Dermatology- Telemedicine, Digital Hospitals, and Robots

Written By Harshil Shah

Author: Katrina Cirone

TLDR

Telemedicine - First Derm connects patients with dermatologists within 24 hours

Digital Dermatology Hospital - AIP Derm uses neural networks for advanced diagnostics

Robotics- Vectra WB360 Imaging System images the whole-body in a single capture

Telemedicine - First Derm manufactured by iDoc24

Real World Application: https://www.firstderm.com

At the intersection of deep technology and healthcare there is a need for technology companies to collaborate with healthcare institutions to deliver and deploy solutions to overcome the deficit of healthcare professionals. Like many other medical specialties, dermatology has undergone and continues to go through various technological changes and advancements to improve patient care and access. Teledermatology is one of such methods and FirstDerm is a teledermatology service founded in California with the ability to connect patients with access to the mobile application or web platform to a dermatologist within 24 hours for a fee of 30 USD. Firstderm has over 500,000 mobile application downloads and has answered over 50,000 dermatology cases. 

This application is primarily intended for patient use to learn about their skin concerns and serve as a teledermatology service. The application has many feature such as the ability to search for nearby clinics and pharmacies, a “Skin Guide” that contains a medical library of diseases, a list of featured physicians, the ability to take a dermoscopic photo using a dermoscopic attachment, and the capability to receive a faster response for their case by paying a higher fee. FirstDerm has various partners that incorporate their products into their workflow/sites which include HealthLine, DermLite, SkyDeck, Your.MD, Philips, Apoteket, Telehealther, Doktorn, Netdoktor.se, iDermo, Lyfep Nd, Intermat, STD Triage, iDoc24, Skin Site, Stroll Health, and Miracle Messages.

This product also has a diagnostic AI-based, CE marked dermatology engine named AutoDerm with the ability to scan 44 skin diseases:

The results show an overall diagnostic accuracy of 56.4% in the top 5, and 22.8% accuracy for the most probable diagnosis. This would probably be interpreted as a poor diagnostic performance if translated into use in clinical practice as a diagnostic aid

This is available to both end users or healthcare organizations with a $500/month subscription. This model, which was trained on over 110, 000 smartphone images of skin conditions generates the top 5 ranked skin diseases based on probability. An additional feature is the Skin Image Search engine which allows users to search a medical image database and match a skin condition; however, this feature has a low diagnostic accuracy. Following submission of 2 images the patient is provided with a top 5 differential diagnosis which will also include information on the condition, symptoms, example photos, and when to seek care or treatment.

Digital Dermatology Hospital  - AIPDerm:

Real World Application: https://www.aipderm.com


Skin conditions and cancers are a leading health concern globally and have a prevalence comparable to that of cardiovascular diseases and other major chronic conditions. The urgent need to address this growing public health concern and the challenges associated with the current system have prompted various companies like AIP to develop technologies to augment and improve the existing process. AIP has developed and implemented a Digital Dermatology Hospital which serves as a fully-integrated end-to-end system that employs artificial intelligence, big data, and analytic technologies that incorporate deep learning, computer vision, and graphical modeling. This system helps provide clinicians with an advanced instant differential, helps triage patients based on condition severity at presentation, prioritizes cases through a skill pyramid that delegates cases appropriately based on internal skill or seniority, and enables remote and asynchronous treatment which benefits both clinicians and patients. 

Recent advances in the field of computer-vision have further enabled development of systems powered by deep-learning and utilization of “neural networks” using layers of artificial neurons to emulate processing of signals in the visual cortex. AIPDerm uses convolutional neural networks (CNNs) which consist of layers that conduct convolutions based on the relationship between functions through an algorithm called “backpropagation” which allows for the extraction of insights to facilitate machine learning.

AIPDerm uses an ensemble of CNN models to classify both macroscopic and dermoscopic (using a dermatoscope) skin lesions and has the ability to accurately detect over 700 unique skin conditions which represent over 95% of the global prevalence. AIPDerm has been able to increase health care efficiency by reducing staff time and needed resources while also increasing patient processing in-person and remotely. AIP has a partnership with Semmelweis University (a research-led medical University in Budapest, Hungary) which has increased the access to images and data needed to build datasets and improve the accuracy of the current system.

It is inevitable that the use of artificial intelligence in the healthcare space will continue to experience rapid growth due to the observed benefits and AIPDerm is only one of the many companies working to achieve this.

Robotics - Vectra WB360 Imaging System:

Real World Application: https://www.canfieldsci.com/imaging-systems/vectra-wb360-imaging-system/

The VECTRA WB360 is a whole body 3D imaging system capable of imaging the entire skin surface with just a single capture designed and developed by Canfield Scientific and HSD. This technology uses 92 cameras to give physicians the ability to map in either 2D or 3D and monitor skin lesion evolution and distributions. The VECTRA WB360 can take a 360-degree scan of the body and can track various conditions such as pigmented lesions,vitiligo, psoriasis, burns, neurofibromatosis, and lymphedema. The generated images are stored in macro quality resolution and allow for the generation of 3D body maps as well as for accurate measurements on the skin. Concerning lesions can be flagged and analyzed over time to track progress and evolution. The ability for longterm tracking allows for the identification of rapidly progressing lesions and prompt treatment to improve patient outcomes. In addition, this system allows for the simplification and standardization of patient photography. Currently, software to compare lesions taken from two different body-maps on different scans is under development.

The VECTRA WB360 which is in use at various hospitals globally  also improves the patient experience and efficiency by decreasing the photography time from 10-20 minutes to less than a second. The system is made from thin walled thermoformed Kydex which allows for light weight parts and assembly by a single technician. These factors indicate the potential for the VECTRA WB360 to become an essential device in the fields of dermatology and plastic surgery. Overtime, the VECTRA WB360 may become the new standard of care in these fields.

Harshil Shah
Previous

Orthopedics- VR, Implants, and Regeneration