Expert comments

Comments on this technology were invited from clinical experts working in the field and relevant patient organisations. The comments received are individual opinions and do not represent NICE's view.

Two experts were familiar with the technology and had both used it in a study. Two experts had not used this specific technology before.

Level of innovation

All experts agreed that Paige Prostate is novel and is proposed as an addition to standard care. One expert highlighted that although the technology is not widely used in the NHS, this is reflective of the position of all AI technology for histopathology and cellular pathology.

Using AI to help histopathologists was highlighted in the UK Government Life Sciences Strategy and by implication in the NHS Long Term Plan as part of digitally enabled care. The aim is to help solve some of the workforce and other resource challenges. Experts acknowledged that there are other systems available in this area. One expert highlighted this technology to be the first AI product in digital pathology to receive FDA approval (see Regulatory information).

Potential patient impact

All experts reported that the system has the potential to reduce missed cancers or areas suspicious of cancer valuable for all patients. All experts also highlighted the potential to increase the efficiencies in the care pathway and as a result the speed of turnaround for patients. One expert highlighted it could reduce staff time, although there would likely be a learning curve before this was seen. One expert highlighted that there is inherent subjectivity to assessment of Gleason scoring by human observers and Paige Prostate has the potential to standardise assessments such as objective grading of the cancer. Two experts reported the technology would benefit all who have a prostate biopsy regardless of the diagnosis.

Potential system impact

Experts identified the complexities in the cost implications for the current care pathway. The balance of cost savings because of greater efficiencies in pathologists time against the cost of the technology is difficult to say. Two experts highlighted a barrier to widespread adoption in the IT infrastructure needed for the deployment in laboratories. One expert reported the deployment of digital pathology to be gaining traction through different initiatives and funding routes including groups of trusts completing successful business cases and the AI centres of excellence programme to allow such technologies across the field.

All experts also highlighted the need for pathologists training on using Paige Prostate and functionality, in particular limitations to make sure it is used correctly. One expert highlighted that those conveying the results to patients (urologists, oncologists, and specialist nurses) may also need brief training in the technology to support patient understanding around the decision making for their diagnosis and management.

General comments

One expert highlighted the studies reported are retrospective and prospective use and audits would be important to inform how valuable the technology could be in practice. One expert highlighted the theoretical possibility that AI may change pathologist reporting profiles influencing diagnostic patterns. This may involve flagging more suspicious areas as atypical small acinar proliferation, leading to more patients being followed up than discharged. However, the rate of atypical small acinar proliferation may be reduced by more diagnostic certainty afforded by AI, allowing more to be definitively categorised as benign or malignant. While these are mitigated by the pathologist having ultimate oversight of the technology the expert highlighted the importance of pathologists training and for professionals to consider and monitor the impact of AI on their reporting patterns.