to the distribution of cytologic diagnosis when analyzed by patient or by duct. Our study had a consistently higher proportion of samples interpreted as showing abnormal cytologic findings than did the study reported by Dooley et al.5 This difference is most likely related to subjectivity in interpretation of cytologic findings of ductal epithelium and not related to the technique of sample acquisition by the NP. The primary difference in cytologic interpretation was related to the proportion of samples interpreted as either benign or mild atypia, and not to the more significant interpretations of marked atypia and malignant. This interobserver variation in cytologic analysis is comparable to results reported by Brogi et al for DL specimens from patients undergoing
1080 Redlich et al
mastectomy for mammary carcinoma.10 In the study by Brogi et al, 3 pathologists reviewed DL samples and independently assigned a diagnosis. The diagnosis of mild atypia was the least reproducible diagnostic category, with 3 of 10 samples having entirely discrepant diagnoses by the 3 cytopathologists and no unanimous diagnosis being reached on the remaining 7 samples. No discordant results were recorded for the diagnoses of marked atypia or benign. In our study, 2 patients with marked atypia were found to have intraductal papillomata, and 1 patient with malignant cytology had ductal carcinoma in situ. Whereas the sensitivity for detecting neoplastic disease within the breast is low using cytologic analysis of DL samples, the specificity is sufficiently high to warrant further diagnostic evaluation in patients in whom cells with marked atypia or malignant characteristics are found.8,9,11,12,13 Increased specificity of DL has been shown in small studies by use of fluorescence in situ hybridization in detecting genetic alterations; however, further studies are required to confirm these findings before the technique of DL can be recommended as a method for detecting neoplastic disease in the breast.12,13 In summary, DL is a safe and generally welltolerated procedure that can be performed by an NP in collaboration with a surgeon. Acceptance and use of this technique as a risk assessment tool may be enhanced by performance of ductal lavage by trained physician extenders. We thank Rachel Davis for her assistance with data management.
REFERENCES 1. Allain D, Gilligan MA, Redlich PN. Genetics and genetic counseling for breast cancer. In: Donegan WL, Spratt JS, editors. Cancer of the Breast. 5th ed. St. Louis: Saunders, 2002. p. 249-68. 2. Gail MH, Brinton LA, Byar DP, Corle DK, Green SB, Schairer C, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 2000;92: 1217-27. 3. Wrensch MR, Petrakis NL, Rei M, King EB, Chew K, Neuhaus J, et al. Breast cancer risk in women with abnormal cytology in nipple aspirates of breast fluid. J Natl Cancer Inst 2001;93:1791-8. 4. Fabian CJ, Kimler BF, Zaller CM, Klemp JR, Damel S, Zeiger S, et al. Short-term breast cancer prediction by random periareolar fine-needle aspiration (FNA) cytology and the Gail risk model. J Natl Cancer Inst 2000;92: 1217-27. 5. Dooley WC, Ljung BM, Veronesi U, Cazzaniga M, Elledge RM, O’Shaughnessy JA, et al. Ductal lavage for
Surgery November 2004
6. 7. 8.
9.
10.
11. 12.
13.
detection of cellular atypia in women at high risk for breast cancer. J Natl Cancer Inst 2001;93:1624-32. Agresti A. Categorical Data Analysis. New York: John Wiley and Sons; 1990. SAS Software: Usage and Reference. Version 8. Cary, NC: SAS Institute Inc; 1997. Morrow M, Vogel V, Ljung BM, O’Shaughnessy JA. Evaluation and management of the woman with an abnormal ductal lavage. J Am Coll Surg 2002;194:648-56. O’Shaughnessy JA, Ljung BM, Dooley WC, Chang J, Duerer HM, Hung DT, et al. Ductal lavage and the clinical management of women at high risk for breast cancer. Cancer 2002;94:292-8. Brogi E, Robson M, Panageas KS, Casadio C, Ljung BM, Montgomery L. Ductal lavage in patients undergoing mastectomy for mammary carcinoma: a correlative study. Cancer 2003;98:2170-6. Lee WY. Cytology of abnormal nipple discharge: a cytohistological correlation. Cytopathology 2003;14:19-26. Yamamoto D, Senzaki H, Nakagawa H, Okugawa H, Gondo H, Tanaka K. Detection of chromosomal aneusomy by fluorescence in situ hybridization for patients with nipple discharge. Cancer 2003;97:690-4. King BL, Tsai SC, Gryga MI, D’Aquila TG, Seelig SA, Morrison LE, et al. Detection of chromosomal instability in paired breast surgery and ductal lavage specimens by interphase fluorescence in situ hybridization. Clin Cancer Res 2003;9:1509-16.
