Should ductal carcinoma in situ be treated?

Remarks given during a session of the ASCO Breast Cancer Symposium titled Ductal Carcinoma in Situ Debate: Treatment vs. Observation

Dr. Kuerer comments: There has been a marked increase ductal carcinoma in situ (DCIS) in what is being called overdiagnosis, and this is leading to concerns of overtreatment, which has been in the news now for several years. We have more than 60,000 cases of DCIS diagnosed per year in the United States. We’re trying to prevent invasive breast cancer and distant metastases. The problem is identifying which patients will go on to develop invasive disease. That’s really unknown.

At M.D. Anderson, looking at 10-year follow-up of 2,449 patients, the rate of developing distant metastases was 0.1%. The problem with this is that no prominent variables were associated with the development of invasive metastatic disease. Overall, we are doing a good job at preventing death from breast cancer when we treat DCIS, with less than 1% of patients dying of breast cancer (Ann. Surg. Onc. 2011;18:2873-8).

Dr. Shelley E. Hwang and her group very elegantly started exploring preoperative systemic therapy as a way to get a clue about which patients might be safely observed alone (that is, a core biopsy diagnosis of DCIS and then just follow-up without surgery or other treatments). At M.D. Anderson, we studied whether or not we could eradicate DCIS in patients with human epidermal growth factor receptor 2 (HER2) overexpression by treatment with trastuzumab prior to surgery, under the hypothesis that if we could eradicate the DCIS, we might eventually use this drug potentially to prevent HER2-positive invasive breast cancers. We saw very dramatic responses in their immune response with antibody-dependent cellular cytotoxicity mediated by natural killer cells, which is exciting, but there were no histopathologic changes. The most important finding was that 42% of our patients who we thought only had DCIS actually had occult invasive breast cancer (Cancer 2011;117:39-47). This is concerning because without surgery, we may be missing or leaving untreated invasive breast cancers in some patients.

So, we have an underestimation of invasive breast cancer at DCIS diagnosis. A very elegant meta-analysis of 7,350 patients reported that even if we take patients with non–high-grade, very-small DCIS, we’re still looking at an upgrade to invasive cancer of about 20% (Radiology 2011;260:119-28). The best and most recent studies of MRI and DCIS again found a diagnostic upgrade of about 27%. There were no MRI features correlating with invasive breast cancer, and this resulted in a 31% increase in procedures and more biopsies with the use of MRI.

So where are we? In active surveillance of 14 patients with DCIS at the University of California, San Francisco, 8 went on to surgery at a median follow-up of 28 months, and 5 of the 8 (62%) had invasive breast cancer (The Breast 2011;20:529-33).

What’s the cost to the health care system of surveillance if we biopsy only the DCIS and follow it without surgery or other adjuvant therapies? The psychological cost to our patients? What is the natural history of DCIS if left in place? What will be the result of microcalcifications? How are we going to follow this? What are the criteria that we will use for repeat biopsy? This really is not known.

Which are the patients that we’re going to select for observation alone? There are trials in Europe that have begun or will begin to address the safety of just observing DCIS without surgery and other therapies.

The United Kingdom Low-Risk DCIS Trial is randomizing patients with low and intermediate grade DCIS of any size to biopsy alone without surgical intervention (active monitoring) or standard therapy. Another trial, a joint Dutch Breast Cancer Research Group and European Organization for the Research and Treatment of Cancer trial is randomizing women with low-risk, low-grade DCIS to active surveillance or standard treatment. If we look at M.D. Anderson’s data and our National Cancer Database, only about 10%-15% of patients would be eligible for the latter study. I don’t think with these numbers we’d have a great impact, but it is a start.

I submit that the best U.S. patients to study will be patients with a diagnosis of atypical ductal hyperplasia (generally the smallest lesions and perhaps the earliest form of DCIS) because we have about 100,000 cases each year.

What’s the patient’s perspective? What about the need for continued repeat biopsies, and how this will affect their overall quality of life? Do patients in the United States really want to observe the DCIS in an era when our patients are requesting more and more mastectomy with reconstruction and contralateral mastectomy at diagnosis of DCIS? I seriously doubt that our patients in the United States would agree to be randomized in DCIS trials to no treatment at all. I really don’t know.

