Brain mets under I.O. Bernard Escudier Gustave Roussy, Villejuif, France
Disclosure Honorarium received from BMS, Novartis, Pfizer, Bayer, Roche, Exelixis, Ipsen, Eisai, Calithera Travel Grant from BMS, Novartis, Pfizer Research Grants from BMS, Novartis, Pfizer
What do we know about Brain Mets in RCC? They represent 8% of all metastases in RCC They increase over time, and probably with longer survival They are poor prognostic factor
What do we know about Brain Mets in RCC? They represent 8% of all metastases in RCC They increase over time, and probably with longer survival They are poor prognostic factor TKIs have (very) little efficacy (Chevreau et al, Ann Oncol ) They have high impact on QoL They are exclusion criteria from most of pivotal trials
What do we know about lymphocyte infiltration in BMs? (Berghof et al, ASCO 2016 educational) Highly heterogenous PDL1 expression tends to be higher in BM from NSCLCC than in paired primary Patients with dense infiltration of effector cells in BM have better prognosis Density of infiltration differs between tumor (melanoma>breast) Interaction between the blood-tumor and the blood-brain barrier swith the infiltration zone of a BM unknown
What do we know about PD1/PDL1 expression in BM? Personal experience will be presented at ASCO 2017 Comparison between BM, pancreatic mets and primary tumors: 180 resected RCC specimen collected 42 primary tumors 138 metastases 87 BM 51 Pancreatic metastases (PM) Derosa L et al, ASCO 2017
Inter and intra-tumor heterogeneity of PD-L1 and MET expression in mrcc Biomarker expression Primary RCC Metastases All specimens PM BM p value All All sample (n=180) PD-L1 TC 12% 19% 23% 0.63 21% 22% PD-L1 IC 51% 49% 47% 0.82 47% 51% MET 0% 2% 35% <0.001 24% 23% Lower expression of PD-L1 and MET in primary RCC compared to metastases and a significant higher expression of MET in BM compared to PM PD-L1 discordance was higher in metachronous metastases (>6months). Biomarker expression Discordance between primary tumor and metasteses PM BM Paired samples (n=35) PD-L1 TC 15% 40% PD-L1 IC 33% 22% MET 0% 67% PD-L1 positivity was correlated to higher Fuhrman grade (p<0.02), PDL-L1 IC to TIL (p<0.02) and MET to TIL (p=0.013) and to necrosis (p<0.0001). Derosa et al, poster ASCO 2017.
Efficacy of IO in BM from different tumors Melanoma: Ipilimumab PD1 antibodies Lung Cancer Renal Cancer
Experience in Melanoma (courtesy of K Margolin and H Kugler) Ipilimumab EAPexperience 51 patients with BM without steroids nor surgery 21 patients with BM and steroids
Proportion Alive Ipilimumab in advanced melanoma: pooled OS, including EAP data; N = 4846 1.0 0.9 0.8 Median OS, months (95% CI): 9.5 (9.0 10.0) 0.7 0.6 0.5 3-year OS rate, % (95% CI): 21 (20 22) 0.4 0.3 0.2 0.1 0.0 Ipilimumab CENSORED 0 12 24 36 48 60 Months 72 84 96 108 120 Patients at Risk Ipilimumab 4846 1786 612 392 200 170 120 26 15 5 0 Schadendorf et al. JCO 2015
Brain mets: Ipi 10 mg/kg x 4 + maintenance Cohort A=no steroid or neuro Sx ~26% 2 yr OS also seen in 165-pt cohort from expanded access trial PLATEAU Margolin et al Lancet Oncology 2012
Disease control at 12 weeks
Nivolumab + Ipilimumab in melanoma BM ASCO 2017
Phase II Trial of Pembrolizumab for Untreated BM Key Eligibility: Advanced NSCLC or melanoma At least 1 untreated brain metastasis 5-20mm No neurologic symptoms or steroid requirement Pembrolizumab 10mg/kg q2 weeks Brain metastasis PD Brain metastasis CR, PR, or SD Consider radiation or surgery to progressing lesions Continue pembrolizumab if systemic control achieved PI: Harriet Kluger Co-PIs: Veronica Chiang, Sarah Goldberg Primary Endpoint: Brain Metastasis Response Rate Secondary Endpoints: Overall response rate, safety, PFS, OS Exploratory Endpoints: PD-L1 expression, TILs, and other predictive biomarkers on T cells, tumor cells, neuronal cells, and in plasma
