Archive for the Breast Cancer category.
Resection Margins
Although the importance of obtaining negative margins is intuitively evident, earlier studies of patients undergoing breast-conserving surgery with radiation therapy were not able to demonstrate that margin status was statistically significant, probably because standardized methods for accurately assessing histologic margins did not exist until the late 1980s and early 1990s. More recent research clearly shows the effect of margin status on local recurrence. In a large retrospective review with a mean follow-up of 70 months, Silverstein stratified patients by margin width (>10 mm, 1 to 9 mm, < 1 mm) and found that even in the irradiated group, the recurrence rate increased from 4% for those with the largest margins to 29% for those with margins smaller than 1 mm. In patients with the widest margins, the addition of radiation therapy did not significantly improve local control, even in those tumors associated with comedonecrosis (7% versus 3%, P = not statistically significant).
Applying Relative Risk Reduction to Treatment Decisions
Formulating an individual treatment algorithm for patients with DCIS is a complex and challenging process. Patients are increasingly interested in participating in treatment decisions and should be encouraged to do so. The survival differences among treatment options are usually small, and the optimal treatment varies, depending on each patient’s risk profile and value system. Often patients seek second or third opinions, all of which may be different. This difference more often reflects the spectrum of thresholds for intervention among surgeons and oncologists rather than indicating “bad” recommendations. Often there is not a clear “best medical practice,” and treatment must be tailored for patient preference and risk management.
Risk Assessment
It is critical to put the risk of recurrence of DCIS into perspective for the patient and to distinguish between relative and absolute risk reduction. Furthermore, the risk of a recurrence of either invasive or intraductal disease must be distinguished from the risk of dying from a consequent invasive cancer. The difference in mortality among the therapeutic choices is likely to range from 0 to 2%, but the risk of recurrence may range from 1% to 30%. Further perspective can be gained by addressing the risk of developing breast cancer in the contralateral breast. It should be remembered that all women face a 5% to 8% lifetime risk of developing breast cancer, and that any woman with a history of breast cancer is at increased risk of developing a contralateral breast cancer. Therefore, the goal should not be to reduce the ipsilateral risk of breast cancer to 0 to 1%, because the woman has accepted a 5% to 10% risk of breast cancer by not opting for bilateral prophylactic mastectomy. The magnitude of difference between the treatment options largely depends on the estimate of the natural history of a particular DCIS. For example, a woman with a 2-mm low-grade DCIS with a wide margin of greater than 1 cm might be estimated by the Van Nuys criteria to have a 3% to 4% ipsilateral long-term recurrence risk. Although irradiation and tamoxifen may be added to her treatment, the absolute benefit of either intervention would be minimal and probably would not be recommended to the patient unless she was personally committed to reducing her risk of ipsilateral recurrence to 1% to 2% with irradiation, or her risk of contralateral breast cancer below 5% at 5 years with tamoxifen therapy. Treatment decisions must be driven by absolute risk; therefore, the understanding of natural history and outcome after lumpectomy is critical.
Other considerations to include in the decision-making process are the time frame for progression of disease and the age of the patient. If the grade of DCIS does indeed affect the time to progression, this may be an important consideration. A risk of recurrence in 15 to 20 years may lead to different choices than a risk of recurrence in 5 years, and this may be affected by whether the woman is 40 or 70 years old. Finally, patients should always be informed about options to participate in clinical trials; this is especially important in a disease such as DCIS in which so many questions remain unanswered (Table 1).
Temporal Considerations
Time is an important variable in the management of DCIS, in which the window for progression of disease probably spans years rather than months.
The Biology of DCIS and the Future
Ductal carcinoma in situ is molecularly indistinguishable from invasive carcinoma, whether one looks at gains and losses over the entire set of chromosomes (CGH), evaluates more specific base pair abnormalities (LOH), or studies particular molecular markers. No patterns have yet been identified that are specific only to DCIS and not to invasive cancer, so from a purely genetic point of view, the transformation to malignancy has already occurred. Ductal carcinoma in situ and invasive cancer are, however, clearly different morphologically from normal cells and also from each other, and the mechanisms responsible for this phenotypic change are the subject of intense research.
Molecular analysis may yield clues for evaluating the extent of tissue at risk for progression to invasive disease, thereby leading to improved surgical intervention. One particular abnormality, LOH at 3p, is accompanied by consistent loss of genetic material from the short arm of chromosome 3. In a careful analysis of specimens from mastectomies performed for DCIS, Smith found that a wide region around the DCIS had evidence of 3p LOH, but the quadrants distant from the DCIS did not. The areas of 3p LOH extended beyond the area of abnormal pathology detected by light microscopy, suggesting that an entire region or ductal system may be affected. If the distribution of the cells demonstrating LOH could be marked preoperatively, it might be possible to resect the entire affected duct system more completely, possibly obviating the need for adjuvant irradiation following surgical excision.
The pattern of angiogenesis may also provide insight into how the biology of DCIS and invasive cancer differ. In DCIS, most increased angiogenesis is found periductally or within the stroma, but never within the malignant areas themselves. The ability to characterize angiogenic patterns on MR imaging may further the understanding of whether these patterns can change. Areas of increased angiogenesis may provide targets for therapy in the near future. By MR imaging, the most common denominator is the failure, in most cases, to see homogeneous uptake as a mass. Because the imaging techniques reflect patterns of intravenous contrast uptake and washout, the authors have investigated the pattern of CD34 staining in DCIS. The authors and others have noted a pattern of perivascular cuffing around the DCIS (both on hematoxylin and eosin stains and on CD34 staining) rather than a pattern of scattered blood vessels throughout the tumor. Further development of these imaging techniques to monitor treatment effects in vivo may make it possible to establish new surrogate end-points that will facilitate clinical testing of new treatments of DCIS and give further insight into the biology of intraductal cancers.
Vaccines
New and innovative treatment strategies are clearly needed to improve outcomes in breast cancer; too frequently, this disease still recurs or progresses despite aggressive multimodality therapy. Indeed, DCIS offers the best chance for truly preventive interventions, such as vaccine strategies, because the burden of disease is low and the patient’s immunocompetence is high. The authors’ data have shown that 50% of all DCIS tumors demonstrate HER-2/ neu over expression, including 85% to 100% of high-grade DCIS, which are the tumors most likely to progress to invasive cancer. Thus, an anti-HER-2 vaccine would be ideally suited to the treatment of DCIS.
Ductal carcinoma in situ offers a unique biological and clinical opportunity for the introduction of novel therapies such as vaccine strategies. Biologically, DCIS represents an early stage in tumor development that is not uniformly destined to progress, suggesting that tumor/host interactions and immune mechanisms may be important in determining the course of these tumors. It has become clear in other systems that the efficacy of immunotherapies, and particularly of vaccine strategies, can be severely limited by the bulkiness of solid tumors, by the emergence of resistant clones, by tumor-associated immunosuppressive effects, and by other factors associated with advanced cancer. The low tumor burden in most patients with DCIS greatly increases the chances that a vaccine strategy may be effective. A vaccine for DCIS that might possibly be used to treat existing cancer and to prevent its recurrence is a promising and exciting avenue of research.
Where Do We Go from Here?
Currently, no precise morphologic or biological prognostic factors reliably identify patients with DCIS who are at high risk of disease progression. Improved understanding of molecular and cellular characteristics of DCIS may help predict outcome. Although correlative studies have focused on protein and genetic markers of DCIS, the small number of cases in most series, the scant amount of tissue available for study, and the long follow-up time needed for analysis have hampered the ability to find clear associations.
With the molecular tools available today, DCIS lesions appear to be indistinguishable from invasive cancer. There are variable levels of ploidy, p53 abnormalities, HER-2/ neu over expression, and LOH and CGH changes in both. The nuclear grade of cells is more likely to predict their molecular characteristics, rather than their characterization as DCIS or invasive disease by light microscopy. Advanced MR imaging of DCIS confirms molecular studies showing that, like invasive breast cancer lesions, DCIS is heterogeneous. The similarities between DCIS and invasive cancer suggest that an important area for future research should be elucidating the processes that can either unleash or contain the invasive potential of DCIS cells. An understanding of the biology of DCIS can assist in the prevention, the assessment, and the treatment of invasive breast cancer.
Patients with DCIS should be reassured that their decisions can be made only with the facts at hand, and that future changes in disease management will not devalue the choices made at the time of their treatment. It is hoped that new knowledge will improve these options in the future and decrease recurrence rates. Physicians and patients should be encouraged to participate in trials and to support further studies to deepen the understanding of this challenging problem and to improve future management.
Imaging
The most common presentation of DCIS is that of abnormal calcifications on routine screening mammography. In the NSABP B-17 randomized trial, 83% of patients had mammographic findings only. Several different patterns of calcifications suggest DCIS. The hallmark of high-grade DCIS is small linear calcifications that follow the pattern of a duct. These tend to represent the comedo necrosis often seen in grade III lesions. An indeterminate or pleomorphic cluster of calcifications or, occasionally, small masses can also indicate DCIS. When intraductal carcinoma is diagnosed on needle-localized biopsy or core biopsy of indeterminate calcifications, it should be carefully noted whether the calcifications are actually associated with the DCIS itself. Often, calcifications occur in benign breast tissue, and DCIS is only incidentally detected in the surrounding parenchyma. Such findings can lead to a clinical dilemma if there are multiple other foci of calcifications, and can erroneously lead to a mastectomy in an effort to remove all such mammographic findings.
In general, mammography remains an excellent diagnostic modality for the detection of DCIS, but it does not provide an accurate assessment of the extent of disease. Holland et al have demonstrated that standard two-view mammography underestimates the extent of DCIS, especially of low-grade lesions, by as much as 2 cm in almost half of patients. Other imaging modalities are being explored in an attempt to better define the pattern of DCIS in the breast. Magnetic resonance imaging with contrast has demonstrated promise in better defining the extent of lesions, thereby refining surgical decision-making. Furthermore, the patterns of contrast uptake and washout on MR imaging probably reflect abnormal angiogenesis. In the future, such radiographic data may help predict which lesions have the potential to progress to invasive disease.
Predictors of Recurrence Following Treatment for Ductal Carcinoma In Situ
The recurrence rate following simple mastectomy for DCIS ranges from 0 to 2%. A critical evaluation of the factors leading to recurrence of intraductal carcinoma following mastectomy has not been possible because of the small number of events in each series. However, recurrence data collected from patients treated with breast-conserving surgery suggest that the events resulting in local failure are multifactorial. There is controversy concerning the relative importance of each variable in determining local control following breast-conserving therapy, but large studies conducted in the past decade have increased the understanding of the complex interactions of patient, tumor, and treatment factors in determining recurrence.
Family History of Breast Cancer
Few studies have been specifically designed to determine whether a history of breast cancer in a first-degree relative places a woman at increased risk of recurrence following breast-conserving surgery for DCIS. At least two reports, however, indicate that family history may have a measurable effect. In a small study of women undergoing breast-conserving surgery and adjuvant radiation for DCIS, 4 of 10 patients with recurrence were found to have a positive family history, compared with 5 of 44 of those women who did not have a recurrence. This finding was statistically significant ( P = 0.03). In another study of patients undergoing treatment for DCIS, the 10-year actuarial rate of local recurrence was found to be 37% for women with a positive family history, compared with 9% for women without a family history of breast cancer. These findings suggest that genetic factors may not only play a role in increasing the incidence of breast cancer but may also lead to greater difficulty in maintaining durable local control.
Age at Diagnosis and Menopausal Status
In studies published to date, the two variables of age and menopausal status have not been separately evaluated in multivariate analysis. In one report of 133 patients with DCIS treated with wide excision and irradiation, Silverstein found that age was not a predictor of recurrence on univariate analysis ( P = 0.3). At least three other studies, however, have demonstrated that postmenopausal status or older age confers a beneficial effect on local recurrence. In one cohort study of 709 women diagnosed with DCIS, patients who were premenopausal at the time of diagnosis were more than twice as likely to have a recurrence than postmenopausal women. This result was statistically significant but did not take into consideration such factors as margin status and patient preference for one treatment over another. It does, however, argue for a more aggressive treatment for younger patients who appear to have a worse tumor biology and more years to be at risk for recurrence. Such considerations may give rise to either more extensive surgery or more frequent use of adjuvant irradiation in these patients.
