As in previous years, the upcoming ASCO meeting will be dominated by immuno-oncology with a particular focus on PD-1 antibodies. The market’s primary focus is expected to be on non-small cell lung cancer (NSCLC) with data from three large randomized trials from BMS (BMY) and Roche. Beyond lung cancer, investors will look for additional indications where PD-1 agents may have clinical utility as monotherapy or in combination with other agents.
Overall, PD-1 programs continue to generate positive data across many indications but to date clinical experience has been sobering. PD-1 antibodies may lead to durable responses in some cases but the vast majority of patients derive limited benefit or don’t respond at all. This is true in most indications with the exception of melanoma where PD-1 antibodies have a dramatic impact and combination with Yervoy appears to lead to further improvement. Below is a recap of some of the data which will be presented over the weekend. Abstract numbers and links are also provided.
PD-1 in NSCLC
Improvement in overall survival is starting to emerge with open questions regarding patient selection. It is already clear that while the survival advantage is clinically meaningful and will make PD-1 the standard of care, the magnitude of effect is limited with the vast majority of patients progressing within 1 year. One can only hope that in some patients PD-1 antibodies will generate durable responses and potentially cures.
In pre-treated unselected patients, the three leading PD-1 inhibitors generate a ~20% response rate. Response rate improves to up to ~40% in patients with high PD-L1 expression (exact Dx tool and threshold varies) whereas PD-L1 negative/low patients have a 10-15% response rate. Importantly, PD-L1+ patients represented 20%-70% of patients depending on cutoff and diagnostic kit.
BMS will present two P3 studies evaluating Opdivo in squamous and non-squamous patients, respectively. Data from the squamous trial (CheckMate 017, #8009) have already been published and demonstrate a 3.2 month (9.2 vs. 6 months) survival benefit, which is clinically meaningful but not dramatic. Response rate was also superior (20% vs. 9%). Interestingly, there was little to no difference in survival based on PD-L1 expression using multiple cutoffs. Results from the non-squamous study will be published at the meeting (#LBA109) and are expected to be positive as well.
Roche will present data from a randomized P2 (POPLAR, #8010), which evaluated MPDL3280A (anti-PDL1) vs. chemo in previously treated patients. In the entire patient population, the antibody demonstrated a similar response rate (15% vs. 15%), a slightly inferior PFS (2.8 vs. 3.4 months) and a modest 2-month survival benefit (11.4 vs. 9.5). In contrast to Opdivo’s data MPDL3280A’s effect was more pronounced in patients with PD-L1 expression with a clear correlation between survival benefit and level of expression (hazard ratio of 0.47 in tumors with the highest expression level). Patients with PD-L1 negative tumors (32% of patients) derived no benefit from MPDL3280A.
It is still unclear whether the difference between BMS’ and Roche’s data sets stems from inherent differences in the antibodies or patient selection tools. BMS’ Opdivo targets PD-1 and blocks interaction with both PD-L1 and PD-L2 whereas Roche’s MPDL3280A binds only PD-L1, which may explain the stronger correlation with PD-L1 expression. Based on the data available to date with all PD-1 antibodies, I tend to ascribe at least some of the discrepancy to biomarkers, which widely vary among the different programs. In the short run, BMS has a clear lead over Roche and Merck (MRK) as Opdivo will likely be used regardless of PD-L1 status. However, Roche’s ability to stratify patients based on sensitivity to PD-1 agents (needs to be validated in additional studies and indications) may become an important tool in developing combination regimens. For example, Roche may be able to pursue PD-L1 negative tumors as a fast and cheap route to market for new immuno-oncology combinations.
Moving to 1st line does not seem to dramatically improve results based on abstracts for Keytruda (#8026) and Opdivo (#8025). For Keytruda response rate in the overall population (all had some PD-L1 expression) was 24% and PFS was 6 months, which is comparable to standard of care treatment. For Opdivo, response rate was 21% in the overall population. Abstract for Roche’s MPDL3280A reports a 29% response rate in first line patients (#8028). Preliminary findings in all abstracts point to a correlation between PD-L1 levels and clinical benefit but numbers are small and data are immature.
