Citations

Abstract Glioblastoma, IDH-wildtype (GBM), and primary central nervous system lymphoma (PCNSL) are two aggressive brain tumors with distinct treatment strategies but overlapping imaging features on MRI. Accurate preoperative differentiation remains a clinical challenge, due to their overlapping imaging features, even when advanced MRI techniques are employed. This study explored deep learning approaches leveraging multi-sequence MRI to distinguish GBM from PCNSL. Models were developed utilizing 146 PCNSL cases, which were matched to146 GBM by age, sex and year of diagnosis. A 3D-DenseNet121 convolutional neural network was trained using two analytical strategies: (i) loss-based model selection and (ii) ensemble prediction using five-fold cross-validation. Model performance was evaluated on an independent validation set of 256 GBM and 37 PCNSL from the same institution as well as 36 PCNSL cases from an external institution. All MRIs were preprocessed using the FeTS pipeline. Using the independent validation cohort, area under the receiver operating characteristic curve (AUC) and 95% confidence intervals (CI) were computed overall and stratified by demographic characteristics. Using both T1 post-contrast (T1Gd) and T2-weighted images, the loss-based model achieved AUC=0.83 (95% CI: 0.77-0.88) and the ensemble model yielded an AUC=0.82 (95% CI: 0.76-0.89). Model performance was similar across males and females; however, model performance was better for subjects older than 50 years of age versus younger than 50. When using only T1Gd images, the loss-based model dropped to AUC=0.55 (95% CI: 0.48-0.63), whereas the ensemble model had AUC=0.79 (95% CI: 0.72-0.85). Models trained with only T2 images achieved AUCs of 0.67 (95% CI: 0.60-0.73) (loss-based) and 0.75 (95% CI: 0.69-0.82) (ensemble). These results highlight the added value of the multi-sequence input. Our results confirm the feasibility of using automated deep learning models to differentiate GBM from PCNSL. Notably, the model generalized well to external MRI data, emphasizing its potential utility as a diagnostic aid in real-world clinical workflows.

Abstract Rural health disparities in the United States remain poorly understood, in part due to inconsistent definitions of “rurality” and the failure to integrate both geographic and socially constructed dimensions of place. This commentary draws from insights shared during the panel “Addressing Diversity in Cancer Prevention, Control, and Survivorship Within Rural Communities” at the 2024 Transforming Community and Rural Healthcare (TCRH) conference. Panelists emphasized the limitations of current rural-urban classification systems, which often overlook the intersectional identities, cultural heterogeneity, and lived experiences of rural populations. They advocated for integrating community-based participatory research (CBPR) and identity-based approaches to define rurality more holistically. Such methods prioritize self-identification, social determinants of health (SDOH), and localized knowledge, allowing for a more accurate assessment of cancer disparities and the development of equitable interventions. The commentary outlines how geophysical and census-based metrics, although convenient, can misrepresent rural communities and inadvertently perpetuate disparities in cancer care access, outcomes, and policy. The authors argue for an identity- and community-centered conceptualization of rurality that is dynamic, inclusive, and responsive to the complexity of rural life.

ABSTRACT Newly diagnosed patients with non‐Hodgkin lymphoma (NHL) often have a history of other diseases, and these comorbidities can impact patient treatment and management options, as well as overall survival (OS). We developed the Lymphoma Epidemiology of Outcomes (LEO) comorbidity index (LCI) as a sum of 10 comorbidities adapted from the Self‐Report‐Generated Charlson Comorbidity Index (SRG‐CCI) for use in the LEO cohort, a national prospective study of newly diagnosed NHL. Of the 5502 participants with self‐reported comorbidity data, 3107 (56.4%) were male and the mean age at diagnosis was 60.9 years (range, 18–99 years). The LCI ranged from 0 to 6, with 48.6% having 0, 30.2% having 1, 21.2% having 2 or more comorbidities. With a median follow‐up of 62.4 months among surviving participants, 2099 patients had an event and 1219 died. Continuous LCI similarly predicted both 1‐year mortality (c‐statistic = 0.654) and OS (c‐statistic = 0.655), while it showed a weaker but still statistically significant predictive ability for lymphoma‐specific (c‐statistics = 0.617) and event‐free (c‐statistic = 0.574) survival. Participants with 1 (HR = 1.21, 95% CI 1.05–1.39) and 2+ (HR = 1.80, 95% CI 1.56–2.08) comorbidities had inferior OS compared to those with no comorbidities after adjustment for age and sex (c‐statistic = 0.654), and performance strengthened after adjustment for the International Prognostic Index (c‐statistic = 0.672). LCI predicted OS most strongly in marginal zone (c‐statistics = 0.748) and weakest in T‐cell (c‐statistic = 0.579) lymphoma. The cumulative incidence of death due to lymphoma, lymphoma treatment, and other causes all increased with increasing comorbidities, with the greatest increase observed for death due to other causes. The LCI performs comparable to other published comorbidity indices, supporting its use in the LEO cohort to better model real‐world outcomes and more generally providing an approach to implementing comorbidity indices in cancer survivorship cohorts.

Abstract Although follicular lymphoma (FL) typically follows an indolent course, patients with FL who experience early events, such as transformation or progression, have increased risk of death related to lymphoma. The FL24Cx is an algorithm based on a 45-target gene expression profiling (GEP) assay, which was developed and trained using 265 formalin-fixed, paraffin-embedded tissue samples on a reliable platform to predict, at the time of diagnosis, whether a patient will experience an event within 24 months. The modeling also confirmed and relied upon previously reported synergy between immune response (IR) gene expression signatures IR1 and IR2. Once locked, the 5-factor logistic regression FL24Cx model was independently validated in a retrospectively assessed cohort of 232 patients from 2 immunochemotherapy-treated arms of SWOG Cancer Research Network S0016 phase 3 clinical trial, in which it assigned 169 patients to the low-risk group with 29 events before 24 months (17.2%) and 63 patients to the high-risk group with 24 events before 24 months (38.1%). The relative risk of an event within 24 months after registration among patients who were classified into the high-risk group relative to patients who were classified into the low-risk group was 2.2 (95% confidence interval, 1.41 to 3.51). An up-front GEP biomarker, such as the FL24Cx, rigorously validated in a clinical laboratory and with a clinically relevant turnaround time, could identify and steer enrollment of patients at high risk for early events in clinical trials, thus enabling timely interpretation of such trials and increasing the pace of innovation.

Abstract Obesity has been associated with non‐Hodgkin lymphoma (NHL), but the evidence is inconclusive. We examined the association between genetically determined adiposity and four common NHL subtypes: diffuse large B‐cell lymphoma (DLBCL), follicular lymphoma, chronic lymphocytic leukemia, and marginal zone lymphoma, using eight genome‐wide association studies of European ancestry ( N = 10,629 cases, 9505 controls) and constructing polygenic scores for body mass index (BMI), waist‐to‐hip ratio (WHR), and waist‐to‐hip ratio adjusted for BMI (WHRadjBMI). Higher genetically determined BMI was associated with an increased risk of DLBCL [odds ratio (OR) per standard deviation (SD) = 1.18, 95% confidence interval (95% CI): 1.05–1.33, p = .005]. This finding was consistent with Mendelian randomization analyses, which demonstrated a similar increased risk of DLBCL with higher genetically determined BMI (OR per SD = 1.12, 95% CI: 1.02–1.23, p = .03). No significant associations were observed with other NHL subtypes. Our study demonstrates a positive link between a genetically determined BMI and an increased risk of DLBCL, providing additional support for increased adiposity as a risk factor for DLBCL.

Oncogene expression can cause replication stress (RS), leading to DNA double-strand breaks (DSBs) that require repair through pathways such as homologous recombination, nonhomologous end-joining, and microhomology-mediated end-joining (MMEJ). Cyclin D1 (encoded by CCND1 ) is a well-known oncoprotein overexpressed in cancer; however, its role in RS is unknown. Using mantle cell lymphoma (MCL) as a naturally occurring model of cyclin D1 overexpression, we examined the impact of cyclin D1 on RS and DSB repair mechanisms. Cyclin D1 overexpression elevated RS, increased DNA damage, especially during mitosis, and caused specific upregulation of MMEJ. Furthermore, cyclin D1 activated polymerase theta ( POLQ ) transcription by binding its promoter loci, driving POLΘ-mediated MMEJ that is essential to withstand cyclin D1–induced RS. Moreover, concurrent ATM deficiency further intensified RS, enhanced POLQ expression, and heightened reliance on MMEJ-mediated DNA damage repair. Consequently, inhibition of POLΘ in cyclin D1–overexpressed settings further exacerbated RS, causing single-strand DNA gap accumulations and chromosomal instability, ultimately leading to apoptosis, an effect amplified in ATM-deficient cells. Targeting MMEJ via POLΘ inhibition is therefore an effective strategy in the context of cyclin D1 overexpression and ATM deficiency and may provide a unique therapeutic approach for treating MCL and other malignancies characterized by similar alterations.