70-Gene Signature Predicts Time to Breast Ca Metastases

SAN ANTONIO—A 70-gene profile developed in Amsterdam to distinguish low-risk from high-risk women with node-negative breast cancer has been shown to be predictive for disease recurrence after surgery. Martine Piccart, MD, chair of the Breast International Group and head of the Medical Oncology Department, Institut Jules Bordet, Brussels, presented the results at the 27th Annual San Antonio Breast Cancer Symposium (abstract 38)

Dr. Piccart noted that the vast majority of women with node-negative breast cancer receive adjuvant chemotherapy, but few—20% to 30%—actually derive significant benefits. No criteria exist to prospectively distinguish patients who can safely avoid chemotherapy, she said. "The traditional approach to clinical research in cancer has been the comparison of treatments through randomized clinical trials, with the extrapolation of an average benefit in a patient population to each individual. This often means that many patients are overtreated, while only a few derive a considerable benefit," she noted.

Dutch researchers identified and validated a 70-gene prognostic signature, which has been developed into the diagnostic test MammaPrint, marketed by Agendia BV, Amsterdam. Prior to initiating the large EORTC prospective randomized clinical trial in node-negative patients, known as MINDACT (Micro-array in Node-Negative Disease May Avoid Chemotherapy), which will evaluate the utility of this tool in clinical practice, the TRANSBIG translational research group undertook a multicenter external validation study of the assay to test its performance in a broader patient population.

The researchers used the MammaPrint assay to analyze frozen archival primary tumor material from node-negative patients aged less than 60 years, diagnosed through 1998, who received locoregional therapy at six non-Dutch cancer centers. Among 291 patients followed for a median of 10 years, there were 100 loco-regional relapses, 75 distant relapses, and 76 deaths.

Hazard ratios were calculated to compare event rates in high-risk vs low-risk groups. Risk groups were defined on the basis of clinical/pathological data according to the 2003 St. Gallen criteria, the Nottingham Prognostic Index (NPI), and the Adjuvant! Online methods, and on the basis of the 70-gene signature. Gene expression profiling was performed in Amsterdam with no knowledge of clinical outcome data, and clinical risk groups were determined by the independent statistical office in Brussels.

Study Results

The results validated the prognostic potential of the MammaPrint assay. The gene signature hazard ratio for time to distant metastases in high-risk vs low-risk groups was 1.8, and for overall survival, 2.5. These hazard ratios are not as large as those first identified in the original validation series. "Nevertheless, the gene signature outperformed standard clinical prognostic models in predicting outcome," Dr. Piccart said.

Cross-classification of gene vs clinical risk groups revealed discordant results (see Table), primarily, overclassification of many women to the high-risk group, compared with their classification according to their gene signatures, Dr. Piccart reported.

By gene signature, out of 291 patients evaluated, 182 women (63%) were classified as high risk and 109 (37%) as low risk. By contrast, Adjuvant! Online assigned 246 (85%) as high risk (eg, predicted 10-year survival less than 90%) and 45 (15%) as low risk (eg, predicted 10-year survival of 90% or more); 80 women (33%) considered high risk by Adjuvant Online! were considered low risk by gene signature criteria, and 16 women (36%) assigned to low risk were considered high risk with the Mamma-Print assay.

The St. Gallen and NPI assigned a similarly larger proportion of patients to the high-risk group, compared with the gene test.

‘Green Light’ for MINDACT

While analyses are underway to determine why the performance in the external validation series was inferior to the original Amsterdam series, "the results are encouraging and provide level-3 evidence for the clinical value of this new genomic tool and ‘the green light’ to mobilize forces for the MINDACT trial," Dr. Piccart said.