14.06.2024
PhD student at DCPT, Casper Dueholm Vestergaard, is honoured with the best poster award at the PTCOG62 conference in Singapore. The study presented evaluates deep learning based synthetic CTs for head and neck cancer patients, and the method is currently being tested for clinical implementation at DCPT with very promising results.
Because of anatomical changes during the proton therapy treatment course, changes that are common for head and neck cancer patients, careful monitoring is required during treatment course. As the in-room cone beam CTs (CBCTs) used at DCPT for daily positioning do not have high enough image quality to evaluate the dose plan on the daily anatomy, this can result in additional CT scans.
Casper Dueholm Vestergaard has developed a deep learning network to generate synthetic CTs from CBCTs, and the results accurately preserve the CBCT anatomy while also improving the image quality. The automated process means much faster evaluation of the daily delivered dose of the patients, and thereby supporting the adaptive workflow while saving the patients for additional control CTs. The study is now awarded with the best poster award at this years’ PTCOG conference, and assistant professor, Vicki Trier Taasti, was present in Singapore to receive the award.
Implementation of AI-developed tools can ease the clinical workflow
For the clinicians, the automation is great news, which is why the method has already been implemented into clinical use for head and neck cancer patients at DCPT. The project is not only a great example of research developments closely connected to the clinic, but also how implementing new AI-developed tools clinically can benefit clinicians and patients alike.
‘The clinical workflow for estimation of the daily dose form CBCTs has been reduced from several hours to less than 10 minutes with the potential for even further reduction and automisation’, says medical physicist, Ulrik Elstrøm.
‘The next steps now are translating these methods to other diagnosis and developing new techniques based on them. We have just started the evaluation of a synthetic CT generation model for prostate cancer patients,’ Casper Vestergaard explains.
The clinical evaluation study of the synthetic CTs was performed in a local collaborative project between the researchers Professor Ludvig Muren, Vicki Taasti and Casper Vestergaard, postdoc Jintao Ren, and the clinical head and neck specialists; medical physicist Ulrik Elstrøm, head of clinical medical physics Ole Nørrevang, and physician and clinical associate professor, Kenneth Jensen