Abstract 2860

Introduction.

Histological lymphoma diagnosis of adult patients with bulky mediastinal mass is either HL or PMBL, two distinct diseases with a specific outcome. PET interpretation is often difficult for mediastinal lymphoma because of the volume of the tumor and the presence of blood pool. No specific criteria for interpretation of interim-PET have yet been defined in this clinical entity. The purpose of this study was to investigate the prognostic value of qualitative and semiquantitative evaluations of interim-PET in mediastinal lymphoma.

Methods.

We retrospectively included 72 patients with either HL (n=48) or PMBL (n=24), previously untreated, aged under 60 at diagnosis and who underwent at least one interim-PET evaluation. Patients with sub-diaphragmatic or medullar localisations of lymphoma were excluded. All PET scans were reviewed. Qualitative evaluation included global visual evaluation (GVE) (positive or negative) and 5 points-scale (5PS). Semiquantitative evaluation consisted in maximum standardized uptake value (SUVmax) and SUVmax reduction between the baseline PET (PET0) and the evaluation performed after two (PET2) or four (PET4) cycles of chemotherapy. Prognostic impact was evaluated on the event-free survival (EFS), defined as disease progression/first relapse (n=18) or death (n=2). ROC (Receiver Operating Characteristic) curve was used to assess the value of SUVmax reduction in discriminating future deaths or relapses, based on area under the curve (AUC). The “best“ cut-off that provides both the lowest false positive and the lowest false negative rates was computed.

Results.

Median age was 29 (24 - 35), 60% male. Tumoral mass was more than 7.5 cm in 70% of the patients, and M/T ratio>= 0.35 in 79%. Median SUVmax at baseline, PET2 and PET4 were 12.8 (4.1;33.2), 1.9 (1.7;3.1), and 2.4 (1.7;3.1), respectively. With a median follow-up at 24 months, 2-year event-free survival (EFS) was 67%, without significant difference between HL and PMBL (p=0.98). Except for ECOG performance status in PMBL, neither clinical nor biological feature was predictive for EFS. Using GVE, a negative PET2 (n=36/59) and PET4 (n=24/34), were achieved in 61% and 71%, respectively. GVE and 5PS have a significant negative predictive value (NPV) at PET2 (HR=3.2, 95%CI: 1.2–8.2; p=0.012, and HR=1.9, 95%CI: 1.1–3.3, p=0.01) and at PET4 (HR=13.9, 95%CI: 3.5–55; p=0.0001, and 2.6, 95%CI: 1.5–4.7, p=0.001 respectively), with a 2-year estimated EFS of 69% for PET2-negative patients versus 51% for PET2-positive patients (p=0.012), and of 86% for PET4-negative patients versus 20% for PET4-positive patients (p=0.001). An optimal cut-off of 81% SUVmax reduction from PET0 to PET2 or PET4 yielded a 2-year estimated EFS of 70% in patients with reduction of more than 81%, versus 47% in those with reduction of 81% or less (p=0.004). In the HL subgroup, GVE and SUVmax reduction higher than 81% at PET2 (p=0.0001, p=0.015, respectively) and PET4 (p=0.004, p=0.015, respectively) showed significant prognostic values for EFS. In the PMBL subgroup, neither qualitative nor semiquantitative evaluation at PET2 was predictive for EFS. At PET4, GVE and 5PS were strongly predictive for EFS (p=0.005, p=0.0001). Patients with SUVmax reduction higher than 81% between PET0 and PET4 reached a 67% 2-year EFS, while patients with lower SUVmax reduction had only a 33% 2-year EFS (p=0.13).

Conclusion.

The cut-off value of SUVmax reduction estimated for predicting EFS with best accuracy in mediastinal lymphoma was 81%. Although the SUV semiquantification helps to reduce the number of false positives, visual global analysis had a significant negative predictive value in interim-PET in mediastinal lymphoma, with a possible very early prediction, as early as PET2 in HL, and a better prediction at PET4 than PET2 in PMBL.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution