Automated Lugano Metabolic Response Assessment in <sup>18</sup>F-Fluorodeoxyglucose-Avid Non-Hodgkin Lymphoma With Deep Learning on <sup>18</sup>F-Fluorodeoxyglucose-Positron Emission Tomography.

For the subset of patients where the method's assessment is in disagreement with the final IRC response, we observe a numerical trend where the OS survival rates of patients assessed as CMR by the model and non-CMR by the final IRC response are numerically better than the survival rates of non-CMR patients by the model and CMR by the final IRC response (Data Supplement, Fig S12).

also in 426 other papers