Profile Including Clinical Data and Biologic Markers Predicts DFS

San Antonio, TX—A better prognostic profile incorporating clinical-pathologic data with biologic markers showed that in the BIG 1-98 trial, high-risk patients did best with 5 years of letrozole (Femara) and low-risk patients did similarly well with any of the trial’s regimens. “Assessing risk by combining multiple factors appears better than any single factor,” said lead investigator Giuseppe Viale, speaking for the BIG 1-98 Collaborative and International Breast Cancer Study Groups (IBCSG) at the CTRC-AACR San Antonio Breast Cancer Symposium 2009 Annual Meeting.

BIG-I-98 had treatment arms evaluating monotherapies (5 years of tamoxifen or letrozole) and sequential therapy (2 years of letrozole followed by 3 years of tamoxifen and 2 years of tamoxifen followed by 3 of letrozole) as adjuvant endocrine therapy for postmenopausal women with hormone receptor—positive breast cancer. After a median follow-up of 71 months, disease-free survival (DFS) was 87.6% for letrozole→tamoxifen and 86.2% for tamoxifen→letrozole. DFS in the monotherapy arms was 87.9% for letrozole and 84.0% for tamoxifen (data from patients in this group who crossed over selectively to letrozole were censored). Dr Viale noted that the modest differences in DFS between groups does not facilitate a clinician’s choices between treatment regimens. To identify patients who may respond better to a particular treatment regimen, the IBCSG took the same BIG 1-98 database (retrospective tissue collection) and first examined individual tumor markers in 5,177 patients using STEPP (Subpopulation Treatment Effect Pattern Plots).

Each of four tumor markers (estrogen receptor [ER], progesterone receptor [PgR], Ki-67 labeling index, and HER2 status) showed a trend favoring letrozole monotherapy at the higher risk end of its spectrum, but none was statistically significant as a predictive factor. Dr Viale and colleagues asked, “Could a composite prognostic risk profile based on St. Gallen risk factors also serve to predict treatment differences?”

A composite risk profile was calculated using a multivariable Cox model for DFS and using the St. Gallen risk factors (ER, PgR, HER2 Ki-67, number of involved metastatic lymph nodes, tumor grade, tumor size and presence of peritumoral vascular invasion [PVI]). The key contributors to high risk were higher number of positive lymph nodes, lower ER%, higher Ki-67, higher grade, larger tumor size, HER2-positivity, presence of PVI and lower pgR%.

Dr Viale reported that, as expected, as composite risk increased, 5-year DFS decreased for all treatment regimens. Also, while 5-year DFS is similar for all four treatment regimens at the lowest level of risk, inferior results for tamoxifen alone are seen earlier in the risk spectrum and separation among letrozole-containing treatments is apparent particularly at the highest risk levels. About a third of the BIG 1-98 population fell into each of low-, intermediate- and high-risk strata.

In individuals with low risk, all treatments including tamoxifen appeared similar; at the intermediate-level all letrozole arms appeared similar and superior to tamoxifen; at the high-risk level, 5-years of letrozole appeared better than either sequence. For high-risk patients, 5-year DFS was ~72% in the letrozole-alone group, ~67% in the tamoxifen→letrozole group, ~65% in the letrozole→tamoxifen group, and ~ 61% in the tamoxifen alone group.

Dr Viale concluded, “In the BIG 1-98 trial, a composite prognostic profile incorporating clinic-pathological data and biological markers was better able to predict the relative treatment benefit.” SABCS Abstract 76.