Tumor Necrosis and Nuclear Grade
The presence of necrosis in DCIS has long been recognized to be associated with poor prognosis. Numerous studies have now demonstrated higher local recurrence rates for intraductal carcinomas containing a comedo component, regardless of whether treatment included adjuvant radiotherapy. Statistical analysis has recently suggested that the presence of necrosis alone may not be as important as cellular architecture and nuclear grade. Lagios reported on a long-term follow-up study of 79 patients treated with complete excision without irradiation and found that 19% of high-grade tumors recurred by 26 months whereas 10% of intermediate tumors recurred within 87 months. There were no recurrences in the low-grade group at 124 months of mean follow-up. Interestingly, the recurrences seen in the high-grade group occurred within a much shorter interval than those in the intermediate group. These findings are corroborated by others who have found nuclear grade to be the most significant predictor of recurrence on both univariate and multivariate analysis.
The concept of disease-free interval has been further explored by data from the Joint Center for Radiation Therapy, which showed no significant difference in long-term recurrence rates between low-grade and high-grade lesions. Solin found that at 5 years following treatment, high-grade DCIS had a 12% recurrence rate compared with 3% for low-grade lesions. By 10 years of follow-up, however, the difference in recurrence rates between the two groups was not statistically significant (18% versus 15%, respectively), again supporting the view that one of the major differences between low- and high-grade lesions may be the time to progression rather than the potential to progress.
The Role of Chemoprevention
Tamoxifen
Tamoxifen, a nonsteroidal compound with mixed estrogenic and antiestrogenic effects, has repeatedly shown effectiveness in decreasing recurrence rates of invasive breast cancer, particularly in women with estrogen receptor (ER)-positive tumors. Tamoxifen has also consistently demonstrated the ability to reduce the incidence of contralateral breast cancer, an observation that has led to the institution of several tamoxifen-based prevention trials. The NSABP trial (P-01), by far the largest prevention study to date, demonstrated that tamoxifen reduces the risk of developing both invasive and noninvasive breast cancer by 40%. Two other European prevention studies have not been able to demonstrate a statistically significant risk reduction, but these trials lack the statistical power of P-01. Additional criticisms of these trials involve poor compliance, high attrition, and the fact that many women, especially in the Italian study, were taking hormone replacement therapy. An important difference between the P-01 and United Kingdom studies was that positive family history was one of the major inclusion criteria for the latter trial. This study is ongoing and may shed light on the use of tamoxifen for women with a strong family history for breast cancer. The role of tamoxifen in treating DCIS is now established. Recent reports from NSABP B-24, which randomly assigned patients undergoing excision and irradiation for DCIS to either placebo or tamoxifen, show that tamoxifen further reduces DCIS recurrence. The effect of tamoxifen on reducing both primary and recurrent breast cancers is remarkably consistent across all trials, with an average relative risk reduction of about 50%.
One would anticipate that tamoxifen therapy alone could also reduce both the recurrence of DCIS and the progression of DCIS to invasive cancer following lumpectomy, even in the absence of radiation therapy, but these studies have not yet been conducted. The NSABP B-24 study found that, for the population studied, the risk of developing contralateral invasive breast cancer (CBC) was nearly as high as the risk of developing ipsilateral invasive breast cancer (IBC) after irradiation. In the placebo group, the cumulative 5-year risk of invasive CBC was 2.4%, and for IBC it was 3.4%. These rates were reduced to 1.8% and 2.1% with tamoxifen. The cumulative incidences of DCIS after 62 months of follow-up in the placebo arm were 5.2% in the ipsilateral breast and 1.0% in the contralateral breast, and these risks were reduced to 4.3% and 0.2%, respectively, for the patients who received tamoxifen.
