Good Morning @matt-s17
Apologies for the delayed response.
In addition to a letter from med/reg confirming that they are actual patients, the presentation would be required to include the statistically significant data from the study for the endpoint being presented. The images should therefore reflect the average response. It would not be acceptable to select images that show the best-case scenario for patients.
The pictures would also have to be disclaimed with “Actual patient. May not be representative of all patients.”. As with patient images, they should only appear in a clinical context and with the supporting data (e.g., if you’re going to place them on a summary page, the quantification will require repeating).