Answers are grounded in this paper and cite their sources.
Primary vitreoretinal lymphoma (VRL) is a rare and aggressive large cell lymphoma that predominantly affects the retina and vitreous body. VRL occurs in 1% of all ocular malignancies, with an incidence of 0.047 cases per 100 000 people.1 The incidence has risen sharply since the 1990s. The prognosis is poor, depending on the extent of spread (e.g. brain involvement). The majority of VRL originate from B lymphocytes, but a small percentage are of T-cell or natural killer (NK) cell origin.1-3 The diagnosis of VRL poses significant challenges due to its nonspecific symptoms and the potential for misdiagnosis. Furthermore, the diagnostic process is complicated by the need to analyse small volumes of fluid from the vitreous or aqueous humour. Traditional diagnostic methods, including clinical examination and imaging, may not provide conclusive results. Consequently, obtaining samples for cytological and flow cytometry (FC) analysis has emerged as a valuable approach to improve diagnostic accuracy.1-3 Cytology and FC are reliable and rapid techniques that can be performed in most biological laboratories, making them accessible tools for early intervention in suspected cases of VRL. Recent advances in these diagnostic techniques have underscored the importance of using vitreous and aqueous humour samples to identify malignant cells. These samples can also be taken in case of suspected central nervous system (CNS) lymphoma with vitreoretinal involvement, avoiding the need for a brain sample.4 These methods not only improve the sensitivity of VRL detection but also help to differentiate VRL from other retinal pathologies. This article reports on our experience with the diagnostic utility of these techniques and their impact on the timely diagnosis of VRL, or ocular localization of CNS lymphoma. Twenty vitreous fluids and one subretinal sample were analysed from 19 patients (9 men and 10 women). One patient underwent vitreous and subretinal puncture during the same surgical procedure (Patient 6) (see Supporting Information). Another patient underwent two vitreous punctures 7 months apart (Patient 11). None patient received corticosteroid treatment prior to undergoing vitrectomy, except for patient 1, for whom a wash out of 30 days was observed before vitrectomy. Vitreous taps were performed from January 2023 to December 2024 in ophthalmology departments of Assistance Publique Hôpitaux de Marseille (Hôpital Nord) and Clinique Juge—Centre Monticelli-Paradis. All biological analysis, systemically including conventional cytological analysis and immunophenotyping by flow cytometry, were conducted at the La Timone Hospital laboratories (see Supporting Information). All tests were performed in the context of routine care. The mean age was 71.8 years. The average transport time from the operating room to the laboratory was 94.6 min. The median volume of vitreous fluid was 0.001 L (range: 0.000350–0.0015 L). The volume of the subretinal sample was 0.00001 L. Nucleated and red blood cell counts were carried out on 18 vitreous fluids, showing a median cellularity of 47×10−6 nucleated cells per litre (range: 13–610 × 10−6cells/L) and 7×10−6 per litre (range: 1–370 × 10−6cells/L). The cellularity of the subretinal sample was higher, with 2675 × 10−6 nucleated cells per litre and 2890 × 10−6 red blood cells per litre. Detailed demographics are available in Table S1. The identification of lymphoma cells in cytology relies on morphological criteria: irregular nucleus shape, less compact chromatin, prominent nucleolus, abundant and hyperbasophilic cytoplasm (Figure 1). FC enables lymphocyte phenotyping, including distribution of different lymphocyte subpopulations and identification of the monoclonal or polyclonal nature of B lymphocytes (Figure 2). In our study, combination of cytological and FC analysis highlighted lymphomatous cells in 57.1% of samples (12/21), corresponding to 57.8% of patients (11/19). Five patients (26%; 5/19) presented an ocular extension of a CNS lymphoma, whereas others (31.5%; 6/19) were primary VRL. Both biological tests were negative for 43% of samples (9/21), corresponding to 42.1% of patients (8/19). For patient no. 2, despite negative findings using cytology and flow cytometry, the diagnosis of VRL was retained due to the interleukin (IL)-10/IL-6 ratio obtained with anterior chamber paracentesis, which is indicative of VRL (IL-6: 180 pg/mL; IL-10: 1320 pg/mL; ratio IL-10/IL-6 = 7.3), in a context of typical lesions on imaging (OCT) showing typical hyalitis and subretinal infiltration. For all other patients (36.8%; 7/19), the diagnosis of lymphoma was excluded and other diagnoses such as idiopathic uveitis, hyalitis and granulomatous panuveitis were retained (Table S1). For patient 11, a brain biopsy performed 11 months after vitrectomy led to the diagnosis of cerebral localization of known extra-ganglionic T/NK lymphoma. However, this is a cerebral localization and vitreous tap is not necessarily the most relevant sample in this case. Interestingly, among the five patients with ocular extension of CNS-lymphoma, brain biopsy was avoided for three patients as lymphoma cells were identified in vitreous tap. Cytological analysis and lymphocyte immunophenotyping can be performed on the same small volume sample. While cytological analysis was inconclusive in three samples due to cell lysis caused by excessive delivery time (80 and 170 min for two of them), FC successfully identified lymphomatous cells in two cases. For one case, cytology did not identify lymphomatous cells, whereas FC identified the presence of only 10 monoclonal B lymphocytes. Cytological analysis showed cellular alterations in 47.6% of samples (10/21) (Figure 1). As expected, delivery time from operating room to the laboratory was significantly higher for these samples compared to samples without altered cells (116 min vs. 79 min, respectively; p = 0.03 unpaired t-test). Interestingly, delivery time did not appear to affect the outcome of FC. Neither sample volume nor cellularity appeared to affect the feasibility of either analysis. Vitreous sampling allows for the direct collection of cells from the vitreoretinal environment, providing privileged access to relevant, albeit low-volume, samples. Main complications following pars plana vitrectomy include glaucoma, elevated intraocular pressure, corneal and cystoid macular oedema, persistent inflammation, retinal detachment, endophthalmitis and vitreous haemorrhage.5 However, the progressive miniaturization of vitreotomes should reduce iatrogenic risks. Our study showed that an interpretable result regarding the presence or absence of lymphoma cells in the vitreous fluid was obtained in 100% of cases. Biological analyses were consistent with the clinical diagnosis in all cases except one, for which the diagnosis of VRL was only based on clinical data and interleukin levels despite the absence of lymphoma cells identified in the vitreous fluid. Depending on the delivery time of the fluid, conventional cytological analysis may be uninterpretable due to cellular lysis. However, this parameter had less impact on FC. Previous studies have already alerted to the need for rapid transport (<60 min).3, 6, 7 Cytology enables the analysis of cell morphology and identification of abnormal lymphocytes. Its sensitivity has been previously evaluated at 31%–87%.6, 8 FC provides precise immunological characterization and the search for B-cell monoclonality, with a higher sensitivity (82%–100%).6 In our study, sensitivity was 81.8% for cytological analysis, among cases with interpretable cytology, and 91.7% for FC. The specificity was 100% for both technics. These two analyses are inextricably linked, as the interpretation of each depends in part on the results of the other. Moreover, cytology, although less sensitive, enables the assessment of the representativeness of the fluid. In addition, it is essential to optimise these low-volume samples. Analysis of MYD88 p.L265P status can be carried out after lymphoma cells have been recovered by scraping May–Grünwald–Giemsa (MGG)-stained slides. Thus, material prepared for cytological analysis can be used and adapted for molecular biology analyses.9 Together, these techniques enhance the sensitivity and specificity of diagnosis and could be proposed as first-line techniques for diagnosing VRL. Our results are consistent with studies proposing the combination of cytology and FC as the gold standard analysis to diagnose VRL, as first-line diagnostic tools.3, 10 Nevertheless, the small cohort of 19 patients remains a limitation of our study, even though it is a very rare disease. Moreover, a wider range of analyses on the same vitreous fluid would be of interest, including IL assays and MYD88 mutation testing, on a larger number of samples. This new study would emphasise the number of false-negative cases in cytology and FC that would be identified by these additional analyses. In our study, IL-10/IL-6 ratio greater than 1 in aqueous humour supported the need for a diagnostic vitrectomy. For patient no. 2, elevated levels of IL-10 and IL-6, in association with typical lesions on OCT, helped with the lymphoma diagnosis. Thus, for ambiguous cases, measurement of ILs in the vitreous fluid and/or aqueous humour may assist in diagnosis. However, these results can be difficult to interpret and the reference values are not the same depending on the sample type.3 Emerging molecular biology technologies, such as MYD88 p.L265P profiling, are promising and should be incorporated into the diagnostic approach for cases where first-line analyses are negative or not interpretable, particularly in the presence of strong clinical suspicion.10-12 The sensitivity of p.L265P MYD88 mutation detection ranges from 67% to 91%, using polymerase chain reaction (PCR) with Sanger sequencing or high resolution melting analysis.6 New technologies such as digital-droplet PCR may improve this performance.13 Nevertheless, these tests are more expensive and less accessible. Early and accurate diagnosis of VRL is crucial for optimal patient management. Our study shows that the combination of cytological and FC analyses is a pertinent tool to identify lymphoma cells in vitreous fluids. We also highlighted that these techniques and this sample are adequate for the diagnosis of ocular localization of CNS lymphoma. This procedure may thus help to avoid cerebral biopsies, which are more invasive, for diagnosing CNS lymphoma with vitreoretinal involvement. Our study demonstrated that the combination of cytology and flow cytometry is effective in detecting lymphoma cells and remains the gold standard analysis to diagnose VRL, as first-line diagnostic tools. The vitreous fluid is the sample of choice for the diagnosis of VRL and ocular localization of CNS lymphoma. This precious and fragile sample must be analysed as quickly as possible for optimal diagnostic yield. Clinical management and follow-up: PG, AC, FM, PRa, JC, TD, LS, VH, OC, ET. Biological analysis and cytological analysis: DF, PRo, EK. Flow cytometry: SF. IL6/IL10: PhR. Sanger sequencing: AA, IN. Writing—original draft: EK, DF, PRo. Writing—review and editing: All authors discussed the results and contributed to the final manuscript. The authors have nothing to report. This study was not supported by any sponsor or funder. No conflicts of interest declared. This study protocol was recorded and approved by the local ethical committee of Assistance Publique—Hôpitaux de Marseille (number PADS23-282). The data for this study were obtained from patients managed at Assistance Publique des Hôpitaux de Marseille for their diagnosis and treatment. This study is a single-institution non-interventional investigation using retrospective data obtained during the initial period of diagnosis workup of patients. In accordance with French law, consent was obtained through an information sheet, and no one declined participation. Not applicable. Data S1. Table S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
No takes yet. Share an insight, caveat, or question.
Gascon et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: