As expected, the major theme this year was (again) Immuno-oncology with a focus on PD-1 antibodies. Another theme that is gaining momentum is segmentation of tumor types to small niches based on high resolution genomic profiling. This approach can be used to identify a drug’s target population already in phase I, as exemplified by multiple presentations I will discuss below. In most cases, these drugs are ineffective in the general population but highly effective in rare subsets of cancer patients.
While the PD-1 arena is dominated by large pharmas, led by BMS (BMY), Merck (MRK) and Roche, investors who would like to get exposure to the personalized medicine paradigm have Foundation Medicine (FMI), which is emerging as the gatekeeper of genomic profiling.
Theme #1 – PD-1 antibodies
As expected, PD-1 antibodies were the focus of attention, with a horse race between BMS, Merck and Roche. As this area has been intensively reviewed elsewhere, below is a “bullet point” summary.
Durable responses – As monotherapy, PD-1 antibodies lead to deep and durable responses in melanoma (30%), renal cancer (20%) and NSCLC (20-30%). Many of the responses are maintained for over a year. Updated response data at ASCO were lower than expected but it appears that response rate does not fully capture the overall benefit of PD-1 drugs.
Updated survival data – The survival figures in melanoma and renal cancer were impressive and likely represent clinical benefit for most patients (including non-responders). In melanoma, median overall survival was 17 (heavily pre-treated) and ~24 months (limited prior therapy, estimated) for nivolumab and pembrolizumab (MK-3475). In renal cancer, nivolumab led to a median survival of ~2 years. In all cases, the survival appears better than historical controls, but the single arm trial design warrants caution in interpreting the data.
As one would expect from effective immunotherapy, trials with longer term follow up demonstrate flattening of the survival curve. This implies that for some patients, PD-1 antibodies may offer long term remission or even a cure.
New indications – Activity seen in new tumors included response rates of 20% in PD-L1+ head and neck cancer, 43% in PD-L1+ bladder cancer and 17% in ovarian cancer. Some of the trials had limited follow up so response rates may improve at the next analysis. Roche announced a large phase II study in bladder cancer, which will likely be used for accelerated approval if the 40%+ response rate is replicated. BMS announced that nivolumab had been awarded breakthrough therapy designation in Hodgkin’s lymphoma, implying there is strong activity there as well.
PD-L1 expression – PD-L1 expression appears to correlate with a higher response rate and is used for patient selection in some of the ongoing pivotal trials. Nevertheless, there is activity in PD-L1 “negative” patients and many speakers emphasized this as a reason for not selecting patients based on the PD-L1 status. Companies use different tests and/or parameters for assigning PD-L1 status.
Combination data – Identifying the right combination regimens for each indication is complex and is perceived as the next stage in the immuno-oncology revolution. Combination data for nivolumab+Yervoy continue to look promising in melanoma and renal cancer but clinical profile in NSCLC is uninspiring with limited efficacy and significant toxicity. PD-1 antibodies are combinable with chemotherapy and targeted therapies with some encouraging signs but safety issues emerge. Yervoy in combination with T-VEC or IDO inhibitor generated preliminary promising results.
Yervoy as adjuvant therapy– Yervoy (Anti-CTLA4) had impressive efficacy in the adjuvant setting in melanoma (25% reduction in risk of recurrence free survival). This is seen as a validation for immune checkpoint blockers and has already prompted adjuvant trials with PD-1 antibodies.
Targets beyond PD-1 – A major ongoing effort is to identify additional immune checkpoint or co-stimulatory molecules that can work in synergy with PD-1. The next hot target is OX40, pursued by Roche (disclosed at its analyst event) and AstraZeneca.
Vaccines are becoming popular again – As PD-1 antibodies modulate the suppressive tumor microenvironment, they are viewed by many as enablers of cancer vaccines. PD-1 antibodies are expected to rime the immune system in general while vaccines will guide the immune system to the right antigens.
Theme #2 – Segmentation of tumors with genomic profiling
This year’s meeting illustrated the future role of next-gen sequencing (NGS) products like Foundation Medicine’s tests in oncology. NGS, which enables genomic profiling of many biomarkers in parallel, will likely become the mainstay of cancer diagnostics given the increasing number of drugs intended for genetically defined tumors.
The underlying hypothesis of identifying driver mutations to which cancer cells are “addicted” has been around for years and culminated in several success stories. These include drugs like Gleevec, Xalkori and Zelboraf that work in tumors with specific mutations (BCR-ABL, ALK/ROS1 and BRAF, respectively).
Today each of these drugs comes with a dedicated companion diagnostic kit that has to be purchased and analyzed separately in order to decide whether a patient is eligible for the drug. As more mutations are identified as biomarkers for patient selection, this single-test/single-gene approach no longer makes sense.
For example, in NSCLC (non-small cell lung cancer), the number of “actionable” mutations is expected to grow from 2 (EGFR, ALK) to >10 in the foreseeable future (see figure below). This means that in order to find out if a patient is eligible for a specific drug, cancer centers will have to run many tests, which will become unfeasible from a logistic as well as economic perspective. Foundation’s tests solve this problem in an elegant way by providing mutational status for many relevant genes in a single test.
This year’s meeting saw the emergence of 4 new niche indications in clinical trials for 4 different classes of drugs.
FGFR1/11q amplification – Clovis (CLVS) presented results for lucitanib demonstrating a 50% (6/12) response rate in breast cancer patients with FGF-aberrant tumors (defined as FGFR1/11q amplification). The remaining 6 patients experienced disease stabilization and median PFS was 9.6 months. (I previously reviewed the results and why they make lucitanib a best-in-class FGFR inhibitor). FGFR1/11q amplification occurs in 20-25% of advanced breast cancer cases, a similar incidence to that of HER2. This efficacy signal, albeit impressive, needs to be corroborated in future studies.
FGFR2/3 fusions – J&J (JNJ) presented phase I data for JNJ-42756493, a selective FGFR inhibitor. Of a total of 45 patients 5 had FGFR fusions or translocations, of whom three achieved a response (ongoing) and one experienced disease stabilization (8 months). Of note, benefit was seen across several tumor types (bladder, GBM, adrenal).
MET amplification – Amgen (AMGN) presented phase I results for AMG337, a selective MET inhibitor. Among 10 gastric/esophageal patients with MET amplification 5 achieved a response including 1 durable CR. Of note, 2 of the responses were ongoing after 133 and 74 weeks. Abbvie’s (ABBV) anti-MET antibody (ABT-700) also led to responses in 50% (4/8) of patients with MET-amplified tumors but durability of response was shorter.
TrkA fusions – Ignyta (RXDX) presented phase I data for RXDX-101, a multi-kinase inhibitor of ALK, ROS and Trk. The trial enrolled patients with relevant genomic alterations and results include responses across all 3 subtypes. While activity in patients with ALK and ROS-mutated tumors may not be as strong as with other drugs, there was one impressive response in a colon cancer patient with a TrkA fusion mutation. As this was the only TrkA+ patient in the trial and given the limited follow up, it is still hard to declare TrkA as a novel target. Nevertheless, the quick and significant response is reminiscent of other genetic niches such as ALK and RET. Loxo Oncology and Array (ARRY) have a selective Trk inhibitor in phase I.
NGS offers higher sensitivity – In addition to the logistic advantages, NGS is potentially superior to available diagnostic methods (FISH, PCR). At ASCO, Foundation Medicine presented several data sets, the most striking one was from a study which suggests that Foundation’s product is more sensitive in identifying ALK+ tumors than the approved method (FISH). The abstract reports 9 patients whose tumors were classified as ALK- by standard assays but as ALK+ using NGS. Of the 9, 5 responded to Xalkori, implying these patients were truly ALK+.
Foundation Medicine as the gatekeeper of genomic profiling
Founation Medicine, through its commercial products and collaborations with drug development companies is the undisputed leader of tumor genomic profiling. The company’s products, FoundationOne and FoundationOne Heme are commercially available and are already generating significant sales with a limited sales force.
Broad reimbursement remains an important gating factor and is still several years away, as HMOs are reluctant to provide automatic coverage without prospective clinical data, FDA approval or incorporation in the medical community’s guidelines. Reimbursement and market adoption will rely on clinical data with approved and investigational drugs, demonstrating FMI’s ability to guide treatment decisions.
Importantly, several new drugs are expected to utilize Foundation Medicine’s technology. The most visible example is Agios’ (AGIO) IDH2 inhibitor (AG-221) which generated phenomenal results in AML. In order to select patients for its clinical trial, Agios uses Foundation One, which will become AG-221’s companion diagnostic test. The phenomenal efficacy of the drug in IDH2+ patients, should make FMI’s test a standard diagnostic test for every newly diagnosed AML patient. Assuming a US incidence of 18,000 and a price of $3000, the AML market represents a $54M opportunity every year in the US alone.
Admittedly, it is hard to justify the company’s market cap ($670M) by ongoing commercial sales ($7M in Q1) but the huge commercial opportunity ($4B in the US) coupled with Foundation Medicine’s dominant position make it an attractive investment. Similarly to the case with biotech stocks, the rationale for holding the stock lies in the huge market potential down the road.
NGS has its limitations and will have to be used in combination with other methods in order to get a full profile of the tumor (for example, PD-L1 status is assessed with standard IHC) but it appears to be the only viable route to cover many of the known actionable mutations. The NGS market is expected to grow exponentially over the coming years, which makes Foundation Medicine an obvious acquisition target given its first mover advantage and market experience.
I will publish a portfolio update and my take on Clovis on Sunday.