Pembrolizumab in BM from melanoma and NSCLCC Goldberg et al, Lancet Oncol 2016
Examples of responders 56 year old male, progressed on 2 prior regimens; response to pembro at 2 months, remains in response, 33+ months 53 year old male, prior progression after adoptive cell therapy at NCI, response to pembro at 2 months, remains in response, 6+ months, still in response
Patient with pseudoprogression Baseline A B Pre-treatment: A) H+E B) S100 C D After 1 dose, biopsy of target-like growing lesion: C) H+E D) HMB45 After 1 dose, confusion E F E) Reactive astrocytosis, macrophages F) CD45 G H G) CD8 H) CD68
Experience in Lung Cancer N=36 (18 melanoma, 18 PD-L1+* NSCLC patients) 1 untreated or progressive brain metastasis (5 and 20 mm in diameter) without associated neurological symptoms or the need for corticosteroids. Pembrolizumab 10 mg/kg every 2 weeks until PD. Brain response in NSCLC patients: 33% Median DoR: 3.2-7 months *>1% tumoral cell + 18 Goldberg -Lancet Oncol 2016
Experience from IGR in NSCLCC (courtesy of B Besse) 38 of 238 patients (16%) treated in GR with PD1/PDL1 inhibitors had BMs 31 pts with BM 4 pts with BM and meningeal carcinomatosis 3 pts with meningeal carcinomatosis 13% progression in the brain 28% (2/7) patients with meningeal carcinomatosis progressed
Experience from IGR in NSCLCC (courtesy of B Besse) Safety is acceptable Overall response in patients with BM: 18 PD= 55% 6 PR= 18% 9 SD= 27%
What do we know in RCC? Preliminary experience from IGR reported at GU ASCO 2017 62 mrcc patients treated with Nivolumab within the Nivoren trial 8 patients with asymptomatic brain metastases (BM) at baseline Albiges L et al. Abstract 520
intracranial PD and lung PR No systemic extracranial PD Intracranial PD No apparent benefit of Nivo on BM Brain imaging before starting Nivo
Larger experience will be reported at ASCO 2017 up to December 2016, 588 pts have been enrolled including 55 pts with BM (35 (67%),6 (12%) and 11 (21%) with 1, 2 or > 2 BM No previous treatment for BM was performed in 67% (n = 37), while 9% had previous brain surgery (n = 5 ;) or brain radiation (n = 17 (31%). Neurologic deterioration requiring steroids was observed in 15 pts (32%)
Efficacy data: dissociated response in the brain is not uncommon Baseline 4 weeks FON.
Efficacy data: dissociated response in the brain is not uncommon Baseline 8 weeks HAL.
Efficacy data: dissociated response in the brain is not uncommon Baseline 8 weeks HOD.
LOU.. Baseline 8 weeks
Baseline 8 weeks 16 weeks MAY
Questions Are these brain progression pseudoprogression or real preprogression? One of our patient had surgery, and had viable tumor cells When a second scan was performed, progression continued Stereotactic radiotherapy was used in most patients with good efficacy
Preliminary conclusions from Nivoren Safety of Nivolumab in RCC patients with BM appears to be acceptable, although some pts do require steroids because of brain progressive disease. Previously untreated BM commonly progress on nivolumab, and requires therapy Overall 23% of patients had objective response, mainly after previous local treatment
CONCLUSIONS Brain mets can be treated with IO Efficacy data might be different according to tumor type Very few data are available in RCC: Preliminary data from Nivoren suggest that it is feasible That it is safe in previously treated patients But that rapid progression in the brain can occur With some dissociated response More data are required to asses the final efficacy