Abstract
Background/Aim: Ductal carcinoma in situ (DCIS) is a non-invasive malignant breast lesion. Patients diagnosed with a DCIS on percutaneous biopsy usually undergo resection, and the final pathology may reveal that the lesion was in fact invasive (upgrading at surgery), this leading to treatment strategy change during its course. The aim of the present study was to identify factors associated with DCIS-upgrading to invasive carcinoma at surgery, and to identify a subgroup of patients more likely to have an invasive cancer. Patients and Methods: A retrospective study was performed in patients diagnosed with DCIS on percutaneous biopsy between April 1997 and December 2010. Based on available data and on previous studies, 21 clinical, radiological and pathological variables were evaluated using univariate analyses. Variables identified in univariate analyses, when p≤0.10, were included in a multivariate model. Results: Among 608 DCIS lesions, 177 (29.1%) were invasive carcinomas after surgery. Using univariate analyses, core needle biopsy (odds ratio (OR)=1.8), physical symptoms (OR=2.9), palpable masses (OR=4.1), number of specimen obtained (1-9 cores, OR=2.2) and a measurable mammographic lesion (OR=1.7) were significantly associated with upgrading at surgery. However, using multivariate analysis, no factor was significantly associated. Conclusion: No characteristic was identified to be independently associated with DCIS upgrading at surgery, and no sub-group of patients could be identified in whom the appropriate surgery could have been performed first.
Ductal carcinoma in situ (DCIS) of the breast is a malignant breast lesion that is confined within the breast's ductal network, without any stromal invasion (1). The widespread use of screening mammography has led to a subsequent increased detection of asymptomatic micro-calcifications needing investigation (2). Percutaneous needle biopsy is the best modality allowing for a minimally-invasive approach to diagnose these mammographic abnormalities (3-8). However, DCIS diagnosis using percutaneous needle biopsy is challenging, since the needle samples only a part of the lesion, and may miss small foci of invasion (3-8). Therefore, an invasive carcinoma might be diagnosed at surgery, even if the initial needle biopsy showed DCIS as the most offensive lesion: this unfortunate event is known as “upgrading at surgery”. Indeed, in a number of published series of DCIS lesions diagnosed using needle biopsy, upgrading occurs in 2-49% of cases (Table I).
The main consequence of upgrading from DCIS to an invasive breast carcinoma is a change in the treatment strategy during its course. Indeed, DCIS is usually treated using breast surgery with or without radiation therapy, while invasive carcinoma is usually treated using breast surgery, lymph node sampling (sentinel node biopsy with or without axillary dissection), radiation therapy, chemotherapy, and adjuvant hormonal therapy (9). However, in cases of upgrading, a second surgery must be undertaken to complete the treatment with lymph node sampling (9), exposing patients to surgery for a second time and anesthesia-related risks (10), as well as in flicting a greater psychological impact on patients to whom the physician first said that they had an easily treatable cancer.
A number of studies have focused on the identification of factors associated with DCIS upgrading at surgery (Table I). These studies were nearly all carried-out in limited patient samples and, most importantly, the results are inconsistent between studies. The Deschênes-Fabia Centre for Breast Diseases (Quebec City, Canada) maintains a prospective database of all needle biopsies performed since April 1997, including patients' clinical, radiological and pathological characteristics. The objective of the present study was to determine our upgrading rate of DCIS at surgery, and to identify factors associated with this upgrading. The identification of these factors might help prevent a second surgery in some women through identification of these with a high likelihood of harboring an invasive carcinoma and providing the adequate surgical treatment from the beginning.
Materials and Methods
Study population. At the Deschênes-Fabia Centre for Breast Diseases, pathology results are prospectively compiled in a database, under pathologist supervision, since the beginning of percutaneous image-guided breast biopsies in 1997. A retrospective study of patients who attended the Centre from April 1997 to December 2010 was performed. During this period, radiologists performed 21,340 consecutive percutaneous breast biopsies.