In a recent meta-analysis of the NSABP trials, the cumulative risk of CBC (both invasive and intraductal) for women with a cancer diagnosis was 5.1% after 5 years, which was reduced to 1.9% in women who took tamoxifen. The NSABP trial showed a cumulative 5-year breast cancer risk of 4.4% for women on placebo versus a risk of 2.4% for women on tamoxifen. It would therefore be reasonable to offer a trial that enables women to use hormone therapy alone to reduce the risk of invasive breast cancer recurrence and of contralateral breast cancer following lumpectomy for DCIS. This therapy may be especially important for small DCIS lesions in which the recurrence rates are low and the risk of contralateral breast cancer is almost as high as the risk of IBC recurrence after lumpectomy without radiation therapy.
The primary concern regarding the widespread use of tamoxifen involves its effect on other end organs. Tamoxifen appears to have a protective effect on bone mineral density in postmenopausal women, although in the P-01 trial its ability to reduce the fracture rate was only marginally significant. The P-01 trial also showed a reduction in the lipid profiles with tamoxifen use, but there was no clear effect on cardiovascular events during the short follow-up interval. Unfortunately, tamoxifen has additional estrogenic effects that are adverse, especially in postmenopausal women. The effect of tamoxifen on the endometrial lining increases the risk of uterine cancer, probably by at least two-fold. In the NSABP P-01 trial, in which more than 13,300 women were randomly assigned to either tamoxifen or placebo, the cumulative risk of developing endometrial cancer in postmenopausal women was 13/1000 in the tamoxifen arm versus 5.4/1000 in the placebo arm. The endometrial cancers were all early stage and treated by hysterectomy, but a 5-year cumulative risk of 1.3 must be placed into the context of other projected benefits. In the tamoxifen arm there was also an increase in venous vascular events of approximately two-fold over the control arm.
New Compounds
The deleterious side effects of tamoxifen have spurred great interest in the development of new agents that have antiestrogenic effects at the level of both breast and uterine tissue while maintaining the beneficial estrogen-like effects on bone mineral density and the cardiovascular system. This group of compounds has been termed selective estrogen receptor modulators, or SERMs, and the best studied is the agent raloxifene (marketed as Evista), which was developed to prevent osteoporosis in postmenopausal women. Raloxifene also lowers serum concentrations of total and low-density lipoprotein cholesterol and does not stimulate the endometrium. Preliminary evidence suggests that raloxifene may be useful in the prevention of breast cancer; however, there is no experience to date in premenopausal women, and data on the effectiveness of raloxifene for women with established breast cancer are very limited.
Initially intended to be an alternative to hormone replacement therapy, raloxifene has been tested for its effect on bone mineral density and fractures but not for its effect on breast cancer. Nevertheless, some striking data were gathered from the Multiple Outcomes of Raloxifene Evaluation Study (MORE trial). In this large randomized study, 7705 postmenopausal women with osteoporosis were treated with raloxifene (60 or 120 mg/d). Raloxifene significantly reduced bone mineral density loss and the fracture rate compared with placebo. Also, after a median follow-up of 33 months, there was a 70% decrease in the risk of both breast cancer (relative risk = 0.26) and endometrial cancer (relative risk = 0.38, P = not statistically significant) compared with women who received placebo. A meta-analysis of nine raloxifene trials involving a total of 10,575 women was recently presented, and after 40 months of follow-up, the drug was associated with a 55% reduction in the relative risk of developing invasive breast cancer compared with placebo use, based on 67 breast cancer events. As in the P-01 trial, the incidence of only estrogen receptor-positive breast cancers appears to be reduced. At present, however, there are few data to show that raloxifene is effective in the treatment of invasive breast cancer, and its use is not recommended except in the context of a clinical trial.
One of the newest “designer estrogens” is LY353381 hydrochloride, a SERM-3 compound that is structurally similar to raloxifene. Like raloxifene, it appears to act as a potent estrogen antagonist in both uterine and mammary tissues while maintaining bone density and lowering serum cholesterol. A clinical trial is currently being planned to test tamoxifen versus SERM-3 in women with DCIS who have chosen breast-conserving surgery, regardless of use of irradiation. Results from this trial should be available 5 years from the start date of accrual.
