An Unusual Presentation of Chronic Myelogenous Leukemia: A Review of Isolated Central Nervous System Relapse

NCCN: Continuing Education

Case Report

A 24-year-old otherwise healthy Caucasian woman presented with several weeks of progressive fatigue, bruising, and eventually the development of fevers. She had no prior medical problems and took no medications. Physical examination revealed only moderate splenomegaly, measuring 14 cm. On initial laboratory evaluation, she was found to have a WBC count of 385,000/mm³, a hemoglobin level of 8.2 g/dL, hematocrit 24%, and a platelet count of 1.23 million/mm³. Differential revealed 25% neutrophils, 3% bands, 10% lymphocytes, 0% monocytes, 7% eosinophils, 1% basophils, 7% metamyelocytes, 24% myelocytes, 15% blasts, and 8% promyelocytes. Her lactate dehydrogenase level was 1938 U/L.

A bone marrow biopsy revealed accelerated phase CML. Additional studies included flow cytometry, which was consistent with an increase in myeloid blasts, with approximately 5% of total cells CD34⁺, CD117⁺, CD13 low, CD33⁺, CD15 ^-, CD64^-, CD2^-, CD56^-, and HLA-DR variable. Peripheral blood flow cytometry revealed an increased population of myeloid blasts, with approximately 3.6% of the cells staining positive for CD34, CD117, and CD7. Bone marrow PCR for the BCR/ABL1 fusion transcript revealed a ratio of 3.240 compared with the internal control transcript (ABL). Chromosome analysis of bone marrow revealed 46,XX,t(9;22)(q34;q11.2) in 15 of 15 unstimulated bone marrow cells. Fluorescence in situ hybridization (FISH) analysis using the LSI BCR/ABL1 fusion probe (Abbott Laboratories; Abbott Park, Illinois) revealed that 97% of cells in the bone marrow contained the gene fusion.

Based on the high white count presentation, the patient was admitted to an outside hospital where cytoreduction consisting of hydroxyurea and imatinib at 800 mg/d was initiated. Similar major molecular response rates at 1 year have been reported with imatinib at doses of both 800 and 400 mg/d; however, rates of major molecular response and complete cytogenetic response have been reported to occur earlier with the 800-mg/d dose. The clinical significance of earlier responses on high-dose imatinib has not been definitively established.⁷ The patient tolerated initial cytoreduction with excellent response in counts. Based on initial response to therapy, consideration of upfront allogeneic hematopoietic stem cell transplantation was deferred. The patient was discharged and followed up as an outpatient, where imatinib was decreased to 400 mg/d. Hydroxyurea was discontinued soon after discharge.

Over the next several months, the patient underwent regular follow-up. At 1 month from diagnosis, an excellent response was noted, with WBC counts decreasing to 1600/mm³ with no left-shifted forms and no blasts. A complete hematologic response was obtained by 2 months from diagnosis, with a WBC count of 2600/mm³, hemoglobin level of 12 g/dL, and a platelet count of 224,000/mm³. The differential failed to show a left shift or any immature forms in peripheral blood.

However, 3 months after initiation of therapy, despite a hematologic response, the patient began to develop fatigue. At this time, the patient presented to the emergency department for new headache evaluation. A CT scan of the head was normal. However, because of worsening symptoms, the patient was admitted to the hematology inpatient service for more definitive evaluation and diagnostics.

An MRI of the brain showed increased signal intensity in the periventricular area of the third ventricle on T2 imaging, although artifact could not be ruled out. A diagnostic lumbar puncture revealed the presence of CML. The WBC count in the cerebrospinal fluid (CSF) was 468 cells/mm³. The cytology report described this finding as consistent with CML blast crisis. Flow cytometry performed on the cytospin specimen was found to be entirely consistent with immature blasts in the CSF, which were of myeloid lineage. FISH analysis also revealed that 100% of cells in the CSF contained the BCR/ABL fusion. Cytogenetic results were also consistent with CML (Figure 1).

Treatment was initiated, consisting of therapeutic lumbar puncture with intrathecal cytarabine and hydrocortisone. The patient tolerated this well with fairly immediate resolution of symptoms. Also, given the patient’s CML involvement of the CNS, the patient was changed from imatinib to dasatinib, because it was reasoned that the dasatinib would have better CNS penetration.

Cytospin preparation of the cerebrospinal fluid showing blasts (top left and bottom center; Wright-Giemsa stain, original magnification x100).

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 7; 10.6004/jnccn.2013.0097

View Full Size

Cytospin preparation of the cerebrospinal fluid showing blasts (top left and bottom center; Wright-Giemsa stain, original magnification x100).

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 7; 10.6004/jnccn.2013.0097

Cytospin preparation of the cerebrospinal fluid showing blasts (top left and bottom center; Wright-Giemsa stain, original magnification x100).

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 7; 10.6004/jnccn.2013.0097

As part of the patient’s restaging, a bone marrow biopsy with aspirate was performed, which was morphologically described as showing no evidence of disease. Flow cytometry likewise showed no evidence of disease. FISH analysis of the bone marrow failed to show any BCR/ABL fusion. Karyotype analysis revealed that 1 of 20 cells was 46,XY,t(9;22), Philadelphia chromosome-positive (Ph⁺), despite negative FISH results noted earlier. Peripheral blood PCR for the BCR/ABL fusion transcript revealed a ratio of 0.02 compared with an internal ABL control transcript. Mutational analysis of the TKI domain was performed on the bone marrow and failed to show any significant or known mutations of the TKI-binding domain. One week later, the patient received intrathecal methotrexate and hydrocortisone. CSF from this procedure revealed that the WBC count in the CSF had decreased from more than 460 to 99 white blood cells, although 39% of these still were blasts. The patient was subsequently discharged after placement of a Holter-Rickham shunt. The patient subsequently received scheduled intrathecal chemotherapy, primarily with liposomal cytarabine (Table 1).

Based on subsequent clearing of the CSF with initial intrathecal liposomal cytarabine, the patient was scheduled to receive monthly liposomal cytarabine for a year as part of her ongoing therapy.

The patient was thereafter maintained on 140 mg/d of dasatinib. She tolerated dasatinib well with the exception of a presentation to the emergency room complaining of new-onset chest pain and shortness of breath, which resolved with ibuprofen. Evaluation did not show fluid retention in the form of edema, pleural effusion, or pericardial effusion.

Subsequently, the patient was referred to the bone marrow transplant program to discuss whether an allogeneic transplantation under these circumstances would be a reasonable option compared with long-duration therapy using intrathecal therapy, plus or minus whole-brain radiation, for her unusual CNS involvement with CML. The patient’s care plan included once-monthly intrathecal cytarabine for 1 year. The patient was referred for allogeneic stem cell transplant from an unrelated donor 10 months after diagnosis. The preparative regimen included total body irradiation along with etoposide and antithymocyte globulin, with tacrolimus and methotrexate for graft-versus-host disease prophylaxis. Plans were made to reinitiate TKI therapy 100 days after stem cell transplant.

Table 1

Representative Response to Intrathecal Therapy as Measured in the Cerebrospinal Fluid While on Dasatinib

View Table

Discussion

Despite achieving an initial hematologic response to imatinib at the 3 month milestone, this patient was found to have significant CNS disease at that time, consistent with CNS relapse. She was then treated with intrathecal chemotherapy using liposomal cytarabine, a therapy shown to provide good local disease control, specifically in patients previously on imatinib.⁸ The imatinib was changed to dasatinib, a second-generation TKI with demonstrated efficacy,^9,10 based on dasatinib’s superior CNS penetration.¹¹

CNS involvement is not typical for CML, although it has been well described.¹² Seeding of the CNS by leukemia cells is seen more often with high WBC counts at presentation.¹³ In this patient, any leukemia cells present in the CNS could have eluded initial imatinib therapy because of the relatively poor CNS penetration of imatinib itself.¹⁴

CML in blast crisis has been shown to have a propensity for CNS involvement.¹⁵ Although also seen in Ph⁺ acute lymphoblastic leukemia, an isolated CNS presentation of CML in chronic phase while on adequate TKI therapy has been less well described.¹⁶ Standard therapies for CNS leukemia often result in significant toxicities and nonsustained responses. These interventions include high-dose systemic therapy, either with methotrexate or cytarabine; intrathecal chemotherapy with the same agents; and radiation to the craniospinal axis.⁸ With the advent of TKIs, with imatinib as the prototype for BCR/ABL tyrosine kinase inhibition, efforts have been made to incorporate them into the treatment strategies of patients with Ph⁺ leukemia with CNS disease.¹⁷ CNS relapses have been documented in approximately 20% of patients with CML in blast crisis or Ph⁺ ALL already treated with imatinib.^17,18 As in this patient, these cases often showed complete responses in the bone marrow and peripheral blood before relapse. This finding bolsters the case for the CNS as a sanctuary site for CML on treatment with imatinib.^19,20

Imatinib has been shown to produce a complete cytogenetic response in approximately 90% of patients in chronic phase CML,³ and is currently included in the NCCN Guidelines for CML along with the second-generation TKIs as an option for first-line therapy²¹ (to view the most recent version of these guidelines, visit NCCN.org). Porkka et al¹¹ showed the inability of imatinib to prevent CNS relapses, resulting from poor drug penetration across the blood-brain barrier. Using a preclinical mouse model, treatment with imatinib and the second-generation TKI dasatinib was compared in intracranial Ph⁺ leukemia. In this experiment, imatinib failed to inhibit intracranial tumor growth, whereas treatment with dasatinib induced clinically meaningful responses and resulted in increased survival. CNS disease showed both stabilization and regression on dasatinib. These results translated into a small human trial of 11 patients with Ph⁺ ALL or blast crisis CML with CNS involvement. On diagnosis of CSF disease, treatment was initiated with oral dasatinib at 140 mg or 70 mg twice daily, with dose escalation in patients showing lack of response. A lumbar puncture was performed every 1 to 4 weeks to assess patient response via quantitative PCR analysis of the BCR/ABL transcript. All evaluated patients showed a response. In addition, dasatinib was shown to have significantly greater CNS penetration than imatinib, as noted by pharmacokinetic measurement and plasma and CSF.¹¹

Conclusions

The possibility of isolated CNS disease in patients with CML previously treated with imatinib should be considered. This may be more important in patients who initially present with hyperleukocytosis, particularly those in blast crisis. As in this patient, isolated CNS relapse may present even in those who have shown a complete hematologic and at least a major cytogenetic response to imatinib therapy. This patient was noted to be in blast crisis, with blasts of myeloid lineage present in the CNS. Aggressive strategies, such as intrathecal chemotherapy, change of TKI to one with increased CNS penetration, and consideration of allogenic stem cell transplantation, are potential therapeutic modalities. CNS prophylaxis in patients deemed at-risk requires further study.