Quantitative Deauville Scoring to Uncover Prognostic Information from 18F-FDG PET–Based Response Assessment: Data from the EuroNet-PHL-C1 Trial

Cure rates are excellent in classical Hodgkin lymphoma. The challenge is tailoring the treatment to the individual patient to avoid not only overtreatment and treatment-related sequelae but also undertreatment. 18F-FDG PET is a major tool for guiding response-adapted treatment. Metabolic response is by default assessed using the visual 5-point Deauville scale (vDS), which is, however, limited by its ordinal nature. In this methodologic paper, we investigate whether a quantitative extension of the vDS, namely qPET, allows for better prognostic discrimination. Methods: In total, 1,447 patients with newly diagnosed pediatric Hodgkin lymphoma of all stages from the multicentric EuroNet-PHL-C1 trial received 2 vincristine, etoposide, prednisone, and doxorubicin cycles before they underwent 18F-FDG PET for early response assessment. The qPET value (quotient of SUVpeak of the lymphoma residual and SUVmean of the liver) corresponding to the lesion with the highest determined vDS was retrospectively measured. Logistic regression used qPET as a continuous biomarker for 60 mo of event-free survival (EFS). Results: The individual qPET value strongly relates to the EFS: EFS is high at about 89% for qPET values of less than 1.3, corresponding to the currently applied threshold for complete metabolic remission (vDS, ≤3). Among patients with a vDS of 4 or 5, qPET further discriminates a larger group with still high EFS, that is, 84.3% (1.3 < qPET ≤ 2; refers to vDS4) and 83.1% (2 < qPET ≤ 3; refers to vDS5), from a smaller, high-risk group with unacceptably low EFS at 47.6% (qPET > 3; refers to vDS5). Thus, using metric measures enabled splitting up the vDS5 category into 5A and 5B, showing totally different EFS rates. Conclusion: qPET extends the ordinal vDS and discriminates prognosis continuously over its whole range. The group of vDS5 patients (qPET > 2.0) has a particularly wide prognostic range (EFS, 47.6%-83.1%). Additional information provided by qPET may justify new treatment strategies for patients with high qPET values greater than 3, reveal that standard treatment (chemotherapy plus radiotherapy) is sufficient for patients with qPET values between 1.3 and 3, and confirm favorable outcome when radiotherapy is omitted if qPET is below 1.3.