Use of primary breast carcinoma characteristics to predict lymph node metastases?reply

Abstract

See also pages 1856–1861 and 1918–1922. We would like to thank Dr. Cady for his kind words of support for our work1 and commend him for his outstanding, thorough, and insightful review of the current axillary lymph node dissection controversy.2 Dr. Cady capsulizes the problems regarding axillary lymph node dissection that oncologists face on a daily basis and makes a strong case for limiting the type and number of patients who must undergo axillary lymph node dissection as part of their treatment for breast carcinoma. At Dr. Cady's suggestion, we analyzed our data by dividing the patients into two groups: those diagnosed from 1979-1987 versus those diagnosed from 1988-1995. As predicted by Dr. Cady, these data (Table 1) revealed an improvement for some categories in the more contemporary group, e.g., a higher percentage of T1a and low nuclear grade lesions, a smaller percentage of T1c lesions, an increase in nonpalpable lesions, and a decrease in the percentage of patients with lymphovascular invasion. Paradoxically, nuclear Grade 3 (high grade) morphology increased and there was a slight increase in the percentage of patients with positive axillary lymph nodes. For those factors that improved, not all the differences were significant, nor were they as large as might be expected. One reason for the small magnitude of these changes lies with the fact that we have been aggressively screening since late 1982 with a single outstanding radiologist, a physician who has been extremely interested in finding early breast carcinoma.3 Hence, the second time period (1988-1995) does not reflect the impact that screening would have made had it been introduced to our group 5 or 6 years later than it was. Invited editorials with author responses are often point-counterpoint discussions, but in this instance, we find ourselves in near complete agreement with Dr. Cady's approach and reasoning. For a procedure that provides little or no measurable benefit in terms of breast carcinoma specific survival, axillary lymph node dissection is performed far too commonly. Our group championed the elimination of axillary lymph node dissection for patients with ductal carcinoma in situ (DCIS) beginning in 1986, when we reported a zero rate of axillary lymph node metastases for the first 100 patients in our DCIS series to undergo axillary lymph node dissection.4-6 Our position was adopted by the National Surgical Adjuvant Breast Project (NSABP) in 1987 when they eliminated the requirement for axillary lymph node dissection for patients entered into protocol B-17, a trial that randomized patients with DCIS into two groups: excision alone or excision with postoperative radiation therapy. Unfortunately, due to surgical preference, 38% of more than 800 patients in that trial underwent an axillary lymph node dissection; all were negative.7 From 1993 through 1996, our group continued to suggest that the threshold for axillary lymph node dissection be limited further by eliminating the procedure for patients with DCIS with microinvasion,8 for patients with T1a tumors,9 and finally for patients with T1b nonpalpable tumors.10 With the publication of this current article,1 we have tried to fine-tune the decision-making process even further by correlating and quantifying specific tumor characteristics with the probability of lymph node involvement, thereby making the axillary lymph node dissection selection process clearer. However, as we move into the next century, we believe that deciding which patients merit an axillary lymph node dissection may become a moot point. In 1996, our group began performing sentinel lymph node biopsy on all patients with primary operable invasive breast carcinoma and all patients with DCIS undergoing mastectomy. We first used a vital blue dye and then a radioactive tracer in combination with the dye. We found this latter procedure to be extremely accurate, allowing removal of the one or two sentinel axillary lymph nodes through a small and precise axillary incision. By using both modalities simultaneously, the surgeon is able to recognize the sentinel lymph node using both visual and auditory senses. In every case that we have studied to date, there were no instances of positive nonsentinel axillary lymph nodes in the face of a negative sentinel lymph nodes. There were of course positive sentinel lymph nodes combined with negative nonsentinel axillary lymph nodes. The major sentinel lymph node studies reported to date confirm a near-perfect correlation with the lymph node status of the dissected axilla.11, 12 In other words, the status of the sentinel lymph node accurately