Aim: Primitive mediastinal B-cell lymphoma (PMBCL) is a relatively rare form of non-Hodgkin lymphoma (NHL), typically concerning the youngster, with an aggressive course and poor prognosis. The therapy generally consists of high dose chemotherapy followed by radiotherapy. PET-CT is used at staging, restaging after chemotherapy and after radiotherapy, or when relapse is suspected. Aim of the study was to compare different criteria in the evaluation of response to chemotherapy in this setting.
Methods: Thirty-eight patients with PMBCL (15 M, 23 F, median age 33 yrs [range 18-79]), all treated with chemo-immunotherapy and radiotherapy, who had undergone baseline (b-PET) and end of chemotherapy (f-CHT-PET) 18F-FDG-PET-CT scans at our institution between July 2004 and September 2014 were retrospectively re-evaluated; the median follow-up was 42 months (range 4-109), at which 4/38 (11%) had died, 5/38 (13%) were in partial response (PR) and 29/38 (76%) were in complete response (CR). The primary endpoint was progression-free survival (PFS), while the secondary one was overall survival (OS), according to the Cheson criteria. SUV max of the mediastinal disease mass at staging, of the residual mass at CT after chemo-immunotherapy, SUV max of the liver and of the mediastinal blood pool (MBP) were calculated for all patients.
Results: In our population, we observed that: 1) visual criteria performs better when positivity-negativity threshold is set at point 3 of the 5-point scale (5-PS); 2) semiquantitative approach by use of Δ SUV max performs better when the threshold is set at 66% decrease: in fact, at Δ SUV max analysis with 66% decrease, 9 patients resulted positive at the test (Δ SUV max ≤66%), 29 negative (Δ SUV max >66%).
Conclusion: In our population Δ SUV max could be working well in these patients because the baseline values are very high and very homogeneous. Our data, though limited in numerosity of patients and events, suggests that in this particular setting the use of the 5-PS reporting system could not be the best tool available; on the other hand, Δ SUV max could prove to be reliable in the evaluation of response to chemotherapy.