Roche will present combination results of MPDL3280A with various chemotherapies in 1st line patients (#8030). The abstract discloses an encouraging response rate of 67%, which is higher than what would be expected with MPDL3280A or chemotherapy alone (25-30%). When added to chemotherapy in the same setting, Opdivo generated a ~40% response rate last year at ASCO 2014 but it is unclear whether there is a real difference.
This year’s meeting will have data for PD-1 antibodies with CTLA-4 antibodies, which are likely to keep the debate regarding this strategy in lung cancer alive. As a reminder, at last year’s meeting BMS presented disappointing results for Opdivo + Yervoy in NSCLC. Response rate was 20% and the addition of Yervoy led to significant toxicities. This year Merck and AstraZeneca will present preliminary findings from their combination studies, respectively. Merck’s abstract (#8011) for Keytruda+Yervoy includes a high response rate (55%) but the sample size is very small (11 patients). AstraZeneca’s abstract for MEDI4736 + tremelimumab demonstrates a lower response rate (26%) in 31 patients (#3014).
PD-1 in other indications
Beyond NSCLC, companies will present data in multiple indications, most of which have already been shown to be relevant for PD-1. Activity is observed across all programs with clear signs of a correlation between PD-L1 expression and clinical benefit. Similarly to the case in NSCLC, most patients do not appear to derive meaningful clinical benefit as response rates in the general population are 20% or lower. In some cases, patient selection results in a higher response rate but so far durability of response and impact on long term survival remain limited.
Melanoma – With both Opdivo and Keytruda approved for melanoma, focus is shifting to combinations. BMS will present results from a P3 data evaluating Opdivo + Yervoy vs. each agent alone (#LBA1). As a late-breaking abstract, no information is disclosed prior to the meeting but expectations for the combination arm are high. BMS will also report a P2 trial evaluating the combination vs. Yervoy alone (#9004). The information in the abstract reveals a dramatically better response rate (60% vs. 11%) and PFS (8.9 vs. 4.7 months). Based on cross trial comparisons, these response rates are numerically superior to the ~35% typically seen with PD-1 antibodies alone.
Renal cancer – BMS will present updated survival data with single agent Opdivo (#4553). Results continue to demonstrate promising survival at the two higher doses as reported at last year’s meeting (although response rates and PFS were underwhelming). This year’s abstract includes analysis based on PD-L1 status which demonstrates a dramatic difference of almost 1 year (29.9 vs. 18.2 months) in favor of PD-L1+ patients. This bodes well for PD-1 antibodies in renal cancer but the promising survival signal should be interpreted with caution in the absence of randomized data. Merck will report combination data for Keytruda+Yervoy (#3009). The abstract does not include efficacy data, which will be presented at the meeting.
Bladder cancer – Bladder cancer is one of the few indications where Roche has a clear lead over BMS and Merck, based on very strong results for MPDL3280A in PD-L1+ patients. This year, Roche is presenting updated data (#4501) which continue to demonstrate an impressive response rate in PD-L1+ vs. a modest response rate in PD-L1- patients (46% vs. 16%). At the meeting, it will be interesting to see updated response durability and PFS, which are immature in the abstract. Merck will report results for Keytruda, which also demonstrate a higher response rate in PD-L1+ patients (38%) (#4502).
Small cell lung cancer – BMS will present data for Opdivo and Opdivo+Yervoy in patients regardless of PD-L1 status (#7503). Response rate in the combination arm (25%) appears better than that observed with Opdivo alone (15%). Merck will present data for Keytruda in PD-L1+ patients (#7502), which demonstrated a 25% response rate according to the abstract.
Ovarian cancer – Merck and Merck Serono/Pfizer will report initial data for Keytruda and avelumab, respectively. Response rate was modest (11-14%) even in PD-L1+ patients. (#5510)
Esophageal cancer – Merck’s Keytruda demonstrated a 23% response rate in PD-L1+ patients according to the abstract (#4010).
Gastric cancer – Merck will report updated results from KEYNOTE-012 in PD-L1+ patients (#4001). The abstract discloses a 22% and 33% response rate by central and investigator review, respectively. 6-month PFS and OS were low (24% and 69%, respectively). Merck Serono/Pfizer’s avelumab will have data from a Japanese study. The abstract reports responses in 27% (3/11) of patients. (#4047).
