GRAALL-2005 study: Impact of CNS involvement on outcomes in adult patients with Philadelphia chromosome-negative ALL

A minority (4–11%) of adult patients with acute lymphoblastic leukemia (ALL) have central nervous system (CNS) involvement at the time of diagnosis.¹ CNS involvement has been associated with lower overall survival (OS) rates in the MRC UKALL XII/ECOG E2993 trial,² and involvement of additional extramedullary sites, T-cell ALL, and a higher white blood cell (WBC) count at diagnosis are more common in this subset of patients.¹ Since the introduction of pediatric-inspired regimens, the outlook for adult patients with Philadelphia chromosome-negative ALL has significantly improved; however, the effects of initial CNS involvement had not previously been reexamined.¹ The ALL Hub previously reported on CNS involvement at diagnosis and its association with CNS toxicities in childhood ALL.

Here, we summarize an article recently published in Haematologica by Orvain et al.¹,  which investigated the impact of CNS involvement in adult patients with Philadelphia chromosome-negative ALL in the GRAALL-2005 study.

Study design and methods

Patients included in the analysis were:

• aged 18–59 years;
• diagnosed with Philadelphia chromosome-negative ALL between 2006 and 2014;
• treated according to the GRAALL-2005 pediatric-inspired protocol; and
• had diagnostic data on CNS involvement by cerebrospinal fluid.

CNS involvement was categorized as:

• CNS-2: <5 WBC/μl and a positive conventional cytospin; or
• CNS-3: >5 WBC/μl and a positive conventional cytospin and/or clinical signs.

Measurable residual disease (MRD) was monitored by standardized quantitative real-time polymerase chain reaction with a sensitivity of ≥10⁻⁴, which was performed after the first induction course.

The aim of the study was to evaluate the impact of CNS involvement on survival outcomes.

Results

Patient characteristics

• Among the 784 adult patients included, 55 had initial CNS involvement, with 47 classified as CNS-3 and 7 as CNS-2 (Figure 1).
• CNS-positive (CNS⁺) patients were more likely to have T-ALL, leukocytes ≥30 G/l, and higher hemoglobin levels.
• Among patients with clinical symptoms, trigeminal anesthesia was the most prominent clinical characteristic, which was reported in 41%.
• CNS-negative (CNS−) and CNS+ patients showed similar rates of first CR, induction death, and achievement of MRD negativity after induction (Ig/TCR <10⁻⁴). Further baseline characteristics are summarized in Table 1.

Table 1. Characteristics of patients with or without CNS involvement*

Association between CNS involvement and patient outcomes

• Median OS was shorter in CNS+ patients compared with CNS− patients (1.9 years vs not reached; p < 0.001).
• CNS involvement at diagnosis was not associated with a statistically different cumulative incidence of relapse (p = 0.11), but it was associated with significantly higher non-relapse mortality (NRM; p = 0.01). Causes of NRM were similar between CNS+ and CNS− patients, and included infections, transplant-related, thrombosis, bleeding, second cancer, and others.
• The rate of 3-year relapse was higher in CNS+ patients compared with CNS− patients (34% vs 26%, respectively).
• Following censoring patients at allogeneic stem cell transplantation (allo-SCT), the cumulative incidence of relapse was significantly higher in CNS+ patients (p = 0.01), whereas NRM was not different between the two groups (p = 0.4)

Factors associated with survival

As shown in Table 2, univariable Cox regression analysis revealed a positive association between lower OS and age (p < 0.001), body mass index (p < 0.001), leukocytes ≥30 G/L (p = 0.1), CNS involvement (p = 0.001), and poor early BM blast clearance (p = 0.02).

In addition, the multivariable Cox regression model showed a significant association between CNS involvement at diagnosis and lower OS (p <0.001). The only other key variable factors associated with lower OS were age (p < 0.001) and leukocytes ≥30 G/L (p = 0.03; Table 3).

Table 2. Univariable analysis of factors associated with NRM, relapse, and overall survival by Cox regression*

Table 3. Multivariable analysis of factors associated with NRM, relapse, and overall survival by Cox regression*

Post-remission treatment for patients with CNS involvement

• CNS+ patients were more likely to receive allo-SCT compared with CNS− patients (55% vs 34%; p = 0.002), with a median interval before proceeding to allo-SCT of 170 days vs 142 days, respectively (p = 0.003). A total of 25 CNS+ patients did not undergo allo-SCT, including six who died before allo-SCT.

• Toxicity was higher in patients randomized to the high-dose cyclophosphamide arm and in those who received allo-SCT
• Exploratory landmark analyses did not show any association between either cranial irradiation (at 160 days) or allo-SCT (at 170 days) and outcome.

Conclusion

This study demonstrated that CNS involvement is significantly associated with poor outcomes in patients with newly diagnosed Philadelphia chromosome-negative ALL. This association was mainly due to excessive toxicity with no benefit from allo-SCT in this patient population. As this condition occurs in a minority of adult patients with ALL, it is unlikely that various treatment modalities will be tested in prospective controlled trials in patients with CNS involvement. Overall, these results can be a foundation for future studies on the allocation of CNS prophylactic measures in adult patients with newly diagnosed Philadelphia chromosome-negative ALL.