Meet ‘Piction Health’, An MIT-Based Startup That Uses Machine Learning And Computer Vision To Help Physicians Identify And Manage Skin Disease

Skin rashes are very common, and there are numerous causes of skin rashes. It refers to any change in the skin caused by an infection or an allergic reaction. A rash may, in extremely rare circumstances, be a sign of malignancy. Knowing the difference can assist someone in getting the required assistance or prevent worry about a noncancerous rash.

People see physicians and dermatologists for a wide variety of skin conditions, including the most common ones— rashes like eczema, acne, and shingles. A survey shows that all these conditions are typically classified as dermatology, while dermatologists prefer to work on skin cancer cases or those needing their help. 

Susan Conover launched a startup called Piction Health (initially a smartphone application) that allows doctors to identify melanoma from photos. When Susan was 22, she had a strange-looking mole examined. She was informed that she would need to wait three months before seeing a dermatologist. The mole was eventually excised, and a biopsy revealed it was malignant. Fortunately, the mole was condoned to one spot. 

This experience introduced her to the field of dermatology and skin conditions. Conover established Piction Health after researching these issues and potential technical answers in the MIT System Design and Management graduate program. 

Conover and Pranav Kuber, who founded the company with her, now concentrate on assisting doctors in recognizing and treating the most prevalent skin conditions, such as rashes like eczema, acne, and shingles. In the future, they hope to collaborate with a company to assist in the diagnosis of skin cancers.

While exploring the field of skin rashes, they discovered that most skin conditions encountered by primary care doctors are skin rashes, including psoriasis, eczema, and rosacea. There are a lot of individuals with skin problems every year, and two-thirds of those people seek assistance from general care. Roughly half of those cases are misdiagnosed because these doctors don’t have as much expertise in dermatology.

Developing a machine-learning model to detect a plethora of different illnesses would be more challenging than training a model to spot cancer. However, with their investigations, they needed to pivot away from skin cancer to assist skin cancer patients in seeing a dermatologist sooner.

During the epidemic, Piction had to build data relationships with hundreds of dermatologists throughout the globe to leap. The team concentrated on collecting images of people with various skin tones because many of them are underrepresented in medical literature and instruction. With more than 1 million images gathered by dermatologists in 18 countries, Conover claims Piction now has the biggest dataset of rashes worldwide.

Piction provides doctors with details on each ailment’s available treatments after identifying the most likely condition. The app displays pictures of skin presentations with comparable characteristics when primary care physicians take a snapshot of a patient’s skin condition. To decide on the best course of treatment for the patient, doctors can distinguish between the illnesses they most suspect.

 According to Conover, Piction can cut the time it takes doctors to examine a case by about 30%. Additionally, it can facilitate faster dermatologist referrals for complicated situations that a doctor is unsure of how to handle.

This app also assists healthcare organizations in lowering expenditures associated with pointless follow-up visits, medicines, and pointless referrals. More than 50 doctors have utilized Piction’s product so far, and the business has partnered with several institutions, including a well-known defense agency that recently had two employees who couldn’t go to a dermatologist immediately diagnosed with late-stage melanoma.

This year, Piction will roll out a number of new pilots. Conover plans to eventually include the ability to recognize and assess wounds and infectious illnesses like leprosy that are more prevalent in other regions. The business also intends to make its solution available to doctors in underdeveloped regions by collaborating with charitable organizations. Conover thinks that this might eventually develop into a complete diagnostic tool.

References:

• https://news.mit.edu/2022/piction-health-skin-app-0706
• https://www.pictionhealth.com/

Please Don't Forget To Join Our ML Subreddit