There has been steady progress with illumination of biology using CRISPR (among other editing tools, such as transcription activator-like effector nucleases [TALENs] and zinc finger nucleases). One example was taking the BRCA1 gene and systematically editing its nucleotides and assessing functional changes. In just one study, we could ascertain functional effects of hundreds of mutations that took more than a decade for a genomics company (Myriad Genetics; Salt Lake City, Utah) to determine via family studies. These approaches are complementary, but the CRISPR «saturation» (as they called it) is a fast-track means of dissecting important loci of the genome, be they in genes per se or in the 98.5% of the human genome that does not reside in genes.
Many clinical trials in rare diseases, such as beta-thalassemia, hemophilia, Hunter and Hurler syndromes, and lysosomal storage disorders, are getting under way with multiple genome editing companies, including Editas Medicine (Cambridge, Massachusetts), CRISPR Therapeutics (Zug, Switzerland), Intellia Therapeutics (Cambridge, Massachusetts), and Caribou Biosciences (Berkeley, California). All of these programs are undertaking genome editing of somatic cells.
This represented an ethical breach, performed unnecessarily, in secrecy, without proper informed consent. The twin girls who were born had unintended mutations introduced into their CCR5 gene and an unknown number of «off-target» mutations throughout their genomes. Not only does this affect the roughly 37 trillion cells of each embryo, but these errant, manmade mutations will be propagated to the offspring of the twins. As might be expected, there was an international outcry. Hopefully, this reckless use of CRISPR will not hold back the field’s remarkable momentum for indications in somatic cells for patients without any treatment that is available.
In December, Apple released their new ECG app for the Apple Watch Series 4. It is the first US Food and Drug Administration (FDA)-cleared consumer deep learning algorithm (a subtype of artificial intelligence [AI]) that will potentially affect millions of users—a watch projected to be purchased by 9 million people by the end of this year. The user’s heart rate is tracked at rest and with physical activity, and when there is deviation from the expected for that individual’s pattern, a haptic alert is given for the person to record their single- lead ECGfor 30 seconds with an immediate interpretation, be it normal or possible atrial fibrillation.
Although this may be useful for individuals at increased risk, the concern is about all the unnecessary and mistaken alerts in large numbers of people, engendering anxiety ; clinical testing; and misdirected treatments, such as blood thinners. We’ll have to see how this plays out, because it is just getting started.
At the clinician level is where most AI is starting to take hold. The first FDA-approved AI system to diagnose diabetic retinopathy, called IDx-DR (IDx LLC; Coralville, Iowa), was based on a clinical trial, prospectively conducted in almost 900 patients with diabetes.[2,3]
Now a nonclinician, such as the receptionist in a doctor’s office, can perform the eye exam, with cloud algorithmic interpretation with approximately 90 % sensitivity and specificity. Most of the studies to date for AI in medicine have been retrospective but have nonetheless led to many FDA approvals, especially for many radiology image interpretation deep learning algorithms. Other areas getting particular attention include pathology, dermatology (for skin cancers, especially melanoma detection), cardiology (for echocardiograms and ECGs), and gastroenterology (for colonoscopy detection of small polyps).
Eventually, no clinician will be spared with respect to the potential for algorithmic support. The implications of AI in healthcare are profound and reviewed in depth in my upcoming book, Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again, which will be available in March 2019.
The Gut Microbiome
The multitude of associations between the gut microbiome’s constituents and so many diseases, from autoimmune conditions to cancers and neurodegenerative disorders, is staggering. But the cause- and- effect relationships are not well established, no less our ability to modulate the microbiome to achieve favorable outcomes. Apart from fecal transplants for Clostridium difficile, which have been unequivocally shown beneficial in randomized trials, we await ongoing efforts to deconvolute the biology and see whether these insights will lead to an improved outlook for any of the many conditions implicated.
In the meantime, there have been a few surprising findings worth mentioning. First, we have assumed that in-hospital, nosocomial infections are due to inadequate handwashing by staff or derived from other patients. Although that certainly occurs, a fascinating study of sequencing the microbiome of hospitalized patients showed that many such infections originate from within the patient—from their gut microbiome.
Finally, for the surprise factor, the number 1 go-to drug for type 2 diabetes—metformin—was shown to work, in large part, via the gut microbiome. This study came in the wake of news that many other drugs used for the treatment of cancer are also affected by the patient’s gut microbiome. Clearly, there’s no shortage of new fascinating insights that we are getting from studies of our gut co-inhabitants, and still so much more to learn.
A Product Manager with expertise in pharma marketing and sales operations