[Lancet Oncol Review] CNS prevention of diffuse large B-cell lymphoma

Involvement of the CNS is an uncommon and usually fatal event, with about 5% of patients with systemic diffuse large B-cell lymphoma (DLBCL) having a CNS recurrence during or shortly after initial treatment, with tumor cells reaching the CNS through hematogenous routes, direct infiltration of neighboring organs, or through neurovascular axis and affecting the brain parenchyma, meninges, cerebrospinal fluid, or eyes

In a review published in Lancet Oncology, the authors critically analyze the available evidence supporting the use of CNS recurrence prognosis models and different CNS prevention strategies, discuss the clinical and biological characteristics associated with elevated risk, summarize the history and current developments in this challenging area, engage in nuanced discussions on the reasons for current and non-adoption of standard preventive measures, outline evidence on the timing of preventive action, and consider possible

CNS recurrence of BL

Clinical risk factors for secondary CNS lymphoma

There are many definitions of patients with high-risk recurrence of CNS, of which the CNS International Prognostic Index (CNS-IPI) is the best-validated prognostic model developed in the era of rituximab, consisting of

Although CNS-IPI helps to compare studies and assess the independent correlation of biomarkers, it is not able to capture all high-risk patients and has low

Overall, the difference in CNS recurrence rate before and after rituximab is unclear from the available literature and may be only marginally reduced

CNS recurrence: biological, genetic, and biomarker effects

MYC translocation with BCL2 translocation, with or without BCL6 translocation (i.

Two separate studies integrated multi-platform genetic analysis to propose a new taxonomy of a DLBCL subclass outside of cell origin, with many

Outcome of secondary CNS lymphoma

Patients with secondary CNS lymphoma have poorer outcomes, even in patients who are suitable for intensive therapy; Prospective studies of high-dose chemotherapy and autologous stem cell transplantation after intensive regimens showed CR rates of 25-63%, but remission is usually not durable, with a 2-year overall survival rate rarely exceeding 50%.

Historical CNS prevention data

Intrathecal chemotherapy

Reports from the 1970s and 1980s showed a risk of secondary CNS lymphoma, with a small number of subsequent uncontrolled studies showing benefit

Other evidence of the effectiveness of intrathecal chemotherapy is mainly extrapolated from patients with high-risk (up to 40%) CNS aggressive B-cell lymphoma, including Burkitt lymphoma and lymphocytic lymphoma, where comprehensive intrathecal chemotherapy is the standard protocol

Intravenous high-dose methotrexate

High-dose methotrexate has become increasingly used as CNS prevention in patients with DLBCL, in part because most (70-80%) CNS relapses into the brain parenchyma in the rituximab era, so it is not possible to prevent

Intrathecal prevention

Although some historical data suggest potential benefits of intrathecal prophylaxis, most clinical studies of intrathecal prophylaxis are difficult to interpret

A systematic analysis of the CD20 monoclonal antibody era evaluated the role of intrathecal prophylaxis alone, which included three post-trial analyses (RICOVER-60, RCHOP-14/21 and GOYA), a prospective database, and 10 retrospective series, analysing a total of 7357 patients

Antimetabolic drug prophylaxis

Medications used and evidence base

Although there is a biological basis for the use of CNS penetrating agents as prophylactic therapy, there are no completed randomized studies

Phase II studies have explored CNS penetrating agents combined with intensive first-line therapy as an alternative prevention strategy and improved systemic disease control of medium- and high-risk invasive B-cell lymphomas, including: 1.

Notably, a randomized study of patients with low CNS-IPI risk and aggressive B-cell lymphoma compared ACVBP with CHOP, resulting in ACVBP reducing the CNS recurrence rate from 8.

In addition, baseline screening techniques for occult CNS lesions (CSF cytology or flow cytometry and neuroimaging) varied in these studies, but the requirements for such procedures limited the comparability with observational studies that did not undergo systematic CNS screening (Table 3).