Dr. Kuerer is a professor of surgery at the University of Texas M.D. Anderson Cancer Center, Houston. He reported financial associations with Gerson Lehrman Group and McGraw-Hill Publishing.

Dr. Hwang comments: Today, about 1/1,300 screening mammograms result in a diagnosis of DCIS. There are two important considerations when we talk about active surveillance or doing less aggressive treatment. First, what is the rate at which progression to invasive cancer can occur, either with or without treatment? Second, what is the fate of these DCIS lesions? These are issues for which we currently don’t have good answers.

We’re treating all disease detected at an early stage, and for the sake of argument, I’ll include DCIS, as if it likely will cause harm if we did absolutely nothing when, in fact, there could be many cancers that progress so slowly and have such a low propensity for developing metastatic disease that they would not likely cause any symptoms or harm during a patient’s lifetime.

In autopsy series, the disease reservoir of unrecognized DCIS is about 9% and the disease reservoir of invasive cancer is about 1%. It’s not that different from prostate cancer, although at a much lower rate. These data show that there certainly are women who die with DCIS rather than of it.

We really don’t have a very solid understanding of the natural history of DCIS. If we do nothing, but maybe surgically biopsy it, what happens to these patients in the long term? In a meta-analysis, the world’s literature included only 151 cases of women who had surgical biopsy of DCIS that initially was misdiagnosed as a benign lesion and therefore didn’t undergo any further therapy, some with up to 31 years of follow-up. The long-term risk of invasive cancer in this cohort is only 22% (Breast Cancer Res. Treat. 2006;97:135-44). The annual risk of breast cancer in women with atypia is 1% per year, so this ends up looking very similar to the risk of progression that you see for atypical ductal hyperplasia or lobular carcinoma in situ.

The most common treatment for DCIS in the United States is lumpectomy with radiation. A meta-analysis by the European Breast Cancer Trialists’ Group found a 50% proportional reduction in local recurrence risk in women treated with lumpectomy and radiation vs. lumpectomy alone. The absolute magnitude of reduction was dependent on baseline recurrence risk. That’s a really important concept, because if your baseline risk is only 5%, then the 50% proportional reduction only translates into a 2.5% reduction in risk in 10 years.

The prospective, randomized Radiation Therapy Oncologists Group 9804 study randomized low-risk women with DCIS, unlike prior randomized trials that included broad eligibility criteria for DCIS. In 5 years of follow-up, the ipsilateral recurrence risk (which includes both invasive cancer and DCIS) was 3.2% in the lumpectomy-only group vs. 0.4% in the lumpectomy and radiation therapy group. The difference is highly statistically significant, however, given the small absolute difference between groups, the clinical significance certainly can be argued. There was no significant difference between groups in contralateral new primary lesions (Radiat. Oncol. 2012;84:S5).

We just presented a study this year looking at DCIS and competing causes of mortality in different age groups with different types of treatment. Among women with DCIS over 70 years of age, there was a significant difference in overall survival but none of the treatments conferred any benefit in disease-specific survival.

Active surveillance alone may be reasonable to consider in some patients. We can learn a lot from our colleagues who treat prostate cancer. They’re at least 10 years ahead of us in thinking about reducing the harms that have been introduced by screening for prostate cancer. When you look at breast cancer–specific and other-cause mortality in the setting of DCIS, the mortality curves look very similar to those with early-stage prostate cancer. Women with DCIS, regardless of what they are treated with, die of other causes and very rarely die of breast cancer.

When we compare such different options as mastectomy to active surveillance only, it’s really hard for a patient to feel that both of those approaches can offer them the same outcomes. But when you look at the actual data and evidence, the differences in recurrence or progression to invasive cancer don’t translate easily into a large difference in breast cancer mortality provided that patients are diagnosed at stage I or II, with more than a 90% survival from breast cancer.

We have an opportunity to take a big step back and redefine our goals of DCIS “treatment” from “curing” DCIS to trying to reduce breast cancer-specific mortality. This will help us reduce the harms that can result from screening and will reserve our aggressive treatments for those most likely to benefit from them.

Dr. Hwang is a professor of surgery at Duke University, Durham, N.C. She serves as a consultant for Genomic Health.

