Getting Personal at ASCO 2017: Precision Therapies, IO, and Autologous Cell Therapy

Posted June 13th, 2017 by Michael Gladstone, in From The Trenches, Personalized Medicine, Science & Medicine, Translational research


This blog was written by Michael Gladstone, Principal at Atlas Venture, as part of the From The Trenches feature of LifeSciVC.

The 2017 American Society of Clinical Oncology (“ASCO”) meeting concluded last week, providing patients, physicians, and industry stakeholders with a big batch of new clinical breakthroughs and disappointments.

This event never lacks media coverage, so rather than run through all of the highlights, I’ll draw attention to several underlying themes I find striking, particularly from my perspective as an early-stage biotech company-builder and investor.

Targeted Therapies are Back in the Spotlight

While immunotherapy has attracted its fair share (or more…) of attention and investment in recent years, 2017 has already been a big year for targeted therapies in genetically stratified patient populations. Two examples are inhibitors of TRK and PARP.

TRK Fusion Inhibitors: Small Biotechs Delivering Personalized Therapies

Loxo Oncology’s larotrectinib is producing striking efficacy in rare patients with fusions of the TRK kinase, validating this target as a druggable driver oncogene across diverse tumor types. Matthew Herper’s Forbes story on this drug calls attention to the unique challenges of identifying and recruiting the very rare patients who have have these mutations and could likely benefit from treatment. As told to Herper by one of Loxo’s clinical investigators:

“You get an email or a phone call: such-and-such patient was identified in Alabama, in Brazil, in wherever. Every one of those cases was an adventure. Every one presented another challenge to getting a patient on the study. We had people coming in the dead of winter in New York from climates where they didn’t own coats.”

It doesn’t surprise me that young, small biotechs (like Loxo, Ignyta, and Blueprint) are leading the charge in developing these personalized therapies, given the unique “hand-to-hand combat” involved in patient identification, trial execution, regulatory strategy, and eventual commercialization. And at least in Loxo’s case, that is all happening fast; Loxo’s Series A financing was less than 4 years ago, and they plan to submit an NDA as soon as this year.

BRCA/PARP Synthetic Lethality: When breast/ovarian cancers go low, we go high

A trio of competitors (AZ, Tesaro, and Pfizer via its Medivation acquisition) are generating compelling late-stage data with PARP inhibitors in BRCA-mutated ovarian and breast cancer. In these patients, oncogenic mutations in the BRCA genes also provide a unique therapeutic opportunity: these mutations impact a key DNA damage repair mechanism, rendering these tumors sensitive to pharmacological inhibition of a compensatory pathway.

PARP inhibitors appear ready to catapult past chemo in a variety of clinical settings, rightfully stoking interest in other DNA damage repair pathways and synthetic lethal opportunities. Yet the BRCA/PARP synthetic lethal story is not particularly young; it took years to deconvolute this biology and successfully deploy these drugs clinically. But in the wake of these breakthroughs we should soon learn much more about the breadth of PARP inhibitor potential, how to address PARP resistance, and (perhaps most intriguingly) how to validate new targets with analogous genetically-defined susceptibilities in other clinical settings. I’m optimistic that a set of nimble young biotechs can leverage new tools (e.g., CRISPR screening and novel bioinformatics approaches) and insights to crack the next wave of opportunities here more rapidly than the first.

Next-gen IO and Cold Tumors: There’s smoke, but is there fire?

Several commentators reacted last week (at times with apparent contrarian glee) to the disappointing early clinical results of trials with several novel IO agents. But attrition in this space should not be surprising; thinning of the clinical herd is an inevitable outcome of the breakneck pace of competition in this field.

But what struck me most wasn’t the reaction to the losers … It was the debate over the therapeutic robustness of the apparent combo trial winners, given the dearth of randomized trials. A glaring message emerging from this ASCO is that it is very difficult to definitively surmise the activity of new agents based upon non-randomized PD-1/-L1 combo studies.

There are now well over 1,000 IO trials running (and seemingly more each day), and most contain at least one arm combining a PD-1 axis inhibitor with another therapy. This pace of investment and clinical innovation provides an unprecedented number of “shots on goal.” But most of this data is being generated in Phase 1/2 studies with some substantial limitations:

  • Many agents in these combo studies have not demonstrated meaningful single-agent activity.
  • Limited follow-up and small N’s: Many studies are presented in “real-time” with data still maturing, so perceived efficacy is gleaned mostly from objective response rates (“ORR”) in relatively small patient cohorts, without mature data on duration of response (“DOR”) or survival.
  • Uncertain “re-responsiveness” of checkpoint-experienced patients: We still don’t have much firm data on “re-responsiveness” of patients who have previously failed to respond to (or relapsed on) PD-1 inhibitors who are re-treated with a PD-1 inhibitor-containing regimen. Given the ongoing evolution of both the tumor and immune system in these patients, it stands to reason that some meaningful proportion of even these “PD-1 progressors/non-responders” may indeed still be responsive to re-treatment. This makes it harder to assess the contributions of a novel agent combined with that checkpoint inhibitor in this population.
  • Uncontrolled, non-randomized studies: All of the challenges above (and others) amplify the usual challenge of interpreting non-randomized oncology studies. There is tremendous heterogeneity both within and across Phase 1/2 oncology studies, with patients often drawn from a diverse pool of tumor types, clinical characteristics, and prior treatment histories. At the risk of stating the obvious, this complicates the utility of cross-trial comparisons, especially when layering novel therapies on top of an active backbone drug (like a PD-1 mAb). This also complicates the design, endpoint selection, and powering of subsequent registration studies.

The clearest takeaway here for a small biotech is “make your drug extremely, undebatably active as a single agent even in highly treatment-refractory N-th line patients.” Because my sophisticated readers are unlikely to find that recommendation insightful, my next clearest takeaway is the importance of a coherent partnership strategy.

A deep-pocketed partner provides more than cash and complementary assets for combo studies. They also facilitate trial design and interpretation by leveraging rich, granular, comprehensive data across many studies in a huge diversity of patients. Your best shot at usefully interpreting non-randomized studies comes from having an exquisite understanding of the activity of each agent (and comparator agents) in monotherapy and in other combos in patient populations similar to the one being your data is generated in.

Admittedly, that does still sound quite complex and challenging. What about that strategy of “make your drug extremely, undebatably active as a single agent even in highly treatment-refractory N-th line patients”?

CAR-T in Multiple Myeloma: BCMA Stands for “Bluebird Crushed My Assumptions”

Sorry. I know bb2121 deserves a better transition and a much better acronym than that, but I am not a natural writer, and this is the best that I can do.

Bluebird’s BCMA CAR-T data astonished me. In my view, these data re-affirm the potential of engineered cell therapies to transform the treatment of cancer.

Bluebird treated patients with a median of 7 lines of prior therapy (and up to 14), with 100% of patients having had prior autologous stem cell transplant. 71% of the treated patients have previously received 2 proteasome inhibitors, 2 IMIDs, and anti-CD38 Darzalex. And in this population, with patients treated with active doses of therapy, Bluebird produced 100% ORR (27% CR; 46% VGPR), with zero DLTs, and a remarkable safety profile (no DLTs, low rates of CRS, and no Grade 3/4 neurotoxicity). The responses are very durable so far and occasionally deepening (with several PRs and VGPRs improving to VGPRs/CRs), with all 4 patients assessed for minimal residual disease found to be MRD-negative.

bb2121 (and other BCMA-directed approaches) obviously have a very long way still to go. But given the risk/reward profile of this therapy so far even in highly refractory patients, it’s tempting to imagine a brave new world where drugs like this are used earlier in the treatment paradigm, when patients have more tractable disease, improved clinical condition, and where a one-time therapy may displace a decade or more of a cocktail of often quite toxic therapies costing $300,000-500,000/year.

We aren’t there yet, and who knows whether bb2121 will ever get there, but this is the potential many of us see for cell therapy, and it is a future worth rooting for while we weather the next (inevitable) round of CAR-T ups and downs. Meanwhile, Novartis and Kite have each presented very impressive results in their respective registrational trials of CD19-targeting CAR-T therapies in relapsed/refractory non-Hodgkin lymphoma. These provide a roadmap for approval for autologous CAR-T in heavily pre-treated patients, and I am optimistic that they will chart a path into earlier lines of treatment with their current therapies or next-generation versions.

Concluding with a Balance of Realism and Audacity

Any column like this requires a big dose of pragmatism alongside its optimism. The battle against cancer is unrelentingly humbling, and we can be sure that much of what we celebrate today may turn out to disappoint us tomorrow. But great strides are surely being made, and this column provides at least a few examples of new medicines (including several entirely new therapeutic classes) providing lasting survival benefit to an appreciable portion of patients. That warrants real excitement and some audacious goals for the next generation of therapies and patient care.

Michael Gladstone

Principal at Atlas
This entry was posted in From The Trenches, Personalized Medicine, Science & Medicine, Translational research and tagged , , , . Bookmark the permalink.