I-SPY Trial Characterizes Tumor Biology and Response to Neoadjuvant Chemotherapy

Preliminary findings from the I-SPY trial of neoadjuvant chemotherapy are a testament to the complexity and heterogeneity of locally advanced tumors and validate the concept that "biology should dictate the 'what' and the 'how' of treatment," according to Laura Esserman, MD, Director of the Breast Cancer Center and Professor of Surgery and Radiation at the University of California, San Francisco School of Medicine. Nora Hylton, MD, was principal author of the study. The I-SPY trial is a multi-institutional study of locally advanced breast cancer integrating serial biopsy and MRI to measure response of tumors to neoadjuvant chemotherapy. The primary objective is to identify surrogate markers of response that are predictive of pathologic complete response (pCR) and survival in women with stage II and III breast cancer. "The I-SPY trial data set is an important resource to help us validate and generate hypotheses about resistance and biology in women with locally advanced breast cancer," Dr. Esserman said. The findings should enable clinicians someday to identify nonresponders early and design effective therapies for them. The data set will eventually be available to the public in the form of an integrative and clinically useful portal. The trial accrued 237 patients from nine centers. Patients underwent an anthracycline-based neoadjuvant regimen and had serial MRI and core biopsies performed at baseline, after one cycle, during treatment, and before surgery to identify markers of tumor response. In the current analysis of 222 patients, mean age was 49 and median tumor size was 6 cm. Seventy percent had positive nodes, 54% were estrogen receptor (ER)-positive, 28% had HER2 overexpression, 38% had intermediate-grade lesions, and 39% had high-grade lesions. "The biology of this population was different from the usual screened group," she said. Patients were younger, tended to have more ER– disease (45%), and tended to be HER+ (27%). Few patients had favorable recurrence scores or prognoses, according to clinical prognostic tools. Distribution of molecular subtypes by 161 cDNA showed the tumors to be luminal A (27%), luminal B (21%), HER2 (9%), basal (35%), and normal (8%). Interestingly, basal tumors had a pCR rate of 45%, while luminal A had a pCR rate of 0%. The overall pCR rate was 27%, and 84% of these patients also had a pCR in the lymph nodes. By HER2 and ER status, pCR rates were 28.1% in HER2+/ER+ patients, 48.1% in HER2+/ER– patients (without trastuzumab), 9.6% in HER2–/ER+ patients, and 35.7% in HER2–/ER– patients. MRI characterized the imaging pattern of the tumors as unicentric mass and well-defined margins (17%), multilobulated mass and well-defined margins (30%), area enhancement and irregular margins (nodules) (31%), area enhancement and irregular margins (no nodes) (14%), and spectal spreading (9%). Change in MRI volume proved to be the optimal imaging measure of tumor response. By MRI phenotype, breast-conserving therapy was most feasible in types 1 and 2. Investigators determined that residual cancer burden (RCB) may be a better method of assessing response than pCR. RCB integrates several pathologic features, including lymph node status, extent of tumor bed, tumor size, and tumor cellularity. Output is a continuous measure (RCB 0 vs 1 vs 2 vs 3) rather than a dichotomous "yes or no" measure, Dr. Esserman said. By pCR and RCB, I-SPY validated molecular predictors of response that clinicians are accustomed to using. The most highly predictive were ER and phosphorylated ER, HER2 and phosphorylated HER2, the Netherlands Cancer Institute 70-gene set, and the OncotypeDX recurrence score. Dr. Esserman said the I-SPY trial validates the need to individualize treatment early, according to the tumor biology. "The neoadjuvant setting may provide the right opportunity to rapidly optimize regimens and improve responses," she said.

Disclosures:

The author(s) have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.