Learning Curves and Breast Cancer Lymphatic Mapping: Institutional Volume Index

Journal of Surgical Research 97, 92–96 (2001) doi:10.1006/jsre.2001.6127, available online at http://www.idealibrary.com on

Learning Curves and Breast Cancer Lymphatic Mapping: Institutional Volume Index 1 Elisabeth Dupont, M.D., 2 Charles Cox, M.D., Steve Shivers, Ph.D., Chris Salud, Keoni Nguyen, Alan Cantor, Ph.D., and Doug Reintgen, M.D. Department of Surgery, H. Lee Moffitt Cancer Center and Research Institute, at the University of South Florida, Tampa, Florida 33612 Presented at the Annual Meeting of the Association for Academic Surgery, Tampa, Florida, November 2– 4, 2000; published online March 28, 2001

Mapping failures were found in all but 7 of the 30 institutions whose data were complete. There were 71 mapping failures among 74 surgeons over 555 cases, which yielded an overall failure rate of 12.79% (71 555). The logistic regression model revealed an inverse relationship between IVI and institutional failure rate. However, the multivariate analysis revealed that the individual surgeon performance was the most significant factor in determining institutional mapping success. Conclusion. Failure to map can be a function of multiple factors including surgical skill, surgical volume index, and injection method of the SLN patient, all under the quality control of an institution. The surgical failure rate on the other hand is a function of surgical skill, surgical volume, and injection methods. While differences in mapping success exist across institutions, this disparity is not due to factors associated with the institution as a whole, but lie with the individual surgeon. © 2001 Academic Press

Introduction. To date, studies of breast cancer lymphatic mapping (LM) have analyzed success with respect to individual surgeons. However, LM and sentinel lymph node biopsy (SLNBx) are procedures that require institutional multidisciplinary cooperation between the departments of radiology, pathology, and surgery. Thus, it is important to evaluate these procedures with respect to the institution. This study examines 30 institutions to clarify the value of the institutional volume index (IVI) (cases/month) to the outcome of LM and SLNBx in breast cancer. Methods. From July 1997 to July 1999, 30 institutions participated in the Department of Defense national breast LM trial. All participants underwent a 2-day training course for surgeons, nuclear medicine physicians, and pathologists. The records for each institution were prospectively accrued and submitted to a database. The false negatives, failure rates, and IVI were calculated for each institution. A logistic regression model plots the relationship between IVI and institutional failure rate. Using a multivariate analysis, mapping failure was analyzed as a function of case number with respect to the individual surgeon and the institution as a whole. Results. False negative results were demonstrated in only 5 (4%) cases among all institutions and were excluded from further analysis due to small numbers.

INTRODUCTION

Outcome analyses based on institutional volume have been conducted for surgical procedures such as angioplasty, Whipple operation for pancreatic cancer, total hip arthroplasty, and cholecystectomy [1, 2]. The initial experiences of many individual surgeons on the topic of sentnel lymph node (SLN) mapping are found in the current literature [3]. However, to date, no studies have analyzed the success of breast cancer lymphatic mapping (LM) and sentinel node biopsy based on institutional volume. Consumers of health care are becoming increasingly aware of outcome measures and patients are actively seeking information about their surgeons and health care institutions performance. Current literature has revealed an inverse relationship between number of cases performed by an individual

1

This study was supported by Grant 30079 from the H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL; Grant R21 CA66553-01 from the National Institutes of Health, Bethesda, MD; the McDonnell Douglas Research Fund, Department of Defense Grant DAMD 17-97-1-7209; and The Joy McCann Culverhouse Surgical Oncology Professorship of The University of South Florida Foundation, Tampa, FL. 2 To whom reprint requests should be addressed at Comprehensive Breast Cancer Program, H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Suite 3157, Tampa, FL 33612. E-mail: [email protected].

0022-4804/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

92

DUPONT ET AL.: BREAST CANCER LYMPHATIC MAPPING

surgeon and mapping failure rate [4, 5]. As LM and sentinel lymph node biopsy (SLNBx) require multidisciplinary cooperation between the departments of nuclear medicine, surgery, and pathology, we thought it necessary to explore the relationship between the number of LM cases performed by an institution and LM failure. METHODS From July 1997 to July 1999, 30 institutions (appendix) participated in a national breast lymphatic mapping trial sponsored by the U.S. Department of Defense and coordinated but not including the H. Lee Moffitt Cancer Center. A purpose for this study was to analyze the institutional influence on breast lymphatic mapping and sentinel lymph node biopsy. All participants underwent a 2-day training course at the H. Lee Moffitt Cancer Center for surgeons, nuclear medicine physicians, and pathologists. The course included a series of didactic lectures, live telemedicine operating room coverage, and hands on training in an animal lab. The participants entered into one of two protocols. Protocol 1: surgeons who had done fewer than 30 cases of breast mapping functioned under a Phase 1 protocol and performed SLNBx followed by complete axillary lymph node dissection. Protocol 2: surgeons who had successfully performed more than 30 cases with one or less false negative results did not perform ALND unless a SLN was positive for metastatic disease. The mapping technique, as it was taught throughout the course, utilizes 450 ␮Ci of filtered (0.2-␮m filter) technetium-99 (Tc 99)labeled sulfur colloid in 6 cc of saline (Syncor International, Tampa, FL). A radiologist or nuclear medicine technician performs the radiocolloid injection approximately 1 to 6 h prior to the operation. State nuclear regulatory mandates require licensed personnel to handle radioactivity for the injection of humans. Blue dye travels much faster through the lymphatic channels. It is injected by the surgeon just 5 min prior to axillary exploration for all breast tumors, regardless of size and location. Approximately 5 cc of isosulfan blue dye (Lymphazurin blue dye, USSC, Norwalk, CN) is injected intraparenchymally at approximately the same location as the radiocolloid. A hand-held gamma detector probe (Navigator System, USSC) is used to locate the SLN and guide the surgery intraoperatively. The SLN is defined as any blue node or node with an ex vivo radioactivity count ratio of the SLN to a neighboring non-SLN of 10:1 andor an in vivo radioactivity count ratio of SLN to background of 3:1. Nodes satisfying these criteria are termed “hot.” Surgeons are also reminded not to abandon clinical judgement. A blue-stained lymphatic channel ending abruptly in a node that is not stained blue or hot, but is clinically positive, is also by definition a SLN particularly when the SLN is chalked full of tumor. This node should be excised and examined by pathology. All sentinel nodes are examined by hematoxylin and eosin staining (H&E) and immunohistochemical staining with cytokeratin (CK). Any patient having positive nodes receives a complete dissection in which a level I and II axillary node dissection is preformed. Previously reported data for the DOD Study reveal the false negative rate of 4% [6]. False negative rate represents the probability of a negative sentinel node sampling among subjects ultimately found to have positive axillary lymph nodes [7]. A mapping failure is defined as the inability to locate a sentinel node by either mapping technique. In the case that a mapping failure occurs, a level I and II axillary dissection is performed, as well. After completing the course, participants were asked to submit data on their lymphatic mapping experience at their institutions via fax or internet. All data were accrued prospectively and maintained in a computerized internet-based database. Patients consented to this IRB approved study. Excluded were pregnant patients, those

93

with multicentric tumors, patients with clinically positive axillary lymph nodes, patients with an allergy to Lymphazurin or sulfur colloid, and those who had previous extensive breast or axillary surgery that may interfere with lymphatic drainage. Each institution was analyzed and an average number of SLNBx cases performed at that institution in a 30-day period defined the institutional volume index (IVI). IVI was measured in cases/month and compared with institutional mapping failure rate over time. Using a logistic regression model, the relationship between IVI and failure rate was plotted. A multivariate analysis was performed to analyze the relationship of other variables to mapping failure.

RESULTS

Mapping failures occurred in all but 7 of the 30 institutions whose data were complete. There were 71 mapping failures among 35 of the 74 participating surgeons over 555 mapping cases. Under Protocols I and II, 467 and 88 cases were accomplished, respectively. The overall failure rate for this study was 12.79% (71/555) with a range of 0 to 40%. False negatives were found in 5 of the 114 cases with positive nodes under the protocol for Phase I. The logistic regression model revealed an inverse relationship between IVI and failure rate; however, there was no statistical trend toward significance (P ⬍ 0.07) between the two variables. Using a multivariate analysis, it was found that the individual surgeon is a more significant factor affecting failure rates than institutional volume (P ⬍ 0.001). The logistic regression model for individual surgeon’s volumes reveals that there is a significant relationship between the surgical volume of an individual surgeon and failure rate. We find that for each increase of 1 case per month in an individual’s surgical volume the expected odds of failure decline by 30%. The overall failure rates for those with surgical volume index (SVI) less than 1 case a month was 19%, between 1 and 2 mapping cases a month 14%, and greater than 2 a month down from 11% (Fig. 1). DISCUSSION

Lymphatic mapping and sentinel lymph node biopsy have been shown to be an effective alternative to complete axillary lymph node dissection (CALND) for the staging of breast cancer patients [3, 8, 9]. Most institutions perform lymphatic mapping with one of three methods: radiolocalization first described by Krag et al. [9], blue dye method first described by Guiliano et al. [10], or the combination of these two methods first described by Albertini et al. [3]. It is clear that with adequate experience, each of the first two techniques can provide accurate identification of the SLN [11]. However, the experience at the H. Lee Moffitt Cancer Center has shown that the combined technique minimizes the learning curve and reduces the false negative rate with adequate multidisciplinary training [4]. The two techniques appear to be complementary in

94

JOURNAL OF SURGICAL RESEARCH: VOL. 97, NO. 1, MAY 1, 2001

FIG. 1.

Logistic regression model of failure rate as a function of individual surgical volume index ( y ⫽ e a⫹bx /(1 ⫹ e a⫹bx )).

that one agent may allow the identification of a SLN that was missed by the other. Thus, it is reasonable to expect that not all SLNs will be located by both blue dye and radiocolloid. However, a SLN identified by either mapping agent should be excised and examined thoroughly by pathology. Therefore, while several controversies continue within the field of breast lymphatic mapping, we have minimized these variables by employing a constant technique and training. Of note, the false negative rate of this study has previously been reported [6]. We utilized cytokeratin analysis of all SLNs as part of a more detailed examination of the SLN to minimize the false negative rate. While mapping failure is unfortunate as it leads to a level I and II axillary node dissection, a false negative mapping is detrimental to the patient who may possibly be left with residual tumor in the nodal basin. We believe that the use of cytokeratin analysis contributes to the low false negative rate. Importantly, this factor is an institutional issue clearly and objectively detected by careful pathology but contingent upon nuclear medicine and the surgeon for proper sentinel node identification. In this evaluation we assess learning and success of the procedure against failure to find the sentinel node. This is an appraisal of the sensitivity and specificity of the mapping and detection method obtained through multidisciplinary cooperation within an institution.

That is, by this analysis consideration is made whether the facility has the multidisciplinary team of nuclear medicine, pathology, surgery, and radiology to consistently perform LM and SLNBx. An exacting standard of training, pathological evaluation, and continuous quality improvement has decreased the false negative SLN biopsy rate to 4% those patients with axillary metastases whose SLN was reported to be negative. Because of the low number of false negatives, no comparable differences could be analyzed between institutions or surgeons with regard to false negative assessment. Our evaluation is of particular importance to both consumers and providers of health care. For the H. Lee Moffitt Cancer Center and other institutions that teach lymphatic mapping and sentinel lymph node biopsy, this study reveals that surgeons, nuclear medicine physicians, and pathologists can be properly trained and in fact perform lymphatic mapping successfully. This is a viable technique even in a community regional hospital setting. While a facility must be properly equipped to perform lymphatic mapping and must have interdisciplinary cooperation, we have demonstrated that the individual surgeon performance is the single most significant factor in institutional mapping success. The current published literature on the topic of SLN mapping in breast cancer has included the

DUPONT ET AL.: BREAST CANCER LYMPHATIC MAPPING

initial experiences of many surgeons. We have previously reported that the odds ratio of mapping failure for a given surgeon decreased by 22% per unit increase in SVI; i.e., for every increase of one case per month performed by that surgeon there was a 22% decrease in probability of their odds of failure [12]. When comparing this current series of surgeons to the previously published series we found no significant difference (P value is 0.49) in the odds of failure for a given change in SVI. This substantiates our claim that surgeonspecific outcomes are the key to success of any given institution and validates our previous observation. In several instances a single surgeon was responsible for all the mapping failures at a given institution. A number of issues may contribute to this. Most center around the surgical technique of the individual surgeon. Attention to detail, patience, and persistence may be the most salient requirements for the technique. The injection technique of radiocolloid and blue dye must be done with great care to avoid the nodal basin especially when the primary site is near the axilla [13]. Injection at the periphery of the tumor or postbiopsy seroma must be achieved. Surgeon experience also becomes important when dealing with radioactive shine through and diffusion of both blue dye and radiocolloid into the surrounding tissues. The placement of the axillary incision is critical. If the incision is made too high in the axilla, the gamma probe and the dissection will be directed toward the primary injection site with shine through becoming more of a problem. If this technical error is made, it is possible as well to miss low axillary nodes, which may be the sentinel nodes. The first few attempts can be frustrating and most surgeons find that, in fact, it initially takes more time to correctly perform the SLN biopsy than complete an axillary lymph node dissection. The procedure should be done in a calm environment and we recommended that the surgeon sit while progressing through the learning phase. Several institutional factors influence the success or failure of lymphatic mapping and SLN yield [14]. The technique used by the nuclear medicine personnel for injecting the radiocolloid directly influences both the ability to image SLNs in the axilla and the surgeon’s success in locating the axillary SLN [4]. Availability of lymphoscintigraphy may have an impact on mapping failure. For patients with inner quadrant lesions lymphoscintigraphy has the potential to identify breast cancer patients with bidirectional drainage or rarely (0.24%) who show no lymphatic drainage into the axilla at all. Anatomic considerations not accounted for may lead to failure. Patient variables such as age and weight may affect outcome. Timing issues, i.e., OR scheduling, nuclear medicine scheduling, often controlled by the institution, are important both to the

95

blue dye and to the radiolabeled materials, as each has its optimum interval of flow migration. Data have recently been reported from the National Cancer Data Base (NCDB) noting that breast cancer survival at 5 years varies according to hospital case volume [15]. We had suspected a similar finding of decreased mapping failures at higher volume centers. However, overall institutional variance was not dramatic. Much has been said recently of the concept of multidisciplinary care for breast patients. Few institutions have actually achieved true multidisciplinary care. This particular technology brings the concept to the forefront and may dramatically change that situation. Lymphatic mapping requires pathology, nuclear medicine, radiology, and surgical coordination and input. The necessity of a multidisciplinary team approach can be either an advantage or a disadvantage depending on the collaborative effort put forth at any given institution. Centers that have established a multidisciplinary approach will be able to incorporate the new technology relatively easily. Those who have not been able to adopt a team approach will find the realization of lymphatic mapping difficult despite of the efforts of an individual surgeon [7, 13, 14]. CONCLUSION

This evaluation reaffirms the importance of outcome reporting in health care. Furthermore, this report demonstrates the value added of simultaneous training experience for all the members of the team. As stated above, the primary role of cytokeratin imunohistochemical analysis by pathologists of lymph nodes was to avoid false negative assessment of the axilla. By incorporating the pathologic assessment within the frame work of the course structure and emphasizing the key objective role of the pathologist in the assessment of the sentinel lymph node the number of false negative nodes detected was minimized. Surgeon-specific outcomes can be reviewed via the internet-based data system to see if they meet community standards or need additional training or mentoring. This methodology provides a framework for continuous quality improvement in surgical training and outcomes. The value of learning and mentoring programs has been reported [4]. Knowledge of a surgeon’s or institution’s failure rate should provide the motivation to continue efforts toward improvement [5, 16]. No longer is the practice of see one, do one, teach one acceptable for learning new techniques. Indeed, this new technology of prospective internet-based data accrual may set a new standard for assessing training and proficiency. For health care consumers, this report emphasizes the importance of the individual surgeon over the institution. The ramifications of volume-outcome associ-

96

JOURNAL OF SURGICAL RESEARCH: VOL. 97, NO. 1, MAY 1, 2001

ation depend on the functional structure of that association. If the likelihood of an adverse outcome continues to decline as caseload increases, then perhaps it is wise to limit care in these instances to a few surgeons and institutions. As well, if there is a threshold effect, such that surgeons who have performed a certain number of cases have a better outcome, then perhaps patients should be directed to those who have attained these threshold outcomes. While outcome and volume are vital issues for the modern medical institution, many have traditionally been unaware of these relationships. We note that while mapping success differences exist across institutions, the disparity is not associated with an institution functioning as a whole but lies with the individual surgeon. Only with objective knowledge of individual outcomes can significant method development and improved efficiency occur.

REFERENCES 1.

Khufi, S. F., Daley, J., Henderson, W., Hur, K., Hossain, M., Soybel, D., Kizer, K. W., et al. Relation of surgical volume to outcome in eight common operations. Ann. Surg. 230, 1999.

2.

