iToBoS

With new hardware and artificial intelligence against skin cancer

The aim of the iToBoS project is to train an AI system  able to integrate information from different sources, assessment of individual moles while considering the specific characteristics of each patient. With systematic  successive explorations of a patient, the system will be able to also robustly determine the changes occurring in  the individual moles, a key feature held as one of the most informative in the detection of skin cancer. The proposed  holistic approach will enable physicians to diagnose skin diseases earlier and with higher accuracy, thus increasing  effectiveness and efficiency in personalized clinical decision making. 

Current state-of-the-art AI systems for detection of melanomas using a single dermatoscopic image of the patient,  fail to “understand” the naevus type of the patient. This naevus type refers to the typical (prevailing) pattern of the  patient’s naevi. Clinical practice has proved that only the evaluation of all naevi allows us to define the “naevus type”  of the patient examined, and more attention should be directed towards naevi not consistent with the prevailing  naevi pattern. 

The proposed novel skin scanner will seamlessly acquire complete skin imagery  comparable to the quality of dermoscopy, but covering most of the patient’s skin. As our AI cognitive assistant  analyses the complete skin of the patient, it will be able to detect all the naevi of the patient, defining the naevus type for that specific patient. This approach enables the recognition of ugly ducklings, i.e., detection of a mole that does  not resemble the average normal mole of that specific patient, which is something current AI systems cannot do. 

iToBoS will provide high prediction quality by fusing different types of data, including medical records, genomics  and temporal evolution of lesions, while using the total body maps to detect ugly ducklings (naevus phenotype) and  measuring the imaging phenotype of that specific patient, i.e. assessment of UV damage, size and number of naevi,  and location of each lesion, thus achieving a highly personalised diagnosis. However, high prediction quality is not  enough to increase effectiveness and efficiency of treatments: the dermatologist requires explanation and insight  for a better understanding beyond standard quantitative performance evaluation. The iToBoS team has proposed in  the literature several AI explanation methods, concerning state of the art computer vision datasets, but also a  few in the medical domain. The project will also develop explainable AI (XAI) methods focused on the generation  of system justifications that will sufficiently and understandably explain the reasoning behind its inference to the  dermatologist.