DISCUSSION Dr Seema A. Khan (Chicago, Ill). I would like to congratulate Dr Redlich on a valuable single-institution experience that corroborates the findings reported so far on this new technique of DL. In terms of the subject’s willingness to undergo this procedure, pain is certainly one barrier, and the 3.2 pain score is a little bit higher than what was reported in the multicenter study. We found that by using periareolar lidocaine infiltration in addition to the topical and intraductal anesthetic, we can reduce this mean score to 2.2. I would like his comments about this. The diagnostic investigation of abnormal cytology is another area that has elicited great discussion and is an important issue. I wonder how the women who had abnormal cytology were evaluated. Finally, I would like to raise the issue of interobserver variation in interpreting the lavage samples because this is the greatest problem, I think, in interpreting the results of this procedure. I will just review what Dr Redlich has shown you already, and compare our results and those of Dr Dooley with the results of Redlich et al (presented in the middle column; Dr Dooley’s are on the left, and our unpublished experience from Northwestern on the right) (Discussion Table). The insufficiency rate has been improved on in experiences subsequent to the multicenter trial. The benign cytology rate is similar in our experience and in Dooley’s, about 50%, but is lower in the Wisconsin study. The mild atypia rate ranges between 17% and 30% by woman, and our
Redlich et al 1081
Surgery Volume 136, Number 5
Discussion Table. DL Cytology by subject across studies Cytology
Dooley n=500
Redlich n=37
Khan* n=100
DL success Insufficient Benign Mild Atypia Marked Atypia
83% 22% 54% 17% 7%
84% 14% 39% 29% 19%
85% 18% 50% 30% 2%
Kappa <.0001. *Unpublished data, updated 1/04.
mild atypia rate is very similar to Dr Redlich’s experience. But the marked atypia rate seems to vary significantly among the 3 series, ranging from 2% (Northwestern) to 19% (Redlich). In looking at this further, I think it is important to point out that interobserver variation is a great problem in this area. We looked at interobserver variability among cytopathologists in a study that we performed at Northwestern and presented at San Antonio a year ago, which is to be published in the Journal of the National Cancer Institute. These samples came from women who were undergoing mastectomy for known disease, and there were 14 mildly atypical samples. Between our readings at Northwestern and the review by a UCSF cytopathologist, Dr Britt Marie Ljung, there was only about a 50% concordance in the diagnosis of mild atypia. This is corroborated also by data from Memorial Sloan-Kettering published recently in Cancer : of 10 mildly atypical samples, none were agreed on by the 3 different readers. So variability in cytopathology interpretation is, I think, the big issue. I would ask, are you contemplating reassessment of your samples by a second cytopathologist? Dr Mark A. Malangoni (Cleveland, Ohio). Do you think that this is a procedure that all general surgeons who treat breast disease to a significant extent should have available in their offices, or should this really be reserved for the center that may treat the occasional case that is difficult to diagnose? My second question relates to the discomfort that you identified in your patients. It just seems that the time to resolution of this seems inordinately long. I wonder if it is a problem associated with the volume of the infusate, or if it has to do with the pH. If it has to do with the pH, you could fix that by using bicarbonate as a buffer rather than using normal saline, the pH of which is rather low. Dr Khan. Just a quick follow-up on the issue of the pH of the infusate. We have in fact switched routinely to using Plasmalyte, which is a neutral buffered electrolyte solution, and we find that it does contribute to lessening the discomfort. Dr Redlich. Dr Khan, the pain score, I agree, was a little higher in our study. We have a smaller number of women; therefore, the percent of our enrollees in the learning curve of this is higher. So I think we probably had some increased pain at the beginning in learning how to probe, how to intubate the small orifices. As we
get better and add more subjects to the study, I would assume that the pain score would decrease. How did we investigate these studies? I can tell you that one patient with malignancy was very interesting. I performed every study I could think of on her. I can tell you what happened as an anecdotal experience. We started with reviewing the physical exam, reviewing the mammograms, and ordering additional films. Then we moved to MRI, which was negative. We did ductography, which showed at best a subtle abnormality of a peripheral duct. Then the only thing I could do was biopsy that subtle abnormality, and the biopsy was needle directed with ductography in the same day. It showed extensive ductal carcinoma in situ with positive margins. So evaluation can be a difficult problem. The two patients that we studied with marked atypia had been found to have papillomata by either ductography or ductoscopy and subsequent biopsy. I think it behooves us not to dismiss results showing marked atypia or malignancy and to evaluate those cases further. Interobserver variation seems to be a constant variable in all the different studies, as you alluded to. In our study, our cytopathologist, who is a trained cytopathologist, did some extra training before we started this project by online training and sending samples to UCSF for review. And 3 of the 10 samples or so he sent at the beginning of our study, for which he had diagnosed mild atypia, were rereviewed and interpreted as benign findings by the outside reviewer. So I think variability is inherent in the cytology. How do we get around it? There are studies looking at ways of measuring abnormalities that have less subjectivity, an example being fluorescent in situ hybridization (FISH), looking at abnormal number of gene copies. In a couple of small studies that have been published so far, FISH demonstrated a higher sensitivity in detecting abnormalities than did cytology. But these are small studies and need to be further corroborated with larger studies. Dr Malangoni, I think that is a good point. Most of the complaints resolved by the following day. We had a 2week follow-up phone call just to confirm that there were no other issues and that they did resolve. I think our learning curve, again, probably had a lot to do with some of the pain and some of the bruising that we observed in our enrollees.
1082 Redlich et al
Ductal lavage cytology should not be used to diagnose breast lesions. If the general surgeon has a high-volume practice in breast and as part of his practice does risk assessment and counseling, DL may fit in that particular
Surgery November 2004
practice. But if you are not doing risk assessment and counseling as a comprehensive program, including BRCA testing, then it may not be helpful to use the technology of DL.
Background. The safety and feasibility of ductal lavage (DL), a risk-assessment tool utilizing a minimally invasive technique that permits sampling of breast duct epithelium, performed primarily by a nurse practitioner (NP), was studied prospectively. Methods. Women at high risk for breast cancer with a normal clinical breast exam and mammogram were enrolled. Nipple aspirate fluid (NAF)-yielding ducts were identified, cannulated, and lavaged primarily by an NP in collaboration with a breast surgeon. Samples with sufficient cellularity were categorized as benign, mild atypia, marked atypia, or malignant. Pain and adverse events were recorded. Results. Thirty-seven women, with a mean age of 51.7 years, were enrolled. Thirty-one (83.8%) women yielded NAF and, of those, 28 (90.3%) had one or more ducts successfully cannulated. Of 65 lavaged ducts in these 28 women, cellularity was adequate for diagnosis in 44 (67.7%) samples. Cytologic findings were as follows: 24 benign, 15 mild atypia, 4 marked atypia, and 1 malignant. The procedure was well tolerated with a mean pain score of 3.2 (SD ± 1.81). The most frequent adverse event was breast fullness, reported by 44.8% of the women. Two women with marked atypia were evaluated further and found to have intraductal papillomata. The woman with malignant cytology had ductal carcinoma in situ. Conclusion. DL is a safe, generally well-tolerated procedure that can be performed successfully by a trained NP. (Surgery 2004;136:1077-82.) From the Departments of Surgery, Pathology, and Biostatistics, Medical College of Wisconsin, Milwaukee, Wis
