Arsenic Trioxide Combined with Mannitol for the Treatment of Central Nervous System Relapse in Acute Promyelocytic Leukemia Patients：a Retrospective Study

Background:

Acute promyelocytic leukemia (APL) has achieved significant advances in curability in recent years; however, central nervous system (CNS) relapse remains a critical concern affecting 3-5% of APL patients. In this context, we propose a treatment regimen incorporating arsenic trioxide (ATO) and intravenous mannitol, demonstrating promising efficacy.

Methods:

From 2013 to 2023, a total of 26 APL patients with CNS relapse were enrolled from Ruijin Hospital and Beizhan Hospital, Shanghai, China. Myeloid blasts were detected by flow cytometry in all patients' cerebrospinal fluid (CSF). The re-induction protocol involved daily administration for 28 days comprising intravenous mannitol and ATO: a 15-minute bolus infusion of 125 mL 20% mannitol followed by a continuous infusion of 125 mL 20% mannitol combined with ATO (0.16 mg/kg/day, maximum 10 mg/day) dissolved in 400 mL normal saline over 10 hours. If achieving complete remission (CR), post-induction protocol was 14 days on and 14 days off for 16 cycles. Additionally, all-trans retinoic acid (ATRA) at 25 mg/m²/day was administered till CR, and then 14 days along with every other cycle of post-induction ATO. For patients with isolated CNS relapse, methotrexate at 3.5 g/m² was given if CR was not reached after 1 course. Intrathecal injections of cytarabine (50 mg), dexamethasone (5 mg), and methotrexate (10 mg) were performed 2-3 times weekly until CSF measurable residual disease (MRD) turned negative, then 1-2 times weekly until obtaining 6 consecutive negative results, followed by monthly administration for 12 months. Patients who failed to achieve 3 consecutive negative CSF MRD or experienced reversion to positive status following initial negativity, were advised to undergo whole cranial and spinal irradiation, followed by autologous hematopoietic stem cell transplantation (HSCT). For patients concurrently experiencing morphological or molecular relapse, Idarubicin at 8 mg/m²/day was administered day 1 to 3. If molecular CR was not reached after 2 courses, allogeneic HSCT was recommended.

Results:

The median age of the 26 patients at diagnosis of CNS relapse was 29.5 years (18-68 years). At the time of initial APL diagnosis, 55% (11/20) of patients were at high risk, 19 patients presented with bleeding symptoms, among whom 4 with cerebral hemorrhage. Upon diagnosis of CNS relapse, 21 patients were in CR1, while 3 in CR2 and 1 in CR3. Initial induction and consolidation therapy predominantly involved ATRA+ATO-based regimens. The median time from APL diagnosis to CNS relapse was 16.86 months (1.6-97.3 months).

During a median follow-up of 27 months (0.3-109.1 months), treatment with this regimen resulted in initial CR from CNS relapse in 92.3% of patients (24/26) within a median of 40.5 days (5-187 days). 79.2% patients (19/24) remained in persistent CR from CNS relapse until the end of follow-up. 4 patients exhibited morphological recurrence, while 68% of patients (17/25) showed molecular recurrence and 32% (8/25) presented with isolated CNS relapse. Additionally, 2 patients had concurrent chloroma. Patients with morphological recurrence (n=4) or received this regimen as salvage therapy following treatment failures with previous therapies for CNS relapse (n= 7) had lower proportion of achieving persistent CR compared to those without morphological recurrence (25% vs 81.8%, P = 0.047) or received this regimen in first line (42.9% vs 84.2%, P = 0.057). In multivariate analysis using binary logistic regression model involving the following 4 factors: morphological recurrence status, first-line use of this regimen, isolated CNS relapse, and combination with chemotherapy, morphological recurrence and not in first-line use were identified as independent risk factors associated with failure to achieving persistent CR (P = 0.026, 0.033, respectively).

3 patients died till the last follow-up, of those 2 underwent allogeneic HSCT. Additionally, 1 patient each underwent allogeneic HSCT and autologous HSCT both achieved long-term survival. The median overall survival (OS) and disease-free survival (DFS) were not reached, while the estimated 5-year OS and DFS were 82.2% and 59.9%.

Conclusion:

The combined therapy of ATO and mannitol has shown robust efficacy even in patients who did not undergo HSCT. We propose this regimen as a first-line treatment for patients diagnosed with CNS relapse, especially without morphological recurrence.

No relevant conflicts of interest to declare.