Highlights of the American Society of Breast Surgeons 2008 Meeting
Highlights of the American Society of Breast Surgeons 2008 Meeting
At 9th Annual Meeting of the American Society of Breast Surgeons, held in New York City April 30-May 4, 2008, experts in the field of breast cancer, including surgeons, oncologists, radiologists, and pathologists, convened to discuss significant updates in breast cancer treatment and diagnosis and to review the evolution of changes in practice.
Pathology has taken on an incrementally larger role in the field of breast cancer since the early 1980s, which was referred to by Melvin J. Silverstein, MD, breast surgeon at Hoag Hospital, Los Angeles, as "the stone age" of breast cancer, when radical or modified radical mastectomies were performed on everybody, and when margins were irrelevant. Today, it is unheard of to consider breast surgery without also thinking about margins. Pathologists and surgeons work hand-in-hand, with surgeons attempting to conserve as much breast tissue as possible while assuring complete removal of cancer so that pathologists can accurately determine malignancies in the excised tissue.
Along with breast-conserving therapy, systemic therapies have evolved, and the options available now extend beyond radiation therapy and standard chemotherapy. The advent of our genomic era has enabled a more individualized diagnosis of disease at the molecular and cellular level, leading to more tailored therapies that now include antiestrogens and biologicals, such as trastuzumab (Herceptin), which form an important component of today's therapeutic armamentarium. As a result, pathology has come to play an increasingly important role in diagnosing breast cancer and determining and predicting the type of treatment most suitable for the individual patient with breast cancer.
In his keynote address, Larry S. Norton, MD, Medical Director of Clinical Oncology of Memorial Sloan-Kettering Cancer Center, New York, stated that molecular information theory has become the predominant concept for explaining metastatic disease, expanding on the steam engine pressure theory of cancer, prevalent since the 19th century, which sought a purely mechanistic explanation for the spread of disease beyond a local point of origin.
"We think we are getting really smart, we are looking at the molecules that are affecting signaling, that affect metastatic behavior and that may affect response to drug therapy. It's molecules that determine the likelihood of spread to an organ," said Dr. Norton. "Molecules equal information and that is probably the current state of the art."
This application of information theory is based on a transformational concept of cellular signals, such as growth factors and their receptors, angiogenic molecules, cytokines, and others, that control critical functions involved in tumor growth and mestastasis, including cell division, angiogenesis, apoptosis, and tissue invasion.
An increasing array of new molecular techniques are available or are being developed that determine individual gene expression profiles, such as those of the HER-2/neu (also known as c-erb-B2) oncogene, a member of the epidermal growth factor receptor and whose amplification is one of the most common genetic alterations associated with human breast cancer.
"We can show an interaction between these gene sets and a patient's responsiveness to therapy," said Dr. Norton. "On the basis of patients' gene expression profiles, patients can be stratified into risk categories and assigned to having a good prognosis or a not so good prognosis. It is known that some patients that fall into a low-risk category on the basis of their gene expression profile may not even benefit from chemotherapy." It is vital for pathologists who assist in determining the molecular characterization of breast cancer tissue to work closely with clinicians in helping with diagnosis, treatment, and prediction of outcome, with specific therapies for individual patients.
Dr. Kenneth J. Bloom, pathologist and Medical Director of CLARiENT, Inc, Aliso Viejo, California, gave an overview of the current molecular characterization techniques of breast cancer used in the pathology labs. The main groupings of molecular techniques are DNA and mRNA-based, but microRNA assays and a variety of methylation assays are also available.
"As pathologists, we are looking for very minimal-type changes, and molecular techniques help us with that," Dr. Bloom said. He mentioned that many of the techniques in use today have a very limited dynamic range and there are also problems with internal reference points --"something you hear with tissue fixation all the time," he says. There are always questions about how the tissue was fixed, how long it was handled, and how long it was devascularized. "We are also limited in the number of measurements that we can do on one tissue section. It would be nice to be able to do many things at the same time," according to Dr. Bloom.
The endpoint of pathologic testing is to improve correlation with outcome and to provide physicians with more clinical information to enable better patient care. "There are a lot of approaches I can take as a pathologist and I can't give a solution until you define the problem," said Dr. Bloom, stressing the general problem of relevance in pathologic measurements in terms of having a direct impact on clinical practice.
Of course, when looking at the inner workings of a cell, things are always more complicated than just one neat little pathway. "So, while it's nice that we can talk about HER2, when you talk out of context of what is going on in the full cell, you don't get the complete picture," says Dr. Bloom. "And when you look at the complicated picture it is actually no great surprise that we get great variability in how patients respond to therapies."
Fluorescent in situ hybridization (FISH) is a commoly used test that identifies the presence or absence of specific DNA sequences, and can be used to count the number of cancer cells in a tumor specimen. How is FISH performed in most labs? A pathologist circles the tumor area on a hematoxylin and eosin-stained slide, which he or she hands over to a technician, who then goes to a dark room and tries to count those cells under a high-power objective lens. "I would contend a significant portion of time they don't get the right cells under their eyes. FISH is a technique that requires that the correct cells get precisely counted." Dr. Bloom commented.
One challenge with FISH is determining the precise cutoff points that would constitute a risk for the overexpression of a particular gene. For testing the HER2 gene, the FISH cut-off level is 2.0, with 0 or 1+ indicating HER2 negative and 2+ suggesting postitive results. Pathology consensus guidelines created an "equivocal range" of between 1.8 and 2.2. Dr. Bloom asked, "Why is there a necessity for an equivocal range when all of the studies showed 2.0 as the cut off with anyone above 2 qualifying for therapy and anyone below not? When lab technicians go to examine the exact same slides and repeatedly try to count the exact same cells by FISH, you can never get more accurate than 0.2." This means that when a result is counted as 1.6, it could be 1.4 or 1.8, which explains the variability and poor sensitivity inherent in many genetic assays. "The single clinical decision point on whether you are going to treat or not has some variability associated with it," said Dr. Bloom.
Another common error, according to Dr. Bloom, is to mistake ductal carcinoma in situ (DCIS) for invasive cancer or to miss cancer entirely. FISH testing is critical, but it is important that it gets done right. "It's not the technique that's wrong, but how it gets applied in reality in a laboratory," he said. He predicts that the next generation of assays are going to be chromogenically based and will be regained by pathologists -- not lab technicians -- who will be making the assessments themselves under a light microscope. "I think we are going to see a wide variety of genes that might become important, as pathologists become familiar and more comfortable with the newer techniques," Dr. Bloom said.
Currently there is a large focus on the HER2 gene in breast cancer because of the development of trastuzumab, which targets this gene. However, a variety of other genes are also being investigated for therapeutic purposes, including the gene for the enzyme topoisomerase II (TOP2A), which is the target of anthrocycline-based chemotherapies.
Polymerase chain reaction (PCR) is considered the flagship method for DNA amplification. Multiplex PCR is an extension of the standard PCR protocol in which multiple loci are amplified simultaneously, They have a number of advantages including saving time, improving throughput, and reducing cost. Multiplex assays allow a detailed assessment of important pathways, provide improved risk stratification, and may help predict the benefit of chemotherapy. They can often outperform single markers, which are subject to day to day variations. However, they are only as good as the validation sets that are used, and the models aren't always intuitive. A big drawback to PCR is that it can't separate out tumor from non-tumor cells. "When you look at the result and look at the patients, you don't always know that things have synced up," said Dr. Bloom.
While these issues can be fixed by using a technique called 'laser capture microdisection' and picking out only the cells of interest, it is "crucial to know the substrate that you are putting into these assays to reliably believe the result," Dr. Bloom said." For example, ductal carcinoma in situ has a lot of heterogeneity in determining the grade with single field ducts."
"The biggest thing you are probably doing in your clinical practice with PRC today is the Oncotype DX assay," he commented and went on to point out that many clinicians are now using it to aid them in decisions involving ER [estrogen receptor]-positive, node-negative tumors, for example whether to give only an antiestrogen or to give an antiestrogen in addition to chemotherapeutic agents."
Oncotype DX is a reverse transcription PCR (RT-PCR) assay that analyzes the expression of 21 specific genes and provides a recurrence score, which correlates with the likelihood of distance recurrence over a 10-year period. It is especially useful for women with early stage (I or II) cancers that are estrogen-receptor positive and lymph node negative. This test is included in both the American Society of Clinical Oncology (ASCO) 2007 Update of Recommendations for the Use of Tumor Markers in Breast Cancer and The National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology Breast Cancer to identify subgroups of patients who may be successfully treated with hormonal therapy alone or who may need adjuvant chemotherapy. The Trial Assigning Individualized Options for Treatment (Rx) (TAILORx ) has been launched by the National Cancer Institute to test prospectively whether the Oncotype Dx can guide treatment selection of breast cancer in the adjuvant setting. Results won't be known, however, until 2013.
The Amsterdam 70-gene profile (MammaPrint) is another PCR-based assay currently used as a prognostic and predictive marker for stratifying risk. A trial using MammaPrint --The MINDACT (Microarray In Node-negative Disease may Avoid ChemoTherapy) -- is currently underway. Like the TAILORX trial, it will use subjects who have lymph node-negative disease. Unlike TAILORx, however, the MINDACT will also allow women who have estrogen receptor-negative disease and HER2-positive disease.
"For all types of new assays, the basic tenet is that it must be built on pre-existing knowledge," says Dr. Bloom. "Many oncologists are under the erroneous impression that new assays signify an improvement over older tested techniques. Improved quantification without localization is really problematic -- to get a FISH result without knowing what cells have been analyzed is problematic, and to get a PCR result with out understanding exactly what cells were analyzed gives you a score but you don't understand what that score really means."
Dr. Bloom pointed out that historically it has been shown that when localization is coupled with less sensitive methodologies, they tend to outperform more sensitive methodology with unlocalized techniques. In other words, it is more important to select the right cells and have a slightly less sensitive detection method than it is to raise the sensitivity with a new test and not understand what you are measuring. "I guess that sounds intuitive -- if it were only so simple," said Dr. Bloom.
Dr. Laura Liberman, Director of Breast Imaging Research Programs at Memorial Sloan-Kettering Cancer Center, New York, gave an overview of some of the salient issues that may lead to discordant and indeterminate core biospies from the radiologist perspective.
Discordance occurs when the histology on a pathologist's report does not provide a sufficient explanation for findings seen by the radiologist using image-guided biopsy. According to Dr. Liberman, when doing a core biospy or vacuum-assisted biopsy, it is rare that the pathologist is going to call it inadequate or insufficient. "If you take an adequate amount of tissue, they usually are going to come back with some diagnosis," she said. And yet, there are instances where the reports by the pathologist do not match up with the radiologist's perspective, leading to missed diagnoses of cancer or erroneous diagnosis of cancer where it isn't actually present.
Numerous studies have looked at the frequency of discordance. Some studies showed overall discordance rates ranging from 1% to 6%, with a median of about 3%. Data presented from a study conducted at Memorial Sloan-Kettering Cancer Center by Dr. Liberman and colleagues looked at almost 2000 lesions that had undergone stereotactic or ultrasound-guided biopsy, either with 14-gauge core automated instruments or 11-gauge vaccuum-assisted instruments. Results showed that a total of 3% were discordant; approximately one fourth of which were cancerous.
There are several factors that affected the discordance rate. Generally, there was a higher frequency of discordance if the person performing the biopsy had less experience, particularly in retrieving calcifications. "One of the common situations in which we consider the result discordant is if we fail to obtain calcifications on the x-rays or the pathologist fails to see calcifications on analysis," said Dr. Liberman.
The study also found a higher frequency of discordance among the lesions that are highly suggestive of cancer than in lesions that are less suspicious of malignancy.
"This is one of those situations where size does matter," she said. "If you're doing a biopsy for calcifications you are less likely to get discordance if you're doing it with an 11-inch vaccuum-assisted instruments, which takes specimens that are 100 mg each, than if you are doing it with a 14-inch automated core gauge instrument, which takes specimens that are about 20-35 mg each," according to Dr. Liberman.
The data showed that the lesion-type guidance modality -- stereo or ultrasound --or method of biopsy for mass lesions had no significant impact on the likelihood of discordance.
Dr. Liberman is in favor of repeat biopsies in the event of discordance, in particular in cases of suspected failure to retrieve calcifications.
Another challenge in diagnosis of breast cancer is with fibrotic changes. "Fibrosis can be a challenging lesion in image-guided core biopsy with any imaging modality," said Dr. Liberman. "Even though the numbers are small of lesions highly suspicious of cancer -- such as spiculated mass in which we got a diagnosis of fibrosis -- 40% of those were cancer at surgery. I would suggest that if you get a diagnosis of fibrosis in a lesion that you consider highly suggestive of malignancy that it be considered discordant and surgical excision is appropriate," she recommended.
Recently there has been an increase in the use of magnetic resonance imaging (MRI) in screening for breast cancer and taking biopsies. Unlike ultrasound-guided biopsy, "...when using MRI-guided imaging, once you're in there it's very hard to see the lesion," according to Dr. Liberman.
MRI's role in various situations is still being studied. At Memorial Sloan-Kettering a total of 300 lesions were examined in a 3-year period. Technically MRI-guided biopsies are more challlenging. "When you're doing an MRI biopsy it's a bit like beating the clock," says Dr. Liberman. "The areas light up, you see them at the beginning of the procedures and then it washes out, making it hard at times to get an adequate amount of tissue," she explains.
Discordance was found at higher rates with MRI, which may in part be attributed to a lack of experience by examiners in performing MRI-guided biopsies. Additionally, there are years and years of collected literature on what a fibroademona, for example, looks like in ultrasound and stereotactic biopsies, but there is a much less experience on what diagnoses look like on MRI. This leads to a low threshold to calling things discordant.
Introduction
At 9th Annual Meeting of the American Society of Breast Surgeons, held in New York City April 30-May 4, 2008, experts in the field of breast cancer, including surgeons, oncologists, radiologists, and pathologists, convened to discuss significant updates in breast cancer treatment and diagnosis and to review the evolution of changes in practice.
Pathology has taken on an incrementally larger role in the field of breast cancer since the early 1980s, which was referred to by Melvin J. Silverstein, MD, breast surgeon at Hoag Hospital, Los Angeles, as "the stone age" of breast cancer, when radical or modified radical mastectomies were performed on everybody, and when margins were irrelevant. Today, it is unheard of to consider breast surgery without also thinking about margins. Pathologists and surgeons work hand-in-hand, with surgeons attempting to conserve as much breast tissue as possible while assuring complete removal of cancer so that pathologists can accurately determine malignancies in the excised tissue.
Along with breast-conserving therapy, systemic therapies have evolved, and the options available now extend beyond radiation therapy and standard chemotherapy. The advent of our genomic era has enabled a more individualized diagnosis of disease at the molecular and cellular level, leading to more tailored therapies that now include antiestrogens and biologicals, such as trastuzumab (Herceptin), which form an important component of today's therapeutic armamentarium. As a result, pathology has come to play an increasingly important role in diagnosing breast cancer and determining and predicting the type of treatment most suitable for the individual patient with breast cancer.
In his keynote address, Larry S. Norton, MD, Medical Director of Clinical Oncology of Memorial Sloan-Kettering Cancer Center, New York, stated that molecular information theory has become the predominant concept for explaining metastatic disease, expanding on the steam engine pressure theory of cancer, prevalent since the 19th century, which sought a purely mechanistic explanation for the spread of disease beyond a local point of origin.
"We think we are getting really smart, we are looking at the molecules that are affecting signaling, that affect metastatic behavior and that may affect response to drug therapy. It's molecules that determine the likelihood of spread to an organ," said Dr. Norton. "Molecules equal information and that is probably the current state of the art."
This application of information theory is based on a transformational concept of cellular signals, such as growth factors and their receptors, angiogenic molecules, cytokines, and others, that control critical functions involved in tumor growth and mestastasis, including cell division, angiogenesis, apoptosis, and tissue invasion.
An increasing array of new molecular techniques are available or are being developed that determine individual gene expression profiles, such as those of the HER-2/neu (also known as c-erb-B2) oncogene, a member of the epidermal growth factor receptor and whose amplification is one of the most common genetic alterations associated with human breast cancer.
"We can show an interaction between these gene sets and a patient's responsiveness to therapy," said Dr. Norton. "On the basis of patients' gene expression profiles, patients can be stratified into risk categories and assigned to having a good prognosis or a not so good prognosis. It is known that some patients that fall into a low-risk category on the basis of their gene expression profile may not even benefit from chemotherapy." It is vital for pathologists who assist in determining the molecular characterization of breast cancer tissue to work closely with clinicians in helping with diagnosis, treatment, and prediction of outcome, with specific therapies for individual patients.
Molecular Characterization of Breast Cancer in the Pathology Lab
Dr. Kenneth J. Bloom, pathologist and Medical Director of CLARiENT, Inc, Aliso Viejo, California, gave an overview of the current molecular characterization techniques of breast cancer used in the pathology labs. The main groupings of molecular techniques are DNA and mRNA-based, but microRNA assays and a variety of methylation assays are also available.
"As pathologists, we are looking for very minimal-type changes, and molecular techniques help us with that," Dr. Bloom said. He mentioned that many of the techniques in use today have a very limited dynamic range and there are also problems with internal reference points --"something you hear with tissue fixation all the time," he says. There are always questions about how the tissue was fixed, how long it was handled, and how long it was devascularized. "We are also limited in the number of measurements that we can do on one tissue section. It would be nice to be able to do many things at the same time," according to Dr. Bloom.
The endpoint of pathologic testing is to improve correlation with outcome and to provide physicians with more clinical information to enable better patient care. "There are a lot of approaches I can take as a pathologist and I can't give a solution until you define the problem," said Dr. Bloom, stressing the general problem of relevance in pathologic measurements in terms of having a direct impact on clinical practice.
Of course, when looking at the inner workings of a cell, things are always more complicated than just one neat little pathway. "So, while it's nice that we can talk about HER2, when you talk out of context of what is going on in the full cell, you don't get the complete picture," says Dr. Bloom. "And when you look at the complicated picture it is actually no great surprise that we get great variability in how patients respond to therapies."
Fluorescent in Situ Hybridization Testing and Its Inherent Errors
Fluorescent in situ hybridization (FISH) is a commoly used test that identifies the presence or absence of specific DNA sequences, and can be used to count the number of cancer cells in a tumor specimen. How is FISH performed in most labs? A pathologist circles the tumor area on a hematoxylin and eosin-stained slide, which he or she hands over to a technician, who then goes to a dark room and tries to count those cells under a high-power objective lens. "I would contend a significant portion of time they don't get the right cells under their eyes. FISH is a technique that requires that the correct cells get precisely counted." Dr. Bloom commented.
One challenge with FISH is determining the precise cutoff points that would constitute a risk for the overexpression of a particular gene. For testing the HER2 gene, the FISH cut-off level is 2.0, with 0 or 1+ indicating HER2 negative and 2+ suggesting postitive results. Pathology consensus guidelines created an "equivocal range" of between 1.8 and 2.2. Dr. Bloom asked, "Why is there a necessity for an equivocal range when all of the studies showed 2.0 as the cut off with anyone above 2 qualifying for therapy and anyone below not? When lab technicians go to examine the exact same slides and repeatedly try to count the exact same cells by FISH, you can never get more accurate than 0.2." This means that when a result is counted as 1.6, it could be 1.4 or 1.8, which explains the variability and poor sensitivity inherent in many genetic assays. "The single clinical decision point on whether you are going to treat or not has some variability associated with it," said Dr. Bloom.
Another common error, according to Dr. Bloom, is to mistake ductal carcinoma in situ (DCIS) for invasive cancer or to miss cancer entirely. FISH testing is critical, but it is important that it gets done right. "It's not the technique that's wrong, but how it gets applied in reality in a laboratory," he said. He predicts that the next generation of assays are going to be chromogenically based and will be regained by pathologists -- not lab technicians -- who will be making the assessments themselves under a light microscope. "I think we are going to see a wide variety of genes that might become important, as pathologists become familiar and more comfortable with the newer techniques," Dr. Bloom said.
Currently there is a large focus on the HER2 gene in breast cancer because of the development of trastuzumab, which targets this gene. However, a variety of other genes are also being investigated for therapeutic purposes, including the gene for the enzyme topoisomerase II (TOP2A), which is the target of anthrocycline-based chemotherapies.
Polymerase Chain Reaction Assays
Polymerase chain reaction (PCR) is considered the flagship method for DNA amplification. Multiplex PCR is an extension of the standard PCR protocol in which multiple loci are amplified simultaneously, They have a number of advantages including saving time, improving throughput, and reducing cost. Multiplex assays allow a detailed assessment of important pathways, provide improved risk stratification, and may help predict the benefit of chemotherapy. They can often outperform single markers, which are subject to day to day variations. However, they are only as good as the validation sets that are used, and the models aren't always intuitive. A big drawback to PCR is that it can't separate out tumor from non-tumor cells. "When you look at the result and look at the patients, you don't always know that things have synced up," said Dr. Bloom.
While these issues can be fixed by using a technique called 'laser capture microdisection' and picking out only the cells of interest, it is "crucial to know the substrate that you are putting into these assays to reliably believe the result," Dr. Bloom said." For example, ductal carcinoma in situ has a lot of heterogeneity in determining the grade with single field ducts."
"The biggest thing you are probably doing in your clinical practice with PRC today is the Oncotype DX assay," he commented and went on to point out that many clinicians are now using it to aid them in decisions involving ER [estrogen receptor]-positive, node-negative tumors, for example whether to give only an antiestrogen or to give an antiestrogen in addition to chemotherapeutic agents."
Oncotype DX is a reverse transcription PCR (RT-PCR) assay that analyzes the expression of 21 specific genes and provides a recurrence score, which correlates with the likelihood of distance recurrence over a 10-year period. It is especially useful for women with early stage (I or II) cancers that are estrogen-receptor positive and lymph node negative. This test is included in both the American Society of Clinical Oncology (ASCO) 2007 Update of Recommendations for the Use of Tumor Markers in Breast Cancer and The National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology Breast Cancer to identify subgroups of patients who may be successfully treated with hormonal therapy alone or who may need adjuvant chemotherapy. The Trial Assigning Individualized Options for Treatment (Rx) (TAILORx ) has been launched by the National Cancer Institute to test prospectively whether the Oncotype Dx can guide treatment selection of breast cancer in the adjuvant setting. Results won't be known, however, until 2013.
The Amsterdam 70-gene profile (MammaPrint) is another PCR-based assay currently used as a prognostic and predictive marker for stratifying risk. A trial using MammaPrint --The MINDACT (Microarray In Node-negative Disease may Avoid ChemoTherapy) -- is currently underway. Like the TAILORX trial, it will use subjects who have lymph node-negative disease. Unlike TAILORx, however, the MINDACT will also allow women who have estrogen receptor-negative disease and HER2-positive disease.
"For all types of new assays, the basic tenet is that it must be built on pre-existing knowledge," says Dr. Bloom. "Many oncologists are under the erroneous impression that new assays signify an improvement over older tested techniques. Improved quantification without localization is really problematic -- to get a FISH result without knowing what cells have been analyzed is problematic, and to get a PCR result with out understanding exactly what cells were analyzed gives you a score but you don't understand what that score really means."
Dr. Bloom pointed out that historically it has been shown that when localization is coupled with less sensitive methodologies, they tend to outperform more sensitive methodology with unlocalized techniques. In other words, it is more important to select the right cells and have a slightly less sensitive detection method than it is to raise the sensitivity with a new test and not understand what you are measuring. "I guess that sounds intuitive -- if it were only so simple," said Dr. Bloom.
Avoidable Discordance From the Radiologist Perspective
Dr. Laura Liberman, Director of Breast Imaging Research Programs at Memorial Sloan-Kettering Cancer Center, New York, gave an overview of some of the salient issues that may lead to discordant and indeterminate core biospies from the radiologist perspective.
Discordance occurs when the histology on a pathologist's report does not provide a sufficient explanation for findings seen by the radiologist using image-guided biopsy. According to Dr. Liberman, when doing a core biospy or vacuum-assisted biopsy, it is rare that the pathologist is going to call it inadequate or insufficient. "If you take an adequate amount of tissue, they usually are going to come back with some diagnosis," she said. And yet, there are instances where the reports by the pathologist do not match up with the radiologist's perspective, leading to missed diagnoses of cancer or erroneous diagnosis of cancer where it isn't actually present.
Numerous studies have looked at the frequency of discordance. Some studies showed overall discordance rates ranging from 1% to 6%, with a median of about 3%. Data presented from a study conducted at Memorial Sloan-Kettering Cancer Center by Dr. Liberman and colleagues looked at almost 2000 lesions that had undergone stereotactic or ultrasound-guided biopsy, either with 14-gauge core automated instruments or 11-gauge vaccuum-assisted instruments. Results showed that a total of 3% were discordant; approximately one fourth of which were cancerous.
There are several factors that affected the discordance rate. Generally, there was a higher frequency of discordance if the person performing the biopsy had less experience, particularly in retrieving calcifications. "One of the common situations in which we consider the result discordant is if we fail to obtain calcifications on the x-rays or the pathologist fails to see calcifications on analysis," said Dr. Liberman.
The study also found a higher frequency of discordance among the lesions that are highly suggestive of cancer than in lesions that are less suspicious of malignancy.
"This is one of those situations where size does matter," she said. "If you're doing a biopsy for calcifications you are less likely to get discordance if you're doing it with an 11-inch vaccuum-assisted instruments, which takes specimens that are 100 mg each, than if you are doing it with a 14-inch automated core gauge instrument, which takes specimens that are about 20-35 mg each," according to Dr. Liberman.
The data showed that the lesion-type guidance modality -- stereo or ultrasound --or method of biopsy for mass lesions had no significant impact on the likelihood of discordance.
Dr. Liberman is in favor of repeat biopsies in the event of discordance, in particular in cases of suspected failure to retrieve calcifications.
Another challenge in diagnosis of breast cancer is with fibrotic changes. "Fibrosis can be a challenging lesion in image-guided core biopsy with any imaging modality," said Dr. Liberman. "Even though the numbers are small of lesions highly suspicious of cancer -- such as spiculated mass in which we got a diagnosis of fibrosis -- 40% of those were cancer at surgery. I would suggest that if you get a diagnosis of fibrosis in a lesion that you consider highly suggestive of malignancy that it be considered discordant and surgical excision is appropriate," she recommended.
Magnetic Resonance Imaging MRI-Guided Imaging for Biopsy.
Recently there has been an increase in the use of magnetic resonance imaging (MRI) in screening for breast cancer and taking biopsies. Unlike ultrasound-guided biopsy, "...when using MRI-guided imaging, once you're in there it's very hard to see the lesion," according to Dr. Liberman.
MRI's role in various situations is still being studied. At Memorial Sloan-Kettering a total of 300 lesions were examined in a 3-year period. Technically MRI-guided biopsies are more challlenging. "When you're doing an MRI biopsy it's a bit like beating the clock," says Dr. Liberman. "The areas light up, you see them at the beginning of the procedures and then it washes out, making it hard at times to get an adequate amount of tissue," she explains.
Discordance was found at higher rates with MRI, which may in part be attributed to a lack of experience by examiners in performing MRI-guided biopsies. Additionally, there are years and years of collected literature on what a fibroademona, for example, looks like in ultrasound and stereotactic biopsies, but there is a much less experience on what diagnoses look like on MRI. This leads to a low threshold to calling things discordant.
Some Tips on Minimizing the Frequency of Discordance
|
Source...