THE ASSESSMENT OF RISK FOR BREAST CANCER development may guide a woman’s choice of risk-reducing interventions.1 Although determination of risk as assessed by the Gail model is useful as a general tool, individualized information can be gained by analysis of breast duct epithelium.2 The detection of cellular atypia was found to correlate with an increased risk of breast cancer in a prospective study of 6,904 women that compared nipple aspirate fluid (NAF) cytology to breast cancer incidence, with a median follow-up of 21 years.3 Similar findings were observed from analysis of breast cytology obtained by random periareolar fine-needle aspiration (FNA).4 Ductal lavage (DL) is a procedure that allows the collection of breast epithelial cells more Presented at the 61st Annual Meeting of the Central Surgical Association, Chicago, Illinois, March 4-6, 2004. Supported by a grant from the Medical College of Wisconsin Cancer Center. Reprint requests: Philip N. Redlich, MD, PhD, Department of Surgery, Froedtert Memorial Lutheran Hospital, 9200 West Wisconsin Avenue, Milwaukee, WI 53226. 0039-6060/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.surg.2004.08.004
effectively and to a greater degree than by NAF and less invasively than by FNA by means of microcatheter cannulation of individual ducts, lavage with normal saline, and collection of ductal effluent for cytologic analysis.5 Furthermore, DL enables localization of abnormal cytology to an individual ductal tree. A multiinstitutional trial by Dooley et al demonstrated the safety and efficacy of DL for collection of ductal epithelial cells, but most of the procedures were performed by surgeons or physicians involved in breast care.5 DL can be a time-intensive procedure, depending on the number of identified NAF-yielding ducts and the ease of the cannulation step. Acceptance of DL by physicians as a risk-assessment tool may be enhanced by successful performance by a physician extender. We sought to corroborate the results of the multiinstitutional trial in our patient population with the DL procedure performed primarily by a trained nurse practitioner (NP) in collaboration with a breast surgeon. METHODS Women were enrolled into an institutional review board--approved protocol after giving their SURGERY 1077
1078 Redlich et al
written informed consent. To be eligible for participation, women must have had an increased risk of invasive cancer as determined by either a Gail index $1.7% at 5 years, known breast cancer gene 1 (BRCA1) or breast cancer gene 2 (BRCA2) mutation, previous diagnosis of lobular carcinoma in situ, or previous diagnosis of invasive or in situ breast cancer in the contralateral breast. All women had a normal clinical breast exam and mammogram with low or no suspicion of carcinoma (breast imaging reporting and data system, or BI-RADS, 1, 2, or 3) before enrollment. Women were excluded if recently pregnant or had lactated within 12 months of enrollment, if they had received chemotherapy or radiation therapy 6 months before enrollment, if they had prior periareolar surgery, or if they had a breast implant. The lavage procedure has been described previously.5 Briefly, topical anesthetic is applied to the nipple and the breast is warmed and massaged. After cleansing with an exfoliant (Nuprep; D.O. Weaver & Co, Aurora, Colo) and alcohol, a nipple aspirator is applied (Cytyc Health Corp, Boxborough, Mass) and NAF-yielding ducts identified. A discharging duct is progressively dilated and cannulated with a microcatheter (Cytyc). Up to 5 mL of 1% lidocaine (Abbott Labs, Chicago, Ill) is infused, followed by up to 20 mL sterile saline in 2 to 5 mL increments. The breast is intermittently massaged and effluent collected, placed in preservative (SurePath Red; TriPath Imaging Inc, Burlington, NC), and processed with standard Papanicolaou staining for cytologic examination. Slides were examined by one cytopathologist (V.B.S.) for all samples and classified as either having inadequate cellular material for diagnosis (ICMD) or having cells that were interpreted as benign or demonstrating mild atypia, marked atypia, or malignant features. Results were provided to the women and, if they desired, their referring physicians. General recommendations regarding risk counseling were provided at the time of result disclosure. Specific recommendations and follow-up occurred under the direction of their health care providers. Training in the use of the microcatheter was provided by a surgeon sponsored by the Cytyc Health Corporation. The cytopathologist completed an Internet-based online training course entitled ‘‘Breast Ductal Lavage Fluid Cytology Training Program,’’ sponsored by the Cytyc Health Corporation. Either the surgeon alone or the NP with the surgeon in attendance performed the first 4 DL procedures, with the remaining 33 lavages performed primarily by the NP with physician oversight.