Historical observations have shown a high CNS recurrence rate (10-30%) for primary testicular lymphoma, but in the IELSG30 study (54 patients with primary testicular lymphoma), the CNS recurrence rate was significantly reduced
after high-dose methotrexate prophylaxis and intrathecal prophylaxis of 1.
5 g/m².

In contrast, several large retrospective studies have not shown that high-dose methotrexate reduces CNS recurrence, with a recurrence rate of 6-12% in patients with high-risk invasive B-cell lymphoma (DLBCL in most studies), regardless of whether high-dose methotrexate is prevented (Table 3); This also included a retrospective cohort
of 2300 patients at high risk of CNS recurrence (CNS-IPI = 4-6 [89.
2%], double-strike, triple-strike, primary testicular lymphoma, or breast DLBCL) treated in the rituximab era.
During or after chemoimmunotherapy, different doses of high-dose methotrexate did not significantly reduce CNS recurrence (8.
4% in the high-dose methotrexate group versus 9.
1% in the non-high-dose methotrexate group, p=0.
1).

In a retrospective cohort study that included 1384 patients all treated with high-dose methotrexate, 600 patients with DLBCL (with a CNS-IPI score of 4-6) had a 3-year CNS recurrence rate of 9.
1%, which is very similar to the cohort that originally developed and validated CNS-IPI (which rarely used CNS for prevention) (10.
2%)
.

Timing of administration of high-dose methotrexate

CNS recurrence usually occurs at a median of 6 to 9 months after the initial DLBCL diagnosis, so it is theoretically given as early as possible with high-dose methotrexate
.
The first study to report embedding high-dose methotrexate between R-CHOP cycles was a small retrospective series (n=65) that resulted in a low CNS recurrence rate (3%), but systemic treatment was delayed
in 12% of patients.

Two single-arm phase 2 studies observed lower CNS recurrence rates in young, high-risk patients receiving high-dose methotrexate-based prophylaxis (Table 3
).
The first study was followed up for 52 months with high-dose methotrexate and cytarabine after R-CHOEP-14, resulting in a CNS recurrence rate of 4.
5%; In the second study, when high-dose methotrexate was given early plus R-CHOP or R-CHOEP-14, the median CNS recurrence rate of 60 months was 2.
5%.

These data also suggest that early preventive administration reduces the risk of CNS recurrence, despite wide confidence intervals and different
timing of high-dose methotrexate infusions.
High-dose methotrexate was given in an initial trial with a 24-hour infusion, while in subsequent studies it was given with a 3-hour infusion (which improves CNS bioavailability
).

A large retrospective study resolved the issue of timing of high-dose methotrexate administration by comparing patients receiving high-dose methotrexate embedding (n=749) or at the end of R-CHOP (n=635) with a significant increase in R-CHOP delayed ≥ 7 days after high-dose methotrexate insertion compared with administration at the end of R-CHOP therapy, and a 20% delay
in subsequent R-CHOP after embedded high-dose methotrexate therapy 。 However, there was no difference in CNS recurrence rates between the two groups, including multivariate analyses and analyses limited to patients at highest risk
.
Overall, embedding large doses of methotrexate can lead to R-CHOP disruption and delay, impairing administration and R-CHOP efficacy
.
Of course, while retrospective data that pioneered high-dose methotrexate has limitations, it supports the administration of methotrexate at the end of R-CHOP
.

Evaluate the evidence for CNS prevention

There is little evidence for the use of CNS prophylaxis in high-risk DLBL, there are no randomized studies evaluating this problem, and a number of factors further complicate the assessment of the utility of CNS prophylaxis, including the heterogeneity of DLBCL and risk assessment of CNS involvement, multiple CNS prophylaxis modalities, and the relative rarity of CNS recurrence, resulting in numerous small and ineffective observational studies
。 As noted above, the data supporting CNS prophylaxis are primarily driven by extrapolations from other lymphoma subtypes, small single-body series studies, and comparisons of CNS recurrence rates with no CNS prophylaxis and expected recurrence rates without CNS, so caution must be exercised when comparing the results of one-arm clinical or observational studies with historical incidence, as DLBCL outcomes may be highly sensitive to selection bias, even if relevant factors (e.
g.
, CNS-IPI)
are controlled.

