(Once Again) It Ain't Necessarily So

By Kevin E. Noonan --

A large part of the debate on patenting genetic diagnostic method and isolated genes has revolved around the effects of such patents on what is loosely termed "personalized medicine."  Personalized medicine can be summarized as a dream/holy grail/GATTACA future of universal genetic information -- every infant having her genomic DNA sequence determined at birth and contained on a medical identity card, to determine what diseases she may get, what drugs she should not and even who she should or should not marry (or at least mate with).  It is the most recent of the promises of the biotechnology or genetic revolution, made possible (in theory) by the fruits of the Human Genome Project; advances in genomic sequencing technology (as described in The $1,000 Genome and elsewhere) have quickened the expectations surrounding the technology.  But many have begun to wonder why this genetic fruit is taking so long to ripen (see, for example, "How Bright Promise of Genetic Testing Fell Apart"), and a recent study published in Science Translational Medicine may provide some clues.

The article, entitled "The Predictive Capacity of Personal Genomic Sequencing," was published on April 2, 2012 by Nicholas Roberts, Kenneth Kinzler, Bert Vogelstein, and Victor Velculescu from the Howard Hughes Medical Institute at Johns Hopkins University Kimmel Cancer Center; Joshua Vogelstein from the Department of Neuroscience, Johns Hopkins University; and Giovanni Parmigiani from the Department of Biostatistics and Computational Biology, Dana Farber Cancer Center and the Department of Biostatistics, Harvard School of Public Health.  The paper presents the results of a bioinformatics study on monozygotic twin pairs, using whole genome sequencing (WGS) that was interrogated for 24 different diseases.  For 23 of the 24 diseases tested, the results were negative, i.e., uninformative results for 19 of the 24, with risk of 50-80% for disease as compared with the general population.  However, 90% of the tested individuals were alerted to "a clinically significant predisposition" to at least one disease thus reflecting a silver lining if not a rainbow from the data.

The paper notes that there is an estimate of three million sequence variants per person (K. A. Frazer et al.,  2009, "Human genetic variation and its contribution to complex traits," Nat. Rev. Genet. 10: 241-51); of these, thousands of genetic variants have been associated with human disease, including Mendelian traits (e.g., sickle cell anemia), SNPs (e.g., Huntington's s disease), or by genome-wide association studies (GWAS) (examples include familial pancreatic cancer and Miller syndrome).

The paper addresses the question of what the benefit of such information would be, defining "benefit" as "receiving information indicating that the risk of disease is increased or decreased to a degree that would alter an individual's lifestyle or medical management."  The authors recognize that it is impossible to assess these benefits generally, but monozygotic twins present the possibility to make this assessment:  as they have in many other instances, the identity of genetic information make it possible that diseases and other traits with a large genetic component should be experienced in common between twins:  "If one twin of the pair has a disease, then the probability of the other twin developing that disease is dependent on the genome whenever that disease has some genetic component."  However, the paper also notes that "[t]he general public does not appear to be aware that, despite their very similar height and appearance, monozygotic twins in general do not always develop or die from the same
maladies," citing Wong et al., 2005, "Phenotypic differences in genetically identical organisms: The epigenetic perspective," Hum. Mol. Genet. 14 Spec No 1, R11-R18, and "Identical Twins Not As Identical As Believed," ScienceDaily, reflecting a limitation in the predictive power of genetic disease assessment even between individuals having almost identical genetic complements.  (Interestingly, even monozygotic twins are not necessarily identical genetic copies of one another, there being copy number variants between them; Bruder et al., 2008, "Phenotypically concordant and discordant monozygotic twins display different DNA copy-number variation profiles," Am. J. Hum. Genet. 82: 763-71.)  The twin populations were selected from "the Swedish Twin Registry, Danish Twin Registry, Finnish Twin Cohort, Norwegian National Birth Registry and the National Academy of Science – National Research World War II Veteran Twins Registry."

The authors also define "heritability" as the difference between the incidence of disease in monozygotic twins compared with dizygotic twins, "reflect[ing] the average genetic contribution to disease" in the population of twins studied.  Averages not as informative as distributions in this regard, since a given average incidence of disease could reflect either "a small fraction of twin-pairs with genometypes [(i.e., a complete genomic DNA sequence from an individual)] conferring high genetic risk or a larger fraction of twin-pairs with genometypes conferring a moderate genetic risk."  This challenge is illustrated by an example:

Suppose a woman receives a whole-genome test result indicating that she has a 90% lifetime risk (the total risk over her entire life) of developing breast cancer.  She may decide to have a prophylactic double mastectomy to prevent this outcome.  Similarly, if the test indicated an 80% or even a 50% lifetime risk of developing breast cancer, she may consider mastectomy.  On the other hand, if the test indicated only a 14% risk of developing breast cancer, then mastectomies would be considered by very few women, given that most women today do not opt for prophylactic mastectomies even though the lifetime risk of developing breast cancer in the general population is 12%.

The authors adopt the threshold of a "positive predictive value" of 10%, meaning that 10% of patients with a "positive" test result are expected to develop a disease, according to Clarke-Pearson, 2009, "Clinical practice. Screening for ovarian cancer," N. Engl. J. Med. 361: 170-77.  However, for several diseases, including chronic fatigue syndrome, gastro-esophageal reflux disorder, coronary heart disease-related death and general dystocia, this threshold is inappropriate due in part to the prevalence of these diseases in the population; for these diseases the threshold is a two-fold higher risk of disease compared with the general population.  In addition to these diseases, the study assessed the risk for coronary artery disease, stroke, cancer (bladder, breast, colorectal, lung, leukemia, ovarian, pancreatic, prostate and stomach), thyroid autoimmunity, diabetes (types 1 and 2), Alzheimer's disease, dementia, Parkinson's disease, irritable bowel syndrome, pelvic organ prolapse, and stress urinary incontinence.

The bulk of paper set forth a mathematical treatment of these biostatistics that is beyond the scope of the discussion here; the results, on the other hand, are informative.  These include:

•  "The fraction of patients that would receive a positive test varies widely from disease to disease."
•  "The majority of patients (>50%) who would ultimately develop 13 of the 27 disease categories would not test positive, even in the best-case scenario."
•  "There were four disease categories -- thyroid autoimmunity, type I diabetes, Alzheimer's disease, and coronary heart disease-related deaths in males -- for which genetic tests might identify more than 75% of the patients who ultimately develop the disease."
•  "The fraction of individuals in the population who would receive positive test results for each disease is small."
•  For 22 of the 27 disease categories studied, "a negative test would not indicate a risk that is less than half that in the general population, even in the best-case scenario" (which is probably not sufficient to warrant changes in behavior, lifestyle, or preventative medical practices).
•  An exception is Alzheimer's disease, where a negative test "could indicate as little as a ~12% relative risk of disease compared to the entire twin cohort . . . .  Knowledge of such a reduced risk might be comforting and relieve anxiety, particularly to those with a family history of Alzheimer's disease."
•  ">95% of men and >90% of women could receive at least one positive test result."
•  Many of these results represent the best-case scenarios and thus the true benefits of genetic disease testing may be overestimated.

The authors' conclusions are a dose of cold water on the hopes and expectations of many in the field (and even more laypersons outside the field, including policymakers, judges, and even Supreme Court justices):

[O]ur results suggest that genetic testing, at its best, will not be the dominant determinant of patient care and will not be a substitute for preventative medicine strategies incorporating routine checkups and risk management based on the history, physical status and life style of the patient.

The authors best state the significance of an accurate assessment of the likelihood of predictive genetic testing:  "Recognition of these merits and limits [for genetic diagnostic testing] can be useful to consumers, researchers, and industry, as they can minimize unrealistic expectations and foster fruitful investigations."  These are words policymakers should no doubt heed when considering changes to established patent and other policies based on perhaps unrealistic prospects for a brighter genetic future.

Comments

[David Koepsell]

None of which, of course, addresses the actual legal (and ontological) question involved: are patents on such genes patents of "products" or "phenomena" of nature. Also, doesn't this challenge the usefulness of getting these sorts of patents in the first place, and thus the value of such patents?

[Suleman Ali, Holly IP]

The authors have investigated the potential of genetics to predict future disease conditions. That does not seem to be the same as using genetic testing to provide the most appropriate medicines. I do not know this area very well, but it seems to me that efficacy of a drug for a condition may have a different basis from the cause of the condition, and so these results might not be directly relevant to 'personalised medicines'.

[lxf001]

I agree, Suleman. There is quite a difference in a genetic marker predicting disease and a genetic marker predicting disease response. Pharmaceutical companies are interested in the latter, i.e., genetic markers that help identify patients that will respond (or respond better) to their drugs. As a result, I do not believe these findings are relevant to personalized medicine, which has, as its goal, maximizing the benefit and minimizing the risk of a drug in a population already afflicted with a particular disease.

[Gary Johnston]

Kevin:

Great post. Not sure what you're driving at with your comment cautioning against changing established patent and other policies in view of the study. Insofar as you are cryptically attacking SCOTUS for its absolute no-brainer opinion in Prometheus, I do not see the connection. I agree with David. If anything, the study suggests that some/many/most/nearly all/all genotype-phenotype associations that are sought to be patented don't have "utility" anyway. So, again, what's the harm from Prometheus?

[Laurie Kellogg]

You seem to have skipped over the best quote: "In sum, no result, including ours, can or should be used to conclude that whole-genome sequencing will be either useful or useless in an absolute sense."

I am in total agreement with Suleman, also. The difference is whether one is talking prediction versus potential treatment.

Where we seem to be right now in the use of genomics in medicine is on the edge of being able to usefully analyze the incredibly large amounts of data from sequencing and just on the verge of the hypothesis generation phase. Using any analysis right now to discount possible uses of genomic information is unwarranted and premature. Until we know what we have, we cannot make any predictions of how to use it. But we also cannot continue drug creation in the old model, where pharmaceutical research centered around possible targets of unknown root causes. The new pharma model will need to be based on the root cause of disease, which is far more complex than just genetics or just environmental factors. Unfortunately, human disease is a very personal thing, which is why individual reactions to medications and treatment can be so, well, individualized.

What I am inelegantly attempting to say is that genomics will play its role, it already is (see Kailos Genetics, a company using markers to shunt breast-cancer patients into the optimal treatment plan). In some cases, such as one-gene-one-disease, it will be THE factor (see Kalydeco being used to treat the 4% of cystic fibrosis patients carrying mutations of the CFTR gene). And in others, there will be combinations along a spectrum.

Aside of this, though, I do not understand how this affects the patent process. Under patent law, one must feasibly show that under known conditions, laws or what-have-you, the invention would theoretically work, not that it absolutely does work, works every time or even that it works well.

[Kevin E. Noonan]

Dear Laurie:

Thanks for a great and informed comment. The gist of what I was driving at was that a great deal of the angst in patent law these days is directed at the possibility that IP protection will stifle innovation, particularly with regards to personalized medicine. My point was that we are sufficiently far from attaining this goal that current patents on genes (which will all expire by about 2020) will not factor into this calculus much, and that decisions based on these fears are misguided.

Thanks for reading.

[Kevin E. Noonan]

Suleman and lxf001:

The kind of companion diagnostics you mention is different, since it is a way to answer a specific question, i.e. will this patient benefit (or be harmed) by this specific drug. Entirely different paradigm and incredibly valuable, when it works.

Thanks for the comments.

[Kevin E. Noonan]

Dear David:

I cannot disagree. But then I don't think any patent claim reciting "an isolated nucleic acid encoding SEQ ID NO. X" is infringed when someone performs a genetic diagnostic test on the gene represented by SEQ ID NO. X.

And at this point, I think the bigger danger is not that human (or any) DNA will be deemed patent ineligible sui generis, but that some unnecessary "product of nature" exclusion will preclude patenting a great many natural products that are not DNA.

Hope you are well.

[Kevin E. Noonan]

Dear Gary:

If the Prometheus opinion was limited to those claims and claims like them I would have little problem with it. I will leave it to others if they want to rehash the analytical mess the decision imposes on patent law. With luck this decision will joint the Funk Bros., Benson, Flook et al. decisions that are rightly ignored day-to-day.

Thanks for the comment.

[David Koepsell]

Dear Kevin,

As you know, I support the court-made exclusion of laws of nature, products of nature, and natural phenomena from monopoly because I think they are a commons by (logical/material) necessity, and simply cannot be ethically monopolized.

But it's glad to see we have some agreement on something, more or less. :)

All is well, best to you and yours.

David

[Laurie Kellogg]

Thanks, Kevin. That cleared it up for me. And I agree with you, definitely on that point.