Clinical needs to address in
diagnosing prostate cancer
Patients
- Faster diagnosis
Waiting times for diagnosis are a particular challenge in the NHS, where patients can sometimes wait 2-3 months.
AI could help by using risk scores to prioritise cases and by reducing clinical work, enabling patients to be reported, have a biopsy, and get a diagnosis earlier.
- Better diagnosis
Many hospitals struggle to provide expert radiologists with prostate cancer specialism.
Research indicates that AI can increase the consistency and accuracy of reporting [1], helping address health equity issues and make the best care available to all.
- Most appropriate treatment options
Patients are often anxious about intrusive and painful investigations or the side-effects of treatment.
By providing additional information, AI can help doctors personalise the decision according to local and patient priorities, whether to avoid biopsy, maximise cancer detection, or choose the best therapy.
Radiology
- Avoid missing cancer
Clinical studies highlight that radiologists can miss around 10% of cancers on prostate MRI.
AI can help satisfaction of search, by highlighting suspicious regions for consideration that otherwise might be overlooked.
- Confirm negative cases or indeterminate cases
One of the benefits of prostate MRI is that it can enable cancer-free patients to avoid further investigations like biopsy, but it can be difficult to make this decision, especially where radiologists work on their own or are less experienced.
AI scores can provide an additional view of risk at the patient and lesion level, assisting decision-making.
- Report at the level of experienced teams
It is well documented that radiology reporting can vary greatly, with potential for significant gaps in sensitivity or high false positive call rates.
Studies indicate that AI can help less experienced or less accurate radiologists report at the same level as experts [1], to offer patients the best possible diagnosis.
- Save time
By calculating volumes automatically and providing segmentations and graphical images, AI could enable radiologists to work more quickly while producing more consistent results. [2]
Urology
- Clear communication
It is often difficult for clinical colleagues to decipher radiology reports, especially in text or zonal form.
AI graphical images and segmentations help ensure that the whole team understands the diagnostic questions.
- Visualisations and segmentations for applications like biopsy and treatment planning
Urologists want the best available information and to make use of state-of-the-art technologies such as ultrasound fusion biopsy, but few radiologists have the time to draw the contours (segmentations) needed.
AI can automatically segment both the anatomy such as the prostate, and regions of interest, for editing as needed and export to systems such as fusion biopsy.
Pathway
- Confidence in consistent, high quality reporting
Few hospitals can have an expert on call at all times. Inconsistent and variable report quality is a common concern.
AI outputs are consistent, and research indicates that AI can help improve the overall accuracy of reporting to benefit the patient, pathway colleagues and provider.
- Prioritise cases or cut waiting times
Many providers face challenges with resourcing or waiting times.
AI risk scores could help radiologists determine where best to focus resources and set priorities to address bottlenecks.
References
[1] Giganti F, et al. Impact of AI on prostate MRI reporting for non-expert and expert readers in a multi-reader, multi-centre evaluation, European Society of Urogenital Radiology Annual Meeting, September 2024
[2] Ward P, et al. The use of an AI software tool for prostate volume measurement on MRI, European Society of Urogenital Radiology Annual Meeting, September 2023