63P Ki67 values and MammaPrint (MP) genomic risk in early breast cancer: Correlation by ductal vs lobular histology

In MINDACT trial, 4 risk categories (RC) by clinical (c) and genomic (g) breast cancer (BC) risk were defined. The study showed that the cHigh/gLow group without chemotherapy (CT) had a 94.8 % (CI 95% 92.9-96-2) 5-year Distant Metastases Free Survival (DMFS), so meeting the trial primary objective. Recently, a MP g group with excellent prognosis, named ultralow (gUL), has been described (15% in MINDACT). However, few data exist regarding the distribution of these 4 MINDACT RC and the proportion of gUL in the invasive lobular carcinoma (ILC) population. Moreover, association of Ki67 values and MP RC is rarely studied, both in general population and, particulary, in ILC pts. We aimed to address all these questions by comparing our series of invasive carcinoma of non-special type (NST) and ILC pts in whom MP was performed. Pts with ER+/PgR±/HER2- early BC diagnosed from 2012 to 2020 in our institution and with MP data were reviewed. Clinical risk was defined as in the MINDACT trial; genomic risk defined by MP score (UL score: 0.355 to 1.00). 152 pts were identified. NST: 85%, ILC: 15%. Median age NST/ICL: 54/59 years. Histological grade (HG) 1/2/3 (NST vs ILC) 19%/73%/8% and 18%/82%/0%, respectively. UL (g) risk: 17 (13.2%) in NST, 3 (13%) in ILC. The distribution of RC is summarized in the table [UL and Low-non-Ultralow (LNUL) combined for small numbers in ILC]. Median Ki67 value by (g) RC (gLow vs High) in the NST group were 12 and 20: (p<0.001). Of note, in ILC pts median Ki67 was 15 for both MP (g) RC (p=0.77). Table: 63POverall (n=151)NST (n=129)CLI (n=22)cLow-gLow (N/%)61 (40.4)51 (39.5)10 (45.5)cLow-gHigh (N/%)39 (25.8)36 (27.9)3 (13.6)cHigh-gLow (N/%)36 (23.8)28 (21.7)8 (36.3)cHigh-gHigh (N/%)15 (9.9)14 (10.8%)1 (4.5) Open table in a new tab Our results suggest a trend for lower genomic RC in ILC population compared to NST pts. As expected, median Ki67 values correlated with MP genomic categories in overall population and in NST but, intriguingly, not in ILC pts. The small number of pts in this latter group limits results. Further data in wider populations are needed to establish the true association of Ki67 with MP categories in ILC population.