Study cohort (n = 65)
A total of 65 patients diagnosed with either PVR-LBCL or SVR-LBCL were identified, including 34 patients from Tübingen and 31 from Graz. PVR-LBCL represented 55.4% (n = 36) of cases, while SVR-LBCL accounted for 44.6% (n = 29). VR-LBCL diagnosis via pars plana vitrectomy (PPV) was performed in 80% of cases (n = 52, Table 1), including four cases with concurrent retinal/choroidal biopsies. The other 20%, all of which SVR-LBCL, were diagnosed through biopsies of other tissues, such as skin, brain, or mammary tissue. All SVR-LBCL cases received systemic treatment before ocular lymphoma diagnosis. The median age at diagnosis across the entire cohort was 72.2 years (range: 49 to 96), and the median follow-up period was 23.2 months (range: 2.3 to 260.0). Detailed demographic and clinical characteristics of the cohort are summarized in Table 1 with center-specific data available in Table S1. The study flow is illustrated in Fig. 1.
Table 1 Characteristics, diagnostic methods, and first-line line intravitreal (itv.) treatment of the study cohort (n = 65).
During initial ophthalmic evaluation, VR-LBCL was considered as a differential diagnosis in 70.8% (n = 46) of cases, with a median time to diagnosis of 31 days (range: 1 to 2805 days). Before the definitive lymphoma diagnosis, 67.7% (n = 44) of patients received topical ocular therapy, including anti-inflammatory agents (e.g., prednisolone acetate, dexamethasone), immunosuppressants (e.g., ciclosporin, mycophenolate mofetil), antibiotics (e.g., moxifloxacin, gentamicin), or lubricating eye drops. Systemic antiviral or glucocorticoid treatments were administered to 20% of patients.
Ocular first-line treatments included Rituximab itv. (52.3%, n = 34), MTX itv. (29.2%, n = 19), or R-MTX itv. (7.7%, n = 5). The remaining 7.7% (n = 5) received systemic chemotherapy, local radiation, or enucleation, whereas for two patients (3.0%) no documentation of treatment was available (Fig. 1). For Rituximab itv., the first-line treatment compromised six itv. injections of Rituximab (1 mg/0.1 mL) administered 4 weeks apart. The staggered protocol for MTX itv., according to Frenkel et al. 25, consisted of 400 µg/0.1 ml MTX itv. twice weekly for 4 weeks, once weekly for 8 weeks, and once monthly for 9 months. R-MTX regimens followed no defined protocol and were given at the discretion of the treating physician. In this study, R-MTX was defined as the administration of both MTX and Rituximab intravitreally within the first six injections. In most cases, combination therapy resulted from the development of MTX-induced keratopathy, prompting a switch to Rituximab. All injections were administered separately, with no instance of concurrent MTX and Rituximab injection at the same visit. The interval between injections was at least one month in all cases. The dosages used for MTX and Rituximab in the R-MTX group followed the respective monotherapy protocols. Despite these guidelines, many patients did not receive the full course of injections, primarily due to perceived individual treatment burden or a transition to systemic therapy regimens. Across all itv. treatments, a median of 5 (range: 1 to 34) injections were administered until local remission was achieved. Patients treated with Rituximab itv. received a median of 4 (range: 1 to 8) injections, whereas those treated with MTX itv. received a median of 7 (range: 1 to 34) injections.
Follow-up cohort (n = 53)
Treatments
For subsequent analyses, only patients with a minimum follow-up period of 9 months were included, resulting in a total of 53 patients. The median follow-up was 34.1 months (range: 9 to 260), and the median age at diagnosis was 73 years (range: 49 to 96). Among these patients, 54.7% (n = 29) had PVR-LBCL, and 45.3% (n = 24) had SVR-LBCL (Fig. 1). Detailed follow-up cohort characteristics are presented in Table 2. Ocular first-line treatment in this cohort compromised Rituximab itv. (52.8%, n = 28), followed by MTX itv. (28.3%, n = 15), R-MTX itv. (9.4%, n = 5), and other regimens (9.4%, n = 5).
Table 2 Characteristics, first-line line intravitreal (itv.) treatment, additional systemic therapies, side effects and outcomes of the follow-up cohort (n = 53).
66.0% (n = 35) of patients experienced a relapse after first-line VR-LBCL treatment and received second-line therapy. Second-line ocular treatment included Rituximab itv. (31.4%, n = 11), local radiation therapy (11.4%, n = 4), R-MTX itv. (5.7%, n = 2) and MTX itv. (2.9%, n = 1). 14.3% (n = 5) of relapsed patients received systemic therapy, while 2.9% (n = 1) transitioned to palliative care. Second-line treatment was undocumented for 31.4% (n = 11), primarily due to transfers to other centers.
At any point during the entire treatment course, systemic therapy was administered to a total of 85.0% (n = 45), including 30.2% as pre-VRL-LBCL diagnosis treatment (all corresponding to SVR-LBCL), 39.6% concurrent with first-line ocular therapy, and 15.0% in subsequent treatment phases (Table 2). Of the 39.6% (n = 21) who received systemic therapy in addition to first-line ocular treatment, of which 13 were PVR-LBCL and eight were SVR-LBCL. Chemo-immunotherapy was the most common systemic approach (84.4%, n = 38), with 71.1% (n = 27) Rituximab-based and 28.9% (n = 11) MTX-based regimens. 13.2% of patients received either systemic radiation or intrathecal therapy alone. HDC/ASCT was used in 17.0% (n = 9) of patients during first-line therapy and 7.5% (n = 4) during second-line therapy (Table 2).
Clinical outcomes
Among 35 patients with relapse following first-line VR-LBCL therapy, 54.3% (n = 19) had isolated, eye-confined relapse, while 45.7% (n = 16) experienced non-eye-confined relapse involving the brain, simultaneous eye and brain, or systemic sites. The median time to relapse at any site was 8.9 months (range: 2.9 to 129.6). No statistically significant difference in time to relapse between eye-confined and non-eye-confined relapses was found (Table 3). Missing data precluded reliable statistical analyses beyond the initial relapse. The following results focus on the PVR-LBCL group (n = 29), while also encompassing data from the entire VR-LBCL follow-up cohort (n = 53) and analyzing the impact of age and sex on relapse characteristics.
Table 3 Information on relapses, as well as second-line treatments of the study cohort (n = 53).
Relapse-free survival
We conducted survival analyses to evaluate local and systemic disease control of different treatment regimens in patients with PVR-LBCL who relapsed either (a) at any site (non-eye-confined) or (b) in the eye only (eye-confined). We first evaluated intravitreal therapy alone, then assessed its combined potential with systemic therapy. Notably, none of the PVR-LBCL patients received systemic therapy before diagnosis.
Rituximab itv., MTX itv. and R-MTX itv. in PVR-LBCL: MTX and R-MTX itv. show a trend toward better local disease control
Survival analyses were performed to compare relapse-free survival among patients with PVR-LBCL treated with Rituximab itv. (n = 13), MTX itv. (n = 9), or R-MTX itv. (n = 4). Systemic therapy administration was not taken into account in these analyses, although groups were well balanced (50% received additional systemic therapy, 50% did not). Comparing Rituximab and MTX itv. regimens, 22 PVR-LBCL patients were available for any site relapse analysis. There was no significant difference in relapse-free survival between patients treated with Rituximab itv. (n = 13) and those treated with MTX itv. (n = 9) (P = 0.3; Fig. 2a). For 16 PVR-LBCL patients with eye-confined relapses, a trend toward improved relapse-free survival with MTX itv. (n = 6) compared to Rituximab itv. (n = 10) was noted but did not reach statistical significance (P = 0.07; Fig. 2b). For patients who received R-MTX itv. (n = 4), a significant difference in relapse-free survival for both any site (P = 0.02; Figure S1a) and eye-confined relapses (P = 0.005; Figure S1b) was observed. Pairwise comparisons of any site relapses showed no significant difference between Rituximab itv. and MTX itv. (P = 0.31) but demonstrated superior outcomes with R-MTX itv. compared to both (P = 0.019 for each). In the eye-confined relapse cohort, R-MTX itv. was also associated with significantly longer relapse-free survival compared to Rituximab itv. (P = 0.021).
Intravitreal +/- systemic therapy in PVR-LBCL: combined intravitreal and systemic therapy demonstrates potential for improved relapse-free survival
To assess the impact of adding systemic immunochemotherapy to itv. therapy, we compared relapse-free survival outcomes between patients who received itv. therapy only (n = 13) and those who received itv. and systemic therapy in parallel (n = 13). Systemic immunochemotherapy, as described above, was predominantly Rituximab- and MTX-based: HD-MTX (n = 5), HDC/ASCT (n = 3), R-MTX (n = 2), R-AraC (n = 1), AraC intrathecal (n = 1), and radiation therapy (n = 1). For relapses at any site, patients receiving both itv. and systemic therapy had a significantly longer relapse-free survival compared to those receiving itv. therapy alone (P = 0.05; Fig. 3a). Notably, patients receiving both itv. and systemic treatment were significantly younger (mean age: 68.8 years) compared to those receiving itv. treatment only (mean age: 75.9 years, P = 0.025).
