Where did this thing come from?
Daiichi Sankyo shared phase 1 data from its investigational HER2-targeted ADC DS-8201a at ESMO this weekend: no DLTs, and among the 20 patients evaluable for response, there was an ORR of 35% and a DCR of 90%.
The NEJM, despite its habit of publishing every precision medicine milestone in the past 15 years (including the Herceptin and Zelboraf papers cited in the first paragraph of this editorial), elected to give voice to the naysayers: per an opinion from physicians at the Princess Margaret Cancer Centre, treating cancer based on genomic profiling is unproven, expensive, toxic, and requires further investigation in controlled trials before we should consider it as a treatment strategy. No shock that this came out of Canada, that hotbed of medical innovation.
One of the the core accusations of this blog is that cancer patients are pushed toward clinical trials in spite of alternatives with demonstrated safety and efficacy, which takes advantage of two patient biases: new (if unproven) means better and your physician has your best interests at heart.
Physicians are paid to participate in clinical trials. Let’s get that out of the way. There is nothing objective about recommending trials to patients; you only need to notice how often a physician points a patient toward a trial outside his or her practice to deduce that trial suggestions are not without bias.
ONT-380 has been in development for a while, first at Array and then
Oncothyreon Cascadian Therapeutics. It’s unique among HER2-targeted small molecules because it does not inhibit HER1 (EGFR), which lends it a more favorable AE profile – it’s the EGFR targeting that is associated with most of the skin rash and GI toxicities we see in this category.
Significantly, ONT-380 is being evaluated for the treatment of CNS mets. In breast, brain mets patients are the ones that actually die – you can live forever with metastatic breast cancer, but brain mets are generally a dealbreaker. (The flip side of the excellent survival stats in breast is why you only see PFS, and never OS, data in breast – with OS as a primary outcome, you’d never get a study analyzed, because you’d wait years collecting mortality events.) If you’re HER2+ positive, and your brain lesions are too multiple or too big for Gamma Knife and you don’t want whole brain radiation (n.b., you do not want whole brain radiation), you’re basically left with lapatinib. We need ONT-380.
There’s not a lot of public data, but there’s signal that the drug works. These results from the Ib trial combining ONT-380 with T-DM1 (Kadcyla) found that in heavily pre-treated MBC patients with and without measurable CNS disease, there was an overall response rate of 47% and a respectable median PFS of 6.5 months. Both of these measures are exceptional for brain cases. This is phase I data summarized in an ASCO abstract, so we don’t have a lot of details on the patient population or their responses, but there were even a couple of complete response (CR) cases in the measurable CNS mix.
The treatment of brain mets is a massive unmet clinical need, but it’s a tough strategy to pursue, because these studies are hard to enroll. The ideal patient would be someone asymptomatic who has never been treated for CNS mets, and in a clinical setting where screening MRI for brain is not standard of care, that profile is not common. Even if you found a perfect patient, you’d have to get them to a site that’s actually enrolling. (I am a strong advocate for de-centralizing clinical trials, but that is an uphill battle and a post for another day.) There are other enrollment challenges for their ongoing Phase II study as well; patients are randomized to receive Herceptin and capecitabine (Xeloda) +/- ONT-380, and prior treatment with capecitabine in the metastatic setting is an exclusion criterion; however, capecitabine is often given with lapatinib, which is permitted in this study as long as it was used >12 months prior to enrollment. A lot of HER2+ metastatic patients have had the Herceptin/Xeloda/Tykerb combo (it was my first-line metastatic treatment), because it’s just about the easiest cancer treatment you can do, and I don’t know how many patients receive Tykerb with just Herceptin or with Herceptin and another cytotoxic agent. At any rate, I doubt eligible ONT-380 patients are just falling into investigators’ laps. And while I have the same complaints about our slow, unwieldy, archaic regulatory body as anyone else, in this case, the FDA does seem to get it: they gave ONT-380 Fast Track status in June.
But there’s something I like about Cascadian that has nothing to do with science (give me some credit; we haven’t even seen PII yet). It’s this, via a screen cap from their website.
Do you see it? How about this?
For contrast, check out what I pulled from the Ixempra site (ixapebilone from BMS; a microtubule-binding agent, similar to a taxane).
WTF is that? Who is that supposed to be?
These are marketing images, but they reflect a tension in the MBC treatment philosophy: are you going for a hike with your dog? Or are you just hanging on to see your grandchild born or whatever nonsense that lipstick mirror crap is supposed to signify? Cascadian is positioning itself as forward-thinking on multiple dimensions. They’re gambling on brain mets patients, who, despite a poor prognosis, are not staring in the mirror, trying to summon the will to hang on another day. Their patients are on a hike. You’re not dead; life still has meaning (this is part of why terrible side effect profiles on useless drugs are so offensive!). This company has balls. And they understand that patients Google, and good for Cascadian that these are the images they want patients to see.
Now bring on the PII results.
Just kidding. There is no wizard, no ruby slippers. Cancer is science, and hospitals are a business, and you’re going to use these facts to your advantage.
This is not your ordinary cancer blog. I don’t care about pink ribbons, or scanxiety, or tear-stricken queries about why everyone around you is dead while you’re still alive (hint: it’s called statistics). I care about outcomes, and treatment optimization, and lifting the veil on the fact that your oncologist does not always have your best interests at heart. This isn’t necessarily malice; he has other patients, she has different priorities, and they’re employed by organizations with their own motives and goals. But I once signed four informed consent forms in one day, all the while hearing the cha-ching of a cash register in the back of my mind, and I happily signed because I had something I wanted too. It was not the same thing as my oncologist wanted. And that’s the art of it: separating fact from preference, optimizing both, and, ideally, performing better than expected.
We’re not in Kansas anymore (definitely not, as Kansas doesn’t have an NCCN-designated cancer center; we’ll get to the fallacy of that soon). It’s a new day, where patients have the opportunity to be meaningful contributors to their own care. And I do mean opportunity. Not right. You have to earn it. So let’s pull back the curtain, shall we?