Väitös (radiologia): MSc Cameron Wright
Aika
24.3.2023 klo 12.00 - 16.00
MSc Cameron Wright esittää väitöskirjansa “OPTIMIZING MRI-GUIDED PROSTATE ULTRASOUND ABLATION THERAPY USING RETROSPECTIVE ANALYSES AND ARTIFICIAL INTELLIGENCE” julkisesti tarkastettavaksi Turun yliopistossa perjantaina 24.3.2023 klo 12:00 (Turun yliopisto, Medisiina D, Alhopuro-auditorio, Kiinamyllynkatu 10, Turku).
Yleisön on mahdollista seurata väitöstä etäyhteyden kautta: https://echo360.org.uk/section/a1bc26ec-76f1-4848-9f0d-dd5436402d8c/public (kopioi linkki selaimeen).
Vastaväittäjänä toimii professori Jurgen Fütterer (Radboud University Medical Center, Alankomaat) ja kustoksena professori Roberto Blanco Sequeiros (Turun yliopisto). Tilaisuus on englanninkielinen. Väitöksen alana on radiologia.
Väitöskirja yliopiston julkaisuarkistossa: https://urn.fi/URN:ISBN:978-951-29-9183-9.
***
Tiivistelmä väitöstutkimuksesta:
Magnetic resonance imaging (MRI)-guided transurethral ultrasound ablation (TULSA) is a novel therapy that has been used to treat prostate cancer (PCa). TULSA destroys prostate tissue using ultrasound heating. Heating is monitored in real-time using MRI thermometry. Despite TULSA’s promise, there are several challenges that have slowed its adoption. Fortunately, MRI images and heating parameters from all TULSA treatments are stored afterwards. By conducting retrospective analyses and applying deep learning on existing TULSA treatments, we can extract valuable information and then leverage it to optimize future TULSA treatments.
One specific patient group that required extensive investigation was those PCa patients who had initially been treated with radiation therapy (RT), later recurred, and were seeking a second therapy with TULSA. These patients often have leftover metal markers in their prostate from RT. These markers can theoretically negatively affect TULSA treatment, both from a safety and efficacy perspective. After detailed retrospective analysis, we have determined that for the large majority of patients with gold markers, they can still be successfully treated, with a few clear exceptions. However, patients with nitinol markers are not ideal candidates.
Artificial intelligence was also used on existing patient data to improve outcomes. Immediately after TULSA heating, MRI contrast agents are used to visualize if the tumour has been fully treated. Unfortunately, even if undertreatment is observed, retreatment is not possible, forcing an additional treatment months later and associated risks of another therapy. By training a deep learning model with existing contrast-free MRI image sets, we have predicted the heating effect after TULSA with high accuracy.
Overall, this work will help daily clinical practice and increase the odds of a successful TULSA therapy.
Yleisön on mahdollista seurata väitöstä etäyhteyden kautta: https://echo360.org.uk/section/a1bc26ec-76f1-4848-9f0d-dd5436402d8c/public (kopioi linkki selaimeen).
Vastaväittäjänä toimii professori Jurgen Fütterer (Radboud University Medical Center, Alankomaat) ja kustoksena professori Roberto Blanco Sequeiros (Turun yliopisto). Tilaisuus on englanninkielinen. Väitöksen alana on radiologia.
Väitöskirja yliopiston julkaisuarkistossa: https://urn.fi/URN:ISBN:978-951-29-9183-9.
***
Tiivistelmä väitöstutkimuksesta:
Magnetic resonance imaging (MRI)-guided transurethral ultrasound ablation (TULSA) is a novel therapy that has been used to treat prostate cancer (PCa). TULSA destroys prostate tissue using ultrasound heating. Heating is monitored in real-time using MRI thermometry. Despite TULSA’s promise, there are several challenges that have slowed its adoption. Fortunately, MRI images and heating parameters from all TULSA treatments are stored afterwards. By conducting retrospective analyses and applying deep learning on existing TULSA treatments, we can extract valuable information and then leverage it to optimize future TULSA treatments.
One specific patient group that required extensive investigation was those PCa patients who had initially been treated with radiation therapy (RT), later recurred, and were seeking a second therapy with TULSA. These patients often have leftover metal markers in their prostate from RT. These markers can theoretically negatively affect TULSA treatment, both from a safety and efficacy perspective. After detailed retrospective analysis, we have determined that for the large majority of patients with gold markers, they can still be successfully treated, with a few clear exceptions. However, patients with nitinol markers are not ideal candidates.
Artificial intelligence was also used on existing patient data to improve outcomes. Immediately after TULSA heating, MRI contrast agents are used to visualize if the tumour has been fully treated. Unfortunately, even if undertreatment is observed, retreatment is not possible, forcing an additional treatment months later and associated risks of another therapy. By training a deep learning model with existing contrast-free MRI image sets, we have predicted the heating effect after TULSA with high accuracy.
Overall, this work will help daily clinical practice and increase the odds of a successful TULSA therapy.