Central nervous system relapse of acute promyelocytic leukaemia in a patient with cerebral haemorrhage at diagnosis.

Abstract

Combination therapy of acute promyelocytic leukaemia (APL) with chemotherapy and all-trans retinoic acid (ATRA) significantly reduces the incidence of relapse. The majority of relapses occur in the bone marrow; however, recent reports suggest an increasing incidence of central nervous system (CNS) relapse, which was rarely reported prior to the widespread use of ATRA (Currie et al, 1997; Evans et al, 1997; Anguita et al, 1999). It has been postulated that ATRA increases the expression of adhesion molecules on leukaemic promyelocytes and their ligands on endothelial cells, thereby increasing their migratory capacity and predisposing to extramedullary leukaemia (Evans et al, 1997). We describe a further case of APL relapsing in the CNS and discuss other possible aetiological factors. A 20-year-old man with hypergranular APL and documented t(15;17) presented with hemiparesis. A computerized tomography (CT) brain scan showed intracerebral haemorrhage, and the white blood cell (WBC) count was 2·3 × 109/l. He was treated with ATRA 45 mg/m2/d, and complete morphological and cytogenetic remission was confirmed on d 40. Four cycles of anthracycline-only consolidation were given according to a modified regimen based on the protocol used by the PETHEMA (Programme for the Study and Treatment of Haematological Malignancies) group (Sanz et al, 2000). This comprised daunorubicin 45 mg/m2 for 3 d for two cycles, idarubicin 10 mg/m2 for 3 d, and mitoxantrone 10 mg/m2 for 3 d. After consolidation the bone marrow was negative for the PML-RARα rearrangement using reverse transcription-polymerase chain reaction (RT-PCR) (sensitivity 10−4). He received maintenance therapy with daily 6-mercaptopurine 100 mg, weekly methotrexate 10 mg/m2 and intermittent courses of ATRA 45 mg/m2/d for 15 d every 3 months. Three months later he developed increasing hemiparesis and headache. A CT brain scan showed an old parietal infarct. Numerous hypergranular promyelocytes were seen on cerebrospinal fluid (CSF) examination, consistent with CNS relapse of APL. Bone marrow examination showed complete morphological and molecular remission. ATRA was restarted and he received intrathecal methotrexate alternating with cytosine arabinoside (ara-C), followed by cranial radiotherapy. Seven months later CSF examination is normal and he remains in morphological remission. The GIMEMA (Gruppo Italiano Malattie Emetologiche dell'Adulto) group recently reported a significantly higher proportion of relapses in the CNS when ATRA was added to idarubicin for induction in APL (Specchia, 1999). Although modulation of adhesion molecules by ATRA may promote entry of leukaemic promyelocytes into the CSF, there may be alternative explanations (Evans et al, 1997). Cerebral bleeding may seed the CSF with leukaemic promyelocytes which are then relatively protected from systemic chemotherapy. This may contribute to later CNS relapse in the presence of bone marrow remission. CNS relapse with early marrow relapse 2 months after consolidation has been reported previously in a man with cerebral haemorrhage at first diagnosis (Currie et al, 1997). It is possible that subclinical cerebral bleeding prior to diagnosis seeded the CSF with leukaemia in other reported cases. This is consistent with a recent report demonstrating an increased incidence of CNS relapse in children with acute lymphoblastic leukaemia (ALL) in whom diagnostic lumbar puncture was traumatic (Gajjar et al, 2000). As the prognosis of APL improves, CNS relapses of APL may be increasingly recognized. Anthracycline-only consolidation has not been associated with reduced antileukaemic efficacy (Sanz et al, 2000). However, ara-C may have a role in treating subclinical CNS leukaemia and a similar case of CNS relapse while in morphological and molecular remission has been reported following anthracycline-only consolidation (Anguita et al, 1999). Therefore anthracycline-only consolidation may not eradicate subclinical CNS leukaemia and its wider use may be associated with an increased incidence of CNS relapses. Furthermore, our case suggests that CNS prophylaxis should be considered for patients who first present with CNS bleeding and receive anthracycline-only consolidation. We would like to acknowledge the role of Dr J. Treleaven, Royal Marsden Hospital, in the initial diagnosis and management of this case and thank her for referring the patient to our department.