USE OF BREAST MRI FOR CHEMOTHERAPY RESPONSE PREDICTION

May 23, 2012 — Breast magnetic resonance imaging (MRI) is more powerful than a conventional clinical assessment for predicting, after a single round of neoadjuvant chemotherapy, whether treatment will lead to a complete response of a patient's locally advanced breast cancer before surgery.

The findings were published online May 23 and in the June issue of Radiology.

Many academic cancer treatment facilities have adopted neoadjuvant chemotherapy guided by MRI response monitoring for locally advanced breast cancers, according to lead study author Nola M. Hylton, PhD, from the Department of Radiology, University of California, San Francisco. Treatments aim at shrinking locally advanced breast tumors to reduce the amount of breast excised during corrective breast surgery while increasing the likelihood of the patient's long-term survival. "Because MRI helps detect breast cancer by demonstrating contrast enhancement associated with tumor angiogenesis," Dr. Hylton and colleagues write, "it provides insight into tumor physiology and could provide an earlier and more accurate marker of tumor response than anatomic imaging or clinical breast examination."

Outperforming a conventional physical assessment was not a big challenge for MRI, Dr. Hylton told Medscape Medical News. It easily won that test while effectively addressing whether it could, after just 2 weeks of chemotherapy, predict whether individual patients would achieve a pathologic complete response before surgery was performed after chemotherapy was completed.

"The data here support the use of MRI for volumetric tumor measurement and capturing changes induced [by] treatment," she said.

Using data from American College of Radiology Imaging Network 6657, the imaging component of the 9-center Investigation of Serial Studies to Predict Your Therapeutic Response With Imaging and Molecular Analysis (I-SPY Trial) breast cancer trial, the current analysis was based on clinical experience between 2002 and 2006 for a total of 216 women (after 21 exclusions). Their average age was 49 years (range, 26 - 68 years).

Women with invasive stage 2 to 3 breast cancer involving large primary tumors of 3 cm or larger received physical assessments and dynamic contrast-enhanced MRI exams at 4 points to monitor treatment. Baseline exams were performed within the 4 weeks before the start of anthracycline-cyclophosphamide chemotherapy. The exams were repeated at least 2 weeks after the first cycle and before a second cycle of anthracycline-cyclophosphamide was initiated. Another assessment was completed before a third cycle of anthracycline-cyclophosphamide treatment and taxane therapy, if taxane was administered. All patients again received a final MRI and physical assessments between the end of chemotherapy and breast-conserving surgery.

Clinical assessments included the recording of tumor size in centimeters, measured in 1 dimension, as well as the tumor location, based on its distance in centimeters from the nipple and clock position. The change in clinical size was based on the largest change in a single dimension.

A complete response involved the disappearance of all lesions. A partial response was at least a 30% decrease in the longest diameter of the primary tumor. Stable disease described cases that were not defined as partial responses or progressive disease. Progressive disease involved at least a 20% increase in the longest dimension of the primary tumor.

The interpretation of dynamic contrast-enhanced MRI included measuring lesion size, shape, extent, distribution, and kinetics. Volume was calculated according to the tumor's detectible anatomic margins, plus physiological information about the full extent of the tumor based on dynamic gadolinium contrast uptake.

The key to the effectiveness of the 2 tests revolved around their predictive power after the first round of chemotherapy, about 2 weeks after the start of the nearly 6-month treatment regime, noted Harry D. Bear, MD, chair of surgical oncology at the Massey Cancer Center of Virginia Commonwealth University in Richmond, who was not involved in the study.

"The most important imaging is done after the first cycle," Dr. Bear explained to Medscape Medical News. "You would like to figure out whether the treatment you are giving is futile and whether you should do something else, such [as] going on to surgery or switch to another therapy."

At the time of surgery, 26% of patients demonstrated a pathologic complete response.

An area under the curve (AUC) statistical analysis revealed that the measures of tumor size drawn from clinical examination were significantly less accurate than MRI for predicting complete response at all points than the measure of the longest tumor dimension, tumor volume, and signal enhancement ratio acquired with dynamic contrast enhancement MRI (P ≤ .05)

A similar pattern appeared in an AUC analysis for the prediction of residual cancer burden. It also showed that the longest tumor dimension, tumor volume, and signal enhancement ratio with dynamic contrast-enhanced MRI were independently better predictors of complete response before surgery than the measurement of clinical size from physical examination (P ≤ .05).

"These findings support the hypothesis that volumetric assessment is a more accurate measure of tumor burden than diameter measurement. It gives the clinician an earlier measure of treatment response," Dr. Hylton said.

However, Dr. Bear noted that the study said nothing about the ability of breast MRI to provide the oncologist with early guidance about whether chemotherapy will lead to a partial response, stable disease, or progressive disease.

"It is not necessarily a good idea to stop treatment on patients who you don't think will have a pathologic complete response because it may be doing them some good anyway," he said.

Dr. Bear, who is cochair of an ongoing American College of Radiology Imaging Network trial that is examining the potential of fluorine-18 thymidine positron emission tomography for predicting the response of localized breast cancers to neoadjuvant chemotherapy before breast surgery, was also critical of Dr. Hylton's summary dismissal of diagnostic ultrasound as an alternative to MRI for response prediction. He noted that early work on biological assays has shown promise in this role.

Dr. Hylton noted the study represents just a starting point for more published research to be derived from I-SPY-1 and I-SPY-2, a similar multicenter clinical trial that is currently recruiting patients. I-SPY-2 will examine the options for new different types of neoadjuvant chemotherapies monitored with breast MRI.

"It will hopefully give us information that can help bolster clinical decision-making in regard [to] if a given patient is having a good response or is not having a good response," she said.

The study was funded by the National Cancer Institute. Dr. Hylton and Dr. Bear have disclosed no relevant financial relationships. Full conflict of interest information for the authors is available in the article.

Radiology. 2012:263:663-672.