AI versus dermatologists in diagnosing melanoma accurately
AI has been used for a variety of purposes in the medical field. It has been used to diagnose skin cancer, identify diseases and provide treatment recommendations. In the future, AI will likely be used to assist in diagnosing melanoma and other cancers.
Significance of diagnostic accuracy of skin cancers
Skin cancer is one of the most common types of cancer with more than 3 million cases diagnosed each year. It is also one of the most preventable types of cancer with early detection being key in minimizing the chance of death.
AI assisted diagnosis of melanoma
AI has been used to assist in this area, and has had a positive impact on what was previously seen as a difficult and frustrating process. Melanoma is an example of a cancer that can be easily detected by AI. When skin cancer is diagnosed correctly, it is often treated quickly and effectively with surgery. Once the diagnosis has been confirmed, AI can be used to turn a skin cancer diagnosis into an accurate treatment recommendation.
A study to compare the accuracy of diagnosing melanoma between AI and dermatologists.
An interesting article was published by the dermatolgoists at the University of Gothenburg titled “Discrimination between invasive and in situ melanomas using a convolutional neural network”.
The research shows that melanoma can be detected by AI technology as accurately as the dermatologists.
To train the AI, dermoscopic images of 1137 histologically verified melanomas were shown to the Ai for machine learning.
A new set of 200 dermoscopic images of melanoma were shown to the AI as well as dermatologists. AI was accurate in 69.2% of cases, while dermatologists were accurate in 73.4%. Although the difference between the two is not statistically significant, the dermatologist was 4% better at diagnosing melanomas.
Real life signficance
The downside of this study is that a dermoscopic image is not the same as a real photo of the skin lesion that can be taken with a smartphone. It is a special photo that is taken under a skin microscope called a dermoscope which is not available to the public.Therefore, the public is not yet ready to have their smartphones diagnose melanoma. However, it could help the General Practitioners.
But imagine if the AI was trained with dermoscopic images of 1 million melanomas. The accuracy of AI dermoscopic image interpretation may surpass that of dermatologists. And furthermore, if dermatologists are no better than AI, would you need to pay expensive fees to see a dermatologist?
Watch this space.
Follow us on Facebook so that you can detect skin cancers early.
If you notice any new spots, it is important that you seek medical advice promptly. Remember, early detection may save your life. The best way to detect early skin cancers is to be able to detect new and abnormal skin spots. Please follow our Facebook to see one new skin cancer image every day. These are skin cancer images taken at the Eastwood Skin Cancer Clinic. It may help you to spot a skin cancer.
