Loading ...

Artificial intelligence (AI) plays an increasingly prominent role in the workplace, and health researchers are exploring its potential to improve the diagnosis and treatment of diseases. In this context, the company Sanguina Inc., based in the US, has developed an app called Ruby Health, designed to help detect anemia using just a photo of a patient’s fingernails.

Currently available for Android and iOS devices in the US, the app’s effectiveness was detailed in a study published by Sanguina’s research team in the journal PNAS.

The study and the companion tool grew out of an academic project launched in 2015 by Wilbur Lam, Erika Tyburski, and Rob Mannino.

“The development was rigorous. The idea came from the fact that, for decades, we have known that patients with anemia have paleness in some parts of the body, such as their fingernails, so we thought about the possibility of a noninvasive method, such as an app, to measure this paleness without having to draw blood,” said Lam, a researcher at Emory University School of Medicine in Atlanta, Georgia.

Since its launch in 2021, Ruby Health has registered 1.4 million users in the US. Users upload fingernail photos to estimate hemoglobin (Hgb) levels and may manually enter the lab results. Researchers then compare estimated Hgb with blood-test values to gauge accuracy. 

Among individuals with Hgb between 7 g/dL and 17 g/dL, the mean absolute error was ±0.72 g/dL. Using an Hgb cutoff of 12.5 g/dL to define anemia, the application demonstrated a sensitivity of 89% and a specificity of 93%. 

For Lam, this tool is promising. “Based on the values ​​entered, when calibrated, our app works similarly to blood-based methods. With ‘homemade’ methods, there is always the concern that they do not translate to what is tested in the lab, but we demonstrated the usability of the platform by replicating data in a closed clinical environment and in the real world,” he said.

The platform’s personal records feature allows for the personalization of treatment tips for patients, and geolocation data helps map cases for the application of public policies and the allocation of treatment resources.

However, Ruby Health has limitations: It does not detect very low Hgb levels (< 7 g/dL) and cannot identify the type of anemia the patient has, which is crucial for determining treatment.

“Anemia is not a disease itself but a condition secondary to many illnesses,” said Helena Zerlotti Wolf Grotto, MD, professor of Clinical Pathology at Universidade Estadual de Campinas, Campinas, Brazil, laboratory-hematology consultant, and member of the Brazilian Association of Hematology. “To determine the cause of anemia, it is necessary to perform tests, evaluate the patient’s history and perform a physical examination, since treatment involves comprehensive care,” she said.

The gold standard for diagnosing anemia remains the complete blood count, which identifies the type of anemia and its possible causes. “Screening tests like this app could be useful in remote or resource‐limited areas,” Grotto added.

However, in the United States, the app was more widely adopted in areas with higher average incomes, better access to primary care, and a higher proportion of Black residents.

In Brazil, Grotto sees limited benefit in urban centers with broad access to laboratory tests, as the public health system already provides robust anemia testing. She is also concerned about accuracy in Black patients, since AI tools often lack adequate training to identify darker skin tones, which could hinder the app’s functionality among Brazilians or other mixed-race populations.

However, Lam stated that the app’s algorithm has been tested on different skin tones and that nails have a lower amount of melanin, which minimizes the variability in skin color from one person to another.

Although not 100% accurate, the app can be used to indicate the possibility of a person having anemia in certain cases, according to experts. The developers themselves have stated that it is not suitable for emergency clinical decision-making.

“The market is full of innovations, but we must remain critical before assuming they will solve every problem,” Grotto said.

This story was translated from Medscape’s Portuguese edition.