Adding a New Dimension to the Diagnosis & Management of Breast Cancer
A new day is dawning for breast cancer diagnosis, treatment and monitoring with the help of molecular imaging.
Anatomical screening mammography will likely remain the diagnostic foundation for initial breast cancer diagnosis. But radiologists will increasingly apply new molecular imaging strategies based on the genetic underpinning of the deadly disease that kills about 460,000 women across the world each year, including about 40,000 women in the U.S. Breast cancer is the leading cancer killer among women aged 20 to 59 years in high-income countries, according to the World Health Organization.
MI researchers and clinicians are responding to the critics of mammography who find diagnosis, based solely on the anatomic presence of disease, too often leads to the over-treatment of indolent cancers and the inability to detect aggressive, interval cancers that often lead to breast cancer mortality.
Although x-ray screening mammography is only 38 percent to 41 percent sensitive to the presence of invasive cancer for all women, its performance is even worse for women with dense breasts. Diagnosis is difficult because the presentation of cancer and normal dense tissue are often identical, according to Deborah J. Rhodes, MD, an assistant professor of medicine at Mayo Clinic in Rochester, Minn. Accurate readings are further complicated by the inconsistent appearance of breast anatomy from woman to woman.
The value of anatomic MRI to address these limitations was demonstrated with its adoption as a screening tool for young women who are susceptible to breast cancer from BRCA1/BRCA2 mutations or strong familial histories of breast cancer. But MRI's high false-positive rate has encouraged researchers to seek out better solutions.
The first, practical, molecular alternatives have taken the familiar form of Tc-99m-sestambi SPECT and F18 FDG PET. After the false start of scintimammography performed on whole-body SPECT cameras in the 1990s, two high-resolution gamma cameras designed for breast imaging have rekindled interest.
Dedicated breast imaging comes in three variations: molecular breast imaging (MBI), breast-specific gamma imaging (BSGI) and positron emission mammography (PEM). MBI and BSGI both employ light breast compression and are paired with technetium-99m-sestamibi for clinical workups. They are positioned as an alternative to breast MRI for screening women with dense breasts or genetic susceptibility to breast cancer. No definitive trial has been published for either technology, but single-center trial experiences have been encouraging.
In January, a proof-of-principle trial involving 936 women with dense or heterogeneous breast demonstrated that MBI detected cancers that screening mammography missed for these at-risk populations. The diagnostic yield for MBI was 9.6 cancers per 1000 compared with 3.2 per 1000 for mammography. In combination, they identified 10.7 per 1000 (Radiology 2010; 258(1): 106-118).
Rachel Brem, MD, at George Washington University Medical Center, reported from a retrospective study of 146 consecutive, symptomatic women in 2008 that BSGI detected 80 to 83 malignant lesions with a sensitivity of 96.4 percent, and correctly classified 50 of 84 nonmalignant lesions as negative for cancer. It was 59.5 percent specific for differentiating between malignant and benign lesions. The positive predictive value for correctly identifying the cancerous status of 80 of 114 malignant lesions was 68.8 percent. The NPV for correctly assessing 50 of 53 nonmalignant lesions was 94.3 percent (Radiology 2008; 247(3); 651-657).
After a positive diagnosis with mammography and ultrasound, BSGI or MBI may be ordered to look for multifocal and multicentric breast cancer before surgery. After treatment, it is performed to differentiate between malignant recurrence and scarring from previous surgery, says Lillian Stern, MD, director of the Women's Health Center at Methodist Hospital in Philadelphia. It also may be used when results from mammography and ultrasound conflict with physical symptoms.
Marcela Böhm-Velez, MD, president of Weinstein Imaging Associates in Pittsburgh, performs BSGI after indeterminate mammography for symptomatic patients. BSGI's extremely high negative predictive value reduces diagnostic uncertainty. "From a negative [BSGI] exam, I can tell a patient with 95 percent confidence that she doesn't have cancer," she says.
Positron emission mammography (PEM) uses a breast-specific PET camera coupled with F18 FDG. In 2006, a multi-center prospective trial involving 94 consecutive patients with known suspicious breast cancers established FDG-PEM as 90 percent sensitive and 86 percent specific for detecting the disease. For three patients, cancer foci were only identified on PEM leading to a substantial changes in their clinical management, according to lead author Wendie A. Berg, MD, PhD (Breast Journal 2006: 12(4): 309-323).
A recent comparative effective study determined that FDG-PEM and MRI were about equally capable for presurgical planning involving the identification of cancers in the ipsilateral breast (Radiology 2011; 258(1): 59-72).
"You will see a whole lot more cancers than with ultrasound or mammography," Berg says. "Compared with MRI, PEM is less likely to lead to unnecessary biopsies."
Radiation exposure may be a limiting factor for all three molecular imaging platforms. The magnitude of the problem was identified in 2010 in a critique by medical physicist R. Edward Hendrick, PhD. He reported that a single BSGI or PEM exam exposes women to as much radiation as all of the screening mammography studies they will probably receive in their lifetimes (Radiology 2010; 257(1): 246-253).
BSGI and PEM involve effective doses of 6.2 mSv and 9.4 mSv, respectively, according the study. Initial protocols for MBI called for 20 mCi of fluorine-18, a dose equivalent to 6.5 mSv, Rhodes says.
Although each manufacturer is modifying their cameras to reduce dose, the issue is most critical for MBI as the hope is it will be used as an alternative to MRI for serial screening.
After two years of development, which began before publication of Hendrick's critique, new collimator and other improvements on the MBI device dropped the required radioisotope dose to 4 mCi, Rhodes says. Results from a large screening trial, scheduled for release this spring, will demonstrate that the new design performs better clinically than the original configuration, while exposing the patient to far less radiation, she says.
"Our preliminary analysis is far better than we had hoped," she says. "Of all the cancers detected with MBI in our trial, mammography has detected none."
A dose reduction also may be on the horizon for BSGI. In April, Böhm-Velez and colleagues presented a study that found a 60 percent dose reduction may be possible by scanning with less than the traditional dose of 20 mCi (note: this is an off-label use).
Such phenotyping has led to new tactics for administering existing chemotherapies and the development of new pharmacological compounds designed to attack specific genetic expressions of the disease, according to David Mankoff, MD, PhD, a professor of radiology at the University of Washington.
The growing array of targeted treatment options has drawn attention to the need for accurate assessments of early response. Numerous MI strategies are under investigation for either affirming that treatment is killing a breast cancer or warning the clinician about the lack of positive response—and thus when a change in therapy is needed.
A larger therapeutic pharmacopeia and noninvasive methods for measuring response have led to a broader adoption of neoadjuvant chemotherapy for locally aggressive or inflammatory breast cancers before surgery. In this setting, therapy is considered definitive, rather palliative because of limited spread, Mankoff says.
The protocol also is good for drug development because the response of primary tumors to therapies is easier to measure than with treated metastatic disease. The excised tumor provides pathological samples for confirming the accuracy of response measures acquired with MI. "From an imager's standpoint, it is rare that you get such a clean gold standard for assessing a response," Mankoff says.
If FDG uptake can serve as a surrogate marker for metabolism, then many MI researchers believe F-18 FLT also may measure response because of its ability track cell proliferation.
Promising preliminary experiences with FLT for monitoring neoadjuvant breast cancer therapy has led to a multi-center trial developed jointly by the National Cancer Institute's imaging program and the American College of Radiology Imaging Network (ACRIN).
ACRIN 6688 hypothesizes that changes in measured FLT uptake observed from a baseline PET study before therapy and a followup PET exam after the first round of treatment will identify patients who will have a complete response and those who will have residual tumor following treatment. Accrual for the trial is under way now for patients with 2 cm or larger primary breast tumors, but no metastatic disease. Mankoff is its principal investigator.
He also is involved with another NCI trial testing F18 alpha-fluorestradiol (FES) for predicting response to first-line hormone therapy in women with hormone receptor-positive metastatic breast cancer.
Fluoroestradiol is an analog of estradiol, a potent estrogen hormone that binds to estrogen surface cell receptors in ER-positive breast cancers. Previous studies have shown that FES binding rates correlate with the amount of estrogen receptor expressed in the tumor.
The trial aims at estimating the ability of F18 FES PET to predict overall response or to first-line endocrine therapy for metastatic breast cancer. More than half of the trial's subjects were recruited by mid-April. Preliminary results may be described this month at the American Society of Clinical Oncology meeting.
Based on this technique, DCE-MRI can measure microstructural remodeling associated with tumor cell death and the status of angiogenic processes that channel oxygenated blood to the tumor.
Completed in 2006, ACRIN 6657 demonstrated the predictive power of this approach. Based on the imaging experience of 237 women with large primary breast cancers, the study concluded that DCE-MRI produces a stronger and earlier prediction of response to therapy than changes observed from conventional anatomic measures tumor size from x-ray mammography.
The Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2 (I-SPY-II ) aims at testing whether the early response experimental model also can be applied during FDA phase II trials to evaluate the early effectiveness of investigational therapies. Five such agents are being tested.
Primary investigator Laura Esserman, MD, a professor of surgery at the University of California, San Francisco, has noted in written descriptions of the trial that less than one of three chemotherapies successfully pass FDA phase III trials. By eliminating the losers with DCE-MRI therapy monitoring at phase II level, Esserman hopes to raise the success rate at phase III to 90 percent.
It has become clear that 3-tesla MRI provides the minimum field strength for MRS breast studies, explains Robyn Birdwell, MD, head of the breast imaging division at the Dana Farber Cancer Center in Boston. Questions remain about the robustness of single-voxel MRS, the effect of differing gadolinium contrast formulations, and the spectral distortions that could arise from clips placed in the breast during core biopsies.
Yet, changes in the spectrum around the choline metabolite peak acquired from breast MRS may provide a noninvasive measure of metastatic spread, Birdwell says.
"There is great excitement about MRS," she says.
Birdwell also is optimistic about the potential of many MI techniques for assessing therapeutic effectiveness. Methods for working with the imaging data are still crude, and more research is needed to build confidence in the procedures' accuracy and reproducibility. Yet, the prospects for noninvasive imaging techniques that assess tumor response within days of the initiation of therapy are extremely attractive, Bird says.
"All of these [molecular] entities have the potential to answer these questions faster than is now possible," she says.
Anatomical screening mammography will likely remain the diagnostic foundation for initial breast cancer diagnosis. But radiologists will increasingly apply new molecular imaging strategies based on the genetic underpinning of the deadly disease that kills about 460,000 women across the world each year, including about 40,000 women in the U.S. Breast cancer is the leading cancer killer among women aged 20 to 59 years in high-income countries, according to the World Health Organization.
MI researchers and clinicians are responding to the critics of mammography who find diagnosis, based solely on the anatomic presence of disease, too often leads to the over-treatment of indolent cancers and the inability to detect aggressive, interval cancers that often lead to breast cancer mortality.
Although x-ray screening mammography is only 38 percent to 41 percent sensitive to the presence of invasive cancer for all women, its performance is even worse for women with dense breasts. Diagnosis is difficult because the presentation of cancer and normal dense tissue are often identical, according to Deborah J. Rhodes, MD, an assistant professor of medicine at Mayo Clinic in Rochester, Minn. Accurate readings are further complicated by the inconsistent appearance of breast anatomy from woman to woman.
The value of anatomic MRI to address these limitations was demonstrated with its adoption as a screening tool for young women who are susceptible to breast cancer from BRCA1/BRCA2 mutations or strong familial histories of breast cancer. But MRI's high false-positive rate has encouraged researchers to seek out better solutions.
The first, practical, molecular alternatives have taken the familiar form of Tc-99m-sestambi SPECT and F18 FDG PET. After the false start of scintimammography performed on whole-body SPECT cameras in the 1990s, two high-resolution gamma cameras designed for breast imaging have rekindled interest.
Dedicated breast imaging comes in three variations: molecular breast imaging (MBI), breast-specific gamma imaging (BSGI) and positron emission mammography (PEM). MBI and BSGI both employ light breast compression and are paired with technetium-99m-sestamibi for clinical workups. They are positioned as an alternative to breast MRI for screening women with dense breasts or genetic susceptibility to breast cancer. No definitive trial has been published for either technology, but single-center trial experiences have been encouraging.
In January, a proof-of-principle trial involving 936 women with dense or heterogeneous breast demonstrated that MBI detected cancers that screening mammography missed for these at-risk populations. The diagnostic yield for MBI was 9.6 cancers per 1000 compared with 3.2 per 1000 for mammography. In combination, they identified 10.7 per 1000 (Radiology 2010; 258(1): 106-118).
Rachel Brem, MD, at George Washington University Medical Center, reported from a retrospective study of 146 consecutive, symptomatic women in 2008 that BSGI detected 80 to 83 malignant lesions with a sensitivity of 96.4 percent, and correctly classified 50 of 84 nonmalignant lesions as negative for cancer. It was 59.5 percent specific for differentiating between malignant and benign lesions. The positive predictive value for correctly identifying the cancerous status of 80 of 114 malignant lesions was 68.8 percent. The NPV for correctly assessing 50 of 53 nonmalignant lesions was 94.3 percent (Radiology 2008; 247(3); 651-657).
After a positive diagnosis with mammography and ultrasound, BSGI or MBI may be ordered to look for multifocal and multicentric breast cancer before surgery. After treatment, it is performed to differentiate between malignant recurrence and scarring from previous surgery, says Lillian Stern, MD, director of the Women's Health Center at Methodist Hospital in Philadelphia. It also may be used when results from mammography and ultrasound conflict with physical symptoms.
Marcela Böhm-Velez, MD, president of Weinstein Imaging Associates in Pittsburgh, performs BSGI after indeterminate mammography for symptomatic patients. BSGI's extremely high negative predictive value reduces diagnostic uncertainty. "From a negative [BSGI] exam, I can tell a patient with 95 percent confidence that she doesn't have cancer," she says.
Positron emission mammography (PEM) uses a breast-specific PET camera coupled with F18 FDG. In 2006, a multi-center prospective trial involving 94 consecutive patients with known suspicious breast cancers established FDG-PEM as 90 percent sensitive and 86 percent specific for detecting the disease. For three patients, cancer foci were only identified on PEM leading to a substantial changes in their clinical management, according to lead author Wendie A. Berg, MD, PhD (Breast Journal 2006: 12(4): 309-323).
A recent comparative effective study determined that FDG-PEM and MRI were about equally capable for presurgical planning involving the identification of cancers in the ipsilateral breast (Radiology 2011; 258(1): 59-72).
"You will see a whole lot more cancers than with ultrasound or mammography," Berg says. "Compared with MRI, PEM is less likely to lead to unnecessary biopsies."
Radiation concerns
 | |
| From left: Dynamic contrast-enhanced MRI (DCE-MRI) tracks a positive therapeutic response before, after one cycle of chemotherapy and after a full course of chemotherapy. DCE-MRI will be performed at four time points during the I-SPY II trial. It will determine if the protocol can investigational chemotherapy during FDA drug trials. Source: Radiological Society of North America |
BSGI and PEM involve effective doses of 6.2 mSv and 9.4 mSv, respectively, according the study. Initial protocols for MBI called for 20 mCi of fluorine-18, a dose equivalent to 6.5 mSv, Rhodes says.
Although each manufacturer is modifying their cameras to reduce dose, the issue is most critical for MBI as the hope is it will be used as an alternative to MRI for serial screening.
After two years of development, which began before publication of Hendrick's critique, new collimator and other improvements on the MBI device dropped the required radioisotope dose to 4 mCi, Rhodes says. Results from a large screening trial, scheduled for release this spring, will demonstrate that the new design performs better clinically than the original configuration, while exposing the patient to far less radiation, she says.
"Our preliminary analysis is far better than we had hoped," she says. "Of all the cancers detected with MBI in our trial, mammography has detected none."
A dose reduction also may be on the horizon for BSGI. In April, Böhm-Velez and colleagues presented a study that found a 60 percent dose reduction may be possible by scanning with less than the traditional dose of 20 mCi (note: this is an off-label use).
Response to therapy
All breast cancers are not created equal because of their genetic makeup. This understanding has led oncologists to reclassify breast cancers by their genetic phenotypes, based on the presence or absence of hormone cell surface receptors (estrogen receptor [ER], progesterone receptor [PR]), and HER2/neu.Such phenotyping has led to new tactics for administering existing chemotherapies and the development of new pharmacological compounds designed to attack specific genetic expressions of the disease, according to David Mankoff, MD, PhD, a professor of radiology at the University of Washington.
The growing array of targeted treatment options has drawn attention to the need for accurate assessments of early response. Numerous MI strategies are under investigation for either affirming that treatment is killing a breast cancer or warning the clinician about the lack of positive response—and thus when a change in therapy is needed.
A larger therapeutic pharmacopeia and noninvasive methods for measuring response have led to a broader adoption of neoadjuvant chemotherapy for locally aggressive or inflammatory breast cancers before surgery. In this setting, therapy is considered definitive, rather palliative because of limited spread, Mankoff says.
The protocol also is good for drug development because the response of primary tumors to therapies is easier to measure than with treated metastatic disease. The excised tumor provides pathological samples for confirming the accuracy of response measures acquired with MI. "From an imager's standpoint, it is rare that you get such a clean gold standard for assessing a response," Mankoff says.
Imaging trials
Quantitative FDG-PET has shown promise for measuring early response in numerous single-center trials. It is particularly adept at predicting complete response and longer-term outcomes, such a disease free-survival and time to progression, Mankoff says.If FDG uptake can serve as a surrogate marker for metabolism, then many MI researchers believe F-18 FLT also may measure response because of its ability track cell proliferation.
Promising preliminary experiences with FLT for monitoring neoadjuvant breast cancer therapy has led to a multi-center trial developed jointly by the National Cancer Institute's imaging program and the American College of Radiology Imaging Network (ACRIN).
ACRIN 6688 hypothesizes that changes in measured FLT uptake observed from a baseline PET study before therapy and a followup PET exam after the first round of treatment will identify patients who will have a complete response and those who will have residual tumor following treatment. Accrual for the trial is under way now for patients with 2 cm or larger primary breast tumors, but no metastatic disease. Mankoff is its principal investigator.
He also is involved with another NCI trial testing F18 alpha-fluorestradiol (FES) for predicting response to first-line hormone therapy in women with hormone receptor-positive metastatic breast cancer.
Fluoroestradiol is an analog of estradiol, a potent estrogen hormone that binds to estrogen surface cell receptors in ER-positive breast cancers. Previous studies have shown that FES binding rates correlate with the amount of estrogen receptor expressed in the tumor.
The trial aims at estimating the ability of F18 FES PET to predict overall response or to first-line endocrine therapy for metastatic breast cancer. More than half of the trial's subjects were recruited by mid-April. Preliminary results may be described this month at the American Society of Clinical Oncology meeting.
I-SPY II coming soon
Dynamic contrast-enhanced (DCE-MRI) typically involves a series of heavily T1-weighted MR images acquired from a specific region of interest after gadolinium contrast injection. During DCE-MRI, each voxel is assumed to consist of two compartments, one vascular and other extravascular, to consider the rate at which contrast medium leaks into extravascular space.Based on this technique, DCE-MRI can measure microstructural remodeling associated with tumor cell death and the status of angiogenic processes that channel oxygenated blood to the tumor.
Completed in 2006, ACRIN 6657 demonstrated the predictive power of this approach. Based on the imaging experience of 237 women with large primary breast cancers, the study concluded that DCE-MRI produces a stronger and earlier prediction of response to therapy than changes observed from conventional anatomic measures tumor size from x-ray mammography.
The Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2 (I-SPY-II ) aims at testing whether the early response experimental model also can be applied during FDA phase II trials to evaluate the early effectiveness of investigational therapies. Five such agents are being tested.
Primary investigator Laura Esserman, MD, a professor of surgery at the University of California, San Francisco, has noted in written descriptions of the trial that less than one of three chemotherapies successfully pass FDA phase III trials. By eliminating the losers with DCE-MRI therapy monitoring at phase II level, Esserman hopes to raise the success rate at phase III to 90 percent.
MR spectroscopy
The promise of proton MR spectroscopy (MRS) to predict response offers a strong rationale for researchers to chip away at the technical problems that complicate its use.It has become clear that 3-tesla MRI provides the minimum field strength for MRS breast studies, explains Robyn Birdwell, MD, head of the breast imaging division at the Dana Farber Cancer Center in Boston. Questions remain about the robustness of single-voxel MRS, the effect of differing gadolinium contrast formulations, and the spectral distortions that could arise from clips placed in the breast during core biopsies.
Yet, changes in the spectrum around the choline metabolite peak acquired from breast MRS may provide a noninvasive measure of metastatic spread, Birdwell says.
"There is great excitement about MRS," she says.
Birdwell also is optimistic about the potential of many MI techniques for assessing therapeutic effectiveness. Methods for working with the imaging data are still crude, and more research is needed to build confidence in the procedures' accuracy and reproducibility. Yet, the prospects for noninvasive imaging techniques that assess tumor response within days of the initiation of therapy are extremely attractive, Bird says.
"All of these [molecular] entities have the potential to answer these questions faster than is now possible," she says.