Technical Considerations and Recommendations
Wide excision for DCIS is appropriate in patients with limited extent of disease. Careful attention must be paid to the margin status, although intraoperative decision-making is hampered because intraductal lesions are for the most part not distinguishable from normal surrounding tissue. Precise anatomic orientation is critical in identifying the location of any positive margins, and careful attention must be paid to inking the specimen correctly. These measures will greatly facilitate any subsequent reexcision. The rate of reexcision for positive or close (<1 mm) margins has been estimated to be as high as 55%, and this number may be even higher in tertiary breast cancer referral centers, where the inspection of margins tends to be more exhaustive.
The accurate determination of the true size of DCIS lesions is challenging. Therefore, both the extent of disease in a given slide and the number of blocks involved with disease should be recorded. Also, it must be noted whether microcalcifications are associated with DCIS within the excised specimen, because this factor is especially important in the evaluation of subsequent residual calcifications. Finally, the authors and others recommend routine follow-up mammography after definitive excision and before starting irradiation. This simple step may guide management and help distinguish between residual incompletely excised calcifications and those that are the result of postsurgical or radiation changes.
Axillary Lymph Node Dissection
Current data indicate that the incidence of axillary lymph node metastases in pure DCIS is 0 to 1%, obviating the need for axillary dissection in these patients. For complicated cases with an associated mass and questions of microscopic invasion, axillary lymph node dissection has a role, but these are a very small fraction of DCIS cases. Axillary lymph node dissection has been shown to be associated with substantial rates of lymphedema, for which treatment options remain limited. Unfortunately, many women still undergo axillary dissection when they are treated for DCIS. Although the number has dropped during the last decade, Winchester et al have shown that as many as 40% of patients still undergo this procedure and that continued efforts to educate women and their physicians are critical. The use of sentinel node biopsy has been proposed in treating DCIS, but when it is undertaken, it should be performed as part of a clinical trial.
The Role of Adjuvant Radiation
The role of irradiation in the treatment of DCIS continues to evolve. Many studies have now conclusively demonstrated approximately a 50% reduction in local recurrence with the addition of radiotherapy to surgical excision. The most compelling data come from The National Surgical Adjuvant Breast and Bowel Project trial (NSABP) B-17, a prospective trial that randomly assigned 818 patients to surgery only or to surgery plus irradiation. The most recent update of this study, with a mean follow-up of 90 months, shows a reduction in noninvasive ipsilateral breast tumors (IBT) from 13.4% to 8.2% ( P = 0.007), with a similar reduction in invasive IBT from 13.4% to 3.9% ( P = < 0.0001). This study has been criticized for suggesting that all patients, regardless of clinical or tumor characteristics, should undergo radiotherapy, rather than providing more selective recommendations based on subgroup analysis. Furthermore, the study has limited applicability, because the majority of lesions in this trial were very small tumors. A prospective randomized study designed by the Radiation Therapy Oncology Group (RTOG) is scheduled to open by mid-2006 to address some of these issues. Included will be patients diagnosed with DCIS of up to 2.5 cm who will be placed on tamoxifen therapy, then randomly assigned to lumpectomy only or to lumpectomy with irradiation. The goals of this study include the identification of those variables that may allow a more individualized use of radiotherapy in the treatment of DCIS.
Natural History of Ductal Carcinoma In Situ
Undoubtedly, the improved availability of mammographic screening has dramatically increased the detection rate of DCIS, resulting in an earlier and possibly overly aggressive intervention at this stage of breast cancer. Autopsy series suggest that the prevalence of undetected DCIS is close to 9% in the overall population, and it is likely that many of these lesions can remain undetected and be clinically insignificant for many decades. The consequences of treating DCIS by observation alone are unclear. Current understanding of the disease suggests that, without treatment, a certain percentage of high-grade lesions will probably progress to invasive disease within 5 to 10 years. The natural history of low-grade intraductal carcinoma is more debatable, and the answer is critical: early surgical intervention in low-grade DCIS may represent over treatment, because the majority of these neoplasms, if they progress to invasive disease, do so over a much longer period of time. Furthermore, ethical considerations make longitudinal studies to address these issues impossible.
