The Lie of Precision Medicine

My next blog post will be entitled "The Lie of Precision Medicine"

— sarcastic_f (@sarcastic_f) June 23, 2018

This post will be my own personalized rant about the false promises of personalized medicine. It will not be about neurological or psychiatric diseases, the typical topics for this blog. It will be about oncology, for very personal reasons: misery, frustration, and grief. After seven months of research on immunotherapy clinical trials, I couldn't find a single [acceptable] one¹ in either Canada or the US that would enroll my partner with stage 4 cancer. For arbitrary reasons, for financial reasons, because it's not the “right” kind of cancer, because the tumor's too rare, because it's too common, because of unlisted exclusionary criteria, because one trial will not accept the genomic testing done for another trial.² Because of endless waiting and bureaucracy.

But first, I'll let NIH explain a few terms. Is precision medicine the same as personalized medicine? Yes and no. Seems to me it's a bit of a branding issue.

What is the difference between precision medicine and personalized medicine?

There is a lot of overlap between the terms "precision medicine" and "personalized medicine." According to the National Research Council, "personalized medicine" is an older term with a meaning similar to "precision medicine."

Here's a startling paper from 1971, Can Personalized Medicine Survive? (by W.M. GIBSON, MB, ChB in Canadian Family Physician).

[it's a defense of the old-fashioned family doctor (solo practitioner) by Gibson]:

...will the solo practitioner's demise be welcomed, his replacement being a battery of experts in the fields of medicine, surgery, psychiatry and all the new allied health sciences, infinitely better trained than their singlehanded predecessor?

We wouldn't want any confusion between a $320 million dollar initiative and the ancient art of medicine. NIH again:

However, there was concern that the word "personalized" could be misinterpreted to imply that treatments and preventions are being developed uniquely for each individual; in precision medicine, the focus is on identifying which approaches will be effective for which patients based on genetic, environmental, and lifestyle factors.

The Council therefore preferred the term "precision medicine" to "personalized medicine." However, some people still use the two terms interchangeably.

So “precision medicine” is considered a more contemporary and cutting-edge term.

Archived from The White House Blog (Obama edition), January 30, 2015.

What about pharmacogenomics? 

Pharmacogenomics is a part of precision medicine. Pharmacogenomics is the study of how genes affect a person’s response to particular drugs. This relatively new field combines pharmacology (the science of drugs) and genomics (the study of genes and their functions) to develop effective, safe medications and doses³ that are tailored to variations in a person’s genes.

At present, precision pharmacogenomics is just a “tumor grab” with no promise of treatment in most cases. There are some serious and admirable efforts, but accessibility and costs are major barriers.

#cancer #oncology #genomics
As of January 18, 2018... pic.twitter.com/4KWpuA3yVh

— sarcastic_f (@sarcastic_f) January 18, 2018

But we've been promised such a utopia for quite a while.

Personalized medicine in oncology: the future is now (Schilsky, 2010):

Cancer chemotherapy is in evolution from non-specific cytotoxic drugs that damage both tumour and normal cells to more specific agents and immunotherapy approaches. Targeted agents are directed at unique molecular features of cancer cells, and immunotherapeutics modulate the tumour immune response; both approaches aim to produce greater effectiveness with less toxicity. The development and use of such agents in biomarker-defined populations enables a more personalized approach to cancer treatment than previously possible and has the potential to reduce the cost of cancer care.

Schilsky (2010)

IT'S 2018, WHERE IS THAT FUTURE YOU PROMISED US?

But wait, let's go back further, to 1999:

New Era of Personalized Medicine 
Targeting Drugs For Each Unique Genetic Profile

Certainly, there are success stories for specific types of cancer (e.g., Herceptin). A more recent example is the PD-1 inhibitor pembrolizumab (Keytruda®), which has shown remarkable results in patients with melanoma, including Jimmy Carter. The problem is, direct-to-consumer marketing creates false hope about the probability that a patient with another form of cancer will respond to this treatment, or one of the many other immunotherapies with PR machines. But if there's a 25% chance or even a 10% chance it'll extend the life of your loved one, you'll go to great lengths to try to acquire it, one way or another. Speaking from personal experience.

But exaggerated claims and the use of the superlatives in describing massively expensive cancer drugs (e.g., “breakthrough,” “game changer,” “miracle,” “cure,” “home run,” “revolutionary,” “transformative,” “life saver,” “groundbreaking,” and “marvel”) are highly questionable (Abola & Prasad, 2016) and even harmful.

It's a truly horrible feeling when you realize there are no options available, and all your hope is gone.


References

Abola MV, Prasad V. (2016). The use of superlatives in cancer research. JAMA oncology. 2(1):139-41.

Gibson WM. (1971). Can personalized medicine survive? Can Fam Physician. 17(8):29-88.

Langreth R, Waldholz M. (1999). New era of personalized medicine: targeting drugs for each unique genetic profile. Oncologist 4(5):426-7.

Schilsky RL. (2010). Personalized medicine in oncology: the future is now. Nat Rev Drug Discov. 9(5):363-6.  {PDF}


Footnotes

¹ 

Oh, well, there was the one trial with a highly toxic combo -- "you cannot travel more than X miles away from our center, because only WE can treat your terrible side effects." And that fun bit was omitted from the consent form.

— sarcastic_f (@sarcastic_f) June 23, 2018

²  But hey, we'll do yet another biopsy of your tumor, and let you know the results in 2-3 months, when you're too ill to be enrolled in any trial. Here's a highly relevant article — The fuzzy world of precision medicine: deliberations of a precision medicine tumor board — but I'm afraid to read it.

³ OMFG, you have got to be kidding me. Here is a subset of the possible side effects from one toxic monoclonal antibody duo:

Very likely (21% or more, or more than 20 people in 100):

• fatigue/tiredness
• decrease or loss of appetite, which may result in weight loss
• cough
• inflammation of the small intestine and / or large bowel causing abdominal pain and diarrhea which may be severe and life threatening

Less likely (5 – 20% or between 5 and 20 people in 100):

• pain and or inflammation in various areas including: muscles , joint, belly, back, chest, headache
• flu-like symptoms such as body aches, fever, chills, tiredness, loss of appetite, cough
• constipation
• dizziness
• shortness of breath
• infection which may rarely be serious and become life threatening
• nausea and vomiting
• dehydration
• skin inflammation causing hives or rash which may rarely be severe and become life threatening
• anemia which may cause tiredness, or may require blood transfusion
• itching
• abnormal liver function seen by blood tests. This may rarely lead to jaundice (yellowing of the skin and whites of eyes) and be severe or life threatening
• abnormal function of your thyroid gland which cause changes in hormonal levels. A decrease in thyroid function as seen on blood tests may cause you to feel tired, cold or gain weight while an increase in thyroid function may cause you to feel shaky, have a fast pulse or lose weight.
• Swelling of arms and/or legs (fluid retention)
• Changes in the level of body salts as seen on blood tests. You may not have symptoms.
• Inflammation of the pancreas that results in increased level of digestive enzymoes (lipase, amylase) seen in bloods and may cause abdominal pain
• Inflammation of the lungs (including fluid in the lungs) which could cause shortness of breath, chest pain, new or worse cough. It could be serious and/or life threatening. May occur more frequently if you are receiving radiation treatment to your chest or if you are Japanese.
• Serious bleeding events leading to death may occur in patients with head and neck tumors. Please talk to your doctor immediately if you are experiencing bleeding.
• Decrease of a protein in your blood called albumin that may cause fluid retention and results in swelling of your legs or arms

You get the idea. I'll skip:

Rarely (1 – 4% or less than 5 in 100 people)

Very Rare (less than 1% or less than 1 in 100 people)

Citric Acid Increases Balloon Inflation (aka sour taste makes you more risky)

from Balloon Analog Risk Task (BART) – Joggle Research for iPad

Risk taking and risk preference¹ are complex constructs measured by self-report questionnaires (“propensity”), laboratory tasks, and the frequency of real-life behaviors (smoking, alcohol use, etc).  A recent mega-study of 1507 healthy adults by Frey et al. (2017) measured risk preference using six questionnaires (and their subscales), eight behavioral tasks, and six frequency measures of real-life behavior.


Table 1 (Frey et al., 2017). Risk-taking measures used in the Basel-Berlin Risk Study.

-- click on image for a larger view --

The authors were interested in whether they could extract a general factor of risk preference (R), analogous to the general factor of intelligence (g). They used a bifactor model to account for the general factor as well as specific, orthogonal factors (seven in this case). The differing measures above are often used interchangeably and called “risk”, but the general factor R only...

...explained substantial variance across propensity measures and frequency measures of risky activities but did not generalize to behavioral measures. Moreover, there was only one specific factor that captured common variance across behavioral measures, specifically, choices among different types of risky lotteries (F7). Beyond the variance accounted for by R, the remaining six factors captured specific variance associated with health risk taking (F1), financial risk taking (F2), recreational risk taking (F3), impulsivity (F4), traffic risk taking (F5), and risk taking at work (F6).

In other words, the behavioral tasks didn't explain R at all, and most of them didn't even explain common variance across the tasks themselves (F7 below).

Fig. 2 (Frey et al., 2017). Bifactor model with all risk-taking measures, grouped by measurement tradition. BART is outlined in red.


Here's where we come to the recent study on “risk” and taste. The headlines were either misleading (A Sour Taste in Your Mouth Means You’re More Likely to Take Risks) or downright false — no lemons were used (When Life Gives You Lemons, You Take More Risks) — and this doozy (The Fruit That Helps You Take Risks – May Help Depressed And Anxious).

To assess risk-tasking, Vi and Obrist (2018) administered the Balloon Analog Risk Task (BART) to 70 participants in the UK and 71 in Vietnam. They were randomly assigned to one of five taste groups [yes, n=14 each] of Bitter (caffeine), Salty (sodium chloride), Sour (citric acid), Umami (MSG), and Sweet (sugar, presumably). They were given two rounds of BART and consumed 20 ml of flavored drink or plain water before each (in counterbalanced order).

[Remember that BART didn't load on a general factor of risk-taking, nor did it capture common variance across behavioral tasks.]

As in the animation above (and a video made by the authors)², the participant “inflates” a virtual balloon via mouse click until they either stop and win a monetary reward, or else they pop the balloon and lose money. The number of clicks (pumps) indicates risk-taking behavior. Overall, the Vietnamese students (all recruited from the School of Biotechnology and Food Technology at Hanoi University) appeared to be riskier than the UK students (but I don't know if this was tested directly). The main finding was that both groups clicked more after drinking citric acid than the other solutions.

Why would this this balloon pumping be more vigorous after tasting a sour solution? We could also ask, why were the Vietnamese subjects more risk-averse after drinking salt water, and riskier (relative to UK subjects) after drinking sugar water?³ We simply don't know the answer to any of these questions, but the authors weren't shy about extrapolating to clinical populations:

For example, people who are risk-averse (e.g., people with anxiety disorders or depression) may benefit from a sour additive in their diet.

Smelling lemon oil is relaxing, but tasting citric acid promotes risk:

Prior work has, for instance, shown that in cases of psychiatric disorders such as depression, anxiety, or stress-related disorders the use of lemon oils proved efficient and was further demonstrated to reduce stress. While lemon and sour are not the same, they share common properties that can be further investigated with respect to risk-taking.

We're really not sure how any of this works. The authors offered many more analyses in the Supplementary Materials, but they didn't help explain the results. Although the sour finding was interesting and observed cross culturally, would it replicate using groups larger than n=14?


Footnotes

¹ From Frey et al. (2017):

The term “risk” refers to properties of the world, yet without a clear agreement on its definition, which has ranged from probability, chance, outcome variance, expected values, undesirable events, danger, losses, to uncertainties. People’s responses to those properties, on the other hand, are typically described as their “risk preference.”

² The video conveniently starts by illustrating risk as skydiving, which bears no relation to being an adventurous eater.

³ The group difference in umami had a cultural explanation.


References

Frey R, Pedroni A, Mata R, Rieskamp J, Hertwig R. (2017). Risk preference shares the psychometric structure of major psychological traits. Science Advances 3(10):e1701381.

Vi CT, Obrist M. (2018). Sour promotes risk-taking: an investigation into the effect of taste on risk-taking behaviour in humans. Scientific Reports 8(1):7987.

 

Frey et al., 2017