Hello dear friends!

Here is an idea - which I got independently and might even be new in some aspects - that goes against intuition but seems to be backed by figures. As the title says, I'd like to treat diseases linked to blood cells by processing all the blood of a patient and remove only the individual sick blood cells.

The challenge is of course the number of cells: about 30e12 red blood cells for an adult human. But semiconductor technology is to help us here. Imagine a micromechanical actuator (Mems) capable of sorting individual cells. Typical blood cells are a bit smaller than 10µm, fine for a Mems. Add some piping and drive: the Mems could measure 100µm*100µm. Now, we won't put a single Mems on a chip, but rather 100*100 Mems on a 10mm*10mm chip as usual; and the machine will consist of 1000 chips, as semiconductor technology is a batch process anyway. So we have 1e7 Mems - now we're getting somewhere.

Let a part of the patient's blood throughput flow through the machine and reinject it, like is done in kidney dialysis. I take 0.2L/min, which means 2e10 blood cells per second. But since we have 1e7 Mems, each one sees only 2000 cells per second, which is a reasonable reaction time for a Mems. Also, the blood throughput is just 3e-13m3/s per Mems, or 0,8mm/s as a mean value in a 20µm*20µm pipe; a viscosity of 10mPa*s without clotting then only drops some 200Pa/mm.

Mems valves exist for two decades - I knew they would have uses beyond bees insemination machines. Other actuators might be smaller and faster: for instance, optic data multiplexers create reflective vapour bubbles in a liquid with tiny electric heaters. I hope such a bubble can push an unwanted blood cell away from the normal outlet. Or a rod can push the unwanted cell out of the pipe - no need for a true valve. Or tiny bellows can suck or blow the unwanted cell as it passes by: the small displacement looks easier and fast.

The chips must also recognize unwanted cells. I believe that drug chemists are able to put tags on many types of unwanted cells - years before they find a means to remove such cells from the patient, and the Mems machine would bridge this gap. Imagine for instance a fluorescent protein that attaches specifically enough to the unwanted cells: the reemitted light would be an easy signal for each Mems.

At 0.2L/min, one may first think that all the blood is cleaned after 30min, but it's less easy. Since we take just a fraction of the heart's throughput, cleaned blood will mix back in the patient's body, so after 30min, we've sunk the proportion of unwanted cells by exp(1) - provided these cells multiply little meanwhile, and provided no organ stores these cells for more than 30 min. The proportion of unwanted cells sinks by exp(16) or nearly 1e7 after 8 hours. And after 20 hours, the number of unwanted cells is zero with a high probability.

Not only blood diseases could be cured with such a machine: potentially all diseases whose cells move mostly through the blood at some time. This hopefully includes many infectious diseases, as well as autoimmunity diseases - which I'd like to be addressed as a personal wish.

I'd be happy that people check this proposal - including physicians and drug chemists, as I've zero knowledge there.
So... Completely crazy? Or just difficult?

Enthalpy, aka Marc Schaefer

I think sooner or later, the idea will catch on. It's one of those inevitable things.

I'm not sure that MEMS is the right technology, though.

My background is lasers and optics, so I'd suggest relatively simple image processing algorithms to spot the "bad" cells or parasites or microbes, whatever, and zapping them with a brief laser pulse.

Some people will get overwhelmed at the numbers you've already cited and laugh in your face, but computers are getting cheaper and faster all the time, and that's one reason I think it's inevitable. It'll work better for some diseases than others, but maybe one of the diseases it works best at might be one of those that has no other cure, like leukemia, so don't let the ignorentia/nay-sayers dismiss the idea too quickly. You'll find that it's just human nature to assume that if he/she can't figure out how to do it, then they'll assume that it's impossible or unrealistic.

Even if you don't personally make the prototype work, if you keep the idea out there, promoting it on the web, sooner or later someone will see your idea and steal it and run with it. You won't get fame or fortune, but you'll get the karma that goes along with curing a disease.

I did something similar, writing an invention disclosure for what later became "photodynamic therapy", a cancer treatment that works. I never received a dime, my company didn't pursue it, but they did spread the word around and a year later somebody tried it and it worked. I figure that I share in at least some of the good karma from that. Expecting more is kinda unrealistic, unless you have enough money to invest in it yourself and still cover mortgage/rent payments.