Successful Molecular Targeted Treatment of AML in Pregnancy with Azacitidine and Sorafenib with No Adverse Fetal Outcomes

Acute myeloid leukaemia (AML) presenting in pregnancy is a rare event. The dilemma faced by the mother, her family and the medical team encompasses a number of difficult and often confounding issues. Here we report the first case of successful molecularly targeted treatment using the hypomethylating agent azacitidine with the multi-kinase inhibitor sorafenib in a pregnant woman with refractory AML resulting in no major foetal morbidity or mortality. A 39-year-old women presented in the 9th week of her first pregnancy with lethargy, easy bruising, circulating blasts and a markedly abnormal haematological profile (haemoglobin 51 g/l, white cell count 2·2 × 109/l, neutrophils count 1 × 109/l, platelet count 34 × 109/l). Bone marrow examination confirmed high-risk myelodysplastic syndrome (refractory anaemia with excess blasts type 2) with normal karyotype and no detectable molecular abnormality by reverse transcription polymerase chain reaction. The patient opted to continue with the pregnancy and close observation. At 22 weeks gestation, a progressive rise in white cell count was accompanied with progression to frank AML (69% myeloid blasts) and the emergence of a FLT3 internal tandem (FLT3-ITD) mutation (17·3%) in the absence of other gross cytogenetic abnormality. Despite standard induction DA (daunorubicin, cytarabine) and FLAG (Fludarabine, Cytarabine, G-CSF) salvage chemotherapy in combination with sorafenib, by 29 weeks gestation a significant blast population remained (30% myeloid blasts). At this stage of pregnancy, fetal monitoring by an ultrasound scan showed a viable fetus with no structural abnormalities. The patient was then given sorafenib (400 mg orally twice daily, off label use) and azacitidine (75 mg/m2 for 7 days, sub-cutanaeously) at 30 weeks gestation. By day 6 of therapy, her neutrophil count started to increment from a pre-therapy baseline of 0·1 × 109/l to 1·4 × 109/l. At 32 weeks gestation, the patient underwent an elective Caesarean section with blood product support and prior corticosteroid administration. A healthy male neonate was delivered with birth weight just above the age-adjusted 10th centile. Newborn clinical examination was unremarkable. Spontaneous respirations were established within one minute, and he was supported by facemask continuous positive airways pressure (CPAP). He was admitted to the neonatal unit and supported on CPAP for 2 days, after which he did not require any further respiratory support. Full enteral feeds were established by day 3. He had a short course of antibiotics for risk of sepsis, but blood cultures were negative after 5 days and these were discontinued. Phototherapy for neonatal jaundice on day 1–2 of life was also administered. Blood counts, an echocardiogram and serial cranial ultrasound scans were noted within the normal range for age at birth and throughout the admission. No clinically significant imbalance was reported in his array Comparative Genomic Hybridization results, along with normal chromosomes. The child was discharged home on day 21 of life fully enterally fed by bottle, at a corrected gestational age of 35+3 weeks. Breast-feeding was not advised due to concurrent azacitidine and sorafenib treatment. Outpatient review noted no concerns at up to a corrected age of 10 months, maintaining developmental milestones appropriately. A maternal bone marrow aspirate following one cycle (28 days) of azacitidine in combination with sorafenib showed 4% blasts by morphology (7% by flow cytometry with original disease phenotype) and undetectable FLT3-ITD. She received 4 cycles of azacitidine and sorafenib resulting in transfusion independence and near normalisation of her neutrophil count (haemoglobin 113 g/l, white cell count 2·6 × 109/l, neutrophils 1·7 × 109/l, platelets 47 × 109/l). She proceeded to a matched unrelated donor allogeneic stem cell transplant (alloSCT) using busulphan-cyclophosphamide conditioning. No evidence of FLT3-ITD was found throughout her treatment with azacitidine and sorafenib. One month post-alloSCT, there was an ongoing morphological remission with >95% chimerism and no FLT3-ITD detected. Detectable FLT3-ITD at 6 months post-alloSCT has led to discussion regarding donor lymphocyte infusion. Experience suggests that past the first trimester, cytotoxic chemotherapy for AML can be safely given with stringent fetal monitoring (Milojkovic & Apperley, 2014; Ali et al, 2015). A recent study (Amant et al, 2015) suggested that paediatric developmental outcomes were not inferior in women exposed to a heterogenous group of chemotherapy regimens after the second trimester compared to normal controls. The combination of azacitidine and sorafenib demonstrated activity in a phase 2 study (Ravandi et al, 2013) of relapsed, refractory FLT3-ITD+ AML. The authors postulated that synergistic FLT3 inhibition by azacitidine and sorafenib facilitates leukaemic cell differentiation and subsequent apoptosis by overcoming C/EBPa function through ERK1/2-mediated phosphorylation (Ravandi et al, 2013). A smaller study of 9 relapsed, refractory FLT3-ITD+ AML patients successfully induced remission in all subjects with single agent sorafenib (Gill et al, 2015). Consolidation with azacitidine was then given (100 mg/day × 4 days every 4 weeks). Despite a reduction in FLT3-ITD allele burden with dual therapy, overall survival was not significantly different from historical controls. Robust human data is lacking for the use of azacitidine and sorafenib in pregnancy in combination or used alone. Azacitidine in mice and rats produced a dose-dependent decrease in offspring survival, fetal weight with an increase incidence of microphthalmia and exencephaly (Cummings, 1994) that was not replicated when azacitidine was given prior to day 4 of pregnancy. Sorafenib caused similar effects in rats and rabbits but at doses considerably below the recommended human dose (http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_Product_Information/human/000690/WC500027704.pdf). Predicting outcome from animal models is, however, difficult; dose comparisons are inaccurate with inter-species differences in drug metabolism likely. Combination therapies may further complicate this issue. There is limited real world experience that reflects the choice of therapy in our patient. A small case series of pregnant women with chronic myeloid leukaemia who continued tyrosine kinase inhibition reported no major adverse paediatric outcomes (Zhou et al, 2013). AML in pregnancy is immensely challenging although targeted molecular therapy is a viable option, especially in refractory cases. No negative fetal outcomes have been noted after early follow-up. The safety and efficacy of azacitidine and sorafenib earlier in pregnancy however is still unknown. The authors would like to acknowledge the advice of Dr Mark Levis in his help with the patient's clinical management. AJM and GD wrote the first draft of the manuscript. CA, CM, RA, CC, KMW and RR jointly managed the patient. CA obtained consent from the patient. CA, CM, RR and KWM critically reviewed the manuscript.