Background and Objective: Assessment of lymphoma treatment response using 2-[18F]FDG PET/CT scans traditionally relies on visual evaluation based on metabolic criteria, which introduces subjectivity and compromises reproducibility. This study aims to develop and evaluate automated approaches that use deep learning or radiomics to standardize treatment response assessment by predicting simplified Deauville scores and detecting disease progression. Methods: We developed a comprehensive analytical framework that incorporates deep learning and radiomic approaches. The methodology focused on two primary tasks: (1) prediction of simplified Deauville scores and (2) detection of tumor progression in paired PET/CT images. Deep learning and radiomic features were extracted from the imaging data and utilized in parallel analytical pipelines to allow performance comparison between the approaches. Results: Despite the constraints of a limited dataset, radiomic methods demonstrated similar performance in both simplified Deauville score prediction and progression detection to deep learning approaches. Both methodological approaches proved effective in performing the designated analytical tasks, with promising results for clinical implementation. Conclusions: The proposed automated framework demonstrates the feasibility of standardizing the assessment of the response to lymphoma treatment using artificial intelligence techniques. The complementary strengths of deep learning and radiomic approaches offer the potential to reduce the complexity of diagnostic evaluation while maintaining precision. These findings establish a foundation for the further development and refinement of automated assessment tools in lymphoma treatment monitoring, with potential implications for improving clinical workflow and patient outcomes.