A pancancer analysis of impact of MDM2/MDM4 on immune checkpoint blockade (ICB).

2630 Background: MDM2/MDM4 are implicated in hyperprogression after immune checkpoint blockade (ICB). Our preclinical studies showed reduced T-cell killing of MDM2-amplified tumor cells that was overcome by an MDM2 antagonist or gene knockdown, and we observed additional tumor killing by T-cells with MDM2 inhibition plus anti-PD1. We hypothesized that MDM2/4 gene amplification/overexpression correlates with resistance to ICB and investigated the association of MDM2/4 alterations to overall survival (OS) following ICB across multiple solid tumors. Methods: Solid tumors tested at Caris Life Sciences (Phoenix, AZ) with NextGen Sequencing on DNA (NGS) were analyzed. MDM2/4 amplification (amp) was tested by NGS and determined as either amp4 (cutoff of >=4.0 copies) or amp6 (>=6.0) or amp8 (>=8.0). Real-world OS was obtained from insurance claims data and calculated from treatment start or tissue collection to last contact. Kaplan-Meier estimates were calculated for molecularly defined groups. X2/Fisher-Exact were used and significance determined as P-value adjusted for multiple comparisons (q<0.05). Results: In a large cohort of TP53-wild type solid tumors, 2785 had MDM2 amp4 (262 were ICB-treated), 2108 had MDM2 amp6 (ICB: 192), 1721 had MDM2 amp8 (ICB: 149); 1040 tumors had MDM4 amp4 (ICB: 59), 293 had MDM4 amp6 (ICB: 8) and 217 had amp8 (ICB: 4). NSCLC, GBM, breast, bladder cancer and liposarcoma had the highest MDM2 amp and breast cancer, GBM, uterine, NSCLC and ovarian cancers had the highest MDM4 amp. When all tumors were considered, MDM2 amp4, 6 and 8 significantly associated with shortened OS (amp4: HR 1.31; amp6: HR 1.31; amp8: 1.29, all p<0.0001); similar results were seen with MDM4 (amp4: HR 1.14, p=0.004; amp6: HR 1.51, p<0.00001 and amp8: HR 1.6, p<0.00001). Of note, MDM4 amp4 but not MDM2 amp4 was associated with significantly worse post-ICB survival (HR: 1.55, p=0.009).When comparing molecular differences between MDM2 amp4 and MDM4 amp4, significantly higher PDL1 expression was associated with MDM2 amp (Ab clone 28.8: 58% vs. 28%; clone 22c3: 48% vs.25%, q<0.05); while higher mutation rates of ARID1A, PIK3CA, PTEN, KRAS and CTNNB1 were associated with MDM4 amp (all q<0.05). When investigating NSCLC, MDM4 amp4 was associated with decreased post-ICB survival in NSCLC (HR: 2.59, p=0.001) but not MDM2; when comparing the molecular alterations, MDM2 amp was associated with significantly higher PDL1 (22c3) (54% vs. 27%), EGFR mutation (36% vs. 14%) but less prevalent KRAS (15% vs. 50%), STK11 (11% vs. 41%), KEAP1 (2.5% vs. 27%) and BRAF (2% vs. 16%) mutations. The association with poor prognosis of MDM2 amp4/6 (HR: 1.2 and 1.3, p<0.05) and MDM4 amp4/6 (HR: 1.3 and 1.6, p<0.05) was seen in breast cancer, but not in NSCLC, GBM, bladder or uterine cancers. Conclusions: MDM2 and MDM4 amplification are negative prognostic factors in TP53-WT breast cancer while MDM4 amp is associated with reduced survival in ICB-treated NSCLC.