Artificial Intelligence Increasingly Aids Dermatologists

MADRID — The growing relevance of the use of artificial intelligence in dermatologists’ day-to-day work, especially in diagnostic processes, was the focus of one of the sessions held during the 49th National Congress of the Spanish Academy of Dermatology and Venereology.

The speakers stated that artificial intelligence is currently used to establish precise differential diagnoses for common diseases, such as psoriasis, atopic dermatitis, and acne, as well as for the development of personalized therapeutic protocols and for predicting medium- and long-term results.

Likewise, teledermatology is proving to be a fundamental tool (as is evident in the major role it has played during the pandemic) in which remote diagnosis is supported by apps that allow image recognition.

“The application of artificial intelligence in the field of dermatology has focused primarily on two areas: image recognition and big data,” Antonio Martorell, MD, PhD, a dermatologist at Manises Hospital in Valencia, Spain, told Medscape Spanish Edition. “Regarding image recognition, its interest focuses on machines’ ability to help dermatologists with teleconsultation prioritization strategies. This improves the early detection of malignant skin processes whose delayed diagnosis puts our patients’ lives at risk, as is the case with malignant melanoma.”

The areas that benefit most from these new technologies are mainly in oncologic dermatology and immune-mediated inflammatory diseases, he added. For oncologic dermatology, these tools enable the early detection of lesions for which there is a high risk of malignancy, he said. In the future, the tools will help reduce the incidence of melanoma and increase the detection of this type of tumor in its earliest stages, thus reducing melanoma mortality.

Early Detection, Accurate Screening

“Artificial intelligence encourages patients with skin cancer to be seen at our clinics earlier and helps to automate skin tumor diagnosis, since it allows us to differentiate between melanocytic and nonmelanocytic lesions, even establishing more precise classifications within these two large tumor groups. This is an advantage for the early diagnosis of skin cancer and helps to generate screening systems capable of prioritizing cases based on their severity,” said Martorell.

Another interesting application in the field of oncologic dermatology is in the development of models that generate prediction algorithms from patients’ anonymized data. “For example, there are computational algorithms that allow us to predict the metastatic progression of cutaneous melanoma based on gene expression and microRNA data from the tumor itself. Big data and its analysis will allow us to better categorize patients and to further personalize treatments.”

Regarding the role of artificial intelligence in immune-mediated inflammatory diseases, Martorell highlighted that the early detection of diseases such as psoriasis and hidradenitis suppurativa will prevent disease progression and lower the risk of developing highly degenerative joint and skin processes in the short and medium terms.

Artificial intelligence is being applied to the quantification of skin diseases that the human eye cannot quantify. With it, objective data on the severity of pathologies such as psoriasis can be collected. “Along these lines, the IMAPSORS project, recently presented at the Spanish Psoriasis Group, was shown to be a promising tool for measuring the severity of this disease, which will facilitate a more objective personalization of its management in the near future, as well as the early detection of patients who may present a higher risk of associated systemic diseases,” Martorell added.

The session addressed the possibility that artificial intelligence could replace dermatologists in the future. “This technology is here to help us, improving the consultation experience and patient care, but machines lack the capacity for intuition and innovation, contrary to the human brain, so it would never be able to replace us,” said Martorell. “Artificial intelligence must be seen as an opportunity, as long as dermatologists get involved in the development of products. It is important that specialists be encouraged to get in touch with the world of artificial intelligence, that they understand it, study it, and see how it can help them solve problems in their clinics.”

Aiding Daily Practice

Martorell addressed the main challenges that specialists face in ensuring that artificial intelligence is implemented in their daily practice, thus optimizing its possibilities. “Essentially, there are three challenges. The first is the integration of platforms dedicated to the health sector. The great diversity of patient management platforms, both in the public and private spheres, makes it difficult to include external software at a global level. In turn, the lack of structure in medical records makes big data analysis difficult to apply when generating reliable predictive algorithms, despite multiple attempts by digital companies aimed at overcoming this limitation with the incorporation of linguistic experts.”

The second challenge is the adaptation of current resources to the needs of the dermatologist. “The fact that most skin-focused digital health companies are not led by dermatology experts means that a lot of time and money are spent on solutions that do not enhance the clinician’s experience in real practice, and therefore, they lack practical use, hence the need to encourage experts in this specialty to take their own steps to generate value propositions,” Martorell added.

A third challenge is the patients’ acceptance of the new tools. While people younger than 25 years are immersed in the digital world, those older than that still need close physical contact with their healthcare provider to feel well taken care of. “Adapting current solutions and prescribing patients future digital tools, such as apps, will greatly help their confidence in these apps to grow and their use to become universal.”

Projects Led by Specialists

Several products that utilize artificial intelligence and that are relevant to dermatologists were presented at the conference. One of them, led by Martorell, is Skiana-Care. It is a smartphone app that allows the evaluation of patients’ juvenile acne. The user photographs himself or herself with a mobile device, and the app determines the extent of the skin lesions and establishes their severity.

The product aims to educate the youth population in skin health and identify patients who may need specialized dermatologic care in the future to prevent scar formation, depending on the level of severity and acne profile detected.

“This solution stems from the need that we have mentioned to generate value propositions in the field of dermatology telehealth from projects led 100% by skin experts,” said Martorell.

“It began in 2018, exploring the different technological and imaging solutions on the market and analyzing uncovered areas that could be of help to the clinician and the patient. In 2019, we obtained grants that allowed us to develop our first project: Skiana-Care. This mobile app, available on iOS and Android, basically seeks to improve the relationship between the patient and their skin specialist,” he added.

Martorell described the benefits of this tool, which he defined as “a social assistant.” First, it provides the patient with a report on the health status of their skin based on three facial images that are processed through native algorithms created for this project. It also provides categorized health education information based on facial analysis. It consists of a personalized cosmetic advisor based on the information generated by the app, which incorporates solutions, for example, for patients with special situations, such as those who are allergic to cosmetic components.

“We are currently in a launching phase of the Skiana-Care app product and seeking funding to optimize this project and accelerate other algorithms that are currently in the development phase,” the specialist concluded.

Martorell has disclosed no relevant financial relationships.

Follow Carla Nieto of Medscape Spanish Edition on Twitter @carlanmartinez.

This article was translated from the Medscape Spanish edition.

Source: Read Full Article