AACR 2018 Summary Highlights
Chicago, IL April 19, 2018 – Another AACR meeting has come to a close and, as always, has yielded a myriad of perspectives, developments, product launches, and clinical advancements. While this is still a cancer research conference, Biopharma continues to have more impact on a yearly basis, and parts of AACR seem to turn into a mini-ASCO from plenary sessions to the exhibit hall.
Immuno-oncology seemed to be by far the most dominant theme at AACR this year, but liquid biopsies, and single-cell genomics were also hot topics. Below are some of our key takeaways from many discussions with various stakeholders across the tools, diagnostics, and pharma landscapes, and from the talks and posters we checked out:
Immuno-oncology and genomics continue to dominate the discussion – 7 / 11 of the plenary sessions at AACR focused on immunotherapies while the remaining 4 sessions focused at least in part on genomics.
The lung cancer plenary session lived up to its high pre-conference expectations, delivering potentially standard-of-care-changing data – In the highly contested battle for the 1L metastatic lung cancer market, each of the major immune-oncology (I/O) players (i.e., Merck, BMS, and Roche) delivered impressive results, but Merck’s data stood out from among the pack (read here, here, and here for detailed data from each trial read-out).
Noteworthy plenary session readouts from Merck, Roche, and BMS – Merck’s Pembro + pemetrexed combo (KEYNOTE-189) more than halved the risk of death for ALK/EGFR- mNSCLC patients vs. chemo alone (1-yr OS HR = 0.49). Benefit was observed across all patient populations (e.g., 1-yr OS HR = 0.59, 0.55, and 0.42 in PD-L1 TPS <1%, 1-49%, and ≥50%, respectively). Roche also showed impressive data in the 1L mNSCLC patient segment with its combination of Atezolizumab + Bevacizumab + Carboplatin + Paclitaxel (IMPower150). While overall survival data is pending, the PFS HR data in certain subgroups is compelling (all 1-yr PFS HR): ITT-WT (EGFR/ALK-) HR = 0.62; Teff high HR = 0.51; PD-L1 TC3/IC3 HR = 0.39; TC2/IC2 HR = 0.56; ALK/EGFR+ only HR = 0.59; EGFR Exon 19 del/L858R HR = 0.41. These HR results are interesting, but we will need to see the OS survival data to determine if/where Roche fits in the 1L mNSCLC landscape. BMS presented data on its Ipilimumab + Nivolumab combination in tumor mutation burden (TMB) high patients (>10 mutations / MB via FoundationOne CDx) (CheckMate-227), demonstrating an impressive 1-yr PFS HR of 0.48 in the TMB-high / PD-L1 <1% population, however, it could not quite match the efficacy of Merck in the PD-L1 ≥1% populations (PFS HR = 0.62 for Ipi + Nivo vs. 0.44 for the Pembro + chemo combo). Similar to Roche, OS data will be needed for a more direct comparison.
Overall, the data presented to date suggests that Merck seems poised to expand on its leadership position in the lung cancer market, and has set a high OS bar for all companies competing for the NSCLC market.
One potentially interesting implication of all the lung cancer trial data is the impact these therapeutic regimens may have on biomarker testing – Merck’s strong results with pembro + chemo across all patient groups (regardless of PD-L1 status) led to murmurs about whether PD-L1 testing is still necessary for 1L lung cancer. In the discussion presentation for KEYNOTE-189, Dr. Roy Herbst re-iterated the need for continued PD-L1 testing as it could distinguish between those who need combo therapy (e.g., in the <50% range) vs. those for whom monotherapy may be sufficient (e.g., in the ≥50% range). Additionally, subsequent presentations revealed potentially better I/O regimens in the PD-L1 <1% population. Based on the data presented and our subsequent discussions with pharma stakeholders, we do not expect to see a reduction in PD-L1 testing in I/O.
As a side note, the data presented demonstrates the resiliency of PD-L1 as a biomarker for immunotherapies. Despite being widely bemoaned as a relatively poor biomarker in our discussions with KOLs and oncology stakeholders over the years, it is apparent that it is still one of the best tools we have to stratify patients for anti-PD-1/PD-L1 immunotherapy, even as data from other markers (e.g., gene expression signatures, TMB) emerges.
The data presented by BMS also confirmed the clinical value of TMB as a biomarker for I/O, and one that is independent and orthogonal to PD-L1, providing clinicians with another tool for patient stratification. What remains to be determined, is how, based on the data presented by both Merck and BMS, TMB will be implemented in clinical practice (e.g., as either a primary marker along with EGFR, ALK, and PD-L1, or as a secondary marker / reflex test to stratify PD-L1- patients). Discussants advocated for TMB testing as a standard of care in lung cancer. Overall, we believe the results from both Roche and BMS, showing differential responses to these different I/O combos in various molecular subgroups (e.g., TMB-H, ALK/EGFR+, EGFR exon 19 del / L858R) continue to strengthen the case for comprehensive genomic profiling in lung cancer. We will be watching closely to see how TMB is ultimately reflected in clinical guidelines and its adoption within the community.
On the diagnostics and tools side, liquid biopsy space continues to develop rapidly. Major trends that we noticed during this year’s AACR include:
Interest in immuno-oncology is spilling over into the liquid biopsy space, but there are challenges – Based on analysis of liquid biopsy abstracts, over 10% investigate at least one of TMB, microsatellite instability (MSI), or PD-L1, with a roughly 60:40 split between commercial and purely academic. Much of the field is trying to overcome challenges associated with applying these concepts from tissue to blood, as ctDNA is inherently shorter than tissue-derived DNA and signals from wild-type DNA can interfere with detection of the relevant variants.
Of note, PGDx presented a poster on its error correction approach for MSI; MSKCC’s poster looked at concordance between its MSK-IMPACT tissue test and MSK-Access cfDNA panel; Foundation Medicine sought to demonstrate clinical utility for a bTMB assay; Northwestern University experimented with bTMB analysis using Guardant360.
Activity and interest in CTC technologies and applications is growing quickly – This trend was actually somewhat of a surprise to us, as stakeholders we interviewed in the past have been highly skeptical of its short- to mid-term utility given the technical difficulties of working with CTCs. The sign for Cell Search advertising it as the first and only clinically actionable CTC platform hung high in the exhibition hall, and the platform appears entrenched (although in a relatively small market) given its early mover advantage. However, the field is moving away from pure enumeration of CTC’s towards more complex genotype and phenotype analysis (including for PD-L1). As a result, many smaller companies, such as CellMax Life, Celsee, Vortex Biosciences, MiCareo and RareCyte, are pushing forward with new methods to isolate and analyze CTCs.
To dominate and differentiate, these players will likely need to move fast to demonstrate validity and utility. Who will lead seems unclear at the moment, but we will be keeping a closer eye on this space, as interest appears to have rebounded faster than we expected. cfDNA maintains strong mindshare, as detailed below.
Larger, more established liquid biopsy players move to differentiate and capture market share – Wanderings around the exhibition floor and interactions with these players gave interesting insights into how some of them are placing bets on how to address various trends in the market, and are positioning themselves to differentiate.
One point of contention or uncertainty amongst commercial players, customers, and observers alike is how much liquid biopsy testing will decentralize in the future. PGDx seems to be readying themselves to be both service and kit providers to guarantee access to the entire market. Guardant Health, seems to be reducing turnaround time to 7 days to reduce incentives for such in-house kit-based testing, and collaborating with academic institutions (e.g., MD Anderson) when need be, to offer its test within on-site labs.
Meanwhile, Archer Dx is carving out a niche by focusing on custom panels, which is well positioned to address the unique research interests of each academic hospital, and Roche’s AVENIO line is positioned to hit a wide range of customers valuing utility (with its 17-gene panel) or research (with its 77- and 197- gene panels).
Single Cell Genomics
Single-cell genomics also remains a hot topic at AACR, with multiple companies pushing their various flavors of single-cell analytical systems. As we previously noted following our attendance at AGBT, “the wave is coming”. Last year, somewhat surprisingly, only a handful of posters mentioned single cell genomics (SCG). In contrast, this year, 35+ posters mentioned SCG.
10x Genomics continues to dominate the SCG space – 10x Genomics was mentioned in more than 30 abstracts this year. At the meeting, they launched a new solution that expands their single cell offering to single cell DNA which accounted for 10 out of the 30 abstracts. One key focus was on understanding tumor heterogeneity, the tumor microenvironment, and CNV. As 10x Genomics is the leader in single cell RNA sequencing, it is well-positioned to take advantage to the I/O wave. With additional solutions on the horizon later this year from their AGBT workshop which includes further expansion into epigenetics with single cell ATAC-seq and multi-omics analysis with feature barcoding, they are well positioned in the overall cancer research and translational space.
Mission Bio has a stronger focus on clinical applications, with a sole focus on DNA Sequencing – The company made multiple announcements, most notably that multiple centers adopted its Tapestri platform since its launch last October. The platform can be used for single cell DNA sequencing (up to 10K cells), and 3 panels are currently available. Interestingly, some (including Mission Bio) may argue that the focus on DNA may be preferred in clinical settings, as the molecule is more stable, and targeted therapies (vs. I/O) still represent the bulk of the market, and with >1,500 trials using combo therapies.
Our discussions suggest that the “killer” clinical application that can be uniquely addressed by single-cell analysis remains elusive – Despite advancements in platform technology and an increasing array of new applications and assays (e.g., RAS-driven FLT3 mutatnt clones), the killer apps are not established. Additionally, the requirement of a fresh-frozen tissue source (vs. FFPE) for solid-tumor analyses poses potential operational challenges in a clinical setting. We see FFPE compatibility as a major opportunity for differentiation in the single-cell analysis space, and a key to moving closer to the clinic. In the meantime, we still expect to see growing adoption of single-cell analysis in the translational research setting.
Fluidigm reports increase traction of mass cytometry and imaging mass cytometry – In the single cell vein, we also noted increasing use of high-parameter profiling of single cells using mass cytometry. There were a >25 abstracts highlighting the use of the technology, including several for I/O research (Fluidigm announced the Human Immune Monitoring Panel at the show). The new Hyperion Imaging System appears well positioned for the interrogation of protein biomarker signatures in tissues and tumors (including FFPE) at subcellular resolution (~1 um). This resolution enables analysis of cellular phenotypes, cell-cell interactions and interrelationships within the spatial context of the tissue microenvironment.
Other Key Oncology Trends (Or Lack Thereof)
Digital spatial profiling – The NanoString booth appeared busy (the company had 60 posters this year!), and was apparently so throughout the conference, with strong interest in their digital spatial profiling (DSP) platform. The company unveiled the DSP prototype instrument and its specs.The DSP can help resolve tumor heterogeneity at high plex in FFPE samples, which makes it attractive. David Rimm showcased data with 44 proteins, and was able to discover 5 biomarkers in his cohort of 55 patient study. Of note, on the biopharma side, Genentech showcased the first RNA and Protein spatial profiling assay that reveal tumor and stroma signatures. We look forward to future development for DSP!
Multi-modality testing – Recent conferences with a more cutting-edge lens (e.g., PMWC, AGBT) highlighted the exciting potential of multi-modality testing, for example incorporating sequencing, IHC and EMR data (e.g., Tempus, NantHealth). This trends was not as prominent as we expected at AACR.
Author | Andrew Aijian
|Andrew Aijian is a Sr. Project Leader at DeciBio Consulting with expertise in the immuno-oncology, companion diagnostics, and molecular diagnostics markets. Andrew leads in a wide variety of consulting engagements in these markets, including market sizing, competitive landscape assessment, product strategy, and commercial due diligence. Connect with him on LinkedIn.|
Author | Susan Zhou
|Susan Zhou is a Senior Analyst at DeciBio with experience in variety of consulting engagements within the molecular diagnostics industry, including market sizing, opportunity assessment, and go-to-market strategy. Susan is passionate about supporting innovative diagnostic technologies, with particular interest in liquid biopsies. Connect with her on LinkedIn.|
Author | Stephane Budel
|Stephane Budel is a partner at DeciBio with over 12 years of combined experience in life science business consulting, entrepreneurship and academic research. Connect with him on LinkedIn.|
Disclaimer: Companies listed above may be DeciBio clients and/or customers.