The Living Image

The omnipresent smartphone may have found a new calling. Researchers at Purdue University are testing a new way to assess blood hemoglobin (Hgb) levels, drawing on a technique known as super-resolution spectroscopy (SSR) to transform any phone into a hyperspectral imager – no hardware modifications or accessories necessary. The researchers hoped to find a noninvasive approach that could improve care in low- and middle-income countries where access to testing laboratories is limited. A patient simply has to pull down their inner eyelid to expose the small blood vessels underneath for the software to compute exact blood Hgb content. The sensing site is important. The easy accessibility and relatively uniform vasculature of the inner eyelid allow for optical reflectance spectroscopy, while the fact it is unaffected by confounding factors of pigmentation negates the need for personalized calibration. So how does the app compare to traditional blood tests? In a clinical study of 153 patients, prediction errors for the smartphone technique were promisingly within five to 10 percent of those measured with clinical laboratory blood. The team say the results support the feasibility of SSR in noninvasive blood Hgb measurements, with the possibility of extending the algorithm to different models of smartphones in the future. We speak to Young Kim, Associate Professor at the Weldon School of Biomedical Engineering, Purdue University in Indiana, USA, to find out more about the app.

What is the role of super-resolution spectroscopy in measuring hemoglobin levels?

Hgb has distinct spectral signatures. If we can use a spectrometer, it is relatively straightforward to measure Hgb content. Super-resolution means high-resolution reconstruction of digital images acquired with low-resolution systems. We successfully extended this concept to the color domain (wavelength) for spectroscopy. This spectral super-resolution spectroscopy allows us to mathematically reconstruct a spectrum from a regular photo that has three-color information (red, blue, and green). In other words, we are transforming the built-in camera of a smartphone into a spectrometer that reliably measures Hgb levels without the need for any hardware modifications or accessories. This is a good example that a data-driven technology can minimize hardware complexity.

Are there any limitations to this method?

Our study is based on 153 patients from Kenya. We need a larger clinical study to conduct a systematic analysis of intra- and inter-blood Hgb measurements for a population of interest.

Is the app currently available?

At the moment it is at a development stage. It does not require any attachments or accessories to the smartphone. It should be noted that this app is not publicly available now, however we are currently working with a collaborator in India to change that. Our progress will be dependent on how much funding we can secure, but hopefully we will make the app widely available in one or two years.

Your paper mentions the use of a separate computer; how long does it take to get a diagnosis?

The computation is fast. The super-resolution and the Hgb spectroscopic analysis take 0.0005 and 0.0006 seconds, respectively, on a personal computer (Windows 10, Intel Core i7-8700 CPU, RAM 16.0 GB).

Is there a timeframe for when the algorithm could be incorporated into the app?

We are currently working on a cloud computing platform. If we can secure enough funding, we are expecting to have a beta version available in a couple of years.

Could the test have applications beyond anemia and sickle cell disease?

Blood Hgb tests are extensively performed for a variety of patient care needs, including assessment of hematologic disorders, transfusion initiation, hemorrhage detection after traumatic injury, and acute kidney injury.