The results demonstrate the ability of an AI device to prioritise studies for clinical review based on AI-determined presence of findings. The highly configurable design of the AI device allows the customisation of findings that warrant the classification of urgent and tuning of thresholds for the subject population. In this instance the sensitivity was deliberately high as the aim was to ensure anything categorized as unremarkable is unremarkable, resulting in a high negative predictive value. ​

Within the data set there were 4 cases of urgent suspected cancer which were all correctly identified. 

On further analysis of the false negatives, a common diagnoses incorrectly categorised was simple pleural effusion accounting for 7 of the 24 cases. In the original configuration, this was deemed as an unremarkable finding for the program, but was flagged by the radiologist to be remarkable in these specific cases. Classifying the finding “simple effusion” as a category 2 finding resulted in an increase in sensitivity to 73.4%, with a corresponding decrease in specificity (83.8%), decrease in PPV (37.9%) and increase in NPV (95.9%). The low PPV in both instances was to be expected, as there was a low low prevalence of remarkable cases (11.9%).

The device performance will be monitored ongoing through a limited user roll-out to confirm usability and governance, before proceeding to rollout across the entire NHS Trust, transitioning the system into prospective real-world operations.​