BMBF/VDI funding program for innovative SMEs, FKZ: 13GW0499 A

?

Basal cell carcinomas are the most common cancers in humans. Up to now, these skin tumors have been detected relatively late, so that they often have to be surgically removed with stepwise margin control when first diagnosed due to their extent. This is because they initially appear in the form of skin-colored plaques and look non-specific. Optical coherence tomography (OCT) is suitable as a non-invasive diagnostic tool for the early detection of basal cell carcinomas. This method can reliably detect even very small, clinically barely visible tumors without a biopsy. Due to their small size and depth of infiltration, these can then be treated with a coagulating laser instead of surgery.

?

In the OCTOLAB project, optical coherence tomography (OCT) for the diagnosis of basal cell carcinomas is integrated into a long-pulsed infrared laser for the treatment of tumors. In addition, the diagnosis and therapy are based on artificial intelligence (AI). The OCT image parameters of basal cell carcinomas are used for a machine learning process from the field of deep learning in order to achieve AI-based automated diagnostics. For this purpose, the subtype, tumor thickness and lateral tumor boundaries are determined. Based on this information, the optimal laser parameters are determined automatically, and therapy is carried out using laser coagulation. This is therefore a closed loop between diagnostics and therapy.

?

This combined device for automated diagnostics and therapy could contribute to resource-saving early detection and individualized minimally invasive therapy of basal cell carcinomas. This can improve patient care and relieve the burden on the health system by eliminating risky and costly operations.

?

Cooperation partners at the 伟德国际_伟德国际1946$娱乐app游戏 of Augsburg:

Prof. Dr. Frank Kramer

To the website

?

?

The line-field confocal OCT is a fusion of optical coherence tomography and confocal laser microscopy. It enables three-dimensional in-vivo microscopy of the skin in real time with high cellular resolution. At the dermatology clinic, this new method is being evaluated and further developed for use in skin cancer and inflammatory skin diseases. A particular focus is on comparing LC-OCT with conventional OCT and confocal laser microscopy. Here, too, explainable artificial intelligence is already being integrated into real-time imaging to support clinical decision-making.

Development of dynamic image parameters, creation and annotation of an image database and the development of a chatbot to create a tool for the AI-supported analysis of parameters of healthy skin

?

Cosmetic users are faced with an almost unmanageable number of products with different mechanisms of action and objectives. It is difficult to make an individualized selection tailored to one's own current needs. Beauticians or employees in pharmacies can help users to assess the condition of their skin, recommend products and explain how to use them. However, this expert advice is always linked to direct personal contact and is therefore inflexible in terms of time and location. As the population becomes increasingly digitized, many people have a strong need to have their skin condition assessed independently, objectively and without contact. There is no image-controlled tool on the market that is initiated via a smartphone and analyses the health and condition of the skin in detail and is also able to provide individualized care and styling advice based on the results of the analysis and to incorporate information from the user's medical history. The tools that can be found on the market are very limited both in terms of their validity and the complexity of their analysis results.

?

The aim of the project is to develop a smartphone-based AI skin analysis that goes far beyond the tools currently on the market, which records comprehensive and valid parameters of the skin's condition in combination with individual requirements and in the context of extrinsic environmental factors and, on this basis, generates individualized skin care consulting services; the aim is to create a portal for users that provides advice on their individual needs. The user should be encouraged to regularly carry out new analyses through the intensive skin analysis described, good usability and intuitive operation of the tool. This allows individual changes, climatic conditions and environmental influences such as UV protection to be taken into account.

?

The sub-project of the project partner Dermatology at the 伟德国际_伟德国际1946$娱乐app游戏 Hospital Augsburg includes the development and testing of the static and dynamic image parameters, the creation and annotation of a large image database and the development of a chatbot (structured questionnaire). For this purpose, static images of faces including dynamic components (facial expressions, elasticity) are taken after the target group has been defined. The determination and quantification of dynamic parameters must be developed as part of the project. The previously defined relevant parameters are annotated on these images, i.e. classified and assigned according to their degree of expression. In parallel, the chatbot is being developed: relevant questions are defined and assigned to the images. This provides a data set for training and validating the artificial intelligence. Clinical expertise is necessary to develop the typing of the skin condition and thus to train the artificial intelligence, which can classify the skin condition and type by pattern recognition in the portrait of a person in conjunction with anamnesis questions (sun exposure, smoking habits, intolerances). In addition, the validity and reliability as well as the relevance and acceptance of the training results are evaluated and assessed.

?

?

Cooperation partners at the 伟德国际_伟德国际1946$娱乐app游戏 of Augsburg:

?

Prof. Dr. Frank Kramer

To the website

?

?

?

Malignant melanoma is one of the most malignant tumor diseases in humans because it metastasizes early. The detection, quantification and analysis of circulating melanoma cells is of great importance for estimating the prognosis, aftercare and therapeutic management. Together with the physics and physiology departments at the 伟德国际_伟德国际1946$娱乐app游戏 of Augsburg, methods for the physical analysis and sorting of melanoma cells are being tested and analyzed. In another project, tumor vessels are analyzed in the context of the microenvironment of various skin cancer types in vivo using LC-OCT and observed under immunotherapy.

Cooperation partner at the 伟德国际_伟德国际1946$娱乐app游戏 of Augsburg:

?

Dr. Christoph Westerhausen

?

To the website

?