We thank Adams and Kwee (1) for their interest in our paper. Imaging of bone marrow infiltration in lymphoma is indeed an interesting and nontrivial topic.
As Adams and Kwee (1) have correctly pointed out, the number of bone marrow involvements as seen on histologic assessment serving as gold standard was relatively low in our cohort of patients (16 of 140 patients; one additional patient had cortical bone involvement; ref. 2). As has already been stated in the article, some sort of selection bias might have existed due to the high number of patients with MALT (mucosa-associated lymphoid tissue) lymphoma included in our series, especially patients with orbital MALT lymphoma. In previous studies, we already demonstrated that in such patients, bone marrow involvement is so rare that the routine use of bone marrow biopsy (BMB) might in fact not be necessary, which has also been suggested in guidelines for MALT lymphoma (EGILS consensus report; ref. 3). For instance, in a series of 140 patients with MALT lymphoma, the percentage bone marrow involvement was 2% (4). The findings in our present study are thus in good accordance with these data, because we observed bone marrow infiltration in only 1 of 38 patients with MALT lymphoma patients. In addition, in the present study, there were 19 patients with “stage 0”—i.e., lymphoma was identified in a very early stage, and no further lesions were observed after surgery (e.g., local lymph node resection; ref. 2).
With regard to the reference standard, unilateral BMB was available in 132 of 140 patients. The remaining 8 patients were diagnosed with either stage IV Hodgkin lymphoma (n = 5) or stage IV DLBCL (n = 3), with extranodal involvement outside of the bone marrow—for these histologies, 18F-FDG-PET/CT can replace BMB, according to the Lugano classification (5). BMB would also not have altered treatment decisions in these patients, as stage IV had already been established.
We have carefully studied the article by Adams and colleagues (6), to determine possible reasons for the differences in results between their study and our own results, and believe that the main reason may be the use of different MRI criteria for bone involvement/bone marrow infiltration:
Adams and colleagues (6) only rated patients with focal bone marrow lesions as positive, whereas those with a diffuse signal increase on T2W or DWI (diffusion-weighted imaging) were rated as negative. We agree with their argument that “diffusely (nonfocal) increased signal intensity at T2W-STIR and DWI may be due to red bone marrow hyperplasia”—this is a well-known pitfall that may sometimes cause diagnostic problems. However, we do not think that simply rating all diffuse bone marrow changes as benign is a good solution, because lymphoma (and other malignancies) can also lead to diffuse infiltration, and thus, diffuse signal changes on MRI (7)—such an approach may therefore decrease the sensitivity of MRI. In our own sample, 4 of 16 patients with BMB-proven bone involvement showed such a diffuse, almost homogeneous pattern on both DWI and T1-weighted MRI—for example, see the MR images and histologic stains of a patient with mantle cell lymphoma and diffuse bone marrow infiltration (URL: http://www.muw.ac.at/radiolehre/Dokumente/BMdiffuse.pdf).
Our strict criterion for diagnosing bone involvement was the presence of both a diffusion restriction on DWI (high signal on the b1000 images and low signal on the ADC (apparent diffusion coefficient) map, in accordance with Taouli and Koh (8), and a low signal on the T1-weighted images (iso- or hypointense relative to skeletal muscle, indicating bone marrow replacement). In contrast, in the study of Adams and colleagues, bone marrow involvement was considered present if both the T1W sequence showed a low signal and the T2W-STIR (short inversion time inversion recovery) sequence (which we did not use in our study) demonstrated a high signal. They also stated that, “In addition, bone marrow infiltration was considered present if DWI showed any focal area with a signal intensity that exceeded that of the surrounding background.”—it is unclear here whether with DWI, they mean the b0 images (which would essentially show the T2 effect), or the b1000 images, or a combination of both—the latter would be necessary to determine whether there was indeed a diffusion restriction, in particular because the authors apparently did not evaluate ADC maps. It is also unclear whether positive on DWI alone was sufficient to diagnose bone marrow infiltration, or whether a low signal on the T1W images was additionally required, like in our own study.
In view of the above, it would be highly interesting to compare the different MR sequence combinations and criteria in future studies, to determine the most meaningful approach.
See the original Letter to the Editor, p. 221
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
This study was supported by funds of the Oesterreichische Nationalbank Anniversary Fund (project number 14587).