Clinical judgment enhances interim PET interpretation in Hodgkin lymphoma: a real-world monocentric experience

Background Interim PET after two cycles of ABVD (PET2) is a strong prognostic tool in Hodgkin lymphoma (HL), yet how clinicians use it to guide treatment in routine practice remains largely unexplored. Methods We retrospectively studied consecutive adults with classical HL treated 2010–2023 at one Italian centre. PET2 was assessed semiquantitatively: Deauville score (DS) 4 was assigned when the SUVmax of the hottest residual lesion exceeded the hepatic SUVmax reference. In PET2-positive patients (DS ≥ 4) the clinician decided whether to intensify (high-dose salvage ± autologous stem cell transplantation, ASCT) or continue. The primary endpoint was complete metabolic response (CMR, DS ≤ 3) at the first PET after PET2 (PETpost). Results Of 373 evaluable PET2, 53 (14%) were positive; 51 were analysed, treatment being modified in 28 (55%) and continued in 23 (45%). CMR at PETpost was more frequent in patients who continued ABVD (15/23, 65% vs. 10/28, 36%; p = 0.036). Residual metabolic activity at PET2 was lower in the continuation group (lesion-to-liver Δ SUV 0.9 vs. 3.5, p = 0.002). In exploratory multivariable analysis the decision to intensify was associated with lesion-to-liver Δ SUV (OR 1.99, p = 0.004) and younger age (OR 0.95, p = 0.033), and was much less frequent from 2015 onwards (OR 0.07, p = 0.009); in the model for CMR the decision was the only variable retained (OR 0.30, p = 0.039). After a median follow-up of 7.9 years, overall survival did not differ (5-year 79.9% vs. 89.3%, p = 0.106), nor the 10-year cumulative incidence of lymphoma-related death (12.7% vs. 7.1%, Gray's test p = 0.76), although patients continuing ABVD received a median of one treatment line and avoided ASCT in 87%. Conclusion In a real-world setting without ultra-specialised PET review, clinicians resolved the grey zone of PET2 by weighting quantitative PET parameters alongside patient characteristics and evidence. This heuristic filter identified patients with borderline positivity responding to ABVD and spared them intensification without a detectable excess of lymphoma-related deaths, supporting the integration of semiquantitative PET parameters and clinical judgment into treatment decisions.