When skin cancer is removed from the face, it often leaves behind a wound that needs to be reconstructed. The desired end result of this reconstruction can vary greatly from patient to patient. Plastic surgeon Maarten Hoogbergen of the Catharina Hospital wants to use artificial intelligence (AI) to help patients make choices.
For some patients, an aesthetically pleasing result is of paramount importance, whilst others are more concerned with a quick recovery or undergoing as few procedures as possible. Hoogbergen aims to give patients a clearer picture of the final result and thus help them make a well-informed choice.
What does the patient consider important?
“As a doctor, you’re always influenced by your training, your experience and your own ideas about what constitutes a good result,” says Hoogbergen. “But what I consider a good result isn’t necessarily the same as what the patient considers important.”
The surgeon wants to use AI as a tool to enrich discussions between doctors and patients. What medical options are available? And which outcome best aligns with the quality of life the patient has in mind?
Outcome measures
Hoogbergen noticed that some patients remained dissatisfied, despite a technically successful operation. Others, on the other hand, were satisfied because they could once again do what was essential to them, even if the medical outcome was not perfect.
Since then, Hoogbergen has focused on outcome measurement in plastic surgery. This involves looking not only at medical results, but also at the impact on the patient’s daily life. Can someone return to social life? Does someone feel comfortable with their appearance? Was the treatment worth it?
Collecting data
The project is still at an early stage. At present, the main focus is on collecting data. In future, the AI will learn from previous patients. A combination of photographs, medical data and questionnaire results should enable the system to predict which treatment offers the best chance of an outcome that meets the patient’s wishes.
“We’re really only just getting started,” says Hoogbergen. “First, you need a good AI system that can predict what the result might look like after reconstruction. Then you need to link that prediction to the results of the questionnaires. Only then can you really start to assess which treatment offers the best chance of achieving the desired quality of life for which patient.”