At the meeting, data are expected for Opdivo+Yervoy in glioblastoma (#3010) as well as two late breakers for Keytruda in colorectal and head and neck cancer.
Immuno-oncology beyond PD-1
In parallel to the unprecedented resources allocated to PD-1 programs, there is an industry-wide effort to find the “next PD-1” or agents that can augment PD-1 inhibitors. These include new immune checkpoint targets (LAG-3, TIM-3, VISTA, KIR etc.), co-stimulatory agents (OX40, CD27, CD40, 41-BB etc.) and other agents designed to stimulate the immune system (IDO, CSF1R, vaccines). In most cases, new programs are pursued in combination with PD-1 antibodies.
Recent deals for novel immuno-oncology assets exemplify their tremendous value in today’s industry even at an early stage. AstraZeneca’s deal with Innate Pharma (IPH.PA) for IPH2201 included a $250M upfront payment, which is quite high for an antibody that has phase I data in healthy volunteers. Earlier this year, BMS paid $800M for Flexus and its preclinical IDO inhibitor. IDO is one of the hottest targets in immuno-oncology but Flexus’ IDO program is years behind Incyte’s (INCY) IDO inhibitor and at least one year behind Newlink’s (NLNK) IDO program (partnered with Roche). Aduro (ADRO) landed another massive preclinical deal for its STING agonists with Novartis (NVS) which agreed to pay $200M upfront (plus $50M in equity investment).
This year’s ASCO includes very few new immuno-oncology programs and limited combination data (most combination studies were only recently initiated).The elephant in the room this year is Incyte’s IDO inhibitor, which demonstrated intriguing efficacy in melanoma last year. IDO is one of the hottest targets in immuno-oncology, evidenced by recent deals around the target. Incyte’s program is in combination studies with three different PD-1 inhibitors which started last year. Data from these trials is expected in 2016.
Cancer vaccines – On the cancer vaccines front, Amgen (AMGN) and Celldex (CLDX) will present updated results for their programs in melanoma and GBM, respectively. Amgen’s T-VEC will have long term follow up for T-VEC+Yervoy. At last year’s meeting, investigators presented a dramatic response rate (56% including 33% CR) compared to ~15% typically seen with Yervoy monotherapy. This year’s abstract does not include additional patients, but results continue to impress with a PFS of 10.6 months (#9063). Celldex will report updated survival data for its EGFRvIII vaccine (rindopepimut) in GBM (#2009). If the survival benefit reported in November is corroborated, the trial may be sufficient for accelerated approval.
Antibodies – Results for new targets will be presented for Roche’s CSF1R antibody (#3005), which is designed to inhibit tumor-promoting macrophages, and Immune Pharma’s/BMS’ anti-KIR antibody (#3065), designed to activate NK cells. Both agents demonstrated little to no efficacy. Pfizer will present data for 4-1BB in combination with Rituxan in non-Hodgkin’s lymphoma (NHL). Results in the abstract include a modest response rate (22%) that can be attributed to Rituxan to some extent but there were 2 intriguing long term CRs in a subset of NHL (#3004).
Other agents – Additional intriguing programs include ARMO’s pegylated IL-10 (#3017), which demonstrated signs of efficacy in renal cancer and ocular melanoma (a tumor type resistant to PD-1) and Roche’s CEA-IL2v, which comprised of an engineered IL-2 fused to CEA to enable better targeting of IL-2 into tumors (#3016).
Next year, investors can expect a flood of clinical data with new immuno-oncology agents. The most promising targets are OX40 (Pursued by AstraZeneca, Roche and Pfizer) and IDO.
Stock related news flow (non immuno-oncology)
ASCO will have important data for some of the companies I follow. Although these are not immuno-oncology programs, they represent significant advances as well as important value drivers.
Clovis Oncology – Clovis (CLVS) will report updated results for rociletinib in EGFR-mutated NSCLC (#8001). Response rate in T790M+ patients was 48-49%, which is lower than previous updates (67%). PFS is not reported in the abstract (was 10.4 months at the last update) and is expected to be a key focus given the neck and neck race with AstraZeneca’s AZD9291. Results will be compared to AZD9291’s recent data update, which included a response rate and PFS of 54% and 13.5 months, respectively.
AZD9291 appears numerically superior based on cross-trial comparisons but this difference may stem from imbalances in patient characteristics, especially geographical distribution. In the AZD9291 study, 61% of patients were Asian whereas the majority of patients in the rociletinib trial were non-Asian. Asian patients are known to derive more benefit from 1st-gen EGFR inhibitors like Tarceva and Iressa but it is not clear whether this is the case with next-gen EGFR inhibitors.
Preliminary data from Astellas’ ASP8273 imply that Asian patients are indeed more sensitive to next-gen EGFR. Astellas is presenting two phase I trials with ASP8273 in the US and Japan respectively. The abstract for the Japanese study (#8014) reports an 80% response rate whereas in the US study the same drug achieved a 25% response rate (#8083). If the difference holds when updated results are presented, comparing rociletinib and AZD9291 will become even more challenging based on available data.
Clovis will also report updated data in T790M- patients, where rociletinib continues to demonstrate a surprisingly high response rate (33%-36%). This is emerging as another differentiator from AZD9291, which reported a 21% response rate with short durability. Rociltenibs’ activity in T790M opens up a new significant commercial opportunity and more importantly, bodes well for the ongoing study in 1st line patients, where most patients are defined as T790M-.
Immunogen – Immunogen (IMGN) will report highly anticipated results with IMGN853 in FR+ ovarian cancer (#5518). Following the monetization of Kadcyla’s royalties, IMGN853 remains Immunogen’s primary value driver in 2015, which makes the data at ASCO extremely important for the stock. Preliminary data in the abstract are aligned with expectations set by management (discussed here) with a 40% response rate but the small sample size (10 patients) and limited follow up make it difficult to interpret the data. A new dosing regimen appears to mitigate ocular toxicities, which had been viewed by the market as a major issue in the past. At the meeting, investors’ focus will be on updated response rate in additional patients and response durability. A >30% response rate may enable Immunogen to pursue this indication with a single arm study based response rate.
Genmab – Genmab’s (GEN.CO) partner, J&J (JNJ) will present P2 data in heavily pre-treated multiple myeloma patients (#LBA8512). Although abstract details are embargoed, top line results have already been announced with a 29% response rate and a 7.4 month durability of response. This study will serve as the basis of an accelerated approval filing later this year.
Interestingly, response rates observed to date with daratumumab may underestimate its true anti-tumor effect based on an abstract from J&J (#8590). As disease activity is measured by antibodies secreted by myeloma cells in the blood, daratumumab may be identified as a sign of disease activity. Genmab’s management already disclosed that distinguishing between daratumumab and myeloma protein leads to increases in CR rate but this is expected to gain more visibility only at ASH.
Foundation Medicine – Foundation Medicine (FMI) and collaborators will present a large body of clinical work which demonstrates the utility of the company’s tumor profiling technology. This year, results include validation of new MET mutations as biomarker for MET inhibitors (#11007), identification of ovarian cancer tumors that are sensitive to PARP inhibitors (#5508) and identification of actionable mutations in certain rare tumors. The most intriguing abstract is for a prospective study comparing patients who received biomarker-matched treatment to those receiving non-matched therapy (#11019). The abstract reports a statistically significant survival difference benefit for patients receiving matched therapy, which is the first large prospective study correlating FoundationOne and clinical outcome. With shares under pressure after a weak quarter and reimbursement concerns, it is up to Foundation Medicine to continue to demonstrate the utility of its tumor profiling system.
ArQule – ArQule (ARQL) will report updated data for its FGFR inhibitor, ARQ087 (#2545). The abstract describes two previously disclosed cases of cholangiocarcinoma patients with FGFR2 fusions who achieved a PR (maintained for ~5 months) and a durable minor response (ongoing). ArQule recently initiated a P2 in this patient subgroup but it is unclear whether additional patients will be included in the data set at ASCO. FGFR2+ cholangiocarcinoma represents a modest commercial opportunity (hundreds of patients in the US annually) but given the high unmet need the regulatory route will be very favorable.