Citations

BACKGROUND AND OBJECTIVES: Aggressive resection for primary tumors of the spine are associated with a high rate of adverse events (AEs), but the impact of AEs on patient-reported outcomes (PROs) remains unknown and is critical to the shared decision-making. Our primary objective was to assess the impact of surgical AEs on PROs using an international registry. Assessing the impact on clinical outcomes and identifying risk factors for AEs were our secondary objectives. METHODS: Patients who underwent surgery for a primary spinal tumor were selected through the Primary Tumor Research and Outcomes Network. Our primary outcome was the impact of AEs on PROs at 3 and 12 months after surgery (measured with Spinal Oncology Study Group Outcomes Questionnaire, Short-Form 36, and EuroQol 5 Dimension). We also assessed the impact on clinical outcomes (local control, surgical margins, readmission, reoperation, and mortality). We stratified our results according to severity of AEs, histology, and type of resection. RESULTS: 374 patients met inclusion criteria (219 males/155 females). The mean age of the cohort was 48.7 years. The most frequent histology was chordoma (37.3%) followed by chondrosarcoma (8.8%). Sixty-seven patients (17.9%) experienced at least 1 intraoperative AE and 117 patients (31.3%) had at least 1 postoperative AE within 3 months. Overall, 159 patients (42.5%) experienced AEs. The readmission rate was significantly higher in patients who experienced AEs (Any AE: 10.1% vs no AE: 1.9% within 3 months; P = <0.001). PROs were not significantly affected by AEs in most questionnaires. Local control, risk of reoperation, mortality, and achieving preplanned margins were similar between AE groups. CONCLUSION: The rate of surgical AEs is considerable in this population. Surgical AEs seem to be associated with a higher number of readmissions, but do not seem to result in significant differences in PROs or in a higher risk of reoperation, mortality, and failure to achieve preplanned margins.

Apart from ancestry, personal or environmental covariates may contribute to differences in polygenic score (PGS) performance. We analyzed the effects of covariate stratification and interaction on body mass index (BMI) PGS (PGS BMI ) across four cohorts of European (N = 491,111) and African (N = 21,612) ancestry. Stratifying on binary covariates and quintiles for continuous covariates, 18/62 covariates had significant and replicable R 2 differences among strata. Covariates with the largest differences included age, sex, blood lipids, physical activity, and alcohol consumption, with R 2 being nearly double between best- and worst-performing quintiles for certain covariates. Twenty-eight covariates had significant PGS BMI –covariate interaction effects, modifying PGS BMI effects by nearly 20% per standard deviation change. We observed overlap between covariates that had significant R 2 differences among strata and interaction effects – across all covariates, their main effects on BMI were correlated with their maximum R 2 differences and interaction effects (0.56 and 0.58, respectively), suggesting high-PGS BMI individuals have highest R 2 and increase in PGS effect. Using quantile regression, we show the effect of PGS BMI increases as BMI itself increases, and that these differences in effects are directly related to differences in R 2 when stratifying by different covariates. Given significant and replicable evidence for context-specific PGS BMI performance and effects, we investigated ways to increase model performance taking into account nonlinear effects. Machine learning models (neural networks) increased relative model R 2 (mean 23%) across datasets. Finally, creating PGS BMI directly from GxAge genome-wide association studies effects increased relative R 2 by 7.8%. These results demonstrate that certain covariates, especially those most associated with BMI, significantly affect both PGS BMI performance and effects across diverse cohorts and ancestries, and we provide avenues to improve model performance that consider these effects.

AbstractBackgroundNormative models (NM) of brain metrics based on large, diverse populations offer novel strategies to detect individual brain abnormalities. To create an age‐dependent statistical model of brain microstructure over the human lifespan, we built the largest multi‐site NM of white matter (WM) diffusion tensor imaging (DTI) metrics based on 54,591 subjects. We used state‐of‐the‐art tools to adjust for site‐dependent effects.MethodWe used Hierarchical Bayesian Regression (HBR) to determine the age trajectory and lifespan centile curves (ages 3‐100 years) by merging data from different sections of the lifespan. We analyzed 19 international public datasets with the ENIGMA‐DTI protocol. Mean fractional anisotropy (FA), mean, axial, and radial diffusivity (MD, AxD, RD) were extracted for 21 bilateral ROIs from the JHU‐WM atlas and the whole WM skeleton. Regressions were run with each metric per ROI as a function of age and sex, and the DTI protocol as the batch effect (37 protocols). Z‐scores were derived for 81 patients with Alzheimer’s disease (AD; mean age: 77y±8.4, 46M/35F) and 225 MCI participants (mean age: 75.1y±8.1, 127M/98F) and extreme deviations (Z>|2|) were quantified for each condition. ROI‐wise areas under the ROC curve were calculated to determine the classification accuracy of the Z‐scores.ResultThe average FA for the overall WM skeleton reached its peak value at 27 years, with minima for MD, RD and AxD at 47, 42 and 54 years, respectively. We found extreme FA deviations (Z>2) for MCI (AUC=0.67) and AD (AUC=0.66) in the genu of the corpus callosum, extreme RD deviations (Z<2) in the tapetum (AUC=0.73) and Extreme AxD deviations (Z>2) for MCI in the superior longitudinal fasciculus (AUC=0.65). See Figures 1 & 2.ConclusionNM of brain microstructure with HBR yielded lifespan trajectories for widely‐used DTI metrics. By pooling large scale multi‐site dMRI to create a statistical reference, single subject deviations are detectable across the lifespan for AD and MCI. These NM will be publicly available to the research community.

AbstractBackgroundNormative models (NM) of brain metrics based on large, diverse populations offer novel strategies to detect individual brain abnormalities. To create an age‐dependent statistical model of brain microstructure over the human lifespan, we built the largest multi‐site NM of white matter (WM) diffusion tensor imaging (DTI) metrics based on 54,591 subjects. We used state‐of‐the‐art tools to adjust for site‐dependent effects.MethodWe used Hierarchical Bayesian Regression (HBR) to determine the age trajectory and lifespan centile curves (ages 3‐100 years) by merging data from different sections of the lifespan. We analyzed 19 international public datasets with the ENIGMA‐DTI protocol. Mean fractional anisotropy (FA), mean, axial, and radial diffusivity (MD, AxD, RD) were extracted for 21 bilateral ROIs from the JHU‐WM atlas and the whole WM skeleton. Regressions were run with each metric per ROI as a function of age and sex, and the DTI protocol as the batch effect (37 protocols). Z‐scores were derived for 81 patients with Alzheimer’s disease (AD; mean age: 77y±8.4, 46M/35F) and 225 MCI participants (mean age: 75.1y±8.1, 127M/98F) and extreme deviations (Z>|2|) were quantified for each condition. ROI‐wise areas under the ROC curve were calculated to determine the classification accuracy of the Z‐scores.ResultThe average FA for the overall WM skeleton reached its peak value at 27 years, with minima for MD, RD and AxD at 47, 42 and 54 years, respectively. We found extreme FA deviations (Z>2) for MCI (AUC=0.67) and AD (AUC=0.66) in the genu of the corpus callosum, extreme RD deviations (Z<2) in the tapetum (AUC=0.73) and Extreme AxD deviations (Z>2) for MCI in the superior longitudinal fasciculus (AUC=0.65). See Figures 1 & 2.ConclusionNM of brain microstructure with HBR yielded lifespan trajectories for widely‐used DTI metrics. By pooling large scale multi‐site dMRI to create a statistical reference, single subject deviations are detectable across the lifespan for AD and MCI. These NM will be publicly available to the research community.

Abstract INTRODUCTION Diffusion‐weighted magnetic resonance imaging (dMRI) is sensitive to the microstructural properties of brain tissues and shows great promise in detecting the effects of degenerative diseases. However, many approaches analyze single measures averaged over regions of interest without considering the underlying fiber geometry. METHODS We propose a novel macrostructure‐informed normative tractometry (MINT) framework to investigate how white matter (WM) microstructure and macrostructure are jointly altered in mild cognitive impairment (MCI) and dementia. We compared MINT‐derived metrics with univariate diffusion tensor imaging (DTI) metrics to examine how fiber geometry may impact the interpretation of microstructure. RESULTS In two multisite cohorts from North America and India, we find consistent patterns of microstructural and macrostructural anomalies implicated in MCI and dementia; we also rank diffusion metrics' sensitivity to dementia. DISCUSSION We show that MINT, by jointly modeling tract shape and microstructure, has the potential to disentangle and better interpret the effects of degenerative disease on the brain's neural pathways. Highlights Changes in diffusion tensor imaging metrics may be due to macroscopic changes. Normative models encode normal variability of diffusion metrics in healthy controls. Variational autoencoder applied on tractography can learn patterns of fiber geometry. WM microstructure and macrostructure are modeled with multivariate methods. Transfer learning uses pretraining and fine‐tuning for increased efficiency.

ObjectiveVagus nerve stimulation (VNS) paired with rehabilitation therapy improved motor status compared to rehabilitation alone in the phase III VNS‐REHAB stroke trial, but treatment response was variable and not associated with any clinical measures acquired at baseline, such as age or side of paresis. We hypothesized that neuroimaging measures would be associated with treatment‐related gains, examining performance of regional injury measures versus global brain health measures in parallel with clinical measures.MethodsBaseline magnetic resonance imaging (MRI) scans in the VNS‐REHAB trial were used to derive regional injury measures (extent of injury to corticospinal tract, the primary regional measure; plus extent of injury to precentral gyrus and postcentral gyrus; lesion volume; and lesion topography) and global brain health measures (degree of white matter hyperintensities, the primary global brain measure; plus volumes of cerebrospinal fluid, cortical gray matter, white matter, each thalamus, and total brain). Eight clinical measures assessed at baseline were also evaluated (treatment group, age, race, gender, paretic side, pre‐stroke dominant hand, time since stroke, and baseline Fugl‐Meyer upper extremity score). Bivariate analyses compared each measure with the primary trial end point (change in Fugl‐Meyer upper extremity score from baseline to end of 6 weeks of treatment) across all subjects, with secondary analyses examining trial groups separately.ResultsMRIs were available from 80 patients (age = 59.8 ± 9.5 years, 29 women). Across all patients, less white matter hyperintensities (r = −0.25, p = 0.028) at baseline was associated with larger Fugl‐Meyer score change. In the VNS group, less white matter hyperintensities (r = −0.37, p = 0.018) and larger ipsilesional thalamus volume (r = 0.33, p = 0.046) were each associated with larger Fugl‐Meyer score change. Analysis of covariance (ANCOVA) analyses tested the interaction that each baseline measure had with treatment group and found that the model examining white matter hyperintensities had a significant interaction term, indicating 2.3 less change in Fugl‐Meyer Upper Extremity (FM‐UE) points in the VNS group relative to the control group for each point increase in modified Fazekas scale.InterpretationNeuroimaging measures are associated with extent of gains on the primary endpoint of a phase III stroke recovery trial. Among the neuroimaging measures examined, a measure of global brain health (extent of white matter hyperintensities) was better at explaining the change in arm impairment as compared with measures of regional injury; this was true when examining all study subjects as well as only those in the VNS group and is consistent with the global mechanism of action that VNS has throughout the cerebrum. Future studies can evaluate additional measures that further predict response to VNS therapy. The current findings suggest that individual patient neuroimaging results may be useful for a personalized medicine approach to stroke recovery therapeutics. ANN NEUROL 2025;97:709–719

Background/Objectives: Lung cancer is a devastating disease with the highest mortality rate among cancer types. Over 60% of non-small cell lung cancer (NSCLC) patients, accounting for 87% of lung cancer diagnoses, require radiation therapy. Rapid treatment initiation significantly increases the patient’s survival rate and reduces the mortality rate. Accurate tumor segmentation is a critical step in diagnosing and treating NSCLC. Manual segmentation is time- and labor-consuming and causes delays in treatment initiation. Although many lung nodule detection methods, including deep learning-based models, have been proposed. Most of these methods still have a long-standing problem of high false positives (FPs). Methods: Here, we developed an electronic health record (EHR)-guided lung tumor auto-segmentation called EXACT-Net (EHR-enhanced eXACtitude in Tumor segmentation), where the extracted information from EHRs using a pre-trained large language model (LLM) was used to remove the FPs and keep the TP nodules only. Results: The auto-segmentation model was trained on NSCLC patients’ computed tomography (CT), and the pre-trained LLM was used with the zero-shot learning approach. Our approach resulted in a 250% boost in successful nodule detection using the data from ten NSCLC patients treated in our institution. Conclusions: We demonstrated that combining vision-language information in EXACT-Net multi-modal AI framework greatly enhances the performance of vision only models, paving the road to multimodal AI framework for medical image processing.

