Category Archives: Artificial Intelligence

Where Pareto Doesn’t Apply – The NICU

Posted on by

One of my favorite posts I did as part of my consulting at Fabric Genomics. Originally posted on December 3, 2020

The use of AI in healthcare is gaining increased attention with the significant advances and widening clinical use in radiology and pathology and now increasingly in genomics. In all of these cases, there are vast quantities of data to consider that should, in fact, be considered as they could be clinically significant.

Beyond the technology in use, we do see certain diagnostic situations that have a need for AI interpretation assistance. To illustrate this, we provide two different diagnostic scenarios. For instance, in adult critical care the doctor’s diagnostic process supplemented by well-researched rubrics has proven resilient. These cases are rarely primarily genetic in nature and are much more impacted by natural aging, environment, infectious agents, and lifestyle. There is a fairly high concentration of causation and a Pareto-like distribution. In contrast, the NICU pattern is quite different, with much more likelihood of a direct genetic cause. NICU genetic conditions are frequently rare and require unique considerations.

Let’s take a prototypical case – an elderly man presents in the ER with shortness of breath. In this case, there are a few highly likely and perhaps 200 possible causes and the doctor has them roughly mentally ranked in order of frequency. The clinician reranks likely causes real-time based on information as it is revealed – history, demographics, test results, etc. The top few causes make up over 95% of the cases and can be selected with reasonable confidence and confirmed via additional testing. Perhaps a few additional low-probability but high-risk causes are tested for, and a diagnosis will be confirmed in the vast majority of cases. A large but manageable dataset is analyzed and iteratively reanalyzed by the doctor in what is essentially a human Bayesian process – adjusting the prior probability based on real-time data. How likely is it heart failure (as opposed to an infection) given that there is no fever? How does the likelihood change given a specific test result? And for the vast majority of diagnoses, the doctor has confirmed and managed that specific diagnosis many times before in their career.

By contrast, let us consider the case of a critically ill child in the NICU. While there are a few common causative elements such as preterm birth, in over 30% of the cases the child’s condition has a genetic basis. The fact that there is a high chance of a genetic cause immediately brings us into a different diagnostic equation implying many thousands of potential causes. The vast majority of genetic causes will not be identified by standard newborn assay screening as they are not on the standard diagnostic panels (and those take weeks). Even the most experienced pediatricians will have only seen and be personally familiar with a tiny minority of those diseases and, on the off chance that they are familiar with a particular genetic disease, the phenotypic presentations are often not fully expressed in the newborn.

The recent Baby Bear study1 led by Rady Children’s Hospital clearly showed the prevalence of rare genetic diseases in NICU cases. Per Appendix A of the study, “Thirty-five of the diagnosed genetic diseases are rare conditions with an incidence of less than one in one million births.Sixty-five of the 71 primary genetic diseases were diagnosed just once in the Baby Bear population.” Of course, given the rarity of these conditions, it is beyond the likely human clinical experience.

We also know that the choice of treatment matters greatly. Genetic, metabolic, and neurological disorders are highly specific, and the wrong or delayed intervention can have life-long consequences.

In the NICU, we ideally would aggressively seek all reasonably accessible diagnostic information and immediately explore all possible genetic causes rather than work our way slowly along a curve that lacks the steep Pareto shape. Time is often the enemy in these cases. Damage from seizures, nutrition acting as a metabolic poison, or invasive procedures can take place that could be avoided with an early, specific diagnosis.

Just as we treat a critically ill febrile patient with broad-spectrum antibiotics to not lose time with narrow therapeutic shots in the dark, we need a broad but fast approach to diagnosis.

Fortunately, with technology lowering the cost of NGS and with the support of AI algorithms such as Fabric GEM, this approach is coming into use. Multiple peer-reviewed studies2including Farnaes et al 2018 show significant clinical efficacy of this approach and even cost savings.3 The technology is here, it’s available and even economical. Now is the time to insist on its use for the NICU babies that depend on us.

References

1. https://radygenomics.org/wp-content/uploads/2020/06/PBB-Final-Report_06.16.20.pdf

2. https://www.nature.com/articles/s41525-018-0049-4

3. https://health.ucdavis.edu/health-news/newsroom/project-baby-bear-shows-genomic-sequencing-for-infants-in-intensive-care-yields-life-changing-benefits-and-medical-cost-savings-/2020/06

Bringing the Promise of Precision Medicine to Rare Disease

Posted on by

My latest post as part of my work at Fabric Genomics.

Precision medicine is a form of medicine that uses information about a person’s own genes or proteins to prevent, diagnose, or treat disease.1 When we think of precision medicine, we often focus on cancer and the recent efforts to specifically target treatment based on the genetic composition of a tumor (somatic) or the person’s germline. While it is early days, the health gains of this approach are quite significant with better outcomes shown in multiple studies2 and even data showing lower overall health costs.3 However, the point of precision medicine is not only to optimize treatment for cancer, but also to target health interventions generally for all conditions. In fact, we see the application of precision medicine to high volume diseases such as diabetes and heart disease and even lifestyle interventions such as diet and exercise.

Nowhere is this more challenging than for patients suffering from a rare disease. Ironically it is common to have a rare disease. More than 25 million Americans and more than 400 million people worldwide suffer from one of over 7000 rare conditions, defined as those conditions having an incidence of 1 in 200,000 or less. Precision medicine has focused on big data approaches to studying more common conditions, thereby leaving out those with rare conditions. Lacking a diagnosis, most of these patients and their families are on a “diagnostic odyssey.” These patients typically spend more than five years seeking accurate diagnosis and might see up to eight doctors, often receiving many misdiagnoses and differing opinions on their journey. Along the way, they may be exposed to harmful treatments and invasive testing. It is clear that comprehensive genetic testing gives the best possibility of getting to the exact diagnosis efficiently. Still, historically, genetic data have been available to a minority of patients: only those referred to a clinical geneticist for testing.

For the complete post see https://fabricgenomics.com/2020/12/bringing-the-promise-of-precision-medicine-to-rare-disease/

Improving Health Outcomes for Infants with Pompe Disease

Posted on by

Fourth post as part of my work at Fabric Genomics

By Martin Reese & Laura Yecies

It was terrific to see this paper by our long time collaborator Arindam Bhattacharjeeon the use of NGS as a second-tier test for Pompe Disease (PD). This is part of an important diagnostic trend of earlier (even to the point of first-line for selected infants) use of NGS as a diagnostic tool that can dramatically improve the newborn’s projected health outcome. Pompe Disease is a perfect example of how this can work.

Background

PD is one of several glycogen storage diseases with variable timing of onset and rates of progressions. According to NORD, “Pompe disease is a rare multisystem disorder caused by pathogenic variations in the GAA gene containing the information for production and function of a protein called acid alpha-glucosidase (GAA). Because of the shortage of this protein (an enzyme), a complex sugar named ‘glycogen’ cannot be degraded to a simple sugar like glucose. This causes the glycogen to accumulate in all kinds of tissues, but primarily in skeletal muscle, smooth muscle, and cardiac muscle, where it causes damage to tissue structure and function. Pompe disease is inherited as an autosomal recessive genetic trait.” Early diagnosis and initiation of treatment are of paramount importance at no later than two weeks of age to minimize muscle damage and avoid significant negative impact on the quality of life.

For the full post please see Fabric’s site here

Enabling NGS Testing and Precision Medicine with Fabric AI Technologies

Posted on by

I was invited to write about NGS on Xifin’s blog…

Next-Generation Sequencing (NGS) testing is experiencing tremendous growth driven at a high level by the promise of precision medicine and the life-changing power of applications in preventive genetic screening, somatic testing, and rare disease diagnosis. In all of these use cases, we see important clinical advances. Preventive genetic screening for risk factors such as BRCA mutations allows people to take preventative measures that save lives every day. Somatic mutation analysis allows for highly targeted therapies, and rare disease diagnosis is improving outcomes for babies in the NICU and providing hope for the 400 million people worldwide suffering from a rare disease.

For the full article please continue on Xifin’s site here

Delivering Better Care at a Lower Cost – a Case Study of Project Baby Bear at Rady’s Children’s Hospital

Posted on by

My first post as part of my work with http://www.FabricGenomics.com

By Martin Reese & Laura Yecies

The power and cost-effectiveness of AI are calling into question many of our assumptions about healthcare.  The most important dichotomy proving to be false is that providing the latest and most thorough diagnostic technology to optimize clinical outcomes is more expensive.  When we use AI to more comprehensively analyze cases we benefit from Moore’s law rapidly and continuously reducing costs.  By contrast, hospital-based care, especially when in an intensive setting such as the NICU is continuously increasing in cost. It is not surprising that when more extensive testing produces clinically actionable results that actually decrease hospital days we can accomplish the holy grail — better care and less expensive simultaneously…

Read more on the Fabric website here