8002 Background: Cytogenetic abnormalities (CA) are one of the most powerful prognostic factors in multiple myeloma (MM). In the FORTE study, carfilzomib-lenalidomide-dexamethasone induction/consolidation with ASCT (KRd_ASCT) significantly improved progression-free survival (PFS) vs KRd without ASCT (KRd12, HR 0.64) or carfilzomib-cyclophosphamide-dexamethasone (KCd) plus ASCT (KCd_ASCT, HR 0.53). KR maintenance significantly improved PFS vs R (HR 0.63). Methods: MM patients (pts) were randomized to KRd_ASCT vs KCd_ASCT vs KRd12 and, thereafter, to KR vs R maintenance. Subgroup analyses according to FISH evaluated the impact of each single high-risk (HiR) CA [del17p, t(4;14), t(14;16), del1p and 1q gain (3 copies) or amp1q (≥4 copies)] and that of combined CA, defining HiR by the presence of ≥1 HiR CA and double-hit (DH) by the presence of ≥2 HiR CA. Pts negative for all the HiR CA were considered at standard risk (SR). The primary objective was the impact of treatment on PFS according to pt risk. Results: 396 out of 474 enrolled pts were included in the analysis with complete FISH data: 243 HiR, 105 DH and 153 SR. Among HiR pts, 60 had del17p, 65 t(4;14), 20 t(14;16), 44 del1p, 126 1q gain and 49 amp1q. SR pts benefited from intensification with KRd_ASCT vs KRd12 (HR 0.47, p = 0.05) and KCd_ASCT (HR 0.38, p = 0.01), with a 4-year PFS of 80%, 67% and 57%, respectively. In HiR pts, KRd_ASCT improved PFS vs KRd12 (HR 0.6, p = 0.04) and KCd_ASCT (HR 0.57, p = 0.01), with a 4-year PFS of 62%, 45% and 45%, respectively. The advantage with KRd_ASCT vs KRd12 (HR 0.53, p = 0.07) and KCd_ASCT (HR 0.49; p = 0.03) was also observed in DH pts (4-year PFS 55%, 31% and 33%, respectively). Analyses by single CA were limited by the small number of pts in each subgroup, but a trend towards a PFS benefit from KRd_ASCT vs KRd12 was seen in pts with del17p (HR 0.61, p = 0.3), t(4;14) (HR 0.59, p = 0.2) and 1q gain (HR 0.45, p = 0.02). In pts with del1p, KRd_ASCT (HR 0.24, p = 0.06) and KRd12 (HR 0.33, p = 0.09) showed superiority over KCd_ASCT, while amp1q pts had the worst outcome regardless of treatment (KRd_ASCT vs KCd_ASCT, HR 1.16, p = 0.73; KRd12 vs KCd_ASCT, HR 1.34, p = 0.45). KR improved PFS vs R in SR (3-year PFS 90% vs 73%, HR 0.42, p = 0.06), HiR (3-year PFS 69% vs 56%, HR 0.6, p = 0.04) and DH pts (3-year PFS 67% vs 42%, HR 0.53, p = 0.1). Despite the small subgroups, a beneficial trend with KR vs R was observed in pts with del17p (HR 0.59, p = 0.37), t(4;14) (HR 0.59, p = 0.3), 1q gain (HR 0.54, p = 0.07) and del1p (HR 0.23, p = 0.08), while amp1q pts showed the worst outcome and no benefit from KR vs R (HR 0.83, p = 0.7). Conclusions: KRd_ASCT and KR maintenance are highly effective in SR and also in HiR and DH pts, with impressive 4-year PFS from diagnosis (KRd_ASCT: HiR 62%, DH 55%) and 3-year PFS from maintenance (KR: HiR 69%, DH 67%), thus supporting their use in HiR pts, who represent an unmet medical need. Clinical trial information: NCT02203643.