Surgery November 2004
Tolerability of the procedure was assessed on a 10-point pain scale, with 10 representing the highest pain rating. Adverse events were recorded on the day of the procedure. Follow-up information was obtained by phone the following day and 2 weeks later. The most significant adverse event and highest pain scale rating per subject in either breast was used in the analysis. Descriptive statistics were used to summarize data. Comparison of data to the multi-institutional trial (Dooley et al) was performed by two-sided chisquare or the Fisher exact test.6 A P value of < .05 was considered statistically significant. Analysis was performed with SAS statistical software version 8 (SAS Institute, Cary, NC).7 RESULTS Thirty-seven white women, with a mean age of 51.7 years (SD ± 7.9), were enrolled. Two women were eligible based on a positive test for a mutation in either the BRCA1 or BRCA2 gene, 5 based on a previous diagnosis of invasive breast cancer in the contralateral breast, 1 based on a diagnosis of ductal carcinoma in situ, 1 based on a diagnosis of lobular carcinoma in situ, and 28 based on an elevated 5-year Gail risk, with a mean value of 3.4% (SD ± 1.5%). One patient who underwent cannulation did not have NAF-yielding ducts identified and was excluded from analysis except for calculations of adverse events. Thirty-one (83.8%) women yielded NAF, and among them, 28 (90.3%) women underwent successful microcatheter cannulation in one or more ducts, which is comparable to a 91.8% success rate calculated from data reported by Dooley et al (P = .73).5 DL yielded samples with adequate cellularity in 85.7% (24/28) of women, not statistically different when compared with the data from Dooley et al (78.1%, P = .48). Among adequate lavage samples, 54.2% (13/24) of women were interpreted as having cytologically abnormal cells in one or more ducts. The proportion of women with abnormal cytology in our study was higher than in Dooley et al (30.8%, P = .02). The distribution of cytologic diagnoses by patient is presented in the Table. In 31 women, 73 NAF-yielding ducts were identified, with a mean of 2.3 ducts per subject. Microcatheter cannulation was successful in 65 ducts (89%), yielding 44 samples (67.7%) of sufficient cellularity for cytologic diagnosis. Neither the success rate of cannulation by duct nor the adequacy of samples obtained was statistically different from data from Dooley et al, 84.2%
Redlich et al 1079
Surgery Volume 136, Number 5
Table. Distribution of cytologic diagnoses of ductal lavage samples Cytologic diagnosis ICMD* Benign Mild Atypia Marked Atypia Malignant
Distribution by subject (n = 28) 4 11 8 4 1
(14%) (39%) (29%) (14%) (4%)
Distribution by duct (n = 65) 21 24 15 4 1
(32%) (37%) (23%) (6%) (1.5%)
*ICMD: Inadequate cellular material for diagnosis.
(P = .38) and 70.7% (P = .67), respectively. Among ductal samples with adequate cellularity for cytologic diagnosis, 45.5% (20/44) were interpreted as demonstrating abnormal cytologic findings, compared with 25.1% in the study by Dooley et al (P = .004). The distribution of cytologic diagnoses by duct is presented in the Table. The primary difference in the distribution of cytologic diagnoses between our study and that of Dooley et al is the proportion of samples diagnosed as benign or mild atypia (P # .018) (Figure). Women found to have mild atypia were offered risk counseling as described8,9 and referred to their or other health care providers for specific treatment recommendations. Of the 4 women with marked atypia, 1 was lost to follow-up after moving to another state. One woman underwent ductoscopy and duct excision by her surgeon, revealing ductal papillomata. Two women underwent galactography, ultrasound, and magnetic resonance imaging (MRI) directed by their surgeons. In 1 of these women, imaging results identified an intraductal mass that proved to be an intraductal papilloma after excisional biopsy. The other woman had a negative imaging evaluation and is currently being closely monitored. The woman with malignant ductal cytology underwent galactography, ultrasound, and MRI. A wire-localized biopsy of narrowed ducts revealed ductal carcinoma in situ. Due, in part, to the extensive nature of her disease, she chose mastectomy as surgical treatment. The DL procedure was generally well tolerated, with a mean pain score of 3.2 (SD ± 1.81). Reported complaints included breast fullness (44.8%), nipple tenderness (22.9%), bruising (17.1%), and erythema (6.9%). Most symptoms resolved within 24 hours, and all were resolved by 2 weeks. DISCUSSION Cytologic analysis of breast duct epithelium in a population of women at high risk for breast
Figure. Comparison of the distribution of cytologic diagnoses between the current study (Redlich et al) and the multi-institutional study (Dooley et al).5
cancer who are otherwise asymptomatic serves as a tool for additional risk assessment at the current time.8,9 Toward this end, DL provides a minimally invasive method of collecting sufficient cells for cytologic analysis in the majority of subjects. In our study, >67% of NAF-yielding ducts sampled yielded sufficient cells for cytologic analysis. On a persubject basis, information from 1 or more ducts was available for risk assessment in >85% of women who had successfully cannulated ducts. The proportion of women who had cytologically evaluable samples, as achieved by a trained NP in all but 4 of the cases, is very similar and not statistically different from data from the multi-institutional study by Dooley et al, in which the majority of procedures were performed by surgeons.5 Our study differed from this large multi-institutional trial with respect to the distribution of cytologic diagnosis when analyzed by patient or by duct. Our study had a consistently higher proportion of samples interpreted as showing abnormal cytologic findings than did the study reported by Dooley et al.5 This difference is most likely related to subjectivity in interpretation of cytologic findings of ductal epithelium and not related to the technique of sample acquisition by the NP. The primary difference in cytologic interpretation was related to the proportion of samples interpreted as either benign or mild atypia, and not to the more significant interpretations of marked atypia and malignant. This interobserver variation in cytologic analysis is comparable to results reported by Brogi et al for DL specimens from patients undergoing
1080 Redlich et al
mastectomy for mammary carcinoma.10 In the study by Brogi et al, 3 pathologists reviewed DL samples and independently assigned a diagnosis. The diagnosis of mild atypia was the least reproducible diagnostic category, with 3 of 10 samples having entirely discrepant diagnoses by the 3 cytopathologists and no unanimous diagnosis being reached on the remaining 7 samples. No discordant results were recorded for the diagnoses of marked atypia or benign. In our study, 2 patients with marked atypia were found to have intraductal papillomata, and 1 patient with malignant cytology had ductal carcinoma in situ. Whereas the sensitivity for detecting neoplastic disease within the breast is low using cytologic analysis of DL samples, the specificity is sufficiently high to warrant further diagnostic evaluation in patients in whom cells with marked atypia or malignant characteristics are found.8,9,11,12,13 Increased specificity of DL has been shown in small studies by use of fluorescence in situ hybridization in detecting genetic alterations; however, further studies are required to confirm these findings before the technique of DL can be recommended as a method for detecting neoplastic disease in the breast.12,13 In summary, DL is a safe and generally welltolerated procedure that can be performed by an NP in collaboration with a surgeon. Acceptance and use of this technique as a risk assessment tool may be enhanced by performance of ductal lavage by trained physician extenders. We thank Rachel Davis for her assistance with data management.
REFERENCES 1. Allain D, Gilligan MA, Redlich PN. Genetics and genetic counseling for breast cancer. In: Donegan WL, Spratt JS, editors. Cancer of the Breast. 5th ed. St. Louis: Saunders, 2002. p. 249-68. 2. Gail MH, Brinton LA, Byar DP, Corle DK, Green SB, Schairer C, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 2000;92: 1217-27. 3. Wrensch MR, Petrakis NL, Rei M, King EB, Chew K, Neuhaus J, et al. Breast cancer risk in women with abnormal cytology in nipple aspirates of breast fluid. J Natl Cancer Inst 2001;93:1791-8. 4. Fabian CJ, Kimler BF, Zaller CM, Klemp JR, Damel S, Zeiger S, et al. Short-term breast cancer prediction by random periareolar fine-needle aspiration (FNA) cytology and the Gail risk model. J Natl Cancer Inst 2000;92: 1217-27. 5. Dooley WC, Ljung BM, Veronesi U, Cazzaniga M, Elledge RM, O’Shaughnessy JA, et al. Ductal lavage for
Surgery November 2004
6. 7. 8.
9.
10.
11. 12.
13.
detection of cellular atypia in women at high risk for breast cancer. J Natl Cancer Inst 2001;93:1624-32. Agresti A. Categorical Data Analysis. New York: John Wiley and Sons; 1990. SAS Software: Usage and Reference. Version 8. Cary, NC: SAS Institute Inc; 1997. Morrow M, Vogel V, Ljung BM, O’Shaughnessy JA. Evaluation and management of the woman with an abnormal ductal lavage. J Am Coll Surg 2002;194:648-56. O’Shaughnessy JA, Ljung BM, Dooley WC, Chang J, Duerer HM, Hung DT, et al. Ductal lavage and the clinical management of women at high risk for breast cancer. Cancer 2002;94:292-8. Brogi E, Robson M, Panageas KS, Casadio C, Ljung BM, Montgomery L. Ductal lavage in patients undergoing mastectomy for mammary carcinoma: a correlative study. Cancer 2003;98:2170-6. Lee WY. Cytology of abnormal nipple discharge: a cytohistological correlation. Cytopathology 2003;14:19-26. Yamamoto D, Senzaki H, Nakagawa H, Okugawa H, Gondo H, Tanaka K. Detection of chromosomal aneusomy by fluorescence in situ hybridization for patients with nipple discharge. Cancer 2003;97:690-4. King BL, Tsai SC, Gryga MI, D’Aquila TG, Seelig SA, Morrison LE, et al. Detection of chromosomal instability in paired breast surgery and ductal lavage specimens by interphase fluorescence in situ hybridization. Clin Cancer Res 2003;9:1509-16.
DISCUSSION Dr Seema A. Khan (Chicago, Ill). I would like to congratulate Dr Redlich on a valuable single-institution experience that corroborates the findings reported so far on this new technique of DL. In terms of the subject’s willingness to undergo this procedure, pain is certainly one barrier, and the 3.2 pain score is a little bit higher than what was reported in the multicenter study. We found that by using periareolar lidocaine infiltration in addition to the topical and intraductal anesthetic, we can reduce this mean score to 2.2. I would like his comments about this. The diagnostic investigation of abnormal cytology is another area that has elicited great discussion and is an important issue. I wonder how the women who had abnormal cytology were evaluated. Finally, I would like to raise the issue of interobserver variation in interpreting the lavage samples because this is the greatest problem, I think, in interpreting the results of this procedure. I will just review what Dr Redlich has shown you already, and compare our results and those of Dr Dooley with the results of Redlich et al (presented in the middle column; Dr Dooley’s are on the left, and our unpublished experience from Northwestern on the right) (Discussion Table). The insufficiency rate has been improved on in experiences subsequent to the multicenter trial. The benign cytology rate is similar in our experience and in Dooley’s, about 50%, but is lower in the Wisconsin study. The mild atypia rate ranges between 17% and 30% by woman, and our
Redlich et al 1081
Surgery Volume 136, Number 5
Discussion Table. DL Cytology by subject across studies Cytology
Dooley n=500
Redlich n=37
Khan* n=100
DL success Insufficient Benign Mild Atypia Marked Atypia
83% 22% 54% 17% 7%
84% 14% 39% 29% 19%
85% 18% 50% 30% 2%
Kappa <.0001. *Unpublished data, updated 1/04.
