Rini Highlights Benefit With Atezolizumab/Bevacizumab Combo in Distinct Gene Signatures in RCC

Brian I. Rini, MD

Brian I. Rini, MD

IMmotion 151, a randomized phase III trial, demonstrated benefit in progression-free survival (PFS) with the combination of atezolizumab (Tecentriq), a PD-L1 inhibitor, and bevacizumab (Avastin), a VEGF inhibitor, compared with sunitinib (Sutent) in previously untreated PD-L1-positive patients with metastatic renal cell carcinoma (mRCC).

According to a presentation at the ESMO 2018 Congress, further analysis of the data from IMmotion 151 was able to validate findings from IMmotion 150, the phase II trial that identified 3 subgroups based on gene expression: an angiogenesis subgroup, T-effector subgroup, and myeloid subgroup.

High T-effector (TeffHigh) subgroups did better with the atezolizumab/bevacizumab combination compared to sunitinib. The combination also led to a numerically better PFS in tumors that had a signature reflecting low expression of genes associated with angiogenesis (AngLow). However, patients with the angiogenesis (AngHigh) signature did better with sunitinib.

Investigators also looked at risk groups in IMmotion 151. Patients with more favorable risk did not do as well with immunotherapy as the intermediate- and poor-risk groups. This supported an idea suggested by previous data that favorable risk patients had a more allogeneic phenotype.

In an interview withTargeted Oncology,Brian I. Rini, MD, professor of Medicine at Cleveland Clinic, discussed these core findings from the IMmotion 151 trial he presented at the conference, as well as the implications of these findings moving forward.

TARGETED ONCOLOGY:Can you provide an overview of the IMmotion 151 trial?

Rini:IMmotion 151 was a phase III trial that randomized previously untreated kidney cancer patients to either of standard of care sunitinib, or atezolizumab, a PD-L1 inhibitor, and bevacizumab, a VEGF inhibitor. The clinical data were reported at the ASCO Genitourinary Cancers Symposium by Robert J. Motzer, MD, and showed PFS advantages for the combination, the VEGF/PD-L1 combination therapy, both in the PD-L1—positive patients – patients whose tumor infiltrating cells expressed PD-L1 – and also in the intent-to-treat in all-comers. That was the main clinical result.

What I’m presenting here at ESMO in 2018 is the core of the data from that trial.

TARGETED ONCOLOGY:What are the findings you are presenting?

Rini:IMmotion 151, as I mentioned, was a randomized phase III [trial]. Preceding that was IMmotion 150, which was a randomized phase II with 2 similar arms [of] sunitinib and atezolizumab [plus] bevacizumab, but also an atezolizumab monotherapy arm. There was core work done in that trial, basically looking at RNA gene expression from baseline kidney tumors from patients who ultimately went on that trial.

What IMmotion 150 showed was that there were different biologic subgroups based on gene expression, and they were labeled as an angiogenesis subgroup, T-effector subgroup with genes obviously having to do with T-cell infiltration and function, and then also a myeloid subgroup, myeloid cells, or immunosuppressive cells.

These data were published inNature Medicinelast year and were really the basis for the 151 analyses. [IMmotion] 150 identified these subgroups and 151 was a much larger group of patients, 300 samples in 150 versus 800 in 151 to validate the subgroups and also look at their association with clinical outcome as some other features.

TARGETED ONCOLOGY:What do these findings mean for patients with metastatic RCC?

Rini:What was found was a few things. One was that patients who were in the patient TeffHigh subgroups — an inflamed subgroup that you would imagine is more susceptible to immunotherapy – indeed did better with atezolizumab plus bevacizumab compared to sunitinib. In fact, that T-effector signature is very much associated with PD-L1 sustaining, not surprisingly.

Somewhat interestingly, within the sunitinib arm, patients who were AngHigh — had gene expression angiogenesis high – did much better than those patients who were AngLow. It seemed sort of obvious, more angiogenic tumors would respond to sunitinib, but we have been looking for a biomarker for sunitinib for over a decade. These are really compelling data in terms of the degree of difference in how patients did over time.

Two other things were looked at. One was the Memorial Sloan Kettering risk groups. Those are risk groups that had been around for 20 to 30 years that look at clinical features and really segregate patients into good, intermediate, and poor. The good risk patients, at least with some other immunotherapy, seemed to do less well. The intermediate and poor seemed to do better with immunotherapy, at least clinically, so 1 thing that was found was that those more favorable risk patients had a more angiogenic phenotype, which really fits with some other data that suggests that VEGF agents have their best effect and maybe immunotherapies don’t work as well in favorable risk patients, at least in traditional risk measures such as response and PFS.

The last interesting biology was that there was a subtype of kidney cancer that is called sarcomatoid.Sarcomatoidis a growth pattern. The cells look like spindle cells under a microscope, and it’s really sort of a nasty histology in kidney cancer. Most patients unfortunately don’t do well. [IMmotion 151] allowed such patients in. They were about 16% of patients. When we looked at their tumors, they have an AngLow profile, which explains why they don’t respond to the standard VEGF agent, but had very much a TeffHigh PD-L1-positive profile. Those patients did especially well with atezolizumab [plus] bevacizumab compared to sunitinib.

I think what it means is that we are finally starting to see the biology behind the tumors that we are treating. We’ve been treating these tumors for decades, certainly [over] 10 years with targeted therapy and a few years with immunotherapy. [However], oncology is very often impaired, [where] we give a drug or drugs and see what happens. Hopefully, these data will allow us to start to differentiate patients when they walk in the door, understand their tumor genotype and phenotype, and treat it most appropriately.

TARGETED ONCOLOGY:Are there any particular subgroups or biomarkers you think should be further investigated after seeing these data?

Rini:There is some controversy in the field about favorable risk patients. Ipilimumab (Yervoy) plus nivolumab (Opdivo), an immune doublet combination, had great effects on intermediate and poor, but not as much, at least by traditional measures, in favorable. Again, I think our findings help explain that. I’m sure that many favorable risk patients would benefit from immunotherapy, so it may be that instead of just clinically categorizing people, we could either genomically categorize them or, probably more realistically, fuse the clinical and genomic data together because favorable, intermediate, and poor are just big buckets. I’m sure within each bucket there’s a lot of heterogeneity.

Again, these are very difficult things to translate these findings into clinical practice, but I think now we are finally starting to get some data sets. I’ll say parenthetically, having been around for the development of targeted therapy, now [we are in] the renaissance of immunotherapy. We’re doing a good job with biomarkers with novel immunotherapies. With VEGF-targeted therapies, we were so happy the drugs worked. We didn’t really get around prospectively to doing biomarkers until years later and we sort of never caught up, but I think the sponsors and investigators are doing a better job and I think that will benefit patients later down the road.

TARGETED ONCOLOGY:What do you hope oncologists take away from this study?

Rini:The take home message from the IMmotion 151 core data is just that tumors obviously have a biologic basis. We can use gene expression to characterize those tumors. It’s not just out of interest, but it’s because it is linked to clinical outcome, both in probably the natural history of the disease but also in predictive biomarker for how they do on 1 therapy or the other. While we are probably not to the step of saying this is going to be routine in clinical practice, it gets us a lot closer than we were.

TARGETED ONCOLOGY:Is there anything else at ESMO that you think was particularly exciting?

Rini:[I presented] a poster on characterization of response for patients who were on CheckMate-214. CheckMate-214 is the phase III of ipilimumab plus nivolumab versus sunitinib that led to the approval and was presented here last year, I believe. Subsequently published, it really became the standard of care, that immune doublet. This subset analysis was looking at responders both in the sunitinib arm but the ipilimumab [plus] nivolumab arm as well, saying what are the characteristics of those responders? There is some really nice swimmer’s plots that look at how patients do over time.

When data is first reported, we might get a complete response (CR) number and partial response (PR) number, maybe a sense of durability, but there’s just not much insight. As time passes and you get more data, you can really dig into how durable are these responses — are CRs better than PRs? Many of these people have gone on therapy, responded, and then come off therapy for some reason, and many of those people will maintain their response.

There’s another poster here by David F. McDermott, MD, looking at treatment-free survival, meaning you’ve gotten treatment, are off treatment, and you are still living with controlled disease, which is really why we give immunotherapy. That data set has particularly matured quicker than the other phase III [trials] that are coming out now, including at this meeting, so we are beginning to understand the drugs better.