High-dose methotrexate does not reduce CNS relapse in patients with DLBCL/HGBCL: findings from two retrospective cohort studies

Although the occurrence of central nervous system (CNS) relapse is rare (5%) in patients with aggressive or diffuse large B-cell lymphoma (DLBCL), it can lead to devastating complications and poorer outcomes. The risk is particularly high (10% at 2 years) in patients with a CNS – International Prognostic Index (CNS-IPI) score of 4–6. High-dose methotrexate (HD-MTX) prophylaxis has been used to mitigate the CNS relapse risk in selected high-risk patients. However, HD-MTX is associated with increased toxicity, prolonged inpatient stays, and delays to frontline systemic therapy. There is currently limited evidence—mainly from underpowered and retrospective studies—on the efficacy and optimal timing of delivery of HD-MTX prophylaxis.

Here we summarize the key findings from two abstracts presented during the 63rd American Society of Hematology (ASH) Annual Meeting and Exposition. Wilson, et al.¹ and Lewis, et al.3 presented findings from their retrospective analyses investigating the influence of HD-MTX prophylaxis in preventing CNS relapse.

Wilson, et al.^1,2

Study design

This was a large, multicenter, retrospective cohort study comprising 47 sites worldwide in patients with DLBCL or high-grade B-cell lymphoma treated with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) or similar frontline therapy and HD-MTX prophylaxis. Patients on intensified regimens such as dose adjusted etoposide, prednisone, vincristine, doxorubicin, and cyclophosphamide with rituximab were excluded.

• The primary endpoint was 3-year cumulative incidence of CNS relapse according to the timing of HD-MTX delivery.
• Secondary endpoints included progression free survival (PFS), overall survival (OS), delays to R-CHOP therapy and risk factors for CNS relapse including use of concurrent intrathecal (IT) CNS prophylaxis.

To mitigate immortality bias, an additional landmark analysis of patients alive and free from progression was conducted at 6 months.

Results

Baseline characteristics

A total of 635 and 749 patients were included in the end of treatment (EOT) and intercalated HD-MTX (i-HD-MTX) groups, respectively, and 1,253 patients were included in the landmark analysis. The i-HD-MTX group had a higher proportion of patients with advanced stage disease whereas the EOT group had a higher proportion of patients receiving IT CNS prophylaxis (Table 1).

Table 1. Baseline characteristics*

CNS relapse

• The 3-year CNS relapse rates for all patients were 5.8% vs 5.7% (hazard ratio [HR], 1.01) in the EOT and i-HD-MTX groups, respectively, with a 3-year difference in CNS relapse rate of 0.04% between the groups.
• The 3-year CNS relapse rates in the landmark analysis were consistent in the EOT and i-HD-MTX groups, respectively (4.7% vs 4.7%; HR, 0.99), with a 3-year difference of 0.03% between the groups.

• The overall 3-year rate of CNS relapse in patients with CNS-IPI 4–6 was consistent with the previously reported rate of 9.1%.

PFS, OS, treatment delay, and risk factors for CNS relapse

Median OS was 5.4 months for all patients experiencing CNS relapse;

• 6 vs 5.2 months in the EOT and i-HD-MTX groups;
• 5.1 vs 7.4 months for CNS relapse occurring at parenchymal and leptomeningeal sites; and
• 5.6 months vs 2.8 months in isolated and synchronous CNS relapse (HR, 1.69; p = 0.069).

In the landmark cohort, 3-year PFS was 82.6% vs 81.2% and 3-year OS was 91.5% vs 88.3% in the EOT and i-HD-MTX groups, respectively. The HD-MTX approach was significantly associated with treatment delays of ≥7 days (odds ratio (OR), 0.44; p < 0.0001) in the EOT group vs the i-HD-MTX group.

Increasing age was associated with R-CHOP delay (median 8 days) in the i-HD-MTX group; however, the risk was significantly reduced if HD-MTX was delivered before Day 10 in the R-CHOP cycle. Patients experiencing any delay of ≥7 days showed inferior PFS (HR, 1.52) compared with those not experiencing treatment delays.

Multivariate analysis showed concurrent IT prophylaxis was not associated with reduction in CNS relapse and isolated CNS occurred in 73% of patients, CNS relapse rates were similar (2.8% vs 2.5%) in the EOT and i-HD-MTX groups, respectively, with a 3-year difference of 0.2%.

Lewis, et al.³

Study design

This was a large, multicenter, retrospective cohort study comprising 21 sites in Australia, Asia, North America, and Europe. Eligible patients were diagnosed between 2000 and 2020 with DLBCL with CNS IPI score of 4–6 or high-grade B-cell lymphoma with rearrangements of MYC + BCL2 and/or BCL6 (double/triple hit) or primary breast or testicular DLBCL irrespective of CNS-IPI score, were aged 18–80 years, and were treated with anti-CD20 based chemo-immunotherapy. HD-MTX was defined as at least one cycle of intravenous MTX at any dose.

The primary endpoint was time to CNS relapse, defined from date of lymphoma diagnosis to date of CNS relapse for all patients and date from completing frontline therapy (6 × 21 days from lymphoma diagnosis) to date of CNS relapse for patients achieving complete response (CR).

To mitigate immortality bias, an additional landmark analysis of living patients was conducted.

Results

Baseline characteristics

A total of 2,267 patients were included, 1,468 of whom comprised the CR cohort. The HD-MTX group had a higher proportion of male patients, patients aged ≤60 years, and patients with ECOG Performance Status of 0–1. Patients in the HD-MTX group also had a higher proportion of patients with ≥3 extranodal sites (48% vs 30%) and high risk extra-nodal sites (51% vs 26%) compared to those in the no HD-MTX group (Table 2).

Table 2. Baseline characteristics*

CNS relapse

At a median follow-up of 5.8 years, the cumulative incidence of CNS relapse was 8.1% vs 9.2% for all patients in the HD-MTX and no HD-MTX groups, respectively.

• In the CR cohort, the cumulative incidence of CNS relapse was 4.8% vs 6.3% for patients in the HD-MTX and no HD-MTX groups, respectively, at a median follow-up of 5.4 years.
• The median time to CNS relapse was 8.5 vs 6.7 months and 11.5 vs 10.3 months in the HD-MTX and no HD-MTX groups for all patients and for patients in the CR cohort, respectively.
• The difference in 5-year risk of CNS relapse between the HD-MTX and no HD-MTX groups was 1.1% (HR, 0.68; p = 0.067) for all patients and 1.5% (HR, 0.77; p = 0.381) for patients in CR cohort, respectively.
• In the CR cohort, 71% of patients received at least one dose of ≥3g/m² of MTX.
• 31% of patients started HD-MTX during primary chemo-immunotherapy, while 53% started after chemo-immunotherapy.
• The most frequent site of CNS relapse was brain parenchyma in the no HD-MTX group (in all patients and in the CR cohort).
• Isolated CNS was more common in all patients (68% vs 61%) and patients in CR cohort (71% vs 69%) receiving HD-MTX and no HD-MTX, respectively.
• Exploratory analyses did not show any difference in CNS relapse risk in any of high-risk subgroups.
• At a median follow-up of 5.4 years, median PFS and OS were not reached.

Conclusion

These two large retrospective studies showed that HD-MTX, including its time of delivery, in patients achieving CR or in any high-risk groups, was not associated with reduction in the risk of CNS relapse. The study by Wilson, et al.¹ showed that i-HD-MTX significantly increased the risk of R-CHOP delay and the overall rates of CNS relapse in high-risk patients remained relatively high, raising concerns around the use of HD-MTX as prophylactic treatment in this population. The study also supports the recommendation to defer delivery of HD-MTX until completion of R-CHOP and to use i-HD-MTX in older patients before Day 10 of R-CHOP cycle to reduce treatment delays. In the absence of any prospective data, these two retrospective studies provide evidence on the efficacy of HD-MTX, and future studies are needed to justify its use in high-risk patients.