Answers are grounded in this paper and cite their sources.
Central nervous system (CNS) involvement at diagnosis for pediatric patients with acute myeloid leukemia (AML) occurs in 10–30% of patients.1-4 Despite this high frequency, controversy still exists regarding the clinical management of such patients. In pediatric acute lymphoblastic leukemia (ALL), diagnosis of CNS-involvement and subsequent therapy have been refined over the last decades, discerning CNS2 (morphologic blasts on spinal fluid cytospins without elevated cell count), traumatic lumbar puncture with blasts (TLP+) status, and implementing additional intrathecal chemotherapy for such patients as well as recommendations to avoid TLP (such as performing the diagnostic lumbar puncture [LP] under anesthesia), although not all studies in pediatric ALL confirmed the negative prognostic impact of CNS2 status.5-8 Studies in pediatric AML are scarce, and as a result the recent expert treatment recommendations for pediatric AML mainly focus on providing intensified therapy to CNS3 patients but do not provide recommendations for CNS2 or TLP+ status.9 Moreover, previous studies have had conflicting results regarding the effect of CNS disease on outcomes.1, 2, 10-13 CNS-directed therapy may include intrathecal chemotherapy, radiation, and components of systemic chemotherapy, such as high dose cytarabine and etoposide. Traditionally, patients with CNS disease at diagnosis (CNS3), as defined by >5 white blood cells (WBC) with blasts on spinal fluid cytospins or with clinical or radiographic signs of CNS leukemia, receive more intensive CNS-directed treatment (either increased intrathecal chemotherapy and/or radiation) than those without evidence of CNS disease (CNS1). In this issue of Pediatric Blood & Cancer, cooperative groups throughout the United States and Europe report on presenting features and outcomes of pediatric AML patients with CNS disease at diagnosis in the current treatment era. Patient characteristics from studies in this issue by Johnston et al., Creutzig et al, and Stove et al. confirm several of the features that have been previously associated with CNS disease at diagnosis (Table 1).3, 4, 14 Some of these characteristics are related, in particular that of inversion (16), which is a low-risk cytogenetic group that is associated with French-American-British classification (FAB) M4 morphologic phenotype and often presents with high tumor burden. Ranta and colleagues reviewed presenting features and imaging in 33 pediatric AML patients with CNS involvement over a 12-year time frame and found that 58% of patients had clinical symptoms related to the CNS disease including exophthalmos, vomiting, cranial nerve palsy, headaches, convulsions, paresis, and irritability.15 Interestingly, upon imaging that was available in 22 cases, seven children had what were considered incidental findings of mastoid bone opacification without any intracranial abnormalities in the report. While two patients had clinical signs of sinonasal inflammatory disease, the other five patients lacked these symptoms but did have facial palsy. The facial nerve contains a mastoid segment that runs in the medial wall of the mastoid antrum before exiting the cranium that can rarely result in facial paralysis in the setting of mastoiditis due to swelling around the nerve. In the absence of signs of sinonasal inflammatory disease, the presence of mastoid bone opacification with facial palsy in these patients suggests leukemic infiltrate into the mastoid and facial nerve. In these cases, in addition to resolution of facial palsy, clinicians may consider monitoring response by repeat imaging. Moreover, this finding suggests that further studies using imaging in pediatric AML as a routine measure in the diagnostic work-up may be warranted. Low-risk cytogenetic/molecular risk group2, 3 FAB, French–American–British classification. The prognostic significance of CNS disease at diagnosis in pediatric AML is uncertain, with reports in the literature of no difference in survival as well as an increase in CNS relapse.1, 2, 16-18 Johnston et al., Creutzig et al, and Stove et al. revisit the issue with analyses of patients treated in the modern era (Table 2).3, 4, 14 In all studies, patients received four to five courses of chemotherapy with two to three blocks that contained high-dose cytarabine. Intrathecal therapy for patients with CNS3 disease at diagnosis consisted of intrathecal “triples” (methotrexate, hydrocortisone, and cytarabine) for all studies, with the exception of the Children's Oncology Group (COG) and Berlin–Frankfurt–Münster (BFM) group that administered intrathecal cytarabine. Cranial radiation was limited to patients treated in the BFM protocols and Dutch Childhood Oncology Group patients greater than 2 years of age included in the MRC12 protocol. In the COG study, patients with CNS2 status also received additional intrathecal therapy, and 4.2% of patients failed to clear cerebrospinal fluid at end of induction I, versus 9.3% of CNS3 patients. For patients without CNS disease at diagnosis, treatment outcomes across all three studies including event-free survival, overall survival (OS), and CNS relapse rates were comparable. Among patients with CNS3 disease at diagnosis, however, all three studies demonstrated an increase in CNS relapse, although OS was not affected suggesting that these patients were salvageable, which raises the question whether control was obtained with stem cell transplantation performed in second complete remission, and whether such patients were irradiated as part of the conditioning regimen. In the COG study, CNS2 patients also had higher rates of CNS relapse when compared to CNS1 patients. Also consistent among all three studies was the absence of an increased incidence of CNS relapse in patients that had a traumatic tap, in contrast to pediatric ALL. In all studies, patients with traumatic taps were treated as CNS positive if blasts were present and in two studies if the WBC/RBC ratio exceeded a defined threshold. The consistency of an increased risk of CNS relapse for patients with CNS3 disease at diagnosis in all three studies and CNS2 disease in the COG report suggests that CNS control needs to be optimized in pediatric AML. The highest incidence of CNS relapse among these reports was seen in CNS3 patients treated on the AAML03P1 and AAML0531 protocols that received seven to nine doses of intrathecal cytarabine.3 A prior retrospective analysis from St. Jude Children's Research Hospital saw a decrease in CNS relapse over successive studies when intrathecal triples were substituted for intrathecal methotrexate.2 Whether a similar improvement would be seen if triples were substituted for cytarabine is uncertain, as cytarabine is felt to be more active in AML when compared to methotrexate. Regardless, both European reports have a higher incidence of CNS relapse in CNS positive patients with the use of triple intrathecals, suggesting that additional measures should be considered. Cranial radiation increases the risk of late toxicities and secondary malignancies and therefore is no longer used by the majority of cooperative groups.11 Optimization of the number of intrathecal triples along with systemic therapy may potentially decrease the risk of CNS relapse further, but requires a large randomized study to answer this question definitively. Further studies are warranted evaluating the use of liposomal cytarabine, ventricular administration of chemotherapy to improve the ventricular dose, or the use of thiotepa, as recently reviewed by Pui et al.19 The authors declare that there is no conflict of interest.
No takes yet. Share an insight, caveat, or question.
Gruber et al. (2017) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: