Here we go again.
Two months ago, I noted that Dr. Stanislaw Burzynski, the Polish expatriate physician who started out as a legitimate medical researcher and then in the late 1970s took a turn away from science-based medicine and towards being a "brave maverick doctor" through his discovery in blood and urine of substances he called antineoplastons with—or so Burzynski claims—major anticancer activity, had finally published the results of one of his clinical trials. Actually, he published the results of 42.5% of one of his clinical trials, given that he had, without adequate explanation, excluded most of the patients in the trial from analysis. As I explained at the time, the one subject who was a long time survivor, almost certainly didn't have an aggressive cancer, but rather had an indolent tumor. Overall, the publication was so bad that it didn't even convince hard-core Burzynski supporters, particularly given that Burzynski's clinical trials were described in the 1990s by his own lawyer as an "artifice," nothing more than a means to administer antineoplastons without pesky interference from the FDA.
When I first learned of Burzynski's promise to publish his results, I made a prediction. I predicted that he might get his studies published in the odd bottom-feeding journal or pay-to-play "open access" journal more interested in collecting fees to publish than in scientific rigor. I was quite sure that he wouldn't publish in a reputable journal. I was mostly right. The last study he published—or, more properly, 42.5% of a study—was published in Child's Nervous System, which, from what I've been able to gather, while not by any means a top-tier journal, is not a particularly disreputable one either. The same can't be said, as far as I've been able to determine, of the Journal of Cancer Therapy. It's a journal to which Burzynski has resorted before, whose articles were not indexed by PubMed. It's to this journal that he's returned with a study hot off the presses (or, at least, the pre-presses), A Phase II Study of Antineoplastons A10 and AS2-1 in Children with High-Grade Glioma. Final Report (Protocol BT-06), and Review of Recent Trials. It's a study that's been touted on Ric Schiff's ANPCoalition website, a Burzynski propaganda site that I've discussed before.
I must say, this is one of the oddest publications I've ever seen. For one thing, only nineteen patients were recruited over a period of twelve years, from August 24, 1993 and closed to accrual on May 19, 2005. It's also odd, because searching for this study in ClinicalTrials.gov under terms such as "BT-06" and "BT-6" lead me to conclude that this is the clinical trial that this paper is reporting, and there's not much detail for the trial. More importantly, the trial start date is listed as being before the year (1997) when Burzynski submitted his 72 clinical trials to the FDA, one of which was BT-06. What's also suspicious is that by June 16, 2006 all subjects were removed from therapy due to various reasons, from death to responses, to patient requests, which makes it odder still that it took Burzynski eight years to publish, given how lethal the brain tumors being studied are. Part of the explanation is likely that "The study was listed by the National Cancer Institute (NCI) and the protocol was initially drawn up by the NIH and later transferred to the BRI and BC." The reference given was this one, which concluded that one of the components of antineoplastons, phenylacetate, "has little activity at this dose schedule in patients with recurrent malignant glioma."
The same appears to be true of antineoplastons.
What first leaps out is that there were only 15 patients evaluated. Three patients were discounted because of brainstem location. (Unlike the previous study, this study examined patients without a brainstem lesion.) Apparently one had no baseline scan, which makes me wonder why the patient was even considered and why he didn't have a baseline scan. Another thing that leaps out is that this is yet another example of Burzynski's treatments not doing particularly well. Certainly, they're no miracle cure. Take a look at the survival graph:
Burzynski spends considerable table space and verbiage comparing his results to other studies, stating that one patient survived ten years from treatment start. As was the case with the previous study, that one patient makes all the difference in the world. Without that one patient, Burzynski's results would likely be indistinguishable from all the other results summarized in Table 3.
There's another issue as well. Burzynski presents two sets of MRI scans of patients who were reported to have had a complete response. Now, I'm not a neuroradiologist, but I do know a bit about reading MRI scans. In Figure 1, scans are presented of a patient with an anaplastic astrocytoma that recurred after surgery, radiation, and chemotherapy. The baseline scan shows an enhancing lesion that's fairly extensive, but to me it's unclear how much is post-surgery artifact, how much is radiation artifact, and how much is tumor. The followup MRI a year later still shows considerable enhancement. I really can't figure out how this could be considered a complete response. Perhaps a neuroradiologist reading this paper can tell me. I'm not proud; I (usually) know the limits of my abilities to read scans. I'm simply more puzzled when I look at Figure 2, which looks far more convincing to me but is billed only as a partial response.
Overall, though, Burzynski's results are singularly...unimpressive. This is particularly true if you take into account all that we know about how he's manipulated "complete responses" and how he's manipulated data in order to categorized subjects inappropriately as having had complete and partial responses. So when Burzynski reports that, out of the 15 subjects, two achieved a complete response, two achieved a partial response, and three had stable disease, I'm hard-pressed to believe any of it, given Burzynski's history, plus his tendency to mistake the normal regression of pseudoprogression for a real response and his counting responses in subjects who are still on high doses of steroids to decrease brain swelling.
Finally, as is so often the case, one of Burzynski's own supporters reveals something that probably would have been better off unrevealed, at least for Burzynski. Ric Schiff writes:
Obviously no article could completely detail all of the nuances involved, so I will share a significant one missing from this partial report. The Atypical Rhabdoid described at the bottom of page 568 was my daughter, Crystin. The article graciously describes her death as a result of pneumonia, which is technically correct. What it does not tell you, nor would it be appropriate to mention in that particular form of media, is that the pneumonia was a direct result of the damage that Crystin had sustained from her previous radiation and chemotherapy treatments.
It is fact that my daughter had progressive disease when she went to BRI for treatment. It is also fact that after being on ANP treatment for about 8 months her MRI’s showed no disease. I took her off ANP treatment and within a month her cancer had returned and was quickly spreading throughout her brain. Within 9 weeks of resuming ANP treatment Crystin’s MRI showed no more visible enhancement. Crystin subsequently died a year later of radiation necrosis……the radiation that she had originally received had been to such a degree that it was lethal.
All of this was documented by the University of California in San Francisco. Their pathologist, their MRI machine and in the end, their autopsy. The article is quite accurate when it cites that Crystin’s autopsy showed no viable cancer cells anywhere in her body.
So, if one child in this study was made cancer free after being treated with ANP, and then died as a result of her previous treatments, how many others would still be alive today had they not received their initial chemo/radiation treatments?
These Phase II clinical trial results are astounding in comparison to other treatments, but it is far more likely that they would even have been better had these patient’s received ANP as their primary treatment. It’s time to start saving lives……
First off, even if Ric Schiff's interpretation of events with respect to his daughter's treatment and death is correct (and there's plenty of reason to doubt it, given the inconsistencies in his story and the lack of medical records presented to back it up, a contrast with a lot of other Burzynski patients), that's not how clinical trials work. If a subject dies while on trial, no matter what the cause, it's a death and contributes negatively to at least the overall survival calculated for the trial. In any case, if you actually read Crystin Schiff's story, you'll see that it's pretty implausible, particularly the part where Ric Schiff seems to be saying that the oncologists treating Crystin had some sort of "prejudice" against her and that's why they wouldn't treat her any more. My educated guess is that, far more likely, given Schiff's insistence on having Burzynski treat his daughter, the oncologists at UCSF had likely decided that, ethically, they couldn't continue to treat her and thus make themselves party to what Burzynski was doing. It's an ethical dilemma that I understand. Later, because of privacy regulations, the physicians can't tell their side of the story, but Schiff can say whatever he likes—and does. Particularly implausible is Schiff's accusation of, in essence, murder by the radiation oncologist, to cover up the amazing success of Dr. Burzynski's treatments:
It seems that the radio-oncologist was so confident that our child would not survive, he gave her a lethal dosage of radiation. From his perspective, Crystin would not live long enough for us to see the long-term effects of his radiation treatments. He felt that we would be grateful for his extending her life a few months, so why tell us. Of course, he wasn’t planning on us going to Dr. Burzynski and having our little girl cured.
While I still feel sympathy for Schiff for having had to watch his daughter die of a horrible tumor, I still can't help but note that his reaction to the tragedy seems to have warped his judgment. He frequently brags of his super awesome police skills that apparently allow him to detect all fraud not finding fraud in Burzynski, as if police skills would be enough to tell him whether Burzynski's treatment had any scientific validity or not (they aren't) but is so hypersensitive to "fraud" elsewhere that he repeatedly accuses the UCSF pediatric oncologists of, in essence, intentionally killing his daughter so that it would appear a Burzynski patient died or out of some sort of revenge (I've seen implications of both). I'm sure that Schiff really and truly believes that that's what happened, but, as they say, believing something doesn't necessarily make it so, and Schiff's interpretation of events has in the past been questionable at best.
I wouldn't have mentioned this at all, but my sympathy for his loss 19 years ago will not keep me silent when I see Mr. Schiff write such irresponsible tripe in which he claims that his daughter would have lived if she had been treated with just antineoplastons first and not received conventional therapy beforehand. That's dangerous misinformation, and it's certainly not supported by this study whose results are not particularly impressive when one takes into account everything that is known about Burzynski, his studies, and how he's long played fast and loose with rules regulating patient protection. One wonders if the editors know that the Burzynski Research Institute IRB is chaired by a crony of Burzynski's from the 1970s who is chair of the board of directors, Carlton F. Hazlewood, PhD, as documented numerous times before in FDA warning letters. That the Journal of Cancer Therapy ever published any clinical trial by Burzynski shows what a bottom-feeding, worthless journal it must be. The editors should be ashamed.
What a mess.
Study Start Date:August 1993"
From the NCI:
BC-BT-6 / NCT00003535
"PROJECTED ACCRUAL: A total of 20-40 patients will be accrued for this study."
... or, you know, 15. Fifteen is like, basically 40. Close enough.
Population size matters. The smaller the population, the more likely we are to confuse outliers for significant differences, something Dr. Burzynski should have learned at Baylor (BCOM) when he was in the Department of Anesthesia there, under the direction of Georges Ungar.
Ungar, it must be remembered, famously published in Nature his research on a chemical basis for memory transfer that has since been discredited as a viable idea based, as it was, on small population studies.
... then again, maybe Dr. Burzynski DID learn something at Baylor.
And the journal is published by Scientific Research Publishing which is on Beall’s List (http://scholarlyoa.com/publishers/) of "Potential, possible, or probable predatory scholarly open-access publishers"
I doubt that publishing there has much of an impact factor unless you can have a negative impact score.
Might be off topic, but I just found an interesting article at the Daily Mail (of all places) about faith healers selling quack treatments, including for cancer. Having just seen Jake's talk before Autism One, I wonder if that group would screen even these people if they promised to cure autism.
I am sorry, but isn't unlawful to go around saying that X has killed somebody? I mean, if somebody said so of me, I would sue them, grieving or not grieving.
SCIRP as an organization is about the same caliber as OMICS. Republishing papers from other journals and listing people as editors without their knowledge? Check. Accepting machine-generated papers? Yup.
Could they publish a decent journal? Depends on the EIC (cf. h__p://scienceblogs.com/aardvarchaeology/2012/04/12/scam-publisher-fools-swedish-c/) and the editorial board. In this case, the EIC listed is Sibu P. Saha, University of Kentucky. A quick look at his publications doesn't show a lot of relevant expertise -- mainly similarly titled papers about pneumonectomy in NSCLC in regional journals.
I am sorry, but isn’t unlawful to go around saying that X has killed somebody? I mean, if somebody said so of me, I would sue them, grieving or not grieving.
It's necessary for the individual to be identifiable, at least by relevant parties. Seems like UCSF itself might have a claim, though.
Might be off topic, but I just found an interesting article at the Daily Mail (of all places) about faith healers selling quack treatments, including for cancer
The general rule at the Daily Mail and similar UK tabloids is that if someone is charging hundreds of pounds for cancer treatment, they are villainous faith healers who deserve to be exposed and prosecuted; if they are charging hundreds of thousands of pounds, they are brilliant maverick scientists fighting the academic establishment, and the only hope of survival for brave little patient X.
@HDB: My trepidations come from the Daily Mail's prior hero worship of Andy Wakefield. They also repeated Sharyl Attkisson's "murder-is-okay" story on Alex Spourdalakis.
@c0nc0rdance: good point. 15 patients (well, 19 to be fair, as they do tell us about 4 patients they excluded) reported is really not the same as 40 patients who were planned.
There are 2 possible explanations here.
One is that the study was stopped early before the planned number of patients were recruited.
Well, that's certainly possible. But if a trial is stopped early, it would be completely standard practice to report that, and the reasons why it was stopped early, in the publication. Burzynski's publication is silent on that matter.
It's interesting to note this statement in the statistics section of the paper (section 3):
"An interim analysis would be conducted after 20 subjects had enrolled in the study. If one or more patients achieved a confirmed radiographic response, an additional 20 subjects would be recruited"
So that certainly leaves the door open to stop after 20 patients, and I'm not going to quibble about whether it's actually 20 or 19. But, crucially, they planned to stop only if there were no responses in the first 20 patients.
So we have a discrepancy. The results section says there were 2 complete responses, which should have meant they carried on after the interim analysis. But clearly they didn't carry on.
I can think of 2 possible explanations for that discrepancy. Either they followed the protocol, in which case there were no responses and they just made shit up for their results section, or they didn't follow the protocol, and have provided no explanation of why.
The second possible explanation for the discordance between the planned 40 patients and the reported 19 is more troubling.
Maybe they really did recruit 40 patients, but have simply not reported the other 21. Perhaps because they died after an embarrassingly short period of time.
Disturbing news just came to me overnight. I have not yet been able to verify it, but it appears that the FDA might have lifted the partial clinical hold on BRI and the Burzynski Clinic. We're looking into it.
That's terrible news, if in fact, the FDA has lifted the partial clinical hold on the BRI and Burzynski's clinic.