Background: Teclistamab (Tec) received regulatory approval for the treatment of relapsed refractory multiple myeloma (RRMM) based on the MajesTEC-1 study. The risk of CRS and ICANS with Tec is well established with a manufacturer's recommendation of inpatient observation for 48-hours for CRS/ICANS monitoring after the initial 3 step up doses. There are limited real-world data on the characteristics of and risk factors for CRS and ICANS with Tec. Such data could inform management of Tec patients (pts), including patient selection for inpatient or outpatient step-up dosing. Methods: RRMM pts treated with standard-of-care Tec across 13 U.S. academic institutions were included in this analysis. The follow-up data cutoff was 30 April 2024. CRS and ICANS classification was based on the American Society for Transplantation and Cellular Therapy consensus grading criteria. Multivariable logistic regression models were fitted to estimate the effects of covariates on CRS and ICANS events, with a particular focus on grade ≥2 CRS, any-grade ICANS, and the combination of CRS grade ≥2 and/or any-grade ICANS. Results: 385 pts were included in the analysis, with a median of 6 lines of prior therapy. Median age was 68 years (range 31-92), 88 patients (25% of pts with available data) had ECOG ≥2, and 301 pts (78%) would have been ineligible for MajesTEC-1. 48 (33%) of 145 evaluable pts with a baseline bone marrow (BM) prior to Tec had &gt;50% BM plasma cells, 139 pts (36%) had extramedullary disease (EMD) (either paraskeletal or true-EMD) and 5 pts (1%) had plasma cell leukemia prior to start of Tec. 224 pts (59%) received Tec with an accelerated step-up dosing regimen, defined as having at least one step-up dose interval of ≥ 2 days. 220 pts (57%) developed any-grade CRS [Grade 1: 46% (N=179), grade 2: 10% (N=37), grade 3: 0.8% (N=3), grade 4: 0.3% (N=1)]. Median time to CRS from start of Tec was 3 days (range, 0-15). Fifty-four pts (14%) developed any-grade ICANS [Grade 1: 7% (N=26), Grade 2: 5% (N=18), grade 3: 3% (N=10), grade 4: none]. There were no grade 5 CRS or ICANS events. Median time to onset of ICANS from start of Tec was 4 days (range, 0-25). There was one case of hemophagocytic lymphohistiocytosis in the entire cohort. 38% of pts received tocilizumab and 17% of pts received corticosteroids. In univariate analysis (UVA), baseline platelet &lt;50, triple-class refractory, penta-refractory status, and EMD prior to start of Tec were associated with increased risk for CRS grade ≥2 (p-values &lt;0.1) and incorporated in MVA. Baseline platelet &lt;50, age &gt;60, ECOG ≥2, Hgb &lt;8 g/dL, BM plasma cells &gt;50%, and ferritin or CRP in the upper quartile were associated with increased risk for any-grade ICANS in UVA (p-values &lt;0.1). Baseline platelet &lt;50, ECOG ≥2, Hgb &lt;8 g/dL, BM plasma cells &gt;50%, ferritin or CRP in the upper quartile, and prior BCMA exposure were associated with increased risk of CRS grade ≥2 and/or any-grade ICANS in UVA (p-values &lt;0.1). Accelerated step-up dosing, renal impairment (CrCl &lt; 30 mL/min) and baseline LDH in the upper quartile were not associated with an increased risk of either CRS or ICANS in UVA. Baseline BM plasma cells &gt;50% was excluded from multivariable analysis (MVA) due to high number of missing data, since most pts (62%) did not undergo BM biopsy prior to start of Tec. In MVA, risk factors for CRS grade ≥2 included baseline platelet &lt;50 at the start of Tec treatment (OR: 2.35, 95% CI 1.08-5.08; p=0.03) and penta-refractory disease (OR: 2.48, 95% CI 1.13-5.45; p=0.02). Risk factors for any-grade ICANS included platelet &lt;50 (OR: 4.18, 95% CI 1.88-9.28; p=0.0004), age &gt;60 (OR: 7.54, 95% CI 1.63-34.89; p=0.01) and CRP in the upper quartile (OR: 2.51, 95% CI 1.16-5.43; p=0.02). Risk factors for CRS grade ≥2 and/or any-grade ICANS included platelet &lt;50 (OR: 2.65, 95% CI 1.42-4.94; p=0.002) and ECOG ≥2 (OR: 2.07, 95% CI 1.19-3.61; p=0.01). Conclusions: This real-world analysis of RRMM pts treated with Tec revealed comparable rