Dear Q Bionauts,
I wanted to thank all of you for being early adopters and share a little about what we have been up to for the past five years and what our larger ambitions look like. I realize from the outside looking in that what we are doing looks mostly like an executive physical platform, and we haven’t really spent the time or effort to change that perception. Mostly because it felt premature to do so until we had made sufficient progress towards our bigger goals to share more.
About 20 years ago, I was doing research in computational physics, which was taking off as Moore’s Law was really hitting its stride. I became fascinated with the idea of “computational physics for biology” and started thinking about questions like “what are the limits of building an A2D converter for the human body?”
Fast forward more than a decade later, the human genome project had been completed, and we started to see huge decreases in the cost/bit per unit time that could be measured about the human body across genetics, epigenetics, transcriptomics, proteomics, metabolomics, microbiomics, wearables, etc. All of these are just different tools that allow us to quantify and measure the state of different layers of our biological information stack, aka the operating system of life.
It was clear we are asymptotically heading towards our ability to construct digital twins of ourselves. But there were lots of open questions: Could this be done non-invasively? Could it be scaled in terms of cost and speed? And most importantly, how reproducible are the measurements? This last one is especially critical if you want to quantify change, and that’s really what Q Bio is about.
Measuring changes in any natural system’s state and modeling them to forecast future states or understand its evolution from past states is fundamental to our understanding of cause and effect and is the foundation of the scientific method. It will also be the foundation that the future of medicine and a new data-driven healthcare system will be built on.
The Missing Pieces
There were two big pieces that were missing in order to bring this paradigm to clinical practice.
Issue #1) Until now, we didn’t have a way to non-invasively, cheaply, and quickly measure changes in our anatomy. “Medical imaging” is optimized for acute/symptomatic diagnostics that require subjective interpretations, so the output doesn’t need to be highly reproducible. Contrary to what some may tell you, non-invasive medical imaging today does not produce reproducible measurements. There is an easy way to tell. Images don’t have error bars; measurements do. One can quantitatively compare cholesterol measurements from different machines, but one can’t quantitatively compare images from different scanners. In physics, we try to parameterize and build models of systems that describe and correlate changes across multiple scales over time to prove we understand the whole picture. We are getting pretty good at measuring the human body at a billionth of a meter (chemistry) and a millionth of a meter (cytometry), but we actually don’t have any tools to measure our bodies effectively at the scale of a thousandth of a meter (anatomical). Because of the way our universe is constructed, things that change at large scales require many changes to have occurred at small scales. For this reason, we believe the future of healthcare will be dependent on longitudinal multi-scale models of health and pathology.
Issue #2) If we are to measure an exponentially increasing amount of information about our body, there needs to be a software platform that can scale to integrate all of this information and summarize the most salient changes in an individual based on their genetics, medical/family history, lifestyle, etc., very similar to how Google is able to summarize the most relevant parts of the internet for us based on what we are looking for. We need this so clinicians can integrate new information at the pace of technological innovation without being restrained or overloaded with information.
At Q Bio, we have spent the past five years developing these missing technologies and vertically integrating them in order to fill these gaps required to deliver the future of clinical medicine in a way that could be made cheap enough for everyone. We call this integrated platform Gemini – the first comprehensive clinical digital twin platform.
Why does it matter?
The first and most immediately impactful capability that will arise as a result of Gemini will be the first “Check Engine Light” for the human body. I believe this is ~2 years away from being deployed clinically. Based on data we will be able to collect in ~30 minutes at a site that would cost less to set up and operate than a car wash, and for an annual cost that is comparable to what insurance currently reimburses for in the current physical, we will be able to answer the question, “Should you have a televisit with a doctor?” And if not, suggest when it is the optimal time for you to come back for another Gemini Exam based on personalized risk factors. This may seem trivial but has massive implications for preventive care, which is, to the first order, a resource optimization problem. Multiple studies have shown that around 70% of doctors’ visits are unnecessary, and there simply aren’t enough doctors in the world to see every patient every year, and this gap is widening. On top of this, doctors in a face-to-face visit collect almost no data about your body which makes the standard annual physical almost worthless. Our preliminary results suggest the “check engine light” for the body can transform primary care from a FIFO/wealth queue into a priority queue where people with the greatest health risks are seen first, improving clinician efficiency by over 10x, improving access, and reducing the costs of care by allowing clinicians to catch the worst things in stages that are cheaper and easier to remediate. After “the check engine light for the human body” and automated triaging in the asymptomatic population will come comprehensive virtual physical exams, novel diagnostics, the first real metrics for value-based care, in-silico clinical trials, better in-vivo clinical trials, better underwriting models, virtual surgical planning and follow-up, municipal & employee policies tailored to population health risks, AI Physician Assistants and more.
How you have helped us
When you can measure something inside a living person for the first time, how do you know if it’s accurate? What is your reference? Your participation in the largest whole MRI reproducibility study ever done over the past years has helped us try to address some of these questions, and for the past year, we have begun discussing with regulators how we address these challenges as we try to bring this rapid anatomical measurement technology to clinical use.
We have also got great feedback from clinicians who have found things in their patients that could only be found by combining all this data and where any single piece of information was not enough to find something that was an existential risk. We feel this strongly supports our hypothesis about longitudinal multi-scale models of pathology and its future role in healthcare.
A major update to the Q Exam
In June, I asked our team to see how far they could get in a one-month sprint to translate some of our technology into practice as part of the existing Q Exam, given the limitation of not having direct access to the scanner hardware. In this short time, we were able to make the following comfort improvements while improving the data quality at the same time:
- reduced the time of the current Q Exam scan from about 50 minutes to about 30 minutes
- reduced the number of breath holds required from 20 to 4
Starting on October 8th, 2022, we will make this faster, more comfortable scan a standard part of the current Q Exam.
What are the limits of this technology?
With full control over existing clinical hardware, we could do a better scan in about 5 minutes with no breath holds. But the real advantage of this technology is that on hardware that is 1000x less ideal, we can do multi-parametric quantitative whole body scans in less than 10 minutes with no breath holds. To prove this out, we have designed and built a prototype of a new kind of whole body scanner where we have direct access and full control of the hardware. This scanner has a much more patient-friendly geometry that can be deployed much more cost-effectively and quickly than any existing scanner. We are actively doing research scans on this prototype scanner under an IRB and are doubling down on our efforts to bring this scanner to market which is the bottleneck to making Gemini accessible to everyone.
When will Gemini be available?
In early Q1 of 2023, we will be launching Gemini v1.0 / Gemini Exams at our research site in Redwood City before rolling out any partner launches, to make sure all of you who helped us get this far get first access to what we believe is a revolutionary new interactive software platform for not just the human body, but your body.
In Summary
I have always been fascinated by the fact that we understand more about the universe outside of our bodies than the one inside of us. At Q Bio, we are on a mission to map innerspace at a depth and scale that has never been attempted or achieved. With this information in hand we believe your digital twin will be like GPS for your health, you will always know where you are and where you are going. There is an enormous amount of suffering in the world today that we believe is preventable. No one should die from a treatable disease by definition. We believe that it is not a matter of if these things come to pass but when, and we hope we can make this a reality sooner rather than later or at the very least pave the way for others to do so.
Thank you for being a part of this journey. We could not have made it this far without your support.
— Jeff