What Changed in CNS5? A Mini-Review on General Changes and Adult Diffuse Gliomas.

The fifth edition of the WHO classification of tumors of the central nervous system (WHO CNS5) was published in 2021 which is the sixth version of the international standard for the diagnostics of CNS tumors. Regular updates of the consortium to inform molecular and practical approaches to CNS tumor taxonomy (cIMPACT-NOW) shaped the WHO CNS5 which continues the trend of incorporating the molecular characteristics of tumors into the histological and immunohistochemical findings. The various updates can be classified into general changes across all tumors and specific changes within the tumor groups. This mini-review highlights the general changes and the major changes in adult diffuse gliomas.

RésuméLa cinquième édition de la classification OMS des tumeurs du système nerveux central (WHO CNS5) a été publiée en 2021, soit la sixième version de la norme internationale pour le diagnostic des tumeurs du SNC. Mises à jour régulières du consortium pour informer les chercheurs moléculaires et pratiques les approches de la taxonomie des tumeurs du SNC (cIMPACT-NOW) ont façonné le CNS5 de l'OMS, qui poursuit la tendance à incorporer la méthodologie moléculaire caractéristiques des tumeurs dans les résultats histologiques et immunohistochimiques. Les différentes mises à jour peuvent être classées en modifications générales dans toutes les tumeurs et les changements spécifiques au sein des groupes de tumeurs. Cette mini-revue met en évidence les changements généraux et les changements majeurs dans gliomes diffus de l'adulte.

NTRK-fused central nervous system tumours: clinicopathological and genetic insights and response to TRK inhibitors.

Background Neurotrophic tropomyosin receptor kinase (NTRK) gene fusions are found in 1% of gliomas across children and adults. TRK inhibitors are promising therapeutic agents for NTRK-fused gliomas because they are tissue agnostic and cross the blood-brain barrier (BBB). Methods We investigated twelve NGS-verified NTRK-fused gliomas from a single institute, Seoul National University Hospital. Results The patient cohort included six children (aged 1-15 years) and six adults (aged 27-72 years). NTRK2 fusions were found in ten cerebral diffuse low-grade and high-grade gliomas (DLGGs and DHGGs, respectively), and NTRK1 fusions were found in one cerebral desmoplastic infantile ganglioglioma and one spinal DHGG. In this series, the fusion partners of NTRK2 were HOOK3, KIF5A, GKAP1, LHFPL3, SLMAP, ZBTB43, SPECC1L, FKBP15, KANK1, and BCR, while the NTRK1 fusion partners were TPR and TPM3. DLGGs tended to harbour only an NTRK fusion, while DHGGs exhibited further genetic alterations, such as TERT promoter/TP53/PTEN mutation, CDKN2A/2B homozygous deletion, PDGFRA/KIT/MDM4/AKT3 amplification, or multiple chromosomal copy number aberrations. Four patients received adjuvant TRK inhibitor therapy (larotrectinib, repotrectinib, or entrectinib), among which three also received chemotherapy (n = 2) or proton therapy (n = 1). The treatment outcomes for patients receiving TRK inhibitors varied: one child who received larotrectinib for residual DLGG maintained stable disease. In contrast, another child with DHGG in the spinal cord experienced multiple instances of tumour recurrence. Despite treatment with larotrectinib, ultimately, the child died as a result of tumour progression. An adult patient with glioblastoma (GBM) treated with entrectinib also experienced tumour progression and eventually died. However, there was a successful outcome for a paediatric patient with DHGG who, after a second gross total tumour removal followed by repotrectinib treatment, showed no evidence of disease. This patient had previously experienced relapse after the initial surgery and underwent autologous peripheral blood stem cell therapy with carboplatin/thiotepa and proton therapy. Conclusions Our study clarifies the distinct differences in the pathology and TRK inhibitor response between LGG and HGG with NTRK fusions.

[Clinicopathologic characteristics and survival analysis of primary large B-cell lymphoma of the central nervous system].