Next, we analyzed the added benefit of systemic treatments to the individual ocular treatment regimens. While overall differences among the different treatment regimens did not reach statistical significance (P = 0.1), the R-MTX itv. + systemic therapy group (n = 3) showed the most promising outcomes, with no observed relapses (Fig. 3b). For eye-confined relapses, significant differences in relapse-free survival were observed across treatment groups (P = 0.003; Fig. 3c). Patients treated with MTX itv. + systemic therapy (n = 3) and R-MTX itv. + systemic therapy (n = 3) tended to have longer relapse-free survival compared to those receiving Rituximab itv. alone (n = 6), MTX itv. alone (n = 3), or Rituximab itv. + systemic therapy (n = 4). Although these comparisons showed a trend toward improved outcomes, the differences did not reach statistical significance after adjustment for multiple comparisons (adjusted P = 0.055) and should be interpreted with caution given the small sample sizes.
PVR-LBCL vs. SVR-LBCL: Relapse-free survival shows no significant differences by VR-LBCL type
We compared relapse-free survival between PVR- and SVR-LBCL patients. No significant differences in time to any site relapse were observed between the PVR-LBCL (n = 28) and SVR-LBCL (n = 22) (P = 0.8; Fig. 4a). Similarly, for eye-confined relapses, no significant differences were found between the two groups (PVR-LBCL n = 22; SVR-LBCL n = 15) (P = 0.7; Fig. 4b). In addition, no significant differences in eye-confined relapses were observed across the different itv. treatments (P = 0.2; Fig. 4c), with the caveat that 40% of patients also received concurrent systemic therapy.
Age and sex show no correlation with treatment regimen or relapse-free survival
Age and sex show no correlation with treatment regimen or relapse-free surviv
As patients age is a known risk factor for adverse outcomes in VR-LBCL, we evaluated the demographic characteristics of patients treated with MTX itv. and Rituximab itv. No significant differences were found in age (P = 0.19) or sex (P = 0.22) between the two treatment groups. To assess the impact of age on time to relapse, patients were divided into two groups based on the median age of 73 years (n = 26 < 73 years, n = 27 ≥ 73 years). No statistically significant difference in time to relapse was observed between the age groups for either VR-LBCL (P = 0.5) or PVR-LBCL (P = 0.2).
Ocular outcomes
To evaluate functional ophthalmic outcomes, BCVA was analyzed at three time points: (a) initial consultation at the treating eye clinic, (b) VR-LBCL diagnosis, and (c) after completion of first-line treatment. Subsequent time points were not analyzed due to missing data. All BCVA values are reported in logMAR.
BCVA shows no significant changes across time points or treatments
For VR-LBCL (n = 53), mean BCVA was 0.54 (range: -0.1 to 2.7) at initial consultation, 0.62 (range: -0.1 to 2.4) at diagnosis, and 0.63 (range: -0.1 to 3.0) after first-line treatment. No significant changes in BCVA were observed between successive time points (P = 0.90 and P = 0.85) or between initial consultation and after first-line treatment (P = 0.84). BCVA outcomes did not differ significantly between patients treated with Rituximab itv. or MTX itv. (P = 0.55). An additional linear mixed model analysis assessed the effects of treatment type (Rituximab itv. and MTX itv.) and time point on BCVA. The model revealed no significant main effects for time point (P = 0.70) or treatment type (P = 0.55). The interaction effect between time points and treatment was also not significant (P = 0.72).
Ocular side effects: keratopathy and IOP decompensation as treatment-associated complications
11 (55%) of 20 patients that received MTX itv. developed keratopathy (corneal epitheliopathy), while two patients who received Rituximab itv. had pre-existing keratopathy that showed no worsening under treatment. 13 (39.4%) of 33 patients who received Rituximab itv. experienced transient intraocular pressure elevation throughout the follow-up period. These were typically short-lived, with normalization or return to tolerable levels within several days. While peak IOP values reached up to 40 mmHg, the majority remained moderate (< 30 mmHg). Management strategies were heterogeneous and included topical and systemic antiglaucomatous agents, cyclophotocoagulation in rare cases, and paracentesis following injection in patients with borderline baseline IOP. In this subgroup, IOP remained well controlled without further intervention. Notably, no cases of sustained or uncontrolled IOP elevation directly attributable to Rituximab were observed. A more detailed analysis was limited by the heterogeneity of the cohort and the frequent presence of previous episodes of IOP elevation. Therefore, for this study, only IOP elevations in patients without elevated IOP at the time of diagnosis were considered for descriptive evaluation. No other itv. treatment-associated side-effects were documented.
