Central Nervous System Relapse in Adults with Acute Lymphoblastic Leukemia: Identifying Key Risk Factors and Survival Outcomes

Background

Despite significant advances in the treatment of acute lymphoblastic leukemia (ALL), central nervous system (CNS) relapses remain a significant challenge. Many novel treatment strategies, such as immunotherapy, have not included patients with CNS involvement. The optimal management of CNS relapse in adults with ALL remains uncertain, as the evidence is limited and primarily based on pediatric studies. Our primary objective was to identify risk factors for CNS relapses, and as a secondary goal, to analyze the outcomes of adults with ALL who experience CNS relapses.

Methods

This case-cohort study from a tertiary care center in Mexico City included all patients diagnosed with ALL between 2009 and 2023. We retrospectively collected clinical and laboratory variables at diagnosis and the time of the first relapse. For outcomes, we separately analyzed isolated CNS (iCNS) and concomitant CNS + bone marrow (CNS+BM) relapses.

Results

We included 201 patients with a median age of 35 years (range, 16-79), and 48% being female. The median body mass index (BMI) at diagnosis was 25.8 kg/m² (range, 13.8-41). The median white blood cell (WBC) count was 13.1 x 10³/μL (range, 0.4-488), and 27.9% had ≥50 x 10³/μL WBC. Of the cohort, 98% had B-cell ALL, and 35% presented with high-risk cytogenetics, including 15% with the Philadelphia chromosome.

Regarding baseline cerebrospinal fluid (CSF) evaluation, 4.5% were classified as CNS-2 or CNS-3. In terms of treatment, 45% of the patients received the Hyper-CVAD regimen, and 35% received a pediatric inspired regimen (mCALGB 10403). During induction, 78.6% received dexamethasone and 21.4% prednisone.

The overall relapse rate was 48.3%, with 12.4% having CNS relapse (56% being iCNS relapses). The median time to CNS relapse was 13.8 months (range, 0.26-55), compared to 10.0 months (range, 0.26-68.6) for systemic relapses.

Risk factors for CNS relapse were WBC ≥50 x 10³/μL and T-cell ALL: 21.4% vs. 9.0% (OR 2.77, 95% CI [1.18-6.52]) and 50% vs. 11.7% (OR 7.56, 95% CI [1.02-56.33]), respectively. Patients with a BMI ≥ 25 kg/m² had a lower probability of CNS relapse (7.5% vs. 18.1%, OR 0.37, 95% CI [0.15-0.89]).

Traumatic lumbar puncture (LP) and the presence of blasts in peripheral blood at the time of the first LP were not associated with CNS relapse.

At the time of CNS relapse, 56% of patients exhibited symptoms, while 44% were asymptomatic, with relapse detected through routine CSF analysis during LPs. Among symptomatic patients, the presentations included headaches (48%), seizures (12%), cranial nerve involvement (8%), and radicular symptoms (4%). CNS imaging was performed in 41.7% of cases, with 70% undergoing MRI, 20% CT scans, and 10% PET-CT scans; notably, 30% of these imaging results were normal.

Treatment for CNS relapse included intrathecal chemotherapy (ITCh) in 88%, with a median of 4 ITCh during the first month (0-7) and a median CSF clearance by the third LP (0-10). Systemic chemotherapy was administered in 80% of cases, with high-dose methotrexate in 61.9% and high-dose cytarabine in 50%. Radiation therapy was used in 52.2% of cases, with a median dose of 24 Gy (range, 20-41). 20% of patients underwent allogeneic cell transplantation (alloHCT). There were no significant differences in treatment/transplantation patterns between patients with iCNS and those with CNS+BM.

A survival analysis from the time of relapse showed that iCNS relapses had significantly better overall survival (OS) compared to other types of relapses, with a median OS of 10.55 vs. 6.44 months and a 2-year OS of 50.0% vs. 14.3% (p=0.023). Asymptomatic relapses were associated with a non-significant better OS than symptomatic relapses (2-year OS 66.3% vs. 27.5%, p=0.209).

Prognostic factors associated with a better OS after relapse included iCNS relapse (HR 0.33, 95% CI [0.120-0.904], p=0.031) and alloHCT (HR 0.539, 95% CI [0.298-0.976], p=0.041).

Conclusions

In a large cohort of ALL patients, we observed a 12.4% incidence of CNS relapse with WBC counts >50x10³ and T-cell ALL associated with a higher risk, whereas having a BMI >25 was a protective factor. Notably, iCNS relapses are associated with better outcomes. Although there are few cases, there is a trend toward better outcomes when CNS infiltration is identified without neurological symptoms. This underscores the importance of vigilant monitoring and tailored therapeutic strategies to improve outcomes in this high-risk population.