Abstract

Aims: To evaluate the diagnostic value of bone marrow biopsy (BMB) with immunohistochemistry studies compared to blood flow cytometry (FC) in response evaluation after treatment in CLL. Patients and methods: 75 previously untreated patients with stage B and C-CLL were treated with oral Fludarabine (30mg/m2/d) + Cyclophosphamide (200mg/m2/d) combination, day 1 through 5, repeated every 28 days for 6 courses. Response evaluation was performed 2 months after the completion of therapy, according to NCI criteria. Sixty patients out of 75 achieved partial or complete remission and underwent BMB. Minimal residual disease (MRD) was simultaneously evaluated on paraffin-embedded BMB by anti-CD5, -CD20 and -CD3 immunostaining and in the peripheral blood using a very high sensitive FC method (Maloum et al, Br J Haematol, 2002, 119:970–5). BMB immunostaining was considered positive for MRD in case of concurrent expression of CD20 and CD5 by the same cells assessed on consecutive sections. Twenty nine patients could have been analyzed so far. The median follow-up was of 49 months [range 19–56].

Results: In all these 29 patients, BMB was considered histologically normal according to the NCI criteria. In 23/29 cases, bone marrow immunostaining did not show expressing CD20 cells. Fifteen out of 23 were in phenotypic remission using blood FC (PhR), of whom only 2 relapsed (at 31 and 37 months). The remaining 8/23 patients did not achieve PhR and all relapsed (at 22, 27, 37, 40, 45 and 51 months). In 6/29 patients, immunochemistry on BMB showed persistent B CD5+ cells : 3 did not achieve PhR of whom 2 relapsed (at 32 and 37 months) while 3 were in complete PhR and displayed a persistent clinical complete remission with a follow-up of 39, 47 and 47 months respectively. Interestingly, in the latter 3 cases, FC analysis showed an expansion of polyclonal BCD5+ cells in blood as assessed, which could explain the presence of such normal BCD5+ cells in the bone marrow. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy (ACC) to predict relapse for BMB and FC are given in the following table:

Sensitivity (%)Specificity (%)PPV (%)NPV (%)ACC (%)
BMB 16.7 76.5 33.3 56.3 51.7 
FC 83.3 94.1 90.9 88.9 89.7 
Sensitivity (%)Specificity (%)PPV (%)NPV (%)ACC (%)
BMB 16.7 76.5 33.3 56.3 51.7 
FC 83.3 94.1 90.9 88.9 89.7 

Conclusion: These results strongly suggest that BMB should not be indicated in the evaluation of treatment response and is not useful for relapse prediction. In contrast, blood flow cytometry constitutes the best method in routine laboratory to predict clinical outcome and should be included in response treatment criteria in CLL.

Author notes

Corresponding author