Remarks given during a session of the ASCO Breast Cancer Symposium titled Ductal Carcinoma in Situ Debate: Treatment vs. Observation

Dr. Kuerer comments: There has been a marked increase ductal carcinoma in situ (DCIS) in what is being called overdiagnosis, and this is leading to concerns of overtreatment, which has been in the news now for several years. We have more than 60,000 cases of DCIS diagnosed per year in the United States. We’re trying to prevent invasive breast cancer and distant metastases. The problem is identifying which patients will go on to develop invasive disease. That’s really unknown.

At M.D. Anderson, looking at 10-year follow-up of 2,449 patients, the rate of developing distant metastases was 0.1%. The problem with this is that no prominent variables were associated with the development of invasive metastatic disease. Overall, we are doing a good job at preventing death from breast cancer when we treat DCIS, with less than 1% of patients dying of breast cancer (Ann. Surg. Onc. 2011;18:2873-8).

Dr. Shelley E. Hwang and her group very elegantly started exploring preoperative systemic therapy as a way to get a clue about which patients might be safely observed alone (that is, a core biopsy diagnosis of DCIS and then just follow-up without surgery or other treatments). At M.D. Anderson, we studied whether or not we could eradicate DCIS in patients with human epidermal growth factor receptor 2 (HER2) overexpression by treatment with trastuzumab prior to surgery, under the hypothesis that if we could eradicate the DCIS, we might eventually use this drug potentially to prevent HER2-positive invasive breast cancers. We saw very dramatic responses in their immune response with antibody-dependent cellular cytotoxicity mediated by natural killer cells, which is exciting, but there were no histopathologic changes. The most important finding was that 42% of our patients who we thought only had DCIS actually had occult invasive breast cancer (Cancer 2011;117:39-47). This is concerning because without surgery, we may be missing or leaving untreated invasive breast cancers in some patients.

So, we have an underestimation of invasive breast cancer at DCIS diagnosis. A very elegant meta-analysis of 7,350 patients reported that even if we take patients with non–high-grade, very-small DCIS, we’re still looking at an upgrade to invasive cancer of about 20% (Radiology 2011;260:119-28). The best and most recent studies of MRI and DCIS again found a diagnostic upgrade of about 27%. There were no MRI features correlating with invasive breast cancer, and this resulted in a 31% increase in procedures and more biopsies with the use of MRI.

So where are we? In active surveillance of 14 patients with DCIS at the University of California, San Francisco, 8 went on to surgery at a median follow-up of 28 months, and 5 of the 8 (62%) had invasive breast cancer (The Breast 2011;20:529-33).

What’s the cost to the health care system of surveillance if we biopsy only the DCIS and follow it without surgery or other adjuvant therapies? The psychological cost to our patients? What is the natural history of DCIS if left in place? What will be the result of microcalcifications? How are we going to follow this? What are the criteria that we will use for repeat biopsy? This really is not known.

Which are the patients that we’re going to select for observation alone? There are trials in Europe that have begun or will begin to address the safety of just observing DCIS without surgery and other therapies.

The United Kingdom Low-Risk DCIS Trial is randomizing patients with low and intermediate grade DCIS of any size to biopsy alone without surgical intervention (active monitoring) or standard therapy. Another trial, a joint Dutch Breast Cancer Research Group and European Organization for the Research and Treatment of Cancer trial is randomizing women with low-risk, low-grade DCIS to active surveillance or standard treatment. If we look at M.D. Anderson’s data and our National Cancer Database, only about 10%-15% of patients would be eligible for the latter study. I don’t think with these numbers we’d have a great impact, but it is a start.

I submit that the best U.S. patients to study will be patients with a diagnosis of atypical ductal hyperplasia (generally the smallest lesions and perhaps the earliest form of DCIS) because we have about 100,000 cases each year.

What’s the patient’s perspective? What about the need for continued repeat biopsies, and how this will affect their overall quality of life? Do patients in the United States really want to observe the DCIS in an era when our patients are requesting more and more mastectomy with reconstruction and contralateral mastectomy at diagnosis of DCIS? I seriously doubt that our patients in the United States would agree to be randomized in DCIS trials to no treatment at all. I really don’t know.