Moore, M. J., and Bennett, C. L. The learning curve for laparoscopic cholecystectomy. Am. J. Surg. 170: 55, 1995.

3.

Albertini, J. J., Lyman, G. H., Cox, C., et al. Lymphatic mapping and sentinel node biopsy in the patient with breast cancer. J. Am. Med. Assoc. 276: 1818, 1996.

4.

Bass, S. S., Cox, C. E., and Reintgen, D. S. Learning curves and certification for breast cancer mapping. Surg. Clin. North Am. 8: 497, 1999.

5.

Cox, C. E., Bass, S. S., Boulware, D., Ku, N. N. K., Berman, C., and Reintgen, D. S. Implementation of new surgical technology: Outcome measures for lymphatic mapping of breast carcinoma. Ann. Surg. Oncol. 6: 553, 1999.

6.

Shivers, S. C., Cox, C. E., and Reintgen, D. S. DOD Breast Cancer Research Program. ERA of Hope Meeting, Atlanta, GA, June 2000.

7.

Cox, C. E., Bass, S. S., McCann, C. R., et al. Lymphatic mapping and sentinel lymph node biopsy in patients with breast cancer. Annu. Rev. Med. 51: 525, 2000.

8.

Veronesi, U., Paganeli, G., Galimberti, V., et al. Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph nodes. Lancet 349: 1864, 1997.

9.

Krag, D. N., Weaver, D. L., Alex, J. C., et al. Surgical resection and radiolocalization of the sentinel lymph node in breast cancer using a gamma probe. Surg. Oncol. 2: 335, 1993.

10.

Guiliano, A. E., Kirgan, D. M., Guenther, J. M., et al. Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann. Surg. 220: 391, 1994.

11.

Morrow, M., Rademaker, A. W., Bethke, K. P., Talamonti, M. S., Dawes, L. G., Clauson, J., and Hansen, N. Learning sentinel node biopsy: Results of a prospective randomized trial of two techniques. Surgery 126: 714, 1999.

12.

Cox, C., Bass, S., McCann, C., et al. Learning curves for sentinel lymph node mapping in breast cancer based on surgical volume analysis. Oral presentation at the Society of Surgical Oncology 53 rd Annual Cancer Symposium, New Orleans, Louisiana, March 2000.

13.

Cox, C. E., Haddad, F., and Bass, S. S. Lymphatic mapping in the treatment of breast cancer. Oncology 12: 1283, 1297, 1998.

14.

Cox, C. E., Pendas, S., Cox, J. M., et al. Guidelines for sentinel node biopsy and lymphatic mapping of patients with breast cancer. Ann. Surg. 227: 645, 1998.

15.

Morrow, M., Stewart, A., Sylvester, J., and Bland, K. Hospital volume predicts outcomes in breast cancer: A National Cancer Date Base (NCBD) study. In Program and Abstracts of the American Society of Clinical Oncology 36 th Annual Meeting; New Orleans, Louisiana, May 20 –23, 2000, Abstract 309.

16.

Cox, C. E., Salud, C. J., Cantor, A., Reintgen, D., and Bass, S. S. Ethical considerations in teaching and learning new surgical technologies. In K. K. Hunt, and M. J. Ross (Eds.), Mastery of Lymphatic Mapping. New York: Springer-Verlag, 2000.

APPENDIX Institutional names Baptist Hospital Bayfront–St. Anthony’s Brown/Rhode Island Hospital Cayuga Medical Center Central Baptist Hospital Centre Community Hospital Dekalb Medical Center Easton Hospital Genesis Medical Center Health System Minnesota Huntsville Hospital Kaiser Permanente Medical Group Memorial Medical Center Mercy Hospital of Pittsburgh Morton Plant Mease Oyster Point Surgical Association Phoenix Baptist Hospital St. Johns Regional Medical Center St. Luke Hospital St. Petersburg–Suncoast Memorial UCSF Mount Zion Cancer Center United Hospital University of Louisville University of Virginia UT MD Anderson Cancer Center Vanderbilt University Virginia Mason Medical Center Warren Cancer Research Western Surgical Group York Health System

No. of surgeons No. of cases 2 3 2 2 1 1 4 2 5 1 4 3 1 2 7 2 1 3 2 1 4 4 2 2 4 2 2 1 3 1

12 8 13 5 12 4 22 37 32 23 11 24 7 11 57 4 14 10 22 8 15 6 20 7 45 13 41 17 16 39