Response to neoadjuvant therapy for breast cancer by magnetic resonance imaging type, estrogen receptor status, grade, and comparative genomic hybridization

L. J. Esserman, D. Sudilovsky, W. I. Kuo, J. Gray, N. Hylton

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


Introduction: Identifying characteristics of breast cancers that respond to neoadjuvant chemotherapy is a critical step toward tailoring therapies to maximize impact and minimize toxicity. Absence of estrogen receptor (ER) and presence of Her-2 has been shown to correlate with a good response to neoadjuvant therapy (nRx). We have identified several distinct MRI patterns in patients with locally advanced breast cancer (LABC) and sought to correlate known and new molecular markers with response to therapy. Methods: 46 patients with LABC had nRx with 4 cycles of Adriamycin (60 mg/m2) and Cytoxan (600 mg/m2). Serial breast MRI was used to estimate change in longest tumor diameter (LD). Tumors fell into 5 distinct categories and were characterized as types 1-5 (circumscribed mass, nodular pattern, diffuse pattern, patchy infiltration, septal spreading). Results: Residual disease at the time of surgical resection (a known predictor of outcome) correlated with post-therapy MRI LD with R2 of 0.94. Tumor grade correlated with MRI response. ER score (% positive cells x staining intensity) predicted response only in context of imaging type. Tumors of type 1 were highest grade, ER negative (72%), and had shrinkage sufficient for breast conservation (92%). Type 1 pattern strongly predicted response. Only 33%. demonstrated a partial or complete MR response (greater than one-third reduction in LD), but 87% of responders (13/15) had Type 1 tumors (p<0.001). 100% of ER negative Type 1 tumors had a partial or complete response, compared to 50% of ER positive Type 1, with least response in presence of strong ER. Her2 Ab staining did not predict for Type I phenotype or response to therapy, but FISH is pending. Type 2 (nodular pattern) were more likely ER positive (80%), but, in contrast to type 1, ER score correlated with a better response. Types 3 and 4 were uniformly ER positive, and a reduction in volume was more likely than a change in LD. Preliminary CGH on a small sample of cases suggests that chromosomal amplification patterns vary by response to therapy. Proliferative indices, Her-2 FISH, E Cadherin staining, and full sequencing on all patients will be presented and correlated to tumor type and response by change in LD as well as shrinkage by volume. Conclusions: Breast MR imaging helps to define distinct breast cancer patterns which correlate with pathology type, ER positively, and patterns of response.

Original languageEnglish (US)
Pages (from-to)245
Number of pages1
JournalBreast Cancer Research and Treatment
Issue number3
StatePublished - 2001
Externally publishedYes

ASJC Scopus subject areas

  • Oncology
  • Cancer Research


Dive into the research topics of 'Response to neoadjuvant therapy for breast cancer by magnetic resonance imaging type, estrogen receptor status, grade, and comparative genomic hybridization'. Together they form a unique fingerprint.

Cite this