Today we are going very briefly discuss some work which was published in the journal Nature Communications titled "Smartphone app for non-invasive detection of anemia using only patient-sourced photos.". In this publication, the authors discuss the development of an application which can be run on a smartphone, take pictures of a person's fingernails, and utilize a correlation between nailbed color and blood hemoglobin levels to quantify those levels with a high degree of accuracy. This technology has the potential to revolutionize diagnosis of anemia, especially in areas where access to the equipment necessary for blood analysis (complete blood count testing, or CBC) is sparse.

---

---

Background

This article revolves around detection of anemia, which is a disease revolving around the loss of red blood cells (or loss of red blood cell functionality). Red blood cells are largely responsible for being the carrier of oxygen in our blood via hemoglobin. Hemoglobin is an iron containing protein which binds and shuttles the oxygen we breathe to where ever it is needed in our bodies, it is also the reason why our blood is red!

It is thought that upwards of 80% of people on earth lack sufficient iron with up to 30% of them maintaining that iron deficient state long enough to develop anemia. [2, 3]. Remember iron is needed for hemoglobin, and it is through the iron in hemoglobin's heme that oxygen can be bound.

With numbers like that it means that upwards of 2 billion people live with anemia, which can result in damage to our bodies and organs, resulting in a drastically reduced lifespan. Many forms of anemia can be treated effectively, however for treatment to begin, diagnosis must first occur. Access to blood tests might be prevalent in the US, Europe and other developed areas of the world, however the vast majority of people on our planet may not live in areas where it is. As such the development of an effective, non blood-test diagnostic could truly be lifechanging for many millions of people around the world. Lets very briefly look at the authors work.

Results

The image to the left depicts how the app itself works. People just need to download the app, take photos of their fingernails and allow the app to calculate the hemoglobin levels. The app is equipped with the ability to discriminate things like reflections from the flash of the camera, or leukonychia white lines on a fingernail, and choose regions of the nail which are appropriate for analysis of the color of the bed. The researchers report that the apps analysis can be performed either directly on the user's phone or on a cloud server depending on needs.

Here we can see a figure comparing the hemoglobin levels that were determined from the analysis of the fingernails of 100 different patients. On the Y-axis is the concentration determined by the app, while on the X-axis is the concentration determined by a traditional complete blood count analysis. The authors report that the app is accurate, with an error +/- 2.4 g dL^-1 of hemoglobin, which is very good considering the ease of the use of the app! Based upon this data the app has more than sufficient resolution to quantitate whether or not someone has anemia (classified as blood hemoglobin levels < 12 g dL^-1 [4]). And this is with out any sort of personalized calibration! Just snap a pic and analyze! Remarkable. The authors report that this margin of error was consistent through out the range of observed blood hemoglobin concentrations.

The authors reported that the app was able to correctly identify a patient with anemia 92% of the time

One important point of consideration the authors took was whether or not skin pigmentation and environmental lighting had any affect on the ability of the app's ability to determine the hemoglobin levels. As you can see in the image to the left, in (a) the authors have the error in quantitation by the app (comparison of app vs blood test quant) relative to the skin tone of the patient. There is a correlation reported here, but that result is near negligible (r = 0.13) and should not interfere with accurate identification of anemia. As for lighting (b) you can see that neither light intensity nor indoor vs outdoor light had any effect on the accuracy of quantitaiton. Both of these results are excellent and illustrate the robustness of the app!

Conclusions

The results and quantitation abilities of this app are downright remarkable. The authors of the paper reported that the app was significantly more accurate then even trained hematologists looking at patients during a physical examination, and the app can be provided for a fraction of the cost of even a doctors visit! Given the massive size of the population of the world with anemia, the potential impact of this app is massive. It also goes to show how we are just only scratching at the tip of the iceberg for what the lies in the future with regards to non-invasive, accurate, and cheap medical diagnostics. Ones that could be performed from the comfort of your own home, and allow you to get the treatment you need as soon as it is necessary.