CNS Prophylaxis in DLBCL: Risk-Stratified Approach Required
CNS prophylaxis is NOT universally required for all DLBCL patients, but IS strongly recommended for specific high-risk subgroups based on clinical and anatomical risk factors. The decision should be driven by systematic risk stratification using validated tools and specific high-risk features 1.
Risk Stratification Framework
The CNS-IPI model is the validated tool for identifying patients who need prophylaxis 1. This stratifies patients into three risk categories:
- Low risk (0-1 factors): 2-year CNS relapse risk 0.6% - prophylaxis NOT indicated
- Intermediate risk (2-3 factors): 2-year CNS relapse risk 3.4% - consider prophylaxis
- High risk (4-6 factors): 2-year CNS relapse risk 10.2% - prophylaxis strongly recommended
The CNS-IPI includes the 5 standard IPI factors (age >60, elevated LDH, ECOG PS >1, stage III/IV, >1 extranodal site) PLUS kidney/adrenal involvement 1.
Absolute Indications for CNS Prophylaxis
Prophylaxis is definitively recommended for 1:
- Elevated LDH AND >1 extranodal site (Level of Evidence B)
- Testicular involvement (primary or secondary)
- Breast involvement (primary or secondary)
- Renal involvement (36% CNS relapse rate reported; highest risk factor)
- MYC rearrangement with BCL2 or BCL6 rearrangement (double/triple-hit lymphomas: 9-45% CNS involvement)
- Dual MYC/BCL2 protein expression by immunohistochemistry (≥40%/≥50%: 2-year risk 9.7% vs 2.2%)
Additional high-risk anatomical sites warranting prophylaxis include: bone marrow involvement, nasal sinuses, epidural space, and paraspinal involvement 1.
Optimal Prophylaxis Method: Critical Evidence
The combination of high-dose methotrexate (HD-MTX) PLUS intrathecal chemotherapy (IT) is superior to either modality alone 2. This recent 2026 study demonstrated:
- HD-MTX + IT: 6.2% CNS relapse rate, 93.2% 5-year overall survival
- HD-MTX alone: 17.4% CNS relapse rate, 73.5% 5-year survival
- IT alone: 48.1% CNS relapse rate, 44.4% 5-year survival
- No prophylaxis: 13.0% CNS relapse rate, 58.0% 5-year survival
The survival benefit with HD-MTX + IT remained significant in multivariate analysis (HR 0.160, p=0.012) 2.
Important Caveats and Controversies
Stand-alone intrathecal prophylaxis lacks strong evidence of benefit 3. A systematic review of 7,357 rituximab-exposed patients found that IT prophylaxis was not associated with CNS relapse reduction in univariable or multivariable analyses 3. This is critical because IT alone appears ineffective against parenchymal brain relapses, which comprise 58% of CNS relapses (vs 27% leptomeningeal) 3.
Recent retrospective studies have questioned the effectiveness of both IT and HD-MTX prophylaxis 4, 5, with the 2025 EHA guidelines noting "the value of CNS prophylaxis remains uncertain" 5. However, the 2026 data showing survival benefit with combined HD-MTX + IT represents the most recent high-quality evidence 2.
Practical Algorithm
For all newly diagnosed DLBCL patients:
- Calculate CNS-IPI score
- Screen for high-risk anatomical sites (testis, breast, kidney, bone marrow, nasal sinuses)
- Assess for MYC/BCL2/BCL6 rearrangements and dual protein expression
If CNS-IPI ≥4 OR any absolute indication present:
- Perform lumbar puncture with CSF cytology and flow cytometry for baseline assessment 1
- Consider brain MRI if clinically indicated 6
- Administer HD-MTX + IT chemotherapy as prophylaxis 2
If CNS-IPI 2-3 without high-risk features:
- Individualized decision based on additional risk factors
- If prophylaxis chosen, use HD-MTX + IT combination (not IT alone)
If CNS-IPI 0-1 without high-risk features:
- Prophylaxis not indicated
Key Pitfall to Avoid
Do not use intrathecal chemotherapy alone - it has poor CNS penetration for parenchymal disease and lacks proven efficacy 3. If prophylaxis is warranted, use the combination approach with systemic HD-MTX to address both leptomeningeal and parenchymal risk 2.
The median time to CNS relapse is 10 months, with 73% presenting as isolated CNS disease 3, emphasizing that prophylaxis must be administered during frontline therapy, not deferred.