predicts the status of the axilla. We would recommend that every surgeon involved in the treatment of patients with breast carcinoma learn the techniques of sentinel lymph node biopsy. Vital blue dye is sometimes a bit messy and takes somewhat longer to learn but it works well and is inexpensive. A radiolabeled isotope is extremely precise but requires an initial investment of approximately $20,000 for a probe capable of detecting tiny amounts of tracer in regional lymph nodes and a staff with the capability to handle radionuclides. Thanks to the efforts of Giuliano et al.13 and Krag et al.14 we have entered a new era. However, in addition to its benefits, this new technology introduces yet another therapeutic dilemma to the complex breast carcinoma treatment selection process: micrometastases (metastases ≤2 mm in greatest dimension). Some micrometastases detected by immunohistochemistry are only a few cells in greatest dimension, which is far smaller than 2 mm. With normal histopathologic hematoxylin and eosin (H & E) processing, a certain percentage of axillary lymph node dissections are found to contain metastases depending on numerous tumor factors (e.g., tumor size, nuclear grade, the presence of lymphatic or vascular invasion, tumor palpability, etc.). If the pathologist is specifically told "This is the sentinel lymph node. Please give it special attention," the lymph positivity rises by approximately 5-10%.11 If the sentinel lymph node is then processed immunohistochemically for cytokeratin, lymph node positivity rises yet another 5-10%.11-14 If more sensitive techniques, such as polymerase chain reaction (PCR) to a variety of tumor antigens are used, an even higher rate of lymph node positivity can be expected.15, 16 Therefore, when Dr. Cady talks about reducing lymph node positivity to 20% overall using standard H & E histopathology, rapidly developing sentinel lymph node technology could easily reverse the trend, taking overall lymph node positivity back above 50% once again due to the discovery of previously undetectable (by H & E) micrometastases. Should a lymph node that is negative by H & E but positive by immunochemistry be coded as N0(H&E) N1(Imm)? Could we go even further with special notations for immunochemically negative but PCR positive lymph nodes, e.g., N0(H&E) N0(Imm) N1(PCR)? What does this mean? How should we handle these patients? Will these data have any clinical use? Does a patient with a small primary tumor with a few breast carcinoma cells per 10,000,000 lymphocytes in a lymph node require chemotherapy? Are these tumors cells a marker of systemic metastatic disease or are these "involved" lymph nodes simply doing their job of eliminating a few circulating carcinoma cells with little metastatic potential? We would like to finish by sharing the clinical details of a patient whom we recently treated. In late 1996, a 49-year-old female presented to the Breast Center in Van Nuys having previously undergone a wire-directed excisional biopsy at another facility for a multifocal DCIS. The lesion measured 29 mm in greatest dimension, margins were focally involved, comedonecrosis was present, and the nuclear grade was high (Grade 3). The excision specimen measured 48 mm in maximum dimension and 10 slides had been prepared, approximately 1 slide for every 5 mm (a smaller number of slides than we would have prepared but certainly adequate by community standards). All slides were reviewed by the Breast Center pathologist, and no evidence of invasion was found. The patient had a relatively small breast and rather than attempt to obtain clear margins at reexcision, she chose to undergo a skin-sparing mastectomy with immediate autologous tissue reconstruction with a free transverse abdominus rectus myocutaneous flap. In the operating room, just prior to mastectomy, 4-5 mL of isosulfan blue dye was injected around the biopsy cavity. During the mastectomy, a single blue sentinel lymph node was identified and removed; four additional lymph nodes near it were also removed. H & E histology yielded five negative lymph nodes. A small amount of residual DCIS was found in the mastectomy specimen near the previous biopsy cavity. With the report of negative lymph nodes by H & E, we followed our protocol and submitted the sentinel (blue) lymph node for immunohistochemistry. This demonstrated a lymph node with three tiny cytokeratin and epithelial membrane antigen positive clusters, each with only two to four abnormal cells. An independent immunohistochemistry expert concurred with the interpretation-this patient with DCIS, without any evidence of invasion by light microscopy whatsoever, had a micrometastasis to the sentinel lymph node. What do we do now?