Description

This study adopted a multicenter retrospective cohort design. A total of 2,000 patients with highly metabolic lymphoma who underwent PET/CT scans and had clinical data available at Ruijin Hospital, affiliated with Shanghai Jiao Tong University School of Medicine, from June 2013 to December 2023, were selected for model development. The model was validated using an external validation cohort of 2,000 patients with highly metabolic lymphoma who underwent PET/CT scans at other institutions, including West China Hospital of Sichuan University. All patients were required to have undergone 18F-FDG PET/CT scans before treatment, during treatment, and after treatment.

Lesions were manually delineated, and software was used to automatically calculate PET/CT parameters such as SUVmax, metabolic tumor volume (MTV), and total lesion glycolysis (TLG) for primary and metastatic lesions based on 18F-FDG PET/CT. Clinical data for all patients were collected from their medical records, including age, gender, LDH levels, B symptoms, extranodal involvement, IPI score, presence of bulky disease, pathological subtype, and treatment regimen.

Each patient's end-of-treatment PET/CT was compared with baseline PET/CT based on the Lugano criteria to evaluate treatment response, and the results were recorded in Group A. Correspondingly, the end-of-treatment PET/CT was also compared with interim PET/CT to derive a new treatment response evaluation, and the results were recorded in Group B. The treatment response evaluations from Groups A and B were then compared. Statistical analysis was performed to identify patients with inconsistent treatment response evaluations between the two groups. Additionally, it was recorded whether the treatment plan for each patient was altered at the end of first-line therapy based on the treatment response evaluation from Group B.