Percutaneous biopsy methods and analyses. The technique used for needle biopsy depends on the lesion's nature. Needle biopsy is preferably performed under ultrasound guidance using 14G core needle biopsy (CNB) or under stereotactic guidance if non-visible by ultrasound. Prior to 2000, stereotactic-guided biopsies were done using 14G core needles (InterV MD Tech, Gainesville, FL, USA). Vacuum-assisted biopsy (VAB) (Mammotome® Breast biopsy system, Ethicon Endo-Surgery, Cincinnati, OH, USA, using 11G needles between 2000 and 2007 and 8G needles from 2007, and SenoRx, Bard Biopsy Systems, Tempe, AR, USA, using 7G needles from 2006) almost became the exclusive needle biopsy technique by 2002.
Needle biopsy specimen obtained for mammographic calcifications are systematically radiographed to confirm the presence of the targeted microcalcifications. The slides containing needle biopsy tissues are stained using standard hematoxylin & eosin (H&E) staining. A minimum of three H&E levels are made from each of these blocks with micro calcifications. The final diagnosis of DCIS is performed using criteria described by the World Health Organization (11, 12). It is important to note that small high-grade lesions are diagnosed as DCIS, even if they encompass only one duct or are less than 2mm.
Selection of biopsies. All patients diagnosed with DCIS as the most advanced lesion on percutaneous biopsy were identified using the Center's biopsy database. Reports were retrieved and reviewed. Patients were excluded if the biopsy showed signs of invasion or micro-invasion. From April 1997 to December 2010, DCIS without any associated invasion was diagnosed in 1,212 out of 21,340 breast biopsies (5.7%). A total of 608 biopsies (out of 1,212) from 604 patients were included in the final analysis. We excluded 604 biopsies for the following reasons: non-conventional breast biopsy (1 or 2 cm diameter biopsy device used in 20 patients); history of DCIS, invasive breast cancer or primary cancer from any other site (230 patients); biopsy or surgery performed in another hospital (254 patients); men (2 patients); and suspicion of invasiveness or micro-invasion on biopsy (98 patients).
Data collection. After Ethical Review Board approval, data from the pathology database, from the Center for Breast Diseases database and from hospital records were reviewed. Several potential variables were collected, including age at-diagnosis, first-degree familial history of breast cancer, indication for mammography, breast symptoms (nipple discharge, pain, itching), mammographic characteristics (microcalcifications, nodule, density, distortion), instrument used for biopsy (CNB vs. VAB), associated diagnosis of papilloma, and the pathological characteristics of the DCIS to name some few. The choice of all variables was made according to factors suggested to influence upgrading at surgery in the published literature (Table I) and according to available data from our databases. Previous hormonal exposition was not analyzed in the present study because of inconsistent and insufficient reporting of these variables.
Statistical analysis. Descriptive statistics were used to present characteristics of the population. A generalized estimating equations (GEE) approach was used in the logistic models to take into account correlations between repeated measures on the same individual (concerning four patients who were biopsied twice) and to evaluate odds ratio (OR) of DCIS upgrading at surgery in relation to selected factors. Variables identified at univariate analyses to be associated with DCIS upgrading at surgery with a p≤0.10 were included in multivariable analyses to account for potential confounders. All analyses were performed using the SAS 9.3 software (SAS Institute, Inc., Cary, NC, USA), where two-sided, and a nominal p-value of 0.05 was considered statistically significant.
Results
Literature review. Table I presents the published series on upgrading following a DCIS diagnosis by percutaneous biopsy. To be included in this Table, studies had to have >50 cases, had to have been published after January 1st, 2000, and had to assess factors associated with upgrading.