Dr. Kuerer is a professor of surgery at the University of Texas M.D. Anderson Cancer Center, Houston. He reported financial associations with Gerson Lehrman Group and McGraw-Hill Publishing.

Dr. Hwang comments: Today, about 1/1,300 screening mammograms result in a diagnosis of DCIS. There are two important considerations when we talk about active surveillance or doing less aggressive treatment. First, what is the rate at which progression to invasive cancer can occur, either with or without treatment? Second, what is the fate of these DCIS lesions? These are issues for which we currently don’t have good answers.

We’re treating all disease detected at an early stage, and for the sake of argument, I’ll include DCIS, as if it likely will cause harm if we did absolutely nothing when, in fact, there could be many cancers that progress so slowly and have such a low propensity for developing metastatic disease that they would not likely cause any symptoms or harm during a patient’s lifetime.

In autopsy series, the disease reservoir of unrecognized DCIS is about 9% and the disease reservoir of invasive cancer is about 1%. It’s not that different from prostate cancer, although at a much lower rate. These data show that there certainly are women who die with DCIS rather than of it.

We really don’t have a very solid understanding of the natural history of DCIS. If we do nothing, but maybe surgically biopsy it, what happens to these patients in the long term? In a meta-analysis, the world’s literature included only 151 cases of women who had surgical biopsy of DCIS that initially was misdiagnosed as a benign lesion and therefore didn’t undergo any further therapy, some with up to 31 years of follow-up. The long-term risk of invasive cancer in this cohort is only 22% (Breast Cancer Res. Treat. 2006;97:135-44). The annual risk of breast cancer in women with atypia is 1% per year, so this ends up looking very similar to the risk of progression that you see for atypical ductal hyperplasia or lobular carcinoma in situ.

The most common treatment for DCIS in the United States is lumpectomy with radiation. A meta-analysis by the European Breast Cancer Trialists’ Group found a 50% proportional reduction in local recurrence risk in women treated with lumpectomy and radiation vs. lumpectomy alone. The absolute magnitude of reduction was dependent on baseline recurrence risk. That’s a really important concept, because if your baseline risk is only 5%, then the 50% proportional reduction only translates into a 2.5% reduction in risk in 10 years.

The prospective, randomized Radiation Therapy Oncologists Group 9804 study randomized low-risk women with DCIS, unlike prior randomized trials that included broad eligibility criteria for DCIS. In 5 years of follow-up, the ipsilateral recurrence risk (which includes both invasive cancer and DCIS) was 3.2% in the lumpectomy-only group vs. 0.4% in the lumpectomy and radiation therapy group. The difference is highly statistically significant, however, given the small absolute difference between groups, the clinical significance certainly can be argued. There was no significant difference between groups in contralateral new primary lesions (Radiat. Oncol. 2012;84:S5).

We just presented a study this year looking at DCIS and competing causes of mortality in different age groups with different types of treatment. Among women with DCIS over 70 years of age, there was a significant difference in overall survival but none of the treatments conferred any benefit in disease-specific survival.

Active surveillance alone may be reasonable to consider in some patients. We can learn a lot from our colleagues who treat prostate cancer. They’re at least 10 years ahead of us in thinking about reducing the harms that have been introduced by screening for prostate cancer. When you look at breast cancer–specific and other-cause mortality in the setting of DCIS, the mortality curves look very similar to those with early-stage prostate cancer. Women with DCIS, regardless of what they are treated with, die of other causes and very rarely die of breast cancer.

When we compare such different options as mastectomy to active surveillance only, it’s really hard for a patient to feel that both of those approaches can offer them the same outcomes. But when you look at the actual data and evidence, the differences in recurrence or progression to invasive cancer don’t translate easily into a large difference in breast cancer mortality provided that patients are diagnosed at stage I or II, with more than a 90% survival from breast cancer.

We have an opportunity to take a big step back and redefine our goals of DCIS “treatment” from “curing” DCIS to trying to reduce breast cancer-specific mortality. This will help us reduce the harms that can result from screening and will reserve our aggressive treatments for those most likely to benefit from them.

Dr. Hwang is a professor of surgery at Duke University, Durham, N.C. She serves as a consultant for Genomic Health.

Remarks given during a session of the ASCO Breast Cancer Symposium titled Ductal Carcinoma in Situ Debate: Treatment vs. Observation

Dr. Kuerer comments: There has been a marked increase ductal carcinoma in situ (DCIS) in what is being called overdiagnosis, and this is leading to concerns of overtreatment, which has been in the news now for several years. We have more than 60,000 cases of DCIS diagnosed per year in the United States. We’re trying to prevent invasive breast cancer and distant metastases. The problem is identifying which patients will go on to develop invasive disease. That’s really unknown.

At M.D. Anderson, looking at 10-year follow-up of 2,449 patients, the rate of developing distant metastases was 0.1%. The problem with this is that no prominent variables were associated with the development of invasive metastatic disease. Overall, we are doing a good job at preventing death from breast cancer when we treat DCIS, with less than 1% of patients dying of breast cancer (Ann. Surg. Onc. 2011;18:2873-8).

Dr. Shelley E. Hwang and her group very elegantly started exploring preoperative systemic therapy as a way to get a clue about which patients might be safely observed alone (that is, a core biopsy diagnosis of DCIS and then just follow-up without surgery or other treatments). At M.D. Anderson, we studied whether or not we could eradicate DCIS in patients with human epidermal growth factor receptor 2 (HER2) overexpression by treatment with trastuzumab prior to surgery, under the hypothesis that if we could eradicate the DCIS, we might eventually use this drug potentially to prevent HER2-positive invasive breast cancers. We saw very dramatic responses in their immune response with antibody-dependent cellular cytotoxicity mediated by natural killer cells, which is exciting, but there were no histopathologic changes. The most important finding was that 42% of our patients who we thought only had DCIS actually had occult invasive breast cancer (Cancer 2011;117:39-47). This is concerning because without surgery, we may be missing or leaving untreated invasive breast cancers in some patients.

So, we have an underestimation of invasive breast cancer at DCIS diagnosis. A very elegant meta-analysis of 7,350 patients reported that even if we take patients with non–high-grade, very-small DCIS, we’re still looking at an upgrade to invasive cancer of about 20% (Radiology 2011;260:119-28). The best and most recent studies of MRI and DCIS again found a diagnostic upgrade of about 27%. There were no MRI features correlating with invasive breast cancer, and this resulted in a 31% increase in procedures and more biopsies with the use of MRI.

So where are we? In active surveillance of 14 patients with DCIS at the University of California, San Francisco, 8 went on to surgery at a median follow-up of 28 months, and 5 of the 8 (62%) had invasive breast cancer (The Breast 2011;20:529-33).

What’s the cost to the health care system of surveillance if we biopsy only the DCIS and follow it without surgery or other adjuvant therapies? The psychological cost to our patients? What is the natural history of DCIS if left in place? What will be the result of microcalcifications? How are we going to follow this? What are the criteria that we will use for repeat biopsy? This really is not known.

Which are the patients that we’re going to select for observation alone? There are trials in Europe that have begun or will begin to address the safety of just observing DCIS without surgery and other therapies.

The United Kingdom Low-Risk DCIS Trial is randomizing patients with low and intermediate grade DCIS of any size to biopsy alone without surgical intervention (active monitoring) or standard therapy. Another trial, a joint Dutch Breast Cancer Research Group and European Organization for the Research and Treatment of Cancer trial is randomizing women with low-risk, low-grade DCIS to active surveillance or standard treatment. If we look at M.D. Anderson’s data and our National Cancer Database, only about 10%-15% of patients would be eligible for the latter study. I don’t think with these numbers we’d have a great impact, but it is a start.

I submit that the best U.S. patients to study will be patients with a diagnosis of atypical ductal hyperplasia (generally the smallest lesions and perhaps the earliest form of DCIS) because we have about 100,000 cases each year.

What’s the patient’s perspective? What about the need for continued repeat biopsies, and how this will affect their overall quality of life? Do patients in the United States really want to observe the DCIS in an era when our patients are requesting more and more mastectomy with reconstruction and contralateral mastectomy at diagnosis of DCIS? I seriously doubt that our patients in the United States would agree to be randomized in DCIS trials to no treatment at all. I really don’t know.

Dr. Kuerer is a professor of surgery at the University of Texas M.D. Anderson Cancer Center, Houston. He reported financial associations with Gerson Lehrman Group and McGraw-Hill Publishing.

Dr. Hwang comments: Today, about 1/1,300 screening mammograms result in a diagnosis of DCIS. There are two important considerations when we talk about active surveillance or doing less aggressive treatment. First, what is the rate at which progression to invasive cancer can occur, either with or without treatment? Second, what is the fate of these DCIS lesions? These are issues for which we currently don’t have good answers.

We’re treating all disease detected at an early stage, and for the sake of argument, I’ll include DCIS, as if it likely will cause harm if we did absolutely nothing when, in fact, there could be many cancers that progress so slowly and have such a low propensity for developing metastatic disease that they would not likely cause any symptoms or harm during a patient’s lifetime.

In autopsy series, the disease reservoir of unrecognized DCIS is about 9% and the disease reservoir of invasive cancer is about 1%. It’s not that different from prostate cancer, although at a much lower rate. These data show that there certainly are women who die with DCIS rather than of it.

We really don’t have a very solid understanding of the natural history of DCIS. If we do nothing, but maybe surgically biopsy it, what happens to these patients in the long term? In a meta-analysis, the world’s literature included only 151 cases of women who had surgical biopsy of DCIS that initially was misdiagnosed as a benign lesion and therefore didn’t undergo any further therapy, some with up to 31 years of follow-up. The long-term risk of invasive cancer in this cohort is only 22% (Breast Cancer Res. Treat. 2006;97:135-44). The annual risk of breast cancer in women with atypia is 1% per year, so this ends up looking very similar to the risk of progression that you see for atypical ductal hyperplasia or lobular carcinoma in situ.

The most common treatment for DCIS in the United States is lumpectomy with radiation. A meta-analysis by the European Breast Cancer Trialists’ Group found a 50% proportional reduction in local recurrence risk in women treated with lumpectomy and radiation vs. lumpectomy alone. The absolute magnitude of reduction was dependent on baseline recurrence risk. That’s a really important concept, because if your baseline risk is only 5%, then the 50% proportional reduction only translates into a 2.5% reduction in risk in 10 years.

The prospective, randomized Radiation Therapy Oncologists Group 9804 study randomized low-risk women with DCIS, unlike prior randomized trials that included broad eligibility criteria for DCIS. In 5 years of follow-up, the ipsilateral recurrence risk (which includes both invasive cancer and DCIS) was 3.2% in the lumpectomy-only group vs. 0.4% in the lumpectomy and radiation therapy group. The difference is highly statistically significant, however, given the small absolute difference between groups, the clinical significance certainly can be argued. There was no significant difference between groups in contralateral new primary lesions (Radiat. Oncol. 2012;84:S5).

We just presented a study this year looking at DCIS and competing causes of mortality in different age groups with different types of treatment. Among women with DCIS over 70 years of age, there was a significant difference in overall survival but none of the treatments conferred any benefit in disease-specific survival.

Active surveillance alone may be reasonable to consider in some patients. We can learn a lot from our colleagues who treat prostate cancer. They’re at least 10 years ahead of us in thinking about reducing the harms that have been introduced by screening for prostate cancer. When you look at breast cancer–specific and other-cause mortality in the setting of DCIS, the mortality curves look very similar to those with early-stage prostate cancer. Women with DCIS, regardless of what they are treated with, die of other causes and very rarely die of breast cancer.

When we compare such different options as mastectomy to active surveillance only, it’s really hard for a patient to feel that both of those approaches can offer them the same outcomes. But when you look at the actual data and evidence, the differences in recurrence or progression to invasive cancer don’t translate easily into a large difference in breast cancer mortality provided that patients are diagnosed at stage I or II, with more than a 90% survival from breast cancer.

We have an opportunity to take a big step back and redefine our goals of DCIS “treatment” from “curing” DCIS to trying to reduce breast cancer-specific mortality. This will help us reduce the harms that can result from screening and will reserve our aggressive treatments for those most likely to benefit from them.

Dr. Hwang is a professor of surgery at Duke University, Durham, N.C. She serves as a consultant for Genomic Health.