| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
EDITORIALS |
From the Department of Surgical Oncology, Princess Margaret Hospital, University of Toronto, Toronto, Ontario, Canada.
Correspondence: Address correspondence to: David R. McCready, MD, Department of Surgical Oncology, Princess Margaret Hospital, University of Toronto, 610 University Ave., Suite 3-130, Toronto, Ontario, M5G2M9, Canada; Fax: 416-946-6590; E-mail: david.mccready{at}uhn.on.ca
The article by Hoorntje et al.1 in the August issue of the Annals of Surgical Oncology contains most of the ingredients for a successful quality improvement project. In this pithy manuscript, a problem is identified, the data reviewed, results tabulated, and possible explanations discussed. To finish the project, further recommendations should be made and a repeat evaluation undertaken.
The problem discussed is clearly clinically relevant: in some cases a core breast biopsy, by its very nature as a sampling procedure, can underestimate the exact nature or amount of disease present within a patient’s breast. The specific issue addressed is the situation when the core biopsy reports duct carcinoma in situ (DCIS), but invasive breast cancer is found at the time of surgical excision. This underestimation leads to initial under treatment because regional nodal assessment by sentinel node biopsy or axillary dissection is not generally performed for pure DCIS2 and, thus, a second procedure will be recommended once the diagnosis of invasion becomes evident.
As in all good quality improvement projects, once the problem is defined, the literature is searched. Many reports have suggested that taking more passes with the 14-gauge core needle3 or excising bigger specimens using larger vacuum-assisted devices would reduce the histologic underestimation and, to a large degree, this is true.4,5 Sometimes, however, the calcifications are too spread out and, on occasion, all sampling procedures can misrepresent the whole.
To help define the causes of underestimation in their hands and to help other practitioners in the same situation, the authors began an evaluative process based on their database of core biopsies. The value of accurate data registration cannot be over emphasized. The reasons for their underestimation are described, but which ones can be corrected? It should be noted that the underestimated invasive disease was not just in areas of microinvasion. Significant cancers were present; the average size was 8.3 mm and 20% (8/41) had nodal metastases with a median of seven nodes positive.
The authors classified the reasons for underestimation: (1) mainly radiologic, (2) mainly histopathologic, and (3) histologic disagreements. Perhaps it might be more appropriate to reclassify the causes of underestimation into either (1) potentially correctable or (2) inherent to a sampling process. The four cases with widespread calcifications and no density, or the three with coincidental invasive cancer could be accepted as a problem inherent with a sampling procedure. The remaining 34 (83%) could possibly have been correctly diagnosed with increased attention to thorough breast imaging, better biopsy targeting, larger sample procurement, proper tissue handling, and expert interpretation.
What can be learned from this paper and others like it? First, women who present with a suspicious breast lesion should have a thorough work-up. The current algorithm should involve a thorough mammographic evaluation of the lesion with magnification or compression views, as necessary, and physician-directed ultrasound in most, if not all, cases. If a concerning mass is seen on ultrasound, then a core biopsy is directed at the mass, because invasive disease will more likely be found there. Women with dense breast tissue containing clustered, worrisome calcifications can be further evaluated by magnetic resonance imaging. Pure mammographically concerning calcifications without a mass will have a stereotactic core biopsy. Consideration should be given to using an 11-gauge vacuum-assisted device and a marker clip deposited into the biopsy cavity.4 It has been suggested that a minimum of 10 cores be obtained for calcifications if the 14-guage device is used.6
Second, the tissue obtained at biopsy should be gently handled, evaluated by mammography, labeled, and transported to pathology. Artifacts caused by crushing or drying can be avoided with appropriate care. The radiologic abnormality and other relevant clinical information should be detailed for the pathologist. We do not advocate frozen section of lesions consisting of only calcifications because some crucial tissue can be lost, even in the best hands.
Third, a surgical, pathologic, and radiologic conference regarding the cases that are not obvious will optimize diagnostic accuracy. Second opinions or collegial confirmation on difficult-to-interpret cases should be the norm, not the exception. We have found that immunohistochemistry using antismooth muscle antibodies (smooth muscle actin, smooth muscle myosin heavy chain), antibasement membrane antibody (collagen IV), and antikeratin antibodies for epithelial cells can help interpret or confirm the diagnosis of stromal invasion in some difficult cases. The absence or presence of a myoepithelial layer around cytologically atypical cells can help distinguish between duct carcinoma in situ or invasive disease.
Although Hoorntje et al. found a 70% rate of subsequent invasive cancer in those poorly differentiated DCIS lesions with a periductal inflammatory response, this should be considered "hypothesis generating" data. For the patient with localized DCIS on core biopsy and the desire for breast preservation, the usual surgical practice should still be wide excision of only the index lesion.
Sentinel node biopsy for pure localized DCIS does not add clinically evaluable information.2 In those who are ultimately diagnosed with invasive disease following a lumpectomy, sentinel node biopsy is still a viable option because previous excision does not reduce lymph node biopsy accuracy.7 For most patients, it is preferable to have two "indicated" procedures than one "unnecessary" operation.
An exception to this plan: If a mastectomy is planned for patients with large areas of disease and only DCIS was diagnosed with core biopsy, it might be prudent to plan a concurrent sentinel lymph node biopsy. The reasons include the fact that large areas of DCIS are more likely to contain invasive disease8 and the sentinel node procedure cannot be reliably performed after mastectomy. Then, if small areas of invasive cancer are diagnosed at final pathology and the sentinel lymph node is negative, an axillary dissection can likely be avoided. The optimal method and injectant for the sentinel node biopsy in this situation still needs to be defined, but the subareolar site9 is appealing.
Finally, we must acknowledge that no sampling procedure will always accurately assess the true state. To complete the quality improvement loop, these recommendations should be applied and the new results checked. It is hoped that the incidence of DCIS underestimation will have fallen and patient care and satisfaction will be measurably improved.
Received for publication August 19, 2003. Accepted for publication August 22, 2003.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
