Safety and efficacy of ferumoxytol for the episodic treatment of iron deficiency anemia in patients with a history of unsatisfactory oral iron therapy: Results of a phase III, open-label, 6-month extension study

only speculate. The second point to which Adams et al. disagree is the prognostic value of PET/ CT-ascertained bone marrow involvement in DLBCL (focal PET/CT lesions), which, as clearly stated in our paper, remains ambiguous. Still Adams et al. feel we discussed the part inappropriately. Khan et al. reported that the outcome of patients with focal PET/ CT lesions in the bone marrow was similar to that of other patients with stage IV disease without bone marrow involvement [10]. Outcome was worse for patients with positive BMB, but these patients often had other high-risk features [10]. Berthet el al found independent prognostic value of PET/CT-ascertained bone marrow involvement in multivariate analysis, while Cerci et al. (not cited in our article) reported adverse outcome in patients with PET/CT-ascertained bone marrow involvement only when BMB was concordantly positive, and vice versa [13,14]. The results of Cerci et al. clearly question the prognostic relevance of searching for bone marrow involvement in PET/CT negative cases [13]. In a retrospective cohort study by Adams’ own group, focally increased FDG uptake in the bone marrow was not prognostic for PFS or OS. However, because of the small number of patients the analysis is likely underpowered and 34/78(44%) of patients had PET/CT-detected bone marrow involvement, which is higher than usually reported [1,10,13–16]. Our study supports an independent prognostic value of PET/CT-ascertained bone marrow involvement, but we do not “claim” prognostic irrelevance of BMB, as accused by Adams et al. The fact that positive BMB was only prognostic in univariate analysis rather reflects the interplay with other important prognostic factors. As in a previous letter, Adams et al. indicate that failure to include imaging-detected bone marrow disease in the NCCN-IPI in contrast to BMB-ascertained bone marrow disease supports prognostic irrelevance of imaging-detected bone marrow disease [3,5]. However, it is worth noting that PET/CT staging was not mandatory in the cohorts that gave rise to the NCCN-IPI [3,17]. At present, we are evaluating the prognostic role of focal PET/CT lesions in further analyses to add further clarity to this ambiguity of PET/CT-ascertained bone marrow lesions in DLBCL. In our discussion, we state that our data may suggest that BMB is not necessary in PET/CT-staged patients, but conclude that more comprehensive exploration is needed. This evokes the third objection by Adams et al. We are aware that some previous studies, including large series and our own data, have shown suboptimal sensitivities of PET/CT to detect bone marrow involvement by BMB [13,16,18]. The Lugano classification recommends BMB in patients without evidence of bone marrow involvement by PET/CT and only if finding of a discordant histology is relevant for patient management [19]. We acknowledge the risk of missing DLBCL and discordant indolent histology in the bone marrow if routine BMB is not performed as part of the routine staging work-up [14,18,20]. However, the guidelines apply a critical view on the added diagnostic value of an invasive diagnostic procedure like BMB. For many patients BMBs are associated with anxiety and pain [21]. Thus, the discussion of PET/CT versus BMB in DLBCL should not be a discussion of sensitivities alone, but need to include all aspects [5]. Staging does not serve its own purpose, but guides treatment decisions and provides baseline information for treatment response assessment. In the current landscape of R-CHOP(-like) therapy for DLBCL, it seems unlikely that indiscriminate use of BMB will alter treatment in a relevant number of patients. Last but not least, future treatment decisions will hopefully be guided by the biological heterogeneity of DLBCL rather than by bone marrow disease detected by BMB (or PET/CT) [22].