As an investor in Advaxis, I follow their scientific development very closely. This company is also blessed (cursed?) with a very active, vocal, intelligent community of cancer investors who are constantly probing deeply into the findings of the company’s studies. These communities are full of laypeople and professionals alike, and it’s a challenging environment to contribute anything of value to. But I do answer questions when I can, and this one I got was particularly valuable for anyone interested in cancer stocks.
hovacre, if you have the time, could you please look at what Jcgmss posted 3 hours ago about Keynote 46 vs Keynote 199? It looks like there is a 50% improvement statistically in #1: PSA declines of greater than 50% comparing PEMBRO alone vs PEMBRO in combo with ADXS-PSA AND and #2: A 50% improvement in ANY % of PSA decline comparing PEMBRO alone vs PEMBRO in combo with ADXS-PSA. I realize these numbers were based on a relatively small sample size and there is still a lot of work that needs to be done but if these numbers hold, are they significant enough for the combo therapy to be the SOC vs PEMBRO alone? Also what is the meaning of the comment about the 1/2 lines of chemotherapy and the impact on the results? That kind of lost me. I realize you are a busy person to already be up at 4:30 AM CST but I value your opinion and would appreciate your thoughts. The closer to “layman terminology” you can make it the better it will be for myself and others to understand. Some of us are downright stupid when it comes to this sort of thing
First, a bit of background. Advaxis is studying Listeria-based immunotherapy for the treatment of various solid tumors. One of its programs is called ADXS-PSA, which is designed to train the immune system to recognize and fight prostate cancer cells that have lots of prostate-specific antigen (PSA). In KEYNOTE-046, ADXS-PSA was combined with Merck’s pembrolizumab (branded Keytruda) to see if the activity would be better than ADXS-PSA alone.
On a separate track, KEYNOTE-199 is assessing pembrolizumab alone in patients with previously treated, metastatic prostate cancer.
Results of both studies were presented at this year’s Annual Meeting of the American Society of Clinical Oncology (ASCO). This is an incredible event drawing a small city’s worth of cancer professionals, and it’s one of the biggest single treasure troves for scientific information surrounding due diligence for cancer stocks.
Comparing the noncomparable: a total taboo for clinical trials of the cancer stocks?
There’s a general rule of thumb in clinical medicine: do not make “cross-trial” comparisons. That is, if the study did not specifically include the two drugs you’re looking to compare, then just don’t do it. In the case of KEYNOTE-046 and KEYNOTE-199, this would be ADXS-PSA/pembrolizumab vs pembrolizumab alone.
Why did this come up? A poster on that discussion board indicated that the results were clearly better for KEYNOTE-046, and that the combo demonstrated a striking, unambiguous improvement in outcomes for patients.
Most notable was the ability to cause a decline in the blood levels of PSA, which is generally a sign that your cancer therapy is working to control tumor burden in prostate cancer.
As we can see here, KEYNOTE-046 looks better at first glance. But the PSA endpoint itself leads to my first issue: we don’t know the baseline levels of patients. At this time, the Advaxis poster is not available to the public in any form I can find. But they divulged in the abstract that the median baseline PSA levels for KEYNOTE-046 was 40.6 ng/mL, which is pretty high. In KEYNOTE-199, the authors reported the mean PSA levels, which were much higher (175.5-346.4 ng/mL), with a huge standard deviation indicative of the wide variation of this particular blood marker.
In essence, without knowing how these compare, the PSA reduction is basically meaningless. Consider if Merck’s study had reported the median and not the mean. A patient with 300 ng/mL achieving a >50% reduction is much more impressive than a reduction from 40 ng/mL. Given that the two studies generally reported similar Gleason scores (lots of patients at 8 or higher, which is poor prognosis), we don’t know if patients were healthier in KEYNOTE-199 or KEYNOTE-046.
Now, we turn to response and survival data. KEYNOTE-199 had more patients, so we can start there. Ten percent of patients saw tumor shrinkage of at least 30%. And a total of 41% had no signs of progression at the time of data cutoff. This was after following up around 8 months. Meanwhile, 21 out of 37 patients in KEYNOTE-046 achieved disease stabilization at least. A total of 56.8% had not progressed, and 5 patients remain on the combo arm of the study.
Overall survival interpretation
Regarding the overall survival data, I had this to say:
And after all that, the patients with visceral metastases had median overall survival of 9.5 months (PD-L1-positive) and 8.0 months (PD-L1-negative) in KEYNOTE-199. We don’t know which, if any, of the patients in KN-046 had bone-only metastases (this is a better sign), but given the treatment history, I presume they are patients with visceral metastases (a bad sign) [NOTE: I have since gotten ahold of the poster, and it was 30% of the patients who had visceral metastases]. And median OS has not been reached after 13 months follow-up. I find this very encouraging, but I view it with caution in light of what we do NOT know about these studies. And, as you state, the patient numbers are smaller than KEYNOTE-199, so we need to be cautious in drawing firm conclusions in any direction, other than pembrolizumab has encouraging efficacy. ADXS-PSA might build on that, but you will absolutely not see the company pursuing approval based on these results; they need to be confirmed in a larger trial.
But another problem is that we cannot even compare the treatments received in the two trials. We know that most patients in KEYNOTE-046 had at least 3 prior lines of therapy. And we know that around 1/4 of the patients in KEYNOTE-199 had 2 or more lines of chemotherapy, and 1/4 had 2 or more lines of hormonal therapy.
We don’t know how much these patient populations overlap! The broad suggestion here is that KEYNOTE-046 enrolled patients who were sicker. But we cannot corroborate that. Therefore, when we see a minor difference in the decline in PSA, is it due to a fluke chance? Is it because the patients are sicker? Is it because of some other factor not considered in the trial reporting? We have no clue.
And this is the big reason why it’s so dangerous to perform cross-trial comparisons. There is a rich and complex layering of information that has to be accounted for. And when you conduct a trial “head to head,” you will enroll patients who are as similar as possible in terms of their health status, their age, their gender, their race, their history of risk factors, etc. This is critical for assuring that your results are not coming from some kind of chance.
And so it goes with this comparison, as well. For Advaxis shareholders, it’s frustrating, since this is the first sign that ADXS-PSA is helping patients. And KEYNOTE-199 was the first-ever report of pembrolizumab alone specifically in patients with prostate cancer. But we were unable to learn very much at all from these two studies and how they might relate to each other!
But so it goes, and these factors are just a few for why any cross-trial comparisons must be made with extreme caution. And we should definitely avoid making definitive statements about these comparisons.
Like what you read? Follow me on Seeking Alpha to receive daily updates and see articles I don't write anywhere else. Just click the button below!