mild atypia rate is very similar to Dr Redlich’s experience. But the marked atypia rate seems to vary significantly among the 3 series, ranging from 2% (Northwestern) to 19% (Redlich). In looking at this further, I think it is important to point out that interobserver variation is a great problem in this area. We looked at interobserver variability among cytopathologists in a study that we performed at Northwestern and presented at San Antonio a year ago, which is to be published in the Journal of the National Cancer Institute. These samples came from women who were undergoing mastectomy for known disease, and there were 14 mildly atypical samples. Between our readings at Northwestern and the review by a UCSF cytopathologist, Dr Britt Marie Ljung, there was only about a 50% concordance in the diagnosis of mild atypia. This is corroborated also by data from Memorial Sloan-Kettering published recently in Cancer : of 10 mildly atypical samples, none were agreed on by the 3 different readers. So variability in cytopathology interpretation is, I think, the big issue. I would ask, are you contemplating reassessment of your samples by a second cytopathologist? Dr Mark A. Malangoni (Cleveland, Ohio). Do you think that this is a procedure that all general surgeons who treat breast disease to a significant extent should have available in their offices, or should this really be reserved for the center that may treat the occasional case that is difficult to diagnose? My second question relates to the discomfort that you identified in your patients. It just seems that the time to resolution of this seems inordinately long. I wonder if it is a problem associated with the volume of the infusate, or if it has to do with the pH. If it has to do with the pH, you could fix that by using bicarbonate as a buffer rather than using normal saline, the pH of which is rather low. Dr Khan. Just a quick follow-up on the issue of the pH of the infusate. We have in fact switched routinely to using Plasmalyte, which is a neutral buffered electrolyte solution, and we find that it does contribute to lessening the discomfort. Dr Redlich. Dr Khan, the pain score, I agree, was a little higher in our study. We have a smaller number of women; therefore, the percent of our enrollees in the learning curve of this is higher. So I think we probably had some increased pain at the beginning in learning how to probe, how to intubate the small orifices. As we
get better and add more subjects to the study, I would assume that the pain score would decrease. How did we investigate these studies? I can tell you that one patient with malignancy was very interesting. I performed every study I could think of on her. I can tell you what happened as an anecdotal experience. We started with reviewing the physical exam, reviewing the mammograms, and ordering additional films. Then we moved to MRI, which was negative. We did ductography, which showed at best a subtle abnormality of a peripheral duct. Then the only thing I could do was biopsy that subtle abnormality, and the biopsy was needle directed with ductography in the same day. It showed extensive ductal carcinoma in situ with positive margins. So evaluation can be a difficult problem. The two patients that we studied with marked atypia had been found to have papillomata by either ductography or ductoscopy and subsequent biopsy. I think it behooves us not to dismiss results showing marked atypia or malignancy and to evaluate those cases further. Interobserver variation seems to be a constant variable in all the different studies, as you alluded to. In our study, our cytopathologist, who is a trained cytopathologist, did some extra training before we started this project by online training and sending samples to UCSF for review. And 3 of the 10 samples or so he sent at the beginning of our study, for which he had diagnosed mild atypia, were rereviewed and interpreted as benign findings by the outside reviewer. So I think variability is inherent in the cytology. How do we get around it? There are studies looking at ways of measuring abnormalities that have less subjectivity, an example being fluorescent in situ hybridization (FISH), looking at abnormal number of gene copies. In a couple of small studies that have been published so far, FISH demonstrated a higher sensitivity in detecting abnormalities than did cytology. But these are small studies and need to be further corroborated with larger studies. Dr Malangoni, I think that is a good point. Most of the complaints resolved by the following day. We had a 2week follow-up phone call just to confirm that there were no other issues and that they did resolve. I think our learning curve, again, probably had a lot to do with some of the pain and some of the bruising that we observed in our enrollees.
1082 Redlich et al
Ductal lavage cytology should not be used to diagnose breast lesions. If the general surgeon has a high-volume practice in breast and as part of his practice does risk assessment and counseling, DL may fit in that particular
Surgery November 2004
practice. But if you are not doing risk assessment and counseling as a comprehensive program, including BRCA testing, then it may not be helpful to use the technology of DL.