Automated Analysis Apps
Overview
The purpose of these apps is to analyze data provided by the patient or user, such as photographs of skin conditions. Since 2013, these apps have been regulated by the FDA. These apps use artificial intelligence to analyze images, providing automated risk assessments and diagnostic suggestions. The majority of AI dermatology apps use machine learning (ML). However, some apps may rely on basic rule-based systems or expert-driven criteria (like the ABCD rule or 7-point checklist) without incorporating ML to assess risks.
The user takes a picture of a suspicious mole using the app’s camera. The app processes the image with AI to identify key features of the lesion and uses machine learning to compare it against a database of skin conditions. Based on the analysis, the app provides a risk score, such as "80% chance this is benign" or "This lesion may be malignant—consult a dermatologist." The user can then follow up with a professional if the app suggests further investigation.
Advantages and Challenges
The automation of the visual examination has great potential. These apps can be helpful in reducing the burden on specialists while also aiding in the early detection of skin cancer, providing convenience for patients. However, when only a selected few lesions (at the user’s discretion) are brought to medical attention, this may cause delay in the diagnosis of early subtle skin cancers that would otherwise be readily detected on a full skin exam. There are significant hurdles to successful automated diagnosis. Pictures may be altered by light quality, user expertise, and camera quality. Inaccurate diagnosis (especially in the context of limited to no history) could lead to a false reassurance and delay medical care. In addition, difficult to reach sites such as the back, buttocks, posterior thighs, and scalp may not get adequate attention from the user. Lastly, security of images, privacy of patient sensitive information, and false advertising remains a concern.
A systematic review focusing on the accuracy of automated skin cancer detection smartphone apps concluded that currently available algorithms appear unreliable in detecting melanoma and other skin cancers (1). A study by Y. Chung et al. (2) aimed to evaluate the performance of one such app by comparing its assessments with those of a dermatologist, using lesions selected by participants. Out of 199 lesions from 125 participants, the app failed to analyze 45% of cases and misclassified many high- and medium-risk lesions as benign. The agreement between the app’s assessments and the dermatologist's diagnoses was poor. These findings highlight the app’s limited diagnostic accuracy and suggest that further research is needed to assess its effectiveness in real-world settings. Overall, these apps are useful but are not recommended to replace a physician-rendered full skin examination for the reasons stated above. In contrast, a study by M. de Carvalho et al. (3) found that the latest version of the SkinVision app (as of 2019) reported 95% sensitivity and 78% specificity for skin cancer detection (3). According to the authors, this was achieved due to improvements in the processing of images taken with the smartphone camera and a large, risk-labeled image database from users, which was used to train a machine learning algorithm.
Examples of automated analysis applications includes:
For healthcare professionals:
VEOS: This app is functional and available on the App Store. It is designed for use with the Canfield ProVEOS dermatoscope and is compatible with specific iPhone models. The app offers features such as skin visualization using standard and cross-polarized light, a 3D surface contour display, and the ability to focus on the depth of lesions. VEOS is intended for general dermatologic conditions, in addition to skin cancer. While VEOS has primarily been used in Europe, it is also expanding into other regions, such as North America and parts of Asia.
For patients:
Skin vision: As of January 2025, SkinVision has over 2.5 million users globally. Currently, it can detect several types of skin cancer (MM, SCC, and BCC), actinic keratosis, and Bowen disease. The app is commercially available globally on iOS and Android devices, except for the United States and Canada.
AI dermatologist: The AI Dermatologist app is available in multiple regions including parts of Europe, Middle East, Africa, and India, but is specifically NOT available in the United States and Canada.
Many of these apps offer free basic versions or trials, with more advanced features or professional-grade tools requiring payment through subscriptions or one-time fees.
Reference:
(1). https://pubmed.ncbi.nlm.nih.gov/32041693/
(2). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7027514/
(3). https://derma.jmir.org/2019/1/e13376/
More eHealth Apps
Automated Analysis Apps
Educational Apps for Clinicians
Educational Apps for General Public
Tele-dermatology Store and Forward Apps
Track and Remind Apps
Automated Analysis Apps
Overview
The purpose of these apps is to analyze data provided by the patient or user, such as photographs of skin conditions. Since 2013, these apps have been regulated by the FDA. These apps use artificial intelligence to analyze images, providing automated risk assessments and diagnostic suggestions. The majority of AI dermatology apps use machine learning (ML). However, some apps may rely on basic rule-based systems or expert-driven criteria (like the ABCD rule or 7-point checklist) without incorporating ML to assess risks.
The user takes a picture of a suspicious mole using the app’s camera. The app processes the image with AI to identify key features of the lesion and uses machine learning to compare it against a database of skin conditions. Based on the analysis, the app provides a risk score, such as "80% chance this is benign" or "This lesion may be malignant—consult a dermatologist." The user can then follow up with a professional if the app suggests further investigation.
Advantages and Challenges
The automation of the visual examination has great potential. These apps can be helpful in reducing the burden on specialists while also aiding in the early detection of skin cancer, providing convenience for patients. However, when only a selected few lesions (at the user’s discretion) are brought to medical attention, this may cause delay in the diagnosis of early subtle skin cancers that would otherwise be readily detected on a full skin exam. There are significant hurdles to successful automated diagnosis. Pictures may be altered by light quality, user expertise, and camera quality. Inaccurate diagnosis (especially in the context of limited to no history) could lead to a false reassurance and delay medical care. In addition, difficult to reach sites such as the back, buttocks, posterior thighs, and scalp may not get adequate attention from the user. Lastly, security of images, privacy of patient sensitive information, and false advertising remains a concern.
A systematic review focusing on the accuracy of automated skin cancer detection smartphone apps concluded that currently available algorithms appear unreliable in detecting melanoma and other skin cancers (1). A study by Y. Chung et al. (2) aimed to evaluate the performance of one such app by comparing its assessments with those of a dermatologist, using lesions selected by participants. Out of 199 lesions from 125 participants, the app failed to analyze 45% of cases and misclassified many high- and medium-risk lesions as benign. The agreement between the app’s assessments and the dermatologist's diagnoses was poor. These findings highlight the app’s limited diagnostic accuracy and suggest that further research is needed to assess its effectiveness in real-world settings. Overall, these apps are useful but are not recommended to replace a physician-rendered full skin examination for the reasons stated above. In contrast, a study by M. de Carvalho et al. (3) found that the latest version of the SkinVision app (as of 2019) reported 95% sensitivity and 78% specificity for skin cancer detection (3). According to the authors, this was achieved due to improvements in the processing of images taken with the smartphone camera and a large, risk-labeled image database from users, which was used to train a machine learning algorithm.
Examples of automated analysis applications includes:
For healthcare professionals:
VEOS: This app is functional and available on the App Store. It is designed for use with the Canfield ProVEOS dermatoscope and is compatible with specific iPhone models. The app offers features such as skin visualization using standard and cross-polarized light, a 3D surface contour display, and the ability to focus on the depth of lesions. VEOS is intended for general dermatologic conditions, in addition to skin cancer. While VEOS has primarily been used in Europe, it is also expanding into other regions, such as North America and parts of Asia.
For patients:
Skin vision: As of January 2025, SkinVision has over 2.5 million users globally. Currently, it can detect several types of skin cancer (MM, SCC, and BCC), actinic keratosis, and Bowen disease. The app is commercially available globally on iOS and Android devices, except for the United States and Canada.
AI dermatologist: The AI Dermatologist app is available in multiple regions including parts of Europe, Middle East, Africa, and India, but is specifically NOT available in the United States and Canada.
Many of these apps offer free basic versions or trials, with more advanced features or professional-grade tools requiring payment through subscriptions or one-time fees.
Reference:
(1). https://pubmed.ncbi.nlm.nih.gov/32041693/
(2). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7027514/
(3). https://derma.jmir.org/2019/1/e13376/