What's the Idea?
A quick summary: The initial goal of this project is to develop an automated imaging cytometer which can provide diagnostic results which will be effective in combating some of the major diseases affecting resource-poor areas, and will be inexpensive enough, portable enough, and easy enough to use that it will reach the people in these areas who need it. More specifically, we are looking at diagnostic methods which will be useful in diagnosing and treating malaria, tuberculosis, and AIDS, and trying to develop a diagnostic instrument with the following five characteristics:
(1) It can be assembled entirely from commercial or open-source components presently available on the market (or, in the case of some of the software, to be written by the participants in the project).
(2) It will cost less than $1,000 per system, for all hardware and software components.
(3) It will be small enough in size and weight to be easily hand-carryable by one person, requiring no more than 12 V DC for power (i.e., one car battery or solar panel).
(4) It will produce an automated analysis for the disease in question (i.e., a T4 lymphocyte count, for AIDS; a parasite count, for malaria; or a bacterial count, for tuberculosis) in less than 1 minute per sample.
(5) It will have sufficiently simple sample-preparation and user-interface characteristics that no more than 2 to 4 hours training of operators should be required.
More details of the plan: An automated cytometer, whether based on flow or imaging, consists of two parts: an “analyzer”, which interacts with the blood sample and produces electronic signals describing it; and a “data station” (a computer), which analyzes those signals to produce clinical results. For this project, I began with the idea of using an optical scanner designed for scanning 35mm slides for the analyzer, since these are readily available commercially. These have a resolution of up to 4000 dpi (about 6µm/pixel), which Dr. Shapiro, in the paper cited above, showed that this is sufficient resolution to obtain multi-pixel images of stained blood cells, and that resolution of intracellular detail (which is not possible at this resolution) is not necessary to perform the relevant diagnostic tests. This resolution, however, proved to be inadequate for cells identified by non-fluorescent markers (Beckman Coulter's Manual CD4 Count Kit). Fluorescent methods would have required modifications of the optical system in a slide scanner in order to work, such as a stronger light source and/or wavelength-selective filters, so I have been looking more recently at some low-cost, rugged, wide-field fluorescent microscopes, intended for global health work, such as the ones developed in Dr. Aydogan Ozcan's laboratory at UCLA, and the company which was recently spun off from that laboratory, Cellmic LLC.
For the data station, I began by investigating several low-cost portable computers, both of the "netbook" type (very low cost) or one of the lower-priced conventional laptops (with a larger screen, more storage capacity and CPU power, but larger and heavier). Both types fell within the cost and size requirements of objectives (2) and (3), above, but the increasing capabilities of cell phones and tablets, more recently, has caused me to shift my investigation to these devices (even more portable and lower in power requirements, although no less expensive). The increasing popularity of Android-based portable devices, in particular, has led to both more choice of features, lower prices, and better software-development tools for these devices, including the one I am currently working with, Xamarin's Mono for Android.
Any of the methods under consideration should require no more than 10-20
µl of
blood (easily obtainable from a finger stick).
The goal of the analyrical methods is to examine a large enough field,
after processing the sample by appropriate methods (depending on which
disease we are looking for) to contain anywhere
from
several hundred to several thousand target cells (for example, T4
lymphocytes,
in the case of AIDS tests). The image will be analyzed using software
developed for
analyzing astronomical images (resolving stars and other objects from
their
background), and the results stored in a database system such as SQLite. All the analytical programs being considered are freely available as open-source software,
so they
will not add to the cost of the instrument. What is not
yet available, free or otherwise, and what I intend to spend much of
the feasibility stage of this project in developing, is a high-level
executive program which will run all of the component programs
mentioned, and will present a sufficiently simple user interface to the
instrument operator that it can be run by people without extensive
technical backgrounds. As mentioned above, my current tools for this
software-development job are based on the Mono version of C#, and the
extensions available under it and Microsoft Visual Studio. The use of a
standard user-interface package should make it easier to translate the
interface into languages other than English, which could be important,
for some countries, in later stages of the project.