Does Detection of <20% Lymphoblasts in Patients with CML Herald Imminent Blast Crisis? a Study from BCR::ABL1 Pathology Group

Background

In chronic myeloid leukemia (CML), low-level myeloblasts (<10%) are generally not considered of prognostic significance. However, a common belief is that any level of lymphoblasts (LBs) in peripheral blood or bone marrow suggests an impending lymphoid blast phase (BP) and need to be managed as such. To better understand the clinical significance of low-level LBs in CML, we conducted this multi-institutional study.

Method

Patients diagnosed with CML since 2009 were retrospectively reviewed for the presence of low-level LBs at initial diagnosis or during therapy, by different count and flow cytometry (FC) performed on bone marrow or peripheral blood, and if possible, confirmed by immunohistochemistry (IHC). Low-level LBs were defined as LBs < 20% by different count and IHC regardless of the % by FC, or <2% by FC if IHC confirmation was unavailable.

Results

The study cohort included 4 groups of patients: Group 1 included 30 Western patients with low-level LBs detected at initial diagnosis; Group 2 included 6 Western patients with low-level LBs detected during therapy; Group 3 included 6 Asian patients with low-level LBs detected at initial diagnosis; Group 4 included 4 Western patients with low-level LBs detected at initial diagnosis by FC but were confirmed to be in BP by IHC initially or retrospectively.

Of 30 patients in Group 1, the median LB% was 0.075% (range, 0.003-3.4%) by FC and 2% (range, 1-15%) by IHC (performed in 19 patients). Twenty-four patients had treatment information available, and 2 of them received upfront induction chemotherapy (LB%: 10% and 15% respectively). Both had resolution of LBs and did not progress to BP. The remaining 22 received TKIs upfront. After a median follow-up of 56 months, 4 patients progressed to BP after a median interval of 11 months, including 2 isolated lymphoid BP in cerebrospinal fluid (CSF), 1 medullary lymphoid BP, and 1 medullary myeloid BP with an interval of 10.7, 7.9, 11.4, and 44.4 months, respectively. The LB levels in these 4 patients were 0.16%, 0.03%, 0.02%, and 0.15% by FC, and 5%, 1%, N/A, and 15% by IHC, respectively.

Of 6 patients in Group 2, LBs were detected at a median interval of 68 months (range: 3.2 - 275 months) after initial diagnosis. The median LB% by FC was 0.66% (range, 0.05-1.21%). IHC was performed in all 6 patients and the median LB% was 5% (range, 2-10%). Five patients had treatment information available prior to LB detection, and all received frontline TKI. However, 4 patients had treatment interruption due to intolerance or non-compliance. After the emergence of LBs, all 6 patients continued on TKI and no BP was observed after a median follow-up of 12 months.

Of 6 Asian patients in Group 3, the median LB% by FC was 0.5% (range, 0.1-1.4%) and no IHC was performed. All patients received TKI consistent with the current standard of care. Surprisingly, all 6 patients progressed to lymphoid BP at a median interval of 5.6 months (range, 1.5-12.4 months).

Of 4 patients in Group 4, 2 had low-level LBs by FC (2.45% and 2.10%), but IHC confirmed both cases to be in BP (LB%: 30% and 25%). Both patients received TKI only and were alive at 10 months and 1 month. One patient with low-level LBs (1.9% by FC, 10% by IHC) had concurrent CSF BP, received chemotherapy followed by allogenic transplant and remained alive in complete molecular response at 16 months. One patient had 15% LB along with 5% myeloblasts, and eventually progressed to myeloid BP after 11 months, received chemotherapy and allogenic transplant, and died of GVHD at 23 months (12 month after BP).

Conclusion

Here we present the largest cohort of CML patients with low-level LBs. The majority of Western patients, whether with low-level LBs detected at initial diagnosis or emerging during therapy, exhibited a good initial response to upfront TKI therapy and didn't progress to BP. These findings suggest that CML cases featuring low-level LBs can be effectively and safely treated with TKIs upfront. However, the disproportionately higher incidence of isolated CSF BP warrants special consideration for CSF evaluation and monitoring. TKIs with better CNS activity such as dasatinib might be preferred over imatinib. The underlying reasons for the notable outcome disparity between Western and Asian patients are unclear and warrant further investigation. Furthermore, our study underscores the importance of IHC in cases where low-level LBs are detected by FC, given the significant underestimation of LBs by FC.