Objective: To retrospectively analyze the clinical and pathologic characteristics, response to treatment, survival, and prognosis of patients with primary large B-cell lymphoma of the central nervous system (PCNSLBCL) . Methods: Clinical and pathologic data of 70 patients with PCNSLBCL admitted to Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine from December 2010 to November 2022 were collected for retrospective analysis. Survival analysis was performed using the Kaplan-Meier method and log-rank test, and prognosis analysis was conducted using the Cox proportional hazards model. Results: Among 70 patients with PCNSLBCL, complete remission (CRs) were achieved in 49 (70.0% ) and partial remission in 4 (5.7% ) after the first-line induction therapy; the overall remission rate was 75.7%. The 2-year progression-free survival (PFS) rate was 55.8% and the median progression-free survival (mPFS) time was 35.9 months, whereas the 2-year overall survival (OS) rate was 79.1% with a median OS time not reached. After CR induced by first-line therapy, cumulative incidence of relapse (CIR) was lower in patients who had received auto-HSCT than in those who had not received consolidation therapy (P=0.032), whose 2-year PFS rate was 54.4% and mPFS time was 35.9 months; comparatively, the 2-year PFS rate in patients having received oral maintenance of small molecule drugs reached 84.4% with a mPFS time of 79.5 months (P=0.038). Multivariant analysis demonstrated that Class 3 in the Memorial Sloan-Kettering Cancer Center (MSKCC) prognostic model is an independent adverse prognostic factor of OS in patients with PCNSLBCL (HR=3.127, 95% CI 1.057-9.253, P=0.039) . Conclusions: In patients with PCNSLBCL achieving CR after the first-line induction therapy, auto-HSCT as consolidation therapy would lead to a decreased CIR, and PFS time could be prolonged by oral maintenance of small molecule drugs. Class 3 MSKCC prognostic model is independently associated with poorer OS.

Primary diffuse large B-cell lymphoma of the central nervous system identified with CSF biomarkers.

BACKGROUND: Diagnosis of primary diffuse large B-cell lymphoma of the central nervous system (PCNSL) is challenging and often delayed. MRI imaging, CSF cytology and flow cytometry have a low sensitivity and even brain biopsies can be misleading. We report three cases of PCNSL with various clinical presentation and radiological findings where the diagnosis was suggested by novel CSF biomarkers and subsequently confirmed by brain biopsy or autopsy. CASE PRESENTATIONS: The first case is a 79-year-old man with severe neurocognitive dysfunction and static ataxia evolving over 5 months. Brain MRI revealed a nodular ventriculitis. An open brain biopsy was inconclusive. The second case is a 60-year-old woman with progressive sensory symptoms in all four limbs, evolving over 1 year. Brain and spinal MRI revealed asymmetric T2 hyperintensities of the corpus callosum, corona radiata and corticospinal tracts. The third case is a 72-year-old man recently diagnosed with primary vitreoretinal lymphoma of the right eye. A follow-up brain MRI performed 4 months after symptom onset revealed a T2 hyperintense fronto-sagittal lesion, with gadolinium uptake and perilesional edema. In all three cases, CSF flow cytometry and cytology were negative. Mutation analysis on the CSF (either by digital PCR or by next generation sequencing) identified the MYD88 L265P hotspot mutation in all three cases. A B-cell clonality study, performed in case 1 and 2, identified a monoclonal rearrangement of the immunoglobulin light chain lambda (IGL) and kappa (IGK) gene. CSF CXCL-13 and IL-10 levels were high in all three cases, and IL-10/IL-6 ratio was high in two. Diagnosis of PCNSL was later confirmed by autopsy in case 1, and by brain biopsy in case 2 and 3. CONCLUSIONS: Taken together, 5 CSF biomarkers (IL-10, IL-10/IL-6 ratio, CXCL13, MYD88 mutation and monoclonal IG gene rearrangements) were strongly indicative of a PCNSL. Using innovative CSF biomarkers can be sensitive and complementary to traditional CSF analysis and brain biopsy in the diagnosis of PCNSL, potentially allowing for earlier diagnosis and treatment.

Patient's HLA Genotype Is Associated With the Risk of Central Nervous System Dissemination and Clinical Disease Presentation in Diffuse Large B-cell Lymphoma.

BACKGROUND/AIM: Central nervous system (CNS) involvement in aggressive B-cell lymphoma, either as a primary or secondary event to systemic disease, portends a poor prognosis. This study sought to identify patients at high risk for CNS relapse by analyzing their human leukocyte antigen (HLA) genotypes. PATIENTS AND METHODS: We retrospectively examined the HLA genotypes of 164 patients with systemic lymphoma, primary CNS lymphoma, and CNS relapse of systemic lymphoma. Patient records were analyzed, and HLA typing was performed by the Finnish Red Cross Blood Service. After excluding patients who received CNS prophylaxis, 131 patients were included in the final analysis. RESULTS: A strong association was found between the HLA-A*31 genotype and CNS disease (p=0.001). Additionally, various HLA genotypes were linked to lactate dehydrogenase levels, extranodal disease, International Prognostic Index score, and disease stage. CONCLUSION: The patient's genetic constitution, rather than solely disease-related factors, plays a role in the tropism of lymphoma for the CNS. If confirmed in a larger study, defining the HLA genotype of a lymphoma patient could provide valuable information for predicting CNS relapse.

Neuroradiological, genetic and clinical characteristics of histone H3 K27-mutant diffuse midline gliomas in the Kansai Molecular Diagnosis Network for CNS Tumors (Kansai Network): multicenter retrospective cohort.

This study aims to elucidate the clinical and molecular characteristics, treatment outcomes and prognostic factors of patients with histone H3 K27-mutant diffuse midline glioma. We retrospectively analyzed 93 patients with diffuse midline glioma (47 thalamus, 24 brainstem, 12 spinal cord and 10 other midline locations) treated at 24 affiliated hospitals in the Kansai Molecular Diagnosis Network for CNS Tumors. Considering the term "midline" areas, which had been confused in previous reports, we classified four midline locations based on previous reports and anatomical findings. Clinical and molecular characteristics of the study cohort included: age 4-78 years, female sex (41%), lower-grade histology (56%), preoperative Karnofsky performance status (KPS) scores ≥ 80 (49%), resection (36%), adjuvant radiation plus chemotherapy (83%), temozolomide therapy (76%), bevacizumab therapy (42%), HIST1H3B p.K27M mutation (2%), TERT promoter mutation (3%), MGMT promoter methylation (9%), BRAF p.V600E mutation (1%), FGFR1 mutation (14%) and EGFR mutation (3%). Median progression-free and overall survival time was 9.9 ± 1.0 (7.9-11.9, 95% CI) and 16.6 ± 1.4 (13.9-19.3, 95% CI) months, respectively. Female sex, preoperative KPS score ≥ 80, adjuvant radiation + temozolomide and radiation ≥ 50 Gy were associated with favorable prognosis. Female sex and preoperative KPS score ≥ 80 were identified as independent good prognostic factors. This study demonstrated the current state of clinical practice for patients with diffuse midline glioma and molecular analyses of diffuse midline glioma in real-world settings. Further investigation in a larger population would contribute to better understanding of the pathology of diffuse midline glioma.

Clinical features, MRI, molecular alternations, and prognosis of astrocytoma based on WHO 2021 classification of central nervous system tumors: A single-center retrospective study.

BACKGROUND: The diagnosis of glioma has advanced since the release of the WHO 2021 classification with more molecular alterations involved in the integrated diagnostic pathways. Our study aimed to present our experience with the clinical features and management of astrocytoma, IDH mutant based on the latest WHO classification. METHODS: Patients diagnosed with astrocytoma, IDH-mutant based on the WHO 5th edition classification of CNS tumors at our center from January 2009 to January 2022 were included. Patients were divided into WHO 2-3 grade group and WHO 4 grade group. Integrate diagnoses were retrospectively confirmed according to WHO 2016 and 2021 classification. Clinical and MRI characteristics were reviewed, and survival analysis was performed. RESULTS: A total of 60 patients were enrolled. 21.67% (13/60) of all patients changed tumor grade from WHO 4th edition classification to WHO 5th edition. Of these, 21.43% (6/28) of grade II astrocytoma and 58.33% (7/12) of grade III astrocytoma according to WHO 4th edition classification changed to grade 4 according to WHO 5th edition classification. Sex (p = 0.042), recurrent glioma (p = 0.006), and Ki-67 index (p < 0.001) of pathological examination were statistically different in the WHO grade 2-3 group (n = 27) and WHO grade 4 group (n = 33). CDK6 (p = 0.004), FGFR2 (p = 0.003), and MYC (p = 0.004) alterations showed an enrichment in the WHO grade 4 group. Patients with higher grade showed shorter mOS (mOS = 75.9 m, 53.6 m, 26.4 m for grade 2, 3, and 4, respectively, p = 0.01). CONCLUSIONS: Patients diagnosed as WHO grade 4 according to the 5th edition WHO classification based on molecular alterations are more likely to have poorer prognosis. Therefore, treatment should be tailored to their individual needs. Further research is needed for the management of IDH-mutant astrocytoma is needed in the future.

BRAF V600E-mutant colorectal cancer with CNS metastases treated successfully with encorafenib, binimetinib and cetuximab.

This report describes a case of BRAF V600E-mutated colorectal cancer with CNS metastases in which treatment with encorafenib, binimetinib and cetuximab was effective. There is limited information on the ability of encorafenib, binimetinib and cetuximab to enter the CNS.The patient was a 53-year-old man was diagnosed with ascending colon cancer (cT3N3M1c stage IVc). BRAF V600E mutation was confirmed. FOLFOX was started, but CNS metastases soon appeared. Encorafenib, binimetinib and cetuximab were administered and had a favorable effect on the CNS lesions. The patient initially responded well, but his disease progressed 2 months later. Further research is needed to improve management strategies for BRAF V600E-mutated colorectal cancer with CNS metastases.

[Box: see text].

Novel insights toward diagnosis and treatment of glioneuronal and neuronal tumors in young adults.

Aim: Glioneuronal and neuronal tumors are rare primary central nervous system malignancies with heterogeneous features. Due to the rarity of these malignancies diagnosis and treatment remains a clinical challenge. Methods: Here we performed a narrative review aimed to investigate the principal issues concerning the diagnosis, pathology, and clinical management of glioneuronal tumors. Results: Diagnostic criteria have been recently overturned thanks to a better characterization on a histological and molecular biology level. The study of genomic alterations occurring within these tumors has allowed us to identify potential therapeutic targets including BRAF, FGFR, and PDGFRA. Conclusion: Techniques allowing molecular sequencing DNA methylation assessment of the disease are essential diagnostic tools. Targeting agents should be included in the therapeutic armamentarium after loco-regional treatment failure.

[Box: see text].

Primary central nervous system lymphomas in immunocompromised patients require specific response criteria.

PURPOSE: Immunosuppression is a well-established risk factor for primary central nervous system lymphomas (PCNSLs), which present in this context distinct radiological characteristics. Our aim was to describe the radiological evolution of treated PCNSL in immunocompromised patients and suggest adapted MRI response criteria. METHODS: We conducted a multicenter retrospective study of patients from the French LOC, K-Virogref and CANCERVIH network databases and enrolled adult immunocompromised patients with newly diagnosed PCNSL. RESULTS: We evaluated the baseline, intermediate, end-of-treatment and follow-up MRI data of 31 patients (9 living with HIV, 16 with solid organ transplantation and 6 with an autoimmune disease under chronic immunosuppressive therapy). At baseline, 23/30 (77%) patients had necrotic lesions with ring enhancement and 28% of the lesions were hemorrhagic. At the end of the first-line treatment, 12/28 (43%) patients could not be classified according to the IPCG criteria. Thirteen of 28 (46%) patients still harbored contrast enhancement, and 11/28 (39%) patients had persistent large necrotic lesions with a median diameter of 15 mm. These aspects were not associated with a pejorative outcome and progressively diminished during follow-up. Six patients relapsed; however, we failed to identify any neuroimaging risk factors on the end-of-treatment MRI. CONCLUSION: In immunocompromised patients, PCNSLs often harbor alarming features on end-of-treatment MRI, with persistent contrast-enhanced lesions frequently observed. However, these aspects seemed to be related to the necrotic and hemorrhagic nature of the lesions and were not predictive of a pejorative outcome. Specific response criteria for this population are thereby proposed.

Comparative Study on the Clinicopathologic and Molecular Characteristics of Primary and Secondary Diffuse Large B-cell Lymphoma in the Central Nervous System.

BACKGROUND/AIM: Diffuse large B-cell lymphoma of the central nervous system (CNS-DLBCL) is an aggressive B-cell lymphoma with clinical and molecular heterogeneity. Primary CNS-DLBCL (PCNSL) affects the brain, eyes, leptomeninges, or spinal cord without systemic involvement. Secondary CNS-DLBCL (SCNSL) manifests concurrently with systemic lymphoma or as an isolated CNS relapse with poor prognosis. MATERIALS AND METHODS: Next-generation sequencing (NGS) was used to identify genomic alterations in 32 PCNSL and 9 SCNSL cases. Single nucleotide variants and copy number variations in addition to the clinicopathologic data and proposed risk predictive values were compared to aid in diagnostic differentiation between the two types of lymphomas. RESULTS: The MCD genotype, characterized by mutations in MYD88 and CD79B, is the most common alteration in PCNSL and is associated with lower survival rates. The frequency of MYD88 mutation was significantly higher in PCNSL compared to SCNSL (75.0% vs. 33.3%; p=0.042). Recurrent copy number loss of 6p21 occurred in 56.1% of cases, more often in PCNSL (65.6%) than in SCNSL (22.2%) (p=0.028). Diagnostic positive predictive values (PPV) of MYD88 mutation and 6p21 loss for PCNSL were 89% and 91%, respectively. PPV of both alterations was 93% for the diagnosis of PCNSL. CONCLUSION: MYD88 mutation and 6p21 loss were significantly higher in PCNSL than in SCNSL, and novel risk prediction models based on these distinct genomic profiles can aid in the clinical differentiation of PCNSL and SCNSL.

Could Cerebral Inflammatory Lesions be the Cellular Origin of Primary Central Nervous System Lymphoma?

INTRODUCTION: Primary central nervous system lymphoma (PCNSL) presents a diagnostic enigma due to the inherent absence of lymphoid tissue in the central nervous system (CNS). The hypothesis posits that lymphocytes infiltrating the CNS during inflammatory responses could represent a cellular source for PCNSL, challenging traditional understandings of its etiology. PATIENT CONCERNS: In 2 illustrative cases, patients presented with neurological symptoms initially misdiagnosed as encephalitis and demyelinating disease, respectively. These diagnoses were established based on clinical assessments and initial biopsy findings. DIAGNOSIS: Subsequent biopsies, conducted months after the first signs of disease, confirmed the diagnosis of PCNSL in both patients. Identifying CD20-positive tumor cells was pivotal, indicating a B-cell lymphoma origin. INTERVENTIONS: Treatment strategies included high-dose methotrexate chemotherapy for both patients. In addition, the second patient underwent adjuvant whole-brain radiotherapy after the chemotherapy regimen. OUTCOMES: The therapeutic approach significantly reduced tumor size in both cases, with no evidence of recurrence observed during the follow-up period. This outcome underscores the potential efficacy of the chosen interventions. CONCLUSION: In response to inflammatory lesions, lymphocyte infiltration into the CNS may serve as a pivotal origin for tumor cells in PCNSL. These cases highlight the complexity of diagnosing CNS disorders and suggest that various forms of encephalitis in the early stages could influence the prognosis of lymphoma. This insight into the cellular origins and treatment responses of PCNSL contributes to a broader understanding of its pathophysiology and management.

Isolated central nervous system relapse in two adolescents with primary mediastinal large B-cell lymphoma after treatment with R-DA-EPOCH.

The addition of rituximab to standard regimens for primary mediastinal large B-cell lymphoma (PMBCL) has significantly improved overall survival. However, the optimal management of isolated central nervous system (CNS) relapse and role of CNS prophylaxis remains undefined. We present cases of two adolescents with PMBCL who developed isolated CNS relapses. While isolated CNS relapse may be managed with high-dose chemotherapy and autologous stem cell transplant with or without CNS radiotherapy, review of these cases and the literature highlight the need for further work to define risk factors for CNS relapse, and identify patients who may benefit from CNS prophylaxis.

Distinct roles of TREM2 in central nervous system cancers and peripheral cancers.

Glioblastomas (GBM) are incurable central nervous system (CNS) cancers characterized by substantial myeloid cell infiltration. Whether myeloid cell-directed therapeutic targets identified in peripheral non-CNS cancers are applicable to GBM requires further study. Here, we identify that the critical immunosuppressive target in peripheral cancers, triggering receptor expressed on myeloid cells-2 (TREM2), is immunoprotective in GBM. Genetic or pharmacological TREM2 deficiency promotes GBM progression in vivo. Single-cell and spatial sequencing reveals downregulated TREM2 in GBM-infiltrated myeloid cells. TREM2 negatively correlates with immunosuppressive myeloid and T cell exhaustion signatures in GBM. We further demonstrate that during GBM progression, CNS-enriched sphingolipids bind TREM2 on myeloid cells and elicit antitumor responses. Clinically, high TREM2 expression in myeloid cells correlates with better survival in GBM. Adeno-associated virus-mediated TREM2 overexpression impedes GBM progression and synergizes with anti-PD-1 therapy. Our results reveal distinct functions of TREM2 in CNS cancers and support organ-specific myeloid cell remodeling in cancer immunotherapy.

A multicenter proof-of-concept study on deep learning-based intraoperative discrimination of primary central nervous system lymphoma.

Accurate intraoperative differentiation of primary central nervous system lymphoma (PCNSL) remains pivotal in guiding neurosurgical decisions. However, distinguishing PCNSL from other lesions, notably glioma, through frozen sections challenges pathologists. Here we sought to develop and validate a deep learning model capable of precisely distinguishing PCNSL from non-PCNSL lesions, especially glioma, using hematoxylin and eosin (H&E)-stained frozen whole-slide images. Also, we compared its performance against pathologists of varying expertise. Additionally, a human-machine fusion approach integrated both model and pathologic diagnostics. In external cohorts, LGNet achieved AUROCs of 0.965 and 0.972 in distinguishing PCNSL from glioma and AUROCs of 0.981 and 0.993 in differentiating PCNSL from non-PCNSL lesions. Outperforming several pathologists, LGNet significantly improved diagnostic performance, further augmented to some extent by fusion approach. LGNet's proficiency in frozen section analysis and its synergy with pathologists indicate its valuable role in intraoperative diagnosis, particularly in discriminating PCNSL from glioma, alongside other lesions.

Clinical, pathological, and molecular features of central nervous system tumors with BCOR internal tandem duplication.

Central nervous system tumor with BCOR internal tandem duplication (CNS tumor with BCOR-ITD) constitutes a molecularly distinct entity, characterized by internal tandem duplication within exon 15 of the BCOR transcriptional co-repressor gene (BCOR-ITD). The current study aimed to elucidate the clinical, pathological, and molecular attributes of CNS tumors with BCOR-ITD and explore their putative cellular origin. This study cohort comprised four pediatric cases, aged 23 months to 13 years at initial presentation. Magnetic resonance imaging revealed large, well-circumscribed intra-CNS masses localized heterogeneously throughout the CNS. Microscopically, tumors were composed of spindle to ovoid cells, exhibiting perivascular pseudorosettes and palisading necrosis, but lacking microvascular proliferation. Immunohistochemical staining showed diffuse tumor cell expression of BCOR, CD56, CD99, vimentin, and the stem cell markers PAX6, SOX2, CD133 and Nestin, alongside focal positivity for Olig-2, S100, SOX10, Syn and NeuN. Molecularly, all cases harbored BCOR-ITDs ranging from 87 to 119 base pairs in length, including one case with two distinct ITDs. Notably, the ITDs were interrupted by unique 1-3 bp insertions in all cases. In summary, CNS tumors with BCOR-ITD exhibit characteristic clinical, pathological, and molecular features detectable through BCOR immunohistochemistry and confirmatory molecular analyses. Their expression of stem cell markers raises the possibility of an origin from neuroepithelial stem cells rather than representing true embryonal neoplasms.

The CNS relapse in T-cell lymphoma index predicts CNS relapse in patients with T- and NK-cell lymphomas.

ABSTRACT: Little is known about risk factors for central nervous system (CNS) relapse in mature T-cell and natural killer cell neoplasms (MTNKNs). We aimed to describe the clinical epidemiology of CNS relapse in patients with MTNKN and developed the CNS relapse In T-cell lymphoma Index (CITI) to predict patients at the highest risk of CNS relapse. We reviewed data from 135 patients with MTNKN and CNS relapse from 19 North American institutions. After exclusion of leukemic and most cutaneous forms of MTNKNs, patients were pooled with non-CNS relapse control patients from a single institution to create a CNS relapse-enriched training set. Using a complete case analysis (n = 182), including 91 with CNS relapse, we applied a least absolute shrinkage and selection operator Cox regression model to select weighted clinicopathologic variables for the CITI score, which we validated in an external cohort from the Swedish Lymphoma Registry (n = 566). CNS relapse was most frequently observed in patients with peripheral T-cell lymphoma, not otherwise specified (25%). Median time to CNS relapse and median overall survival after CNS relapse were 8.0 and 4.7 months, respectively. We calculated unique CITI risk scores for individual training set patients and stratified them into risk terciles. Validation set patients with low-risk (n = 158) and high-risk (n = 188) CITI scores had a 10-year cumulative risk of CNS relapse of 2.2% and 13.4%, respectively (hazard ratio, 5.24; 95% confidence interval, 1.50-18.26; P = .018). We developed an open-access web-based CITI calculator (https://redcap.link/citicalc) to provide an easy tool for clinical practice. The CITI score is a validated model to predict patients with MTNKN at the highest risk of developing CNS relapse.

Prediction of DNA methylation-based tumor types from histopathology in central nervous system tumors with deep learning.

Precision in the diagnosis of diverse central nervous system (CNS) tumor types is crucial for optimal treatment. DNA methylation profiles, which capture the methylation status of thousands of individual CpG sites, are state-of-the-art data-driven means to enhance diagnostic accuracy but are also time consuming and not widely available. Here, to address these limitations, we developed Deep lEarning from histoPathoLOgy and methYlation (DEPLOY), a deep learning model that classifies CNS tumors to ten major categories from histopathology. DEPLOY integrates three distinct components: the first classifies CNS tumors directly from slide images ('direct model'), the second initially generates predictions for DNA methylation beta values, which are subsequently used for tumor classification ('indirect model'), and the third classifies tumor types directly from routinely available patient demographics. First, we find that DEPLOY accurately predicts beta values from histopathology images. Second, using a ten-class model trained on an internal dataset of 1,796 patients, we predict the tumor categories in three independent external test datasets including 2,156 patients, achieving an overall accuracy of 95% and balanced accuracy of 91% on samples that are predicted with high confidence. These results showcase the potential future use of DEPLOY to assist pathologists in diagnosing CNS tumors within a clinically relevant short time frame.

Patterns of Treatment Failure in Primary Central Nervous System Lymphoma.

OBJECTIVES: Progression of PCNSL remains a challenge with salvage therapies, including the risk of substantial morbidity and mortality. We report patterns of first tumor progression to inform opportunities for improvement. METHODS: This is an institutional retrospective review from 2002 to 2021 of 95 consecutive patients with pathologically confirmed PCNSL, of whom 29 experienced progressive disease. Kaplan-Meier method, log-rank test, and Cox proportional hazard models are used to characterize associations of patient, tumor, and treatment variables with LC, PFS, and patterns of first failure. RESULTS: Most patients were below 65 years old (62%) with KPS >70 (64%) and negative CSF cytology (70%). In 70 patients with MRIs, the median tumor volume was 12.6 mL (range: 0.5 to 67.8 mL). After a median follow-up of 11 months, 1-year PFS was 48% and 1-year LC was 80%. Of the 29 patients with progression, 24% were distant only, 17% were distant and local, and 59% were local only. On MVA, LC was associated with age (HR: 1.08/y, P =0.02), KPS (HR: 0.10, P =0.02), completion of >6 cycles of HD-MTX (HR: 0.10, P <0.01), and use of intrathecal chemotherapy (HR: 0.03, P <0.01). On UVA, local only first failure trended to be increased with >14 mL tumors (OR: 5.06, P =0.08) with 1-year LC 83% (<14 mL) versus 64% (>14mL). There were no significant associations with LC and WBRT ( P =0.37), Rituximab ( P =0.12), or attempted gross total resection ( P =0.72). CONCLUSIONS: Our findings reaffirm the importance of systemic and intrathecal therapies for local control in PCNSL. However, bulky tumors trend to fail locally, warranting further investigation about the role of local therapies or systemic therapy intensification.