Ibritumomab tiuxetan plus BEAM in refractory diffuse large B-cell lymphoma

We have read with great interest the results of the recently published prospective analysis of the Grupo Español de Linfomas/Trasplante Autólogo de Médula Ósea (GELTAMO) phase II trial.1 This paper may have an extraordinary impact on decision-making in clinical practice worldwide. Javier Briones and colleagues analyzed treatment efficacy of autologous stem cell transplant (ASCT) after Y-ibritumomab tiuxetan plus BEAM conditioning regimen in relapsed or refractory patients with diffuse large B-cell lymphoma (DLBCL). We fully agree with their statement that Y-ibritumomab tiuxetan is safe and its application leads to high treatment response rates and promising survival in DLBCL patients with a very poor prognosis. However, we are not sure about the treatment response evaluation and survival analyses.

First, this study included 30 evaluable patients who received treatment. Three patients died during or shortly after ASCT and did not undergo the final re-staging procedures three months after ASCT. We understand that treatment response is analyzed on an intention-to-treat basis (i.e. calculated for all 30 patients) but it does not reflect the efficacy of Y-ibritumomab tiuxetan plus BEAM conditioning. It is our opinion that a more detailed analysis of only those patients (n=27) who were evaluable for the response should have been made.

Second, the sum of all treatment responses is 31 (18 complete, 3 partial and 10 progressions), probably due to the inclusion of one primarily progressive, and thus untreated, case. Furthermore, it is not completely clear whether all patients were evaluated for response using PET-CT. Data are shown for only 21 of them despite the statement that treatment response was to be evaluated using PET-based Cheson criteria.2

Third, we have some doubts about the survival curves. Figure 1 shows the overall survival (OS) of all patients (n=30) and Figure 2 progression-free survival (PFS) of all patients. The OS curve drops very quickly compared to the PFS curve. If both curves are superimposed, the PFS curve is always above, or at least overlaps, the OS curve. This is surprising because, by definition, PFS includes all deaths of any cause plus all progressions/relapses.2 From this point of view, the PFS curve should lie below that showing OS, also because of the fact that relapse/progression precedes death-related events.

In conclusion, we believe that results obtained from this trial are very impressive and optimistic. But the above inaccuracies should be clarified to enhance the broad scientific impact of this paper.