Artificial intelligence (AI) has been in the spotlight in recent years and is strongly believed to be able to impact all areas of healthcare in the future. Bladder cancer is one area where extensive AI research is being undertaken in diagnosis, recurrence and survival.
A subfield of AI called machine learning has been a major focus where it reads enormous amounts of data and, using statistical models, is able to analyse the information in a way that mimics human intelligence. The large language model is one such development of machine learning that has accelerated these advances in AI, culminating in the release of ChatGPT in late 2022 (known as a type of generative AI), which was the first introduction of AI to the general population. In terms of medicine and bladder cancer, much research has been undertaken using machine learning to make current processes more effective in order to deliver better outcomes and improve patient care.
For the diagnosis of bladder cancer, AI-assisted cystoscopy, radiomics, urine cytology, biomarkers are all different areas that AI has been analysed. Cystoscopy is the investigation of choice in diagnosing bladder cancer for its ability to provide direct visualisation of the tumour. However, challenges exist especially in the detection of carcinoma in situ (CIS) due to its flat appearance, which is an early form of bladder cancer that has a high risk of becoming invasive cancer. Multiple studies have looked at improving cancer recognition with the assistance of AI through analysing large volumes of cystoscopy images.
Imaging is an integral part of bladder cancer management ranging from diagnosis, staging to surveillance. Machine learning has been utilised in various imaging modalities (known as radiomics) such as computed tomography (CT) or magnetic resonance imaging (MRI) to improve the accuracy of bladder cancer detection.
AI has been applied to urine cytology to see if it can improve upon the detection of malignant bladder lesions. Similarly, biomarkers are being tested that are specific to bladder cancer in the hope of improving diagnosis.
Survival outcomes in bladder cancer is another area of AI research. Models are being developed that can calculate morbidity and mortality post-cystectomy to improve risk assessment and optimise patient selection. Predicting recurrence of cancer post-cystectomy is another example given the poor long-term survival rates currently.
There is significant hope that AI will impact healthcare greatly and improve the lives of many patients. Extensive research is being conducted in this field and bladder cancer is one condition that will benefit from it. There are also concerns with this rapidly developing technology however and it does not come without its risks, which has been recognised by governments around the world, many of which have signed the Bletchley declaration that AI must be developed in a safe manner at the recent AI Safety Summit held in the United Kingdom. Appropriate governance of the use of this technology will be the key to optimising outcomes for patients in the future.