The Moore's Law of blood testing

The amount of information extracted from a single drop of blood using a digital microchip has been increasing exponentially and has surpassed the glucometer.

Our bodies consist of approximately 40 trillion cells in constant communication and interaction. The vascular system is an information super highway and the drop of blood is our access point. Yet out of the gigabytes of information in a drop, the glucometer - one of the few devices that can collect information from a single drop of blood - can only provide a single measurement with 8 bits of resolution. And in the 40 years since its introduction, there’s been little to no improvement.

All efforts to combine additional indications on a single platform have invariably consumed more than one drop of blood, and therefore require phlebotomy in a clinical setting. Elizabeth Holmes correctly identified comprehensive testing from a finger-stick as a worthwhile vision, but she failed to develop a minimally viable roadmap for achieving it, namely a blood testing technology that can be progressively miniaturized, or scaled, to incorporate an increasing number of indications without increasing the volume of blood needed.

It must be noted that some indications simply can't be measured from a small drop of blood, but these are exceptions. Troponin, for example, is a heart attack biomarker with one of the lowest limits of detection at 1ng/L, corresponding to approximately 1 copy per nano-liter. At those scales, the laws physics as we understand them no longer apply and analytical performance increases dramatically as a result of sensor quantization, incidentally the same reason your iPhone camera gets better every generation.

The first bit is the hardest

I began developing microchip blood tests in 2004 in the hopes of co-opting Moore’s Law for our purposes. By 2010, I demonstrated a digital microchip capable of detecting Dengue. The Yes/No results it provided can be represented by a 1 or 0, equivalent to 1 bit of information from a drop of blood. By 2017, a team of us spun out Xip Diagnostics and placed 4 sensors on a microchip that provided 4 Yes/No results or 4 bits of data. More recently, in 2023 we launched a liver enzyme test that measures two indications, ALT and AST, with 8 bits of resolution from 20 micro-liters of whole blood – equivalent to 16 bits of total data and exceeding for the first time the amount of information available from a glucometer. More startling, however, is that by plotting the amount of data over time, we notice an exponential rise. The amount of data we can extract from a single drop of blood using digital microchips is doubling every 36 months, or half the rate of Moore’s Law.

Slow in the short-term, fast in the long-term

The spectrophotometer is the workhorse of modern blood testing. To date, we have miniaturized it by over 1000x, from 100 micro-liter cuvette volumes to less than 100 nano-liters! And by integrating it onto a digital microchip, the microfluidics are reduced to passive capillaries that are both modular and scalable.

We are working on a combined ALT, AST, BUN and Creatinine test for quantitative liver and kidney indications from a single drop. Looking beyond, advances in micro-injection molding will enable miniaturization well past the comprehensive metabolic panel, and magnetic particle immunoassays and isothermal nucleic acid testing will be integrated alongside metabolic panels for multi-modal operation.

FDA approval can be a long and expensive process. In addition, placing bulky or expensive instruments with customers has historically been a barrier to adoption and a reverse compatibility anchor to innovation. The traditional blood testing business model has been overhauled to keep pace with the need for rapid product iteration. Metabolic indications are identical across all species so products are launched in the large and unregulated veterinary space prior to FDA submission. And to guarantee rapid and universal adoption while retaining the freedom to scale, the instrument is gone, replaced by your smartphone.

A little help from my friends

Moore’s Law was never preordained. In fact, it’s technically not even a law, merely an observation of Intel’s pioneering progress scaling transistors. Yet it quickly became the de-facto standard for the semiconductor industry, aligning investment and R&D and ultimately spawning all of Silicon Valley as we know it.

The initial and very difficult work of pioneering blood test scaling is nearing completion. The burden of persisting that trend lies squarely with us, a responsibility we take both seriously and with focused dedication.?Just like Intel though, we won't be able to do it alone. We welcome partners in developing the ecosystem to ensure not just success but inevitability of blood test scaling. It’s impossible to predict the ultimate impact this approach will have. But who would have predicted the internet, smartphones and AI when Gordon Moore published a paper in 1965 with 64 transistors?