Fraunhofer researchers have developed a digital solution using AI technology and a mobile app to dramatically speed up skin cancer diagnosis.
Skin cancer is a particularly misleading form of cancer. In the early stages, it visually resembles a harmless mole or birthmark and does not cause any pain. According to data from the German Cancer Society, more than 200,000 people suffer from skin cancer in Germany every year. The earlier a patient is diagnosed with skin cancer, the more treatable it is.
The Fraunhofer Center for Assisted Information and Communication Solutions (AICOS) in Porto and Lisbon has developed a solution to speed up early recognition. The Derm.AI solution combines smartphone photos of the skin lesion with image analysis software and artificial intelligence. It provides a rapid first assessment of potentially dangerous skin surface changes. Dermatologists can access this decision support platform and prioritize cases with an increased risk of skin cancer. The Derm.AI solution aims to improve the existing teledermatology processes in the Portuguese national health system.
“In recent years, GPs have become increasingly concerned with detecting skin cancer early. People who notice dark spots or other noticeable changes on their skin need clarity quickly. But in areas where there are few specialists, it often takes a long time to get an appointment for the initial assessment, and often patients also have to travel long distances for these appointments. Derm.AI solution,” said Maria Vasconcelos, Senior Scientist at Fraunhofer AICOS.
Accessible smartphone app
The Fraunhofer team has developed a specialist app for GPs to photograph the potential skin cancer area with a smartphone – it will be available for iPhones and Android smartphones. The app ensures that photos are aligned correctly, taken from the correct distance, and have the correct resolution. The application takes two photos: one close-up of the area in question and another further away to show the area in context. This also helps align and position the camera correctly. This creates standardized photos with consistent settings for resolution, color, brightness, and contrast.
“Standardized plans are easy to compare and can be reliably analyzed by specialists,” Vasconcelos explained.
The images obtained at the general practitioner’s office are sent to the dermatology department of a hospital. AI software will analyze the images of the skin lesion, compare them to reference data and data from other patients, and then provide a risk assessment. The lesion in question is labeled as “normal”, “priority” or “high priority”. This is not a definite diagnosis of skin cancer; it is an initial assessment that allows prioritization of the order in which potential cases are examined. Physicians can prioritize cases that the AI software thinks are high risk, as these need to be quickly confirmed or clarified.
“The software does not make decisions; it simply provides a preselection based on probability. The actual examination and diagnosis is always in the hands of dermatologists or skin cancer specialists,” Vasconcelos explained.
Once the images have been analyzed, as well as patient data such as age, sex or previous conditions, the hospital dermatologist can either provide feedback to the general practitioner in charge of the patient by teleconsultation or schedule a face-to-face consultation. with the patient.
During the Derm.AI project, the AICOS researcher and her team developed the algorithm for the image analysis software. The deep learning software was fed with image data and information from around 4000 cases. The algorithm also used the specialist knowledge of dermatologists in the subsequent prioritization.
“We had a lot of discussions with general practitioners and dermatologists to understand their needs. We have received very good feedback from doctors for Derm.AI,” said Vasconcelos.
Diagnosis of skin cancer
A dermatologist will schedule an in-person appointment if there is any uncertainty about the risk level of the skin lesion. An examination of the skin lesion under a reflected light microscope will take place, or a tissue sample for analysis will be taken, which will allow the specialist to accurately issue a diagnosis of skin cancer.
Approximately 80% of cases where patients present to the GP’s office with suspicious skin changes turn out to be harmless moles or birthmarks, after image analysis and consultation between doctor and dermatologist . This allows GPs to give patients the green light quickly, saving them long waiting times and an often long journey to a hospital appointment.
However, the new digital solution speeds up the process of getting an accurate diagnosis of skin cancer and enables case prioritization to ensure high-risk cases are treated first.
For patients whose skin changes are not identifiable as harmless or for those that indicate a less dangerous form of skin cancer, the GP asks the patient to return, for example, in three months and have another picture taken of the area.
Recommended related articles