Clinical and radiological characteristics. Table II presents the clinical and radiological characteristics of the 608 breast biopsies performed in 604 women. Most women diagnosed with DCIS on percutaneous biopsies were 50-59 (43.9%) or 60-69 years old (29.0%), and had no 1st degree family history of breast cancer (53.5%). More than half were diagnosed in a screening context (61.7%). Most lesions were microcalcifications-alone (88.5%), and only 22.7% of lesions had a measurable size on mammography. Lesions were biopsied using CNB in 25.3% of cases, and using VAB in 74.7%. The number of cores obtained was highly variable. Most women had no physical symptoms (82.5%). Among those who had physical signs (n=106), a palpable mass was the most common (64/106, 60.4%).
Histological characteristics. Table III shows the histological characteristics of 608 breast biopsies performed in 604 women. Most micro-calcifications were observed in the biopsy sample (91.1%). Most lesions displayed a nuclear grade of II (54.8%) or III (35.0%). Necrosis was present in 73.7%. Concerning architecture, 23.2% of lesions displayed a micropapillary, 23.7% a cribriform, 2.8% a papillary, 15.0% a comedocarcinoma and 59.7% a solid architecture. Some lesions presented more than one architecture. Seventy-seven percent of lesions were positive for estrogen receptors and 53% for progesterone receptor.
Among the 608 lesions, 177 were upgraded to invasive carcinoma at surgery, for an upgrading rate of 29.1% (Table III).
Univariate analysis. Table IV shows the univariate associations between upgrading at surgery and clinical and radiological characteristics in 608 breast biopsies in 604 women. A measurable lesion on mammography was associated with an increased risk of upgrading (OR=1.67, 95%CI: 1.11-2.51, p=0.01), the use of CNB (OR=1.80, 95%CI: 1.22-2.65, p=0.003), sampling of 1-9 cores (OR=2.15, 95%CI: 1.27-3.64, p=0.009), the presence of a physical symptom (OR=2.91, 95%CI: 1.89-4.48, p<0.0001), the presence of a physical sign (OR=2.17, 95%CI: 0.98-4.79, p=0.05), and the presence of a palpable mass (OR=4.08, 95%CI: 2.39-6.97, p<0.0001). All remaining characteristics were not significantly associated with upgrading.
Table V shows the univariate associations between upgrade at surgery and histological characteristics in 608 breast biopsies in 604 women. The presence of necrosis in the biopsy was associated with a lower upgrade rate (OR=0.64, 95%CI: 0.43-0.93, p=0.02), while the presence of a cribriform architecture had a tendency towards a lower upgrade rate (OR=0.75, 95%CI: 0.53-1.06, p=0.10). All remaining characteristics were not significantly associated with upgrade.
Multivariate analysis. All characteristics displaying an univariate association with upgrade with a p-value ≤0.10 were included in a multivariate model. Thus, in the model we included: type of biopsy, number of sampled cores, physical symptoms, physical signs, palpable mass, lesion size on mammography, necrosis and cribriform architecture (Table VI). Among all these characteristics, only the presence of a palpable mass (OR=2.11, 95%CI: 0.99-4.47, p=0.05) and the presence of a cribriform architecture (OR=0.72, 95%CI: 0.49-1.06, p=0.09) had a tendency to associate with upgrade at surgery. All remaining characteristics were not significantly associated with upgrade on multivariate analyses.
Discussion
Identification of invasion in a malignant breast disease is crucial, since the surgical axillary approach will be different, and since the axillary status is one of the most important prognosis factors in breast cancer (9). Thus, the aim of the current study was to assess the upgrade rate at surgery in a tertiary breast cancer care Center, and to identify factors associated with the presence of invasion in percutaneous needle biopsies diagnosed with DCIS only. Results from the present study revealed an upgrading rate of 29.1%. Univariate analyses suggest that the use of CNB, a small number of sampled cores, the presence of physical symptoms, the presence of physical signs, a palpable mass, a measurable lesion on mammography, absence of necrosis and absence of a cribriform architecture were associated with a higher risk of upgrade. However, using multivariate analyses, no characteristic was identified to be statistically associated with the presence of invasiveness in women diagnosed with DCIS on percutaneous needle biopsy. The percutaneous image-guided needle biopsy approach is less traumatic for the patient and provides good results, but is associated with some disadvantages. Indeed, the diagnosis of many lesions is based on the size and/or on the invasiveness of the lesion, and these factors might be difficult to estimate using limited sampling (3-8). Furthermore, DCIS and invasive carcinoma are frequently concomitant, and the needle may simply miss the invasive region. Thus, previous studies report an upgrading rate ranging from 2%-49% (Table I). However, these previous studies all reported a variety of different factors associated with upgrade at surgery and, taken together, they mostly failed to identify common characteristics. Furthermore, some studies did not identify any factor associated with upgrading at surgery (12-15). Nevertheless, a few characteristics were more common across studies than others.
A measurable (mass image) lesion on mammography, compared with microcalcifications alone, has been associated with an increased risk of upgrading in a number of studies and is suggestive of a more advanced lesion (16-23). The use of CNB has been demonstrated to sample a lesser amount of lesion than vacuum-assisted techniques. Since fewer tissues are sampled, the risk of missing an invasive focus is increased (16, 24-26). In the same way, sampling a smaller number of cores also increased the risk of upgrading at surgery (27, 28).
Since DCIS is, by definition, confined within the mammary ducts without invasion of the stroma, it usually does not form a palpable mass in the same way as an invasive tumor does (1). Thus, the presence of a palpable mass diagnosed as DCIS on percutaneous biopsy might be indicative of a more advanced disease. Indeed, a number of studies reported that a palpable mass concomitant with a DCIS diagnosis on percutaneous needle biopsy was associated with upgrading at surgery (16, 18, 20, 22, 29, 30). In the present study, even if not significant, the presence of a palpable mass had a tendency to be associated with DCIS upgrading at surgery.
DCIS is a member of a continuum starting with atypical ductal hyperplasia, going through DCIS, to end with invasive ductal carcinoma (31). Thus, a number of histopatological factors observed on DCIS might be associated with a higher risk of upgrading. Indeed, some studies identified a high DCIS grade as a risk factor for upgrading (16, 20, 24, 32-35). However, there is a controversy regarding the DCIS architecture associated with a higher risk of upgrading. Indeed, comedonecrosis has been associated with upgrading (24, 36-38), as well as the solid type (18). Beside these, a number of other pathological risk factors were observed in one or a few studies, preventing any conclusion on the matter (Table I). In the present study, presence of necrosis was significantly associated with lower upgrade rate, which is in contradiction with previous studies. Indeed, necrosis is usually indicative of hypoxia within a larger tumor, and is usually associated with a more advanced disease. For this reason, this statistical result must be validated with caution.
However, when all factors identified by univariate analyses were analyzed together in a multivariate model, no factor remained statistically associated with upgrading at surgery. The same conclusion was reached by four studies (12-15) in smaller numbers of patients. Furthermore, when considering all studies that identified at least one factor associated with upgrading (Table I), no factor is common to all studies.
The present study may suffer some limitations. Ours was a retrospective study in a clinical setting, and we had to work with the data available. A prospective study would allow us to assess characteristics that are not routinely assessed in clinical practice. On the other hand, we assessed characteristics that are more likely to be evaluated in any breast cancer care Center.
In conclusion, the present study was unable to identify a sub-group of patients diagnosed with a DCIS on percutaneous image-guided breast biopsy in whom an invasive cancer is more likely to be present. Thus, patients with DCIS-upgrading after surgery will still have to undergo a second surgery.
Acknowledgements
This work was funded by the Enfant-Jésus-St-Sacrement Hospitals Foundation. CD is a Junior Investigator of the Canadian Cancer Society (2011-700657).
Footnotes
-
↵* These Authors contributed equally to the study.
-
Conflicts of Interest
The Authors declare that there are no competing financial interests.
- Received December 16, 2013.
- Revision received January 13, 2014.
- Accepted January 14, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved