The quack view of preventing cancer versus reality and Angelina Jolie, part 4

Why, oh why, did I look at GreenMedInfo again?

You remember GreenMedInfo? It’s yet another wretched hive of scum and quackery, but with a twist. Its proprietor, Sayer Ji, thinks he’s an expert at interpreting the biomedical literature. Unfortunately, as he demonstrates time and time again with depressing regularity, he is nothing of the sort. In reality, what Ji is an expert at is cherry picking medical studies and torturing them until they confess agreement with whatever quack idea he’s currently espousing. In the wake of the news coverage of Angelina Jolie’s decision to have her ovaries removed because she had a dangerous mutation in the gene for BRCA1, a mutation that gave her a very high risk of breast and ovarian cancers, he demonstrated that rather well. Indeed, last week, I deconstructed his criticism of Jolie’s decision and the incredibly faulty scientific and logical reasoning behind it. Other than a minor Twitter confrontation with Ji the day after I posted, I moved on after having refuted his nonsensical and scientifically ignorant arguments because, well, you have to move on in blogging.

Unfortunately, Ji revisited the topic of BRCA1 mutations just yesterday with an equally fallacious article about BRCA1 mutations entitled Is BRCA (“Breast Cancer Gene”) A Death Sentence? I could have told Ji the answer if he had asked me and saved him a lot of trouble. The answer is no. BRCA1 mutations predisposing to breast and ovarian cancer are bad, but they are hardly a “death sentence.” Of course, Ji, demonstrating that the intended correct answer to any question used as the title of an article is no, goes beyond that basic answer to try to argue again that BRCA1 mutations don’t matter and that “genes are not destiny,” basing his claims on a recent systematic review and meta-analysis by van den Broek et al entitled Worse Breast Cancer Prognosis of BRCA1/BRCA2 Mutation Carriers: What’s the Evidence? A Systematic Review with Meta-Analysis. It is exactly what its title says, a meta-analysis of studies examining the effect of BRCA1 mutations on breast cancer survival. As is his wont, Ji begins the article very confidently, amazingly overconfidently, actually:

What we think we know about the BRCA (Breast Cancer Susceptibility Associated) genes causing cancer is patently false, according to a new meta-analysis on the extant literature on the subject of these gene variations on breast cancer survival prognosis.

No. Not exactly. Let’s just say that this meta-analysis doesn’t say what Ji thinks it does, at least nowhere near to the level the he does. Ji read this meta-analysis and concluded that it tells us that “what we think we know about BRCA1 is false.” It’s not, except in Ji’s mind. First off, this meta-analysis says nothing about BRCA1 mutations causing cancer. That’s just the underlying assumption of the review, because it’s so well-established that certain BRCA1 mutations confer a vastly elevated risk of breast and ovarian cancer that no cancer doc seriously argues against that contention any more. What this meta-analysis looks at is one question: Whether breast cancers associated with BRCA1 are deadlier than basic run-of-the-mill breast cancers, which is a common belief among physicians, one not without evidence to support it. Not surprisingly, Ji’s interpretation of the study at one point is so spectacularly wrong that it’s downright hilarious:

A groundbreaking new meta-analysis published in PLoS titled,”Worse Breast Cancer Prognosis of BRCA1/BRCA2 Mutation Carriers: What’s the Evidence? A Systematic Review with Meta-Analysis”, calls into question the value of using BRCA1/2 gene status to determine breast cancer survival prognosis, as is common practice today. This implications of this research may have wide-ranging effects as the present climate, following Angelina Jolie’s high profile decision to have prophylactic breast, ovary and fallopian tube removed due to her perceived “genetic inheritance,” is to equate BRCA status with bona fide and mathematically calculable disease risk certainty.

Jolie’s decision to subject herself to multiple prophylactic organ removal was based on the premise that her BRCA mutations would result in an 87 percent lifetime risk of developing breast cancer and up to 54 percent chance of ovarian cancer, as prognosticated by her doctors. The notion that BRCA genes have full or near full penetrance (the ability of a mutation to cause clinically identifiable disease) has profound implications for the health of millions of women who rely on these predictions to make life and death medical decisions.

I’ve dealt with this nonsense before, of course, but this is the same nonsense with a different spin. So I consider it useful to deconstruct that different spin. First off, this meta-analysis does not in the least bit call into question the finding that certain BRCA1 mutations can confer an 87% lifetime risk of breast cancer and a 50% lifetime risk of ovarian cancer. What it calls into question is whether the breast cancers caused by BRCA1 are actually deadlier than cancers not associated with such mutations. It’s been commonly thought that they are, and, in fact, there’s a fair amount of evidence that they are. The conclusion of this meta-analysis is also not a new finding. For example, a year and a half ago, Huzarski et al published a study examining ten year survival in cancers with BRCA1-positive breast cancer. Basically, 3,345 women with stage I to III breast cancer under the age of 50 were tested for three founder mutations in BRCA1. It was found that 7% carried a BRCA1 mutation and that the ten year survival rate in BRCA1 carriers and non-carriers was very similar, 80.9% and 82.2%, respectively, with no statistically significant difference between the two. Basically, what the meta-analysis shows is that the evidence is inconclusive as to whether BRCA1-associated breast cancers result in worse survival.

So why is Ji harping on this?

Basically, it’s the same nonsense that I discussed last time and expanded upon elsewhere. Ji is heavily invested in “proving” that cancer is not a genetic disease and that whatever woo du jour he favors, be it diet, lifestyle, supplements, or whatever, can prevent or reverse virtually any cancer. Cancers due to a genetic predisposition, such as BRCA1-induced breast and ovarian cancers are a direct challenge to that concept, and Ji cannot allow his readers to accept the scientific consensus with respect to them because then they might start doubting that what he’s selling will do them any good. So, like last time, he goes on the attack, ignoring the fact that the very meta-analysis that he uses as his jumping off point didn’t even conclude that BRCA1-associated breast cancers are deadlier, only that the evidence that they are is inconclusive.

Last time around, Ji conflated absolute risk of cancer with relative risk in BRCA1 mutation carriers, shamelessly quoting studies of sporadic breast and ovarian cancer as though they had any more than minimal relevance to cases of BRCA1-induced cancers. He’s even more shameless in spinning the question of overdiagnosis and overtreatment. For instance, he states that, according to a New England Journal of Medicine study, 1.3 million US women were “wrongly diagnosed and treated in the past 30 years.” Uh, no. Not exactly. The study to which Ji refers is, of course, Archie Bleyer and H. Gilbert Welch’s 2012 study that found that approximately 30% of mammographically detected breast cancers were overdiagnosed. I’ve discussed Ji’s characterization of this particular study before. “Overtreatment” doesn’t necessarily mean “wrongly treated.” In fact, arguably it does not. As we like to say in medicine, the retrospectoscope is 100% accurate. The problem was (and is) that as yet we do not and cannot know for mammographically-detected breast cancers which lesions will progress to threaten the life of the woman and which will not. Knowing what we know, we therefore are obligated to treat all screen-detected cancers as though they are potentially deadly. Until we have the tools to do that, we either have to treat them this way, or stop or decrease screening.

Ji continues to use loaded language here:

And so, previous studies on BRCA1/2 gene variations and the incidence of breast cancer have not taken this massive statistical inflation of non-cancerous “breast cancer diagnoses” into account, further feeding the illusion that having an identified BRCA mutation equates to having a inexorably higher risk of a deadly cancer, when in fact, in cases where BRCA was linked to so-called early stage or ‘stage zero’ lesions such as Ductal Carcinoma In Situ (DCIS), this condition was recently determined to be intrinsically benign by a NCI-commissioned expert panel and therefore should not be lumped together with other truly deadly forms of breast cancer, as is still common practice today despite the growing body of evidence against it.

DCIS “intrinsically benign”? That’s a rather obvious misrepresentation of what the NCI panel actually concluded. This is the paper to which Ji refers. It’s by Laura Esserman, Ian Thompson, and Brian Reid. Yes, Esserman et al, summarizing the NCI panel’s findings, argue that DCIS and precancerous prostate lesions should not be called cancer, but they hardly argue that they are “intrinsically benign.” In fact, they write:

Change cancer terminology based on companion diagnostics. Use of the term “cancer” should be reserved for describing lesions with a reasonable likelihood of lethal progression if left untreated. There are 2 opportunities for change. First, premalignant conditions (eg, ductal carcinoma in situ or high-grade prostatic intraepithelial neoplasia) should not be labeled as cancers or neoplasia, nor should the word “cancer” be in the name. Second, molecular diagnostic tools that identify indolent or low-risk lesions need to be adopted and validated. Another step is to reclassify such cancers as IDLE (indolent lesions of epithelial origin) conditions. An example is the reclassification of grade 1 papilloma to urothelial neoplasia of low malignant potential. Presciently, the rationale for reclassifying papilloma and grade 1 carcinoma as “papillary urothelial neoplasia oflow malignant potential” was “to take the lowest grades of tumor, the most benign-appearing lesions, and remove the word carcinoma.” A multidisciplinary effort across the pathology, imaging, surgical, advocate, and medical communities could be convened by an independent group (eg, the Institute of Medicine) to revise the taxonomy of lesions now called cancer and to create reclassification criteria for IDLE conditions.

In other words, lesions with a relatively low chance of progressing to cancer should be reclassified. This is a completely reasonable suggestion. However, Esserman et al make it clear that, while these lesions should probably not be called “cancer,” they are by no means “intrinsically benign,” as Ji falsely characterizes them. In any case, Ji claims that the existence of overdiagnosis somehow calls into doubt the high lifetime risk of breast cancer conferred by high risk BRCA1 mutations. It’s a profoundly stupid argument, and I’ll show you why. Let’s, for instance, assume an 87% lifetime risk of breast cancer from Angelina Jolie’s BRCA1 mutation, as her doctors told her she had. Now, let’s do something as simplistic as Ji’s reasoning but, in the context of his argument that overdiagnosis “inflates” the risk of breast cancer due to BRCA1 mutations. Let’s take a rate of overdiagnosis of 30%, meaning that any cancers detected in our hypothetical situation examining BRCA1 carriers are 70% likely to be “real.” Now, 0.70 x 0.87 = 0.61, or a 61% lifetime risk of developing a life-threatening breast cancer. Even using Ji’s simplistic model, that’s still pretty damned high. In brief, Ji’s handwaving and invoking of overdiagnosis are nothing more than an obvious smokescreen. Overdiagnosis does not even come close to explaining the high risk of breast cancer in BRCA1 carriers, and I’m not even taking into account how overdiagnosis declines as disease prevalence goes up. It is, however, known that DCIS is roughly as prevalent in BRCA1 mutation carriers as in non-carriers, but occurs at an early age, evidence in favor of the premalignant character of a lot of DCIS.

Ji also resurrects his claim that the existence of Borderline Ovarian Tumors (BOTs), which can also be overdiagnosed and overtreated, somehow inflate the number of cancers attributed to BRCA1. He cites a JAMA study he cited last time that “found that 5 times more women are diagnosed and treated with ovarian cancer than actually have it — indicating a massive problem that is not being taken into account by most literature on the role of BRCA mutations in cancer risk because these studies accept diagnosed cancer uncritically as actual cancer which is simply not the case due to the still largely unaccounted for issue of overdiagnosis.” This study, of course, is one big reason why we don’t screen the general population for ovarian cancer using transvaginal ultrasound and CA-125 blood tests. The study found that for ovarian cancer screening using these tests doesn’t work; it doesn’t decrease mortality from ovarian cancer and results in a lot of overdiagnosis.

Again, however, screening the general population is a very different case than observing women with high risk BRCA1 mutations for the development of ovarian cancer. Let’s just put it this way, even if, in general, overdiagnosis due to screening didn’t decline precipitously with increasing prevalence of a disease, we have plenty of evidence that prophylactic surgery is effective, that intervening, as was done with Angelina Jolie, to prophylactically remove the tissue at high risk for cancer dramatically improves overall survival of these women. Bilateral prophylactic mastectomy reduces the risk of breast cancer in women with high risk BRCA1 mutations by at least 90%. (See, for example, Hartman et al, 1999; Domchek et al, 2010; Rebbeck et al, 2004; Meijers-Heijboer et al 2001.) Removal of the ovaries is also highly effective in reducing mortality from ovarian cancer in such patients, lowering the risk by 70-90%. (See, for example, Kauff et al, 2002; Kauff et al, 2008; Finch et al 2014.) That prophylactic oophorectomy to prevent ovarian cancer and prophylactic bilateral mastectomy to prevent breast cancer in carriers of high-risk BRCA1 mutations are highly effective is not even controversial. Such recommendations are “baked” into major evidence-based guidelines, like those of the NCCN, as a significant option for women found to have such mutations to consider. Other options exist, such as enhanced screening (which has not been shown to be effective in ovarian cancer) or chemoprevention.

If you want to get an idea how mendacious Ji is, look no further than this passage:

Furthermore, prophylactic removal of the ovaries before age 45 (Jolie is 39) has been linked to 67% increased mortality risk, according to a 2006 study published in Lancet Oncology, indicating that organ removal as a generic form of “cancer prevention” may be doing the opposite of ‘saving lives’ as widely claimed.

That 2006 study? Take a look at just the abstract, and you’ll see why Ji is full of it. Specifically, the Method part:

From an existing cohort of all women who underwent unilateral or bilateral oophorectomy while residing in Olmsted County, MN, USA, in 1950-87, we analysed those who had received an oophorectomy for a non-cancer indication before the onset of menopause. Every member of the cohort was matched by age to a referent woman in the same population who had not undergone oophorectomy.

Yes, these were women who had their ovaries removed for non-cancer indications before 1987. I note that BRCA1 was not identified as a gene until 1990 and was not cloned until 1994. I also note that this study examined women who had their ovaries removed for non-cancerous reasons. Finally, Ji left a big finding of this study out:

However, mortality was significantly higher in women who had received prophylactic bilateral oophorectomy before the age of 45 years than in referent women (hazard ratio 1.67 [95% CI 1.16-2.40], p=0.006). This increased mortality was seen mainly in women who had not received oestrogen up to the age of 45 years. No increased mortality was recorded in women who underwent unilateral oophorectomy in either overall or stratified analyses.

In other words, the increased mortality of women who had their ovaries removed before age 45 was mainly seen in women who didn’t undergo estrogen replacement therapy until age 45 after their surgery. The authors also noted:

Although prophylactic bilateral oophorectomy undertaken before age 45 years is associated with increased mortality, whether it is causal or merely a marker of underlying risk is uncertain.

So, yes, Ji misrepresented another study. Quelle surprise. He also resurrects his observation that BRCA1-associated ovarian cancers might be somewhat less aggressive and more sensitive to chemotherapy than sporadic cancers, to which I again say: So what? It’s still a plenty deadly cancer, and certainly no advantage is seen in BRCA1-associated breast cancers. In fact, another 2015 meta-analysis published in a better journal and more comprehensively looking at both breast and ovarian cancer found that BRCA1-associated ovarian cancer does have a better prognosis but that BRCA1-associated breast cancer portended a significantly worse overall survival, in disagreement with the meta-analysis Ji cited.

Ji finishes with his real message:

And therefore, what do we really understand about BRCA gene ‘mutations’ when there are over 500 that have been identified, and whose complexity and role in health and disease are still not yet understood? The truth is that the linear and deterministic gene > trait > disease risk/prognosis way of thinking is archaic, and reflects the type of hubris that should have been sloughed off after the first draft of the human genome project in 2005 found that the ‘holy grail’ of molecular biology was not to be found in the genome, but in the interstitial space of its interactions with the myriad factors ‘beyond the control of the gene,’ the realm of epigenetics, which involves everything from the food your mother ate, your in utero exposures, to your breast feeding duration, the toxins and toxicants you were and are exposed to, your way of thinking, attitudes and beliefs and the downstream physiological effects they have, ad infinitum.

Ironically, the notion that genes determine destiny is more than just an idea but a reality for those who believe it and act on the meme, putting ideology into practice in their biology and medical decisions. This is why acknowledging the research that calls into question biological determinism and medical fatalism is so powerful and why we hope our readers continue to explore the primary literature itself as it expands and transforms the often out-dated knowledge base that conventional practice is still under the illusion is reflective of the truth.

Ji sounds rather like a creationist, doesn’t he, harping on the “interstitial space” of the interactions of the genome with various other factors? Oh, and the attacks on genetic “determinism,” too. He also makes a bald-faced appeal to ignorance. Yes, there are hundreds of BRCA1 mutations. Yes, we don’t know the significance of a lot of them. Those two observations, however, completely ignore the fact that we do know a lot about the significance and cancer risk associated with quite a few of those hundreds of BRCA1 mutations. Just because we don’t know the significance of a lot of them doesn’t mean we don’t know the significance of any of them. We do. Then Ji just repeats the same quack view of epigenetics, which is represented as a magic process by which humans can completely control their health and biologic destiny, genes be damned. It’s a dangerous delusion.

There’s a reason why Sayer Ji is dangerous. In my opinion, he advocates quackery, but he tries very hard to put a sheen of science on it. Because he is skilled at cherry picking the literature and seemingly “talking the talk” of biomedical research, he can sound quite convincing to lay people, even as real scientists who actually know something about the topics he’s writing about cringe as they read his simplistic and misleading misrepresentations. Actually, we either cringe or we laugh (or both), but it’s hard to laugh for too long at Ji’s ignorance. The reason is that, as ridiculous as I find Sayer Ji, I know that that his advice, if followed, will lead to the preventable deaths of women with BRCA1 mutations.