Fig. 2: Overall workflow of the PETAL trial and the AI-based re-evaluation.

In case of a negative interim PET result, patients were randomized to receive either 6xR-CHOP or 6xR-CHOP plus two additional rituximab administrations. A negative interim PET scan was defined as an SUV decrease >66% of the most FDG-avid lesion compared to the baseline scan. Patients with a positive interim scan were randomized to receive either 8x(R-)CHOP or 2x(R-)CHOP followed by 6 blocks of the Burkitt protocol (R was restricted to CD20-positive lymphomas) (A). The interim PET scans were re-evaluated using an AI-based software to automatically quantify FDG uptake (B). Patients with a positive interim PET had statistically significantly higher mean-SUVAI compared to those with a negative interim PET (C).