Despite the lack of strong data to support its use, CNS prevention has been widely used, even though large retrospective studies have shown limited utility (Table 3
).
Of course, these studies are complicated by the heterogeneity of the treatment physician or institution using CNS prophylaxis, and there may also be treatment choice bias, because patients receiving CNS prevention tend to be younger and more fitting than patients who have not received prevention, and careful analysis to adjust clinical factors can partially correct this choice bias
.
Broadly speaking, the clinical protocol and evidence base suggest that losses should be avoided, so even if there is little evidence to support the potent role of CNS prophylaxis, the cognitive risk of not treating CNS recurrence usually outweighs the lack
of evidence (prevention using CNS).

Arguments for high-dose methotrexate prophylaxis

Adverse outcomes for secondary CNS lymphoma are attributable to poor CNS penetration of chemotherapy agents, impaired neurocognitive function and the patient's energy status, increased treatment-related toxicity, and recurrent genetic aberrations leading to treatment-resistance
.
It is therefore widely accepted that the risk of CNS recurrence should be minimized, but a key question is whether patients are truly benefiting from prevention, and if the answer is yes, what are the most effective treatment strategies?

Unfortunately, the practice-inspired data are largely retrospective, there are significant differences in choice criteria, types, and primary treatment regimens, and studies comparing patients receiving or not receiving CNS prophylaxis often do not exactly match high-risk features, so there may be additional biases
in guiding treatment decisions.
Therefore, caution should be exercised in interpreting retrospective studies that are meaningless in CNS prevention, and their limitations
should be acknowledged.

In some cases, current practice results in a large proportion of patients with DLBCL receiving CNS prophylaxis unnecessarily, including that approximately 90% of patients with a CNS-IPI score of 4-6 may never have a CNS recurrence
.
In addition, the CNS-IPI scoring model is less sensitive, and a significant proportion of CNS events occur in the middle risk group
.
Therefore, other risk factors should be considered, including those described herein (i.
e.
, testicles, kidneys, adrenal glands, ≥ 3 extranodal lesions), and it is recommended that CNS prevention strategies focus on
these high-risk groups.
In contrast, the evidence for infrequent DLBCL subtypes, such as double strikes, whose risk of CNS recurrence may be due to accompanying risk factors rather than biological causes; In fact, however, there are few
large-scale studies on the CNS risk in DLBCL histology double-strike patients.

A comparison between two prospective IELSG studies of isolated primary testicular lymphoma supports high-dose methotrexate prophylaxis
.
Primary testicular lymphoma is an important model because the 5-year CNS recurrence rate in one study of patients who did not receive preventive treatment was 19%.

In the ISGEL10 study, 53 patients with stage I.
-II primary testicular lymphoma received R-CHOP-21, contralateral testicular radiotherapy, and 4 sessions of intrathecal methotrexate, resulting in a recurrence of CNS in 3 patients (2 pipsum), resulting in a long-term CNS recurrence rate of 6% (median follow-up of 65 months).

In a subsequent ISGEL30 study, 54 patients with primary testicular lymphoma were treated with intrathecalocytic cytarabine liposomes and two additional doses of 1.
5 g/m² methotrexate; No CNS recurrence
was observed at median follow-up of 73 months.
Although cross-comparisons of one-arm trials are limited, these results may suggest that high-dose methotrexate prophylaxis may be beneficial
in the treatment of primary testicular lymphoma.
The risk of CNS involving the testes in advanced DLBCL is higher, and although evidence is limited, CNS prophylaxis
may be reasonably considered.

In one study of 103 patients with invasive B-cell lymphoma, high-dose methotrexate 3 g/m² could be given on an outpatient basis for 2 cycles without monitoring serum methotrexate, but with fixed-dose folinic acid rescue and oral hyperhydration
.
Only 8 patients did not receive a second methotrexate administration due to toxicity, suggesting the feasibility
of this method.

Arguments that do not support high-dose methotrexate prophylaxis

There are insufficient studies on the frequency of asymptomatic CNS involvement at the time of initial DLBCL diagnosis, and there are no large studies to continuously screen all high-risk patients
with CSF flow cytometry and full CNS imaging, including MRI.
A single-center study evaluated 154 patients with newly diagnosed DLBCL, of whom 93 (60%) had baseline CSF flow cytometry, but MRI was
performed only in symptomatic patients.
Of the 101 samples obtained in patients without neurologic symptoms, 12 were positive for CSF flow cytometry, and most of them were not considered high-risk CNS involvement
.
Although only a few patients with positive flow cytometry have CNS recurrence, the use of CNS prophylaxis and a short follow-up period limit the understanding
of the clinical relevance of infiltration for flow cytometry testing.
In another study, 11 of the 51 patients with high-risk DLBCL had occult CNS involvement at the time of diagnosis, defined as positive flow cytometry, and 5 of these 11 patients later had a CNS recurrence (no description of whether or not high-dose methotrexate was used).

The above data suggest that a large number of patients without significant neurologic symptoms may have occult CNS involvement at baseline, which is consistent with early manifestations of CNS recurrence; These patients may also represent a significant proportion of patients
who are later diagnosed with a recurrence of CNS.
Therefore, the first step in preventing secondary CNS involvement is to consider systematic screening of very high-risk patients (imaging and CSF cytology and flow cytometry) and to develop more sensitive techniques to capture minimal CNS involvement
.
According to the medical consensus, these patients can be better managed with intensive treatment with CNS penetrating drugs instead of R-CHOP
.

Based on the true significance of CNS prophylaxis, considering CNS prophylaxis is only appropriate for patients without CNS involvement at baseline, and high-dose methotrexate is unlikely to be cost-effective for all patients at high risk of CNS-IPI: the CNS-IPI high-risk group accounts for 12 to 23% of all patients with DLBCL and the CNS 2-year recurrence rate is about 12%.

。 Although outpatient administration is feasible, most patients receiving high-dose methotrexate therapy remain hospitalized due to the complexity of hydration and rescue treatment options; High-dose methotrexate takes about 6 days to complete, placing a huge administrative and financial burden on
hospitals and patients.
In addition, high-dose and intrathecal methotrexate is rarely associated
with potentially severe leukoencephalopathy and myelopathy.

Two large studies showed a similar
incidence of secondary CNS lymphoma in patients at high risk of CNS-IPI receiving or not receiving high-dose methotrexate and after receiving different high-dose methotrexate regimens.
In one of the studies, 2300 patients with high-risk CNS relapse (CNS-IPI 4-6: 89.
2%) were analysed based on the use of high-dose methotrexate, who received R-CHOP-like therapy (93.
8%)
.
A total of 410 patients (17.
8%) received high-dose methotrexate, of which 32 (7.
8%) had CNS relapse, compared with 169 (8.
9%) of 1890 patients who did not receive high-dose methotrexate
.
The 5-year CNS recurrence rate corrected for high-dose methotrexate was 8.
4% in the high-dose methotrexate group and 9.
1%
in the group without high-dose methotrexate.
Since high-dose methotrexate will certainly introduce a temporal bias of CNS recurrence until a high-dose methotrexate is given, patients with CR were analyzed
。 Of the 1455 patients with CR, 284 (19.
5%) received high-dose methotrexate, 16 of 284 patients (5.
6%) had CNS recurrence, compared with 68 (5.
8%) of 1171 patients who did not receive high-dose methotrexate treatment; Similarly, there was no difference in the risk of 5-year CNS recurrence between the two groups (5.
0% vs 6.
0%): if the real difference was a 1% reduction in CNS recurrence, in favor of high-dose methotrexate, this meant that 100 patients would need to receive high-dose methotrexate to avoid a single CNS recurrence
.
The above results should be interpreted with caution as the proportion of patients with high-risk features in the high-dose MTX group is higher, including ≥ 2 external nodal sites (44% and 30%) and high-risk external junction sites (47% and 24%)
.
Overall, these data suggest that although high-risk patients can be identified, the preventive measures currently available may be insufficient
.

Even in the absence of strong evidence of efficacy, the continued use of CNS prophylaxis is partly due to the dismal
outcomes of secondary CNS lymphoma.
Adverse outcomes were primarily driven by patients with concurrent systemic and CNS relapses, and as shown in the MARIETTA/IELSG42 study, subgroups including isolated patients with CNS recurrence (arguably the only secondary CNS lymphoma class that CNS prophylaxis may prevent) also showed unsatisfactory
outcomes with intensive immunochemotherapy.
All patients still need to continue to improve patient outcomes, but patients with concurrent systemic or CNS recurrence are of greater concern because failure of both systemic therapy and CNS prophylaxis is problematic and may require different treatment strategies
.

It is also important
to recognize the effects of increased CAR-T cell therapy and novel oral therapy in patients with secondary CNS lymphoma.
The immunomodulator lenalidomide and the BTK inhibitor ibratinib have clear clinical activity in CNS lymphoma, and results from multiple case series have shown that CAR-T is active in both primary and secondary CNS lymphomas, suggesting a possible significant improvement in outcomes
in these patients 。 While further research is needed on CAR-T, this treatment strategy may still affect the risk-benefit balance when making CNS prevention-related decisions, but it should also be recognized that fitness status may hinder eligibility and that CAR-T is not readily available
in many countries due to cost and resource implications.

Advances in the biological understanding of molecular DLBCL have identified genetic subgroups with high recurrence of CNS, which are meaningful for both diagnosis and treatment, with the potential to identify high-risk patients with higher specificity, and to investigate the use of novel targeted drugs
that enhance systemic disease control and CNS penetration.
A more personalized strategy to use this targeted agent in patients with known high-risk molecular subtypes can improve the specificity and effectiveness of traditional CNS prophylaxis
.

Future directions

Novel treatment options to prevent CNS recurrence

Lenalidomide and ibratinib are CNS permeable in primary CNS lymphoma and active in both systemic ABC or non-GCB DLBCL subtypes, leading to first-line studies
integrating these drugs.
A postmortem analysis of a randomized, phase 3 REMARC study showed that patients with DLBCL aged 60 to 80 years (CNS-IPI ≥1) with nalide maintenance therapy (CNS-IPI 1) did not reduce CNS recurrence rates (2-year CNS recurrence rate: lenalidomide 3.
3% vs placebo 0.
9%)
.
CNS prevention was performed according to local practice or investigator judgment without changing outcomes
.
In addition, the Phase 3 ROBUST study evaluating R-CHOP-lenalidomide versus R-CHOP and the Phase 2 ECOG-ACRIN E1412 study evaluating R-CHOP-lenalidomide have not reported on the incidence
of CNS.
A Phase 3 PHOENIX study comparing R-CHOP-ibratinib with R-CHOP for ABC DLBCL showed that CNS recurrence rates were generally low and similar (2.
4% vs 3.
8%)
.

The immune escape phenotype is evident in some DLBCL subtypes, particularly the MCD or C5 subtype DLBCL, thus highlighting the potential role
of PD1 inhibitors.
Overall, although these studies do not show benefits for biologic drugs in preventing CNS recurrence, they are models for hypothesis generation, opening up new options
for future CNS prevention research.
In this setting, CAR-T cell therapy may play a role because it is effective for refractory DLBCL, remission in half of patients with CNS lymphoma, and CAR-T cells can be expanded peripherally and metastasized to the CNS without active disease
at the time of infusion.
The broader application of CAR-T therapy as part of first- and second-line therapy in patients with DLBCL may theoretically help prevent CNS recurrence
.
In the recent phase II ZUMA-12 study, 40 patients with DLBCL with a CNS-IPI score of ≥3 and 2 courses of PET⁺ with anthracycline chemotherapy did not experience CNS recurrence at a median follow-up of 15.
9 months after receiving axicabtagene ciloleucel
.
This general hypothesis deserves further study
.

CSF circulating tumor DNA analysis

In addition to imaging tests, patients at high risk of DLBCL often undergo CSF cytology and flow cytometry analysis; But sensitivity is still poor
.
Cell-free circulating tumor DNA (ctDNA) has recently become a non-invasive prognostic biomarker for patients with lymphatic malignancies and may play a role
in CSF analysis.
A recent study of 67 patients with central nervous system lymphoma, including 12 isolated secondary central nervous system lymphomas, detected ctDNA in all pretreatment samples of cerebrospinal fluid (100%) and showed a significant correlation
between plasma ctDNA concentration and tumor volume and prognosis.

Two studies also evaluated the potential role
of CSF ctDNA in predicting the recurrence of CNS in high-risk B-cell lymphoma.
Bobillo et al.
analyzed specific tumor-derived mutations in the cerebrospinal fluid of 12 newly diagnosed patients with B-cell lymphoma, and concluded that ctDNA
was detected in cerebrospinal fluid samples taken 3 months prior to the recurrence in 12 patients with CNS recurrence.
Another study analyzed 22 patients with high-risk B-cell lymphoma using NGS-MRD and detected clonal DNA in the CSF of 8 patients (36%) at the time of diagnosis, including 2 cases of CNS recurrence, a cumulative risk of 29% for 12 months of CNS recurrence, and a 0%
risk for patients with negative CSF.
In addition, about 70% of primary testicular lymphomas have MYD88 mutations, and this information may play a role
in CSF CNS testing in oligocellular patients.

Decisions regarding the use of CNS for prophylaxis must be practical and need to take into account the expected risk of secondary CNS lymphoma, the effectiveness and toxicity of currently available prophylaxis strategies, available treatment options for secondary CNS lymphoma (if any), patient preferences, and medical resource utilization, but reliable prospective data
on risk estimation and clear benefits of prevention strategies to date are lacking 。 Future developments should focus on integrating molecular (i.
e.
, ABC, DLBCL genetic subtypes) and clinical risk factors (i.
e.
, CNS-IPI, number of extranodal lesions, and high-risk sites, as described) to identify extremely high-risk patients and expanding hypersensitivity techniques to detect occult CNS involvement at presentation, including ctDNA or CSF MYD88 mutation detection (or both), to guide patients in CNS treatment decisions
.
Importantly, anti-DLBCL active biologics with good CNS bioavailability improve first-line therapeutic efficacy and reduce CNS spread
.
If secondary CNS lymphoma continues to occur after these strategies, several new protocols are currently being developed, with CAR-T cell therapy already showing potential promise
in high-risk patients.

Overall, there is still an urgent need for prospective international collaborations with full efficacy in patients at high risk of CNS recurrence (homogenization therapy) to address this important clinical issue
.

References

Toby A Eyre ,et al.
CNS prophylaxis for diffuse large B-cell lymphoma.
Lancet Oncol .
2022 Sep; 23(9):e416-e426.
doi: 10.
1016/S1470-2045(22)00371-0.

The label is not dotted when previewing

： ，
。 Video Mini Program like, tap twice to cancel like in viewing, tap twice to cancel in viewing