EXPERIMENTAL DESIGN: To determine whether the clinical benefit of complete remission (CR) may depend on prognostic subgroups of patients with multiple myeloma. PATIENTS AND METHODS: Newly diagnosed patients with myeloma received a tandem autotransplant regimen. Using multivariate regression analyses, we examined the prognostic implications of time-dependent onset of CR on overall survival and event-free survival in the context of standard prognostic factors (SPF) and gene expression profiling-derived data available for 326 patients. RESULTS: CR benefited patients regardless of risk status when only SPFs were examined. With knowledge of gene array data, a survival (and event-free survival) benefit of CR only pertained to the small high-risk subgroup of 13% of patients (hazard ratio, 0.23; P = 0.001), whereas the majority of patients with low-risk disease had similar survival expectations whether or not CR was achieved (hazard ratio, 0.68; P = 0.128). CONCLUSIONS: Access to gene expression information permitted the recognition of a small very high-risk subgroup of 13% of patients, in whom prolonged survival critically depended on achieving CR. Absence of such benefit in the remainder should lead to a reassessment of clinical trial designs that rely on this end point as a surrogate for long-term prognosis.