| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
EDITORIAL |
From St. Mary’s Medical Center (MDL), San Francisco, and USC/Norris Cancer Center (MJS) Los Angeles, California.
Correspondence: Address correspondence to: Michael D. Lagios, MD, St. Mary’s Medical Center, 450 Stanyan Street, San Francisco, California 94117; Fax: 415-750-4833.
The natural history of DCIS, a disease with a breast cancer-specific mortality rate of only 1% after mastectomy, suggests that DCIS is exactly what we think it is: a non-obligate precursor of invasive breast cancer with little or no metastatic potential while it is in the in situ phase. Why pursue immunohistochemically detected cytokeratin positive material in patients with such minimal disease?
For invasive breast cancer, sentinel lymph node biopsy technology has numerous advantages over standard level I and II axillary node dissection, a procedure which is rapidly being replaced by this newer method. As part of its evolution, the sentinel lymph node technique became wedded to immunohistochemical methods, a technology which can identify a single epithelial cell in an otherwise negative sentinel lymph node. While axillary metastases identified by hematoxylin and eosin (H&E) staining in patients with invasive breast cancer have prognostic significance supported by decades of outcome data, the prognostic significance of immunohistochemically identifiable axillary metastases, often represented by fewer than 100 cells in a sentinel node, is less certain.
Among patients with invasive breast carcinoma, the percentage of positive sentinel lymph nodes by virtue of immunohistochemical techniques is significantly greater than that reported by conventional H&E staining.
Dowlatshahi, et al.,1 reported that 52% of H&E node-negative T1a and T1b carcinomas were positive by immunohistochemical techniques when the sentinel nodes were methodically examined by a serial sectioning technique which, incidentally, yielded no additional node positivity by H&E. In an update of the Ludvig trial, Cote et al.,2 in examining node-negative patients, were also able to significantly increase the yield by immunohistochemical techniques, yet neither immunohistochemical or H&E identifiable "occult" metastases were significant for disease-free survival or overall survival in premenopausal patients. More recently, Carter et al.,3 have shown that prior operative procedures can artifactually displace both benign epithelium and epithelium from noninvasive carcinoma into lymphatics, resulting in a positive sentinel lymph node. In light of this, the propriety of sentinel lymph node technology applied to patients with DCIS is even more problematic because nearly all will have had a prior diagnostic biopsy (image-directed or open) which can potentially dislodge and transplant cytokeratin-positive cells into the sentinel lymph nodes.
Data derived from the time during which a level I and II axillary dissection was routinely performed for patients with DCIS show nodal positivity to be < 1%.4 It is highly unlikely that sentinel lymph node technology will increase this 10 to 15-fold. It is far more likely that an increase of this magnitude is iatrogenic.
The purpose of node dissection in breast carcinoma is largely to establish prognostic groups for which a specific therapeutic intervention is appropriate, for example, chemotherapy. Sentinel lymph node biopsy in patients with DCIS will lead to overtreatment. We have repeatedly had the experience of consulting with patients with DCIS in whom a positive sentinel lymph node was followed by a recommendation for aggressive chemotherapy, the sentinel lymph node biopsy having incorrectly upstaged the patient from Stage 0 to Stage 2A. In that unhappy example, speculative ideas about the possible significance of an immunohistochemical-positive sentinel lymph node in a DCIS patient were compounded by an even more speculative thesis that chemotherapy is appropriate and beneficial for such patients. One only has to look at the natural history of DCIS to see the folly of that recommendation.
The article by Klauber-DeMore et al.,5 assumes identical clinical significance for immunohistochemical and H&E identifiable metastases, which is an unproven thesis. It would have been helpful to know what is being described as an immunohistochemical "micrometastasis." There is a difference, alluded to in the works of Cote et al.2 and Braun et al.,6 between micrometastases detected by immunohistochemistry of fewer than 100 scattered positive cells and a cohesive colony within a sentinel lymph node of 1000 cells or more. Both could be described as "micrometastases."
The assertion that sentinel lymph node technique may be a good screen for occult invasion in breasts with unresected or "unsampled" DCIS is at odds with pathologic experience. In fact, invasive foci are far more commonly detectable by standard H&E examination if it is thorough. As many as 48% of patients with DCIS with an extent > 5.5 cm may exhibit identifiable microinvasion at mastectomy. The majority of such occult invasive foci will lie within the T1mic to T1b size groups and almost all are node-negative by conventional H&E staining.7
Thorough examination of the tissue in DCIS is mandatory. Although the authors processed all of the "tumor" when it measured 2 cm or less in size, we are unclear about what type of tissue sampling occurred for the many "high-risk" patients who had more extensive disease; six of eight DCIS patients with a positive sentinel lymph node had "mammographic" sizes of > 2 cm and three of eight had sizes of > 5.5 cm.
The importance of careful review of the pathology is exemplified by the authors own retrospective review. Five of the combined sentinel lymph node-positive group of DCIS and DCIS with microinvasion cases had evidence of more extensive invasive disease on retrospective review. Three of nine sentinel lymph node-positive DCIS cases were invasive on review (1 T1mic and 2 lymphatic invasion only). An additional sentinel lymph node-positive DCIS case had a contralateral invasive carcinoma with a positive sentinel node. Recalculating the DCIS sentinel lymph node-positive rate after deletion of these four cases reduces the rate of sentinel lymph node positivity in DCIS patients from 12% to 6.5%. Additionally two of three DCIS microinvasive cases (foci of invasion 
1 mm) were found to be T1a on review and one exhibited lymphatic invasion as well. Of the 76 cases of pure DCIS, 24% were entered into the sentinel lymph node program because the pathology findings were "suspicious but not diagnostic" of microinvasion. We presume that those were reviewed and found to be noninvasive, otherwise the 6.5% true-positive sentinel lymph node rate might be reduced even further.
The authors suggest that the 1.5% of patients who died of metastatic breast cancer reported in the National Surgical Adjuvant Breast and Bowel (NSABP) Protocol B178 reflect the natural history of DCIS. On the contrary, these 1.5% developed metastatic disease as a first event without an initial or synchronous in-breast recurrence. None of almost 900 patients in the Van Nuys database have exhibited a similar distant first event, although certainly a small fraction developed metastatic disease after local recurrence.9 In like fashion, Solin et al.10 have recently presented an update of the collaborative trial in which only one of 270 patients exhibited a distant first event at 15 years of followup.11 The high metastatic rate in NSABP Protocol B17 and in the more recently published European Organization for Research and Treatment of Cancer (EORTC) DCIS trial12 reflect inadequate pathologic examination resulting in undetected areas of invasion in the tissue resected or left unsampled in the breast. This is reflected in the patients who develop metastatic disease within 1 year of treatment as reported in NSABP Protocol B17.8 This is decidedly not the natural history of DCIS!
Sentinel lymph node techniques will not be useful in predicting which patients will develop invasive recurrence. As Anderson noted in his editorial,13 many attempts to define a preinvasive subgroup in DCIS have been made and have failed. The best way to avoid an invasive recurrence is to adequately resect the initial DCIS with a 1 cm margin.14 In nearly all studies of breast conservation therapy for DCIS, 50% of local recurrences are invasive and the remainder are in situ disease. Therefore, anything that reduces the local recurrence rate will reduce subsequent invasive events and therefore will decrease the potential of distant disease. Sentinel lymph node technique will not impact this because it does not reflect the status of margins. Patients with DCIS will be better served by more thorough pathologic evaluation and more careful surgical attention to margins than by the questionable value of a sentinel lymph node procedure, particularly one that employs immuno-histochemistry.
Received for publication February 20, 2001. Accepted for publication February 26, 2001.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  F. J. Dominguez, M. Golshan, D. M. Black, K. S. Hughes, M. A. Gadd, R. Christian, B.-A. Lesnikoski, M. Specht, J. Michaelson, and B. L. Smith Sentinel Node Biopsy is Important in Mastectomy for Ductal Carcinoma In Situ Ann. Surg. Oncol., January 1, 2008; 15(1): 268 - 273. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. S. Cody III Sentinel Lymph Node Biopsy for DCIS: Are We Approaching Consensus? Ann. Surg. Oncol., August 1, 2007; 14(8): 2179 - 2181. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Intra, N. Rotmensz, G. Viale, and U. Veronesi Re: Trends in the Treatment of Ductal Carcinoma In Situ of the Breast J Natl Cancer Inst, July 21, 2004; 96(14): 1110 - 1111. [Full Text] [PDF]
|
|
|
|
 |
|
 P. C. Stomper, J. Geradts, S. B. Edge, and E. G. Levine Mammographic Predictors of the Presence and Size of Invasive Carcinomas Associated With Malignant Microcalcification Lesions Without a Mass Am. J. Roentgenol., December 1, 2003; 181(6): 1679 - 1684. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. R. McCready Quality Assurance: The Value of Data and the Will to Improve Ann. Surg. Oncol., October 1, 2003; 10(8): 837 - 838. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Intra, P. Veronesi, G. Mazzarol, V. Galimberti, A. Luini, V. Sacchini, G. Trifiro, O. Gentilini, G. Pruneri, P. Naninato, et al. Axillary Sentinel Lymph Node Biopsy in Patients With Pure Ductal Carcinoma In Situ of the Breast Arch Surg, March 1, 2003; 138(3): 309 - 313. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Bonnema and C. J. H. van de Velde Sentinel lymph node biopsy in breast cancer Ann. Onc., October 1, 2002; 13(10): 1531 - 1537. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. S. Cody III, N. Klauber-DeMore, P. I. Borgen, and K. J. Van Zee Is It Really Duct Carcinoma In Situ? Ann. Surg. Oncol., September 1, 2001; 8(8): 617 - 619. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