Indirect evidence pertaining to the natural history of DCIS can be obtained from early studies examining the long-term recurrence rates in women treated for DCIS with biopsy only. Page and Dupont reported a small series of 28 women who had undergone surgery only for cribriform or micropapillary DCIS between 1952 and 1968 and found that 7 of 28 patients developed invasive cancer within 10 years, all within the same quadrant as the index lesion. Furthermore, 2 additional women were diagnosed with invasive breast cancer at 15 and 31 years following their diagnosis of DCIS, suggesting that the increased risk for developing invasive disease may continue long after the diagnosis of low-grade DCIS has been made. Betsill reported similar results on a cohort of 24 women who underwent biopsy only for DCIS. He concluded that in long-term follow-up (mean 21.6 years) 39% of women with DCIS treated with biopsy alone developed invasive cancer in the ipsilateral breast.
These studies demonstrate that low-grade DCIS also progresses to invasive cancer but do not elucidate which patients are at highest risk for progression to more aggressive disease. What is clear, however, is that the risk of developing invasive cancer may persist even two decades after diagnosis of DCIS if an initial complete excision is not performed. One of the most challenging areas for research and intervention will be to identify those factors responsible for the transition from DCIS to invasive cancer, to avoid the over treatment of the large percentage of women with DCIS who will probably never progress to clinically significant disease.
Treatment Considerations and Management of Ductal Carcinoma In Situ
Surgery
Mastectomy
The recurrence rate following mastectomy for DCIS has been shown to be between 0 and 2% on long-term follow-up. There is thus no role for adjuvant irradiation following mastectomy for DCIS. This procedure remains the standard against which the outcomes of all other therapy must be compared but is the most aggressive of the treatment options. Mastectomy still, however, remains the procedure of choice in those patients with large or high-grade lesions or in women with multicentric disease. Advances in breast reconstruction techniques have allowed many women to undergo mastectomy without the disfiguring sequelae. The most significant advance in recent years has been the introduction of skin-sparing mastectomy, in which the breast tissue is excised by a periareolar incision. Large retrospective studies have now shown that this technique does not result in higher local breast cancer recurrence rates.
Breast Conservation
The widespread successful use of breast-conserving treatment for invasive cancer has focused efforts to identify which women with DCIS may be appropriately treated with wide excision rather than mastectomy. Attempts have been made to classify patients with DCIS into prognostic categories, as for invasive cancer. The Van Nuys Prognostic Index (VNPI) originated by Silverstein is perhaps the best known among these classification systems. The VNPI assigns patients an overall score taking into account three factors: the size of the DCIS, the width of the closest margin, and the differentiation (grade) of the tumor cells.
Silverstein evaluated 333 patients who were treated with breast-conserving surgery and divided them into three different recurrence-risk groups based on the VNPI score (3-4, 5-7, or 8-9). With a median follow-up of 70 months, he found that the 8-year actuarial local disease-free survival rates for the three groups were 97%, 77%, and 20%, respectively. For the purposes of analysis, patients were stratified by treatment (wide excision or wide excision and irradiation). Silverstein found that adjuvant irradiation offered a significant therapeutic benefit only in the VNPI intermediate group. The low-risk group had an excellent prognosis, so irradiation conferred very little absolute benefit. Irradiation reduced the likelihood of recurrence in high-risk patients, but even those treated with adjuvant radiotherapy had very high failure rates, with only 30% of patients in this group enjoying long-term local control. These data, although retrospective, are compelling because they represent the experience of a large number of patients treated at a single institution with identical tissue handling and analysis. Thus, although the prospective randomized trials eliminate individual physician bias, Silverstein’s data are valuable in that they control for process.
The Van Nuys data suggest that surgical treatment alone can lead to a lower risk of recurrence. Patients with a low VNPI score may be treated with wide excision, because the recurrence rates are in the 3% to 4% range. The use of adjuvant irradiation in addition to wide excision is suggested for intermediate-score tumors, for which the long-term local recurrence rates may be reduced from about 28% to 15%. A high VNPI score would merit consideration of simple mastectomy because of the excessive local recurrence rates (25% and 45%, with and without irradiation at 4 years, respectively) seen with breast-conserving treatment in this group. This classification system allows the generalized formulation of a treatment plan, but it takes into consideration only three prognostic factors. It should be remembered that other variables that may affect recurrence rates for DCIS are not included in the VNPI analysis.
Significance of Cellular and Genetic Changes in Ductal Carcinoma In Situ
The last decade has seen an explosive growth in the understanding of the changes that occur at the cellular and genetic level in all cancers, and increasing attention is being focused on such changes in DCIS. p53 has long been known to be an important tumor-suppressor gene in lung, colon, and breast cancers. Recently, p53 mutations have been identified in 7% to 37% of intraductal carcinomas but have not been found in areas of epithelial hyperplasia. In some studies, aberrant p53 expression appears to be associated with higher nuclear grade. Microdissected specimens of DCIS both with and without invasion have been studied, and the identical p53 mutation was identified in both the DCIS and the invasive component. This finding indicates that the genetic alterations seen in infiltrating cancers have already occurred at the preinvasive level, but the mechanism by which the phenotypic transformation occurs has not yet been elucidated. It is known, however, that the prevalence of p53 mutations is greater in DCIS than in invasive breast cancers, and thus it is probable that not all of these lesions progress to invasive disease. This has also been noted in HER2/ neu expression: 90% of high-grade DCIS lesions over express HER-2/ neu, but only 40% to 50% of high-grade invasive lesions show over expression.
It has been shown that a greater degree of apoptosis is seen in DCIS than in invasive cancers, and, furthermore, that higher grade DCIS exhibits more abundant apoptosis than lower grade lesions, probably reflecting the greater growth fraction seen in high-grade disease. The degree of apoptosis appears to be independent of p53 expression and has not yet been fully explained. These data, however, strongly suggest that aberrant cell cycle regulation plays a role in the transition between atypia and in situ carcinoma.
PRAD-1/cyclin D1 has been implicated as the relevant oncogene on chromosome 11q13 and is seen frequently in breast and other epithelial cancers. This gene may be a marker of early malignant transformation, because its expression is found in 18% of atypical ductal hyperplasia, in 76% of low-grade DCIS, in 87% of high-grade comedo-type DCIS, and in 83% of invasive breast cancer. To date, however, the precise role of cyclins and cyclin-dependent kinase inhibitors such as p21 (waf-1), p15, and p27, which may act as tumor-suppressor genes, has not been well studied for DCIS.
The presence of genetic abnormalities in DCIS has been clearly demonstrated in studies examining loss of heterozygosity (LOH) and microsatellite instability (MI) in microdissected specimens of DCIS. Loss of heterozygosity is associated with 70% of noncomedo DCIS and with 79% of comedo DCIS. Interestingly, DCIS associated with an invasive breast cancer has been found to share LOH at one or more loci with the infiltrating component. Microsatellite instability has been reported in one study in 22% of DCIS at two or more loci and in 30% at a single locus. Such genetic changes were correlated with higher grade and c- erbB-2 positivity. Further research will be needed to define the relationship between such genetic alterations and the clinical behavior and prognosis of DCIS.
The presence of increased angiogenesis has also been explored. Studies have shown that high-grade DCIS exhibits an increase in both stromal and periductal vascularity. Evidence shows that VEGF mRNA expression is greater in DCIS than in adjacent normal cells in 96% of cases. High-grade tumors showed increased VEGF expression more often than low-grade tumors. Thus, even at the preinvasive stage, alterations at the genetic or translational level allow DCIS to establish and maintain the neovascularization necessary for tumor growth. These alterations may provide a target for therapeutic interventions in the future.
The compilation of research on molecular markers leads to the compelling observation that DCIS and invasive cancer are clearly distinct from normal cells and from each other but appear to be virtually identical at the molecular and genetic levels. Possible factors leading to progression from DCIS to invasive cancer include the interaction with the surrounding stromal environment. Promising avenues for further investigation include the study of myoepithelial cells, stromal interactions, angiogenic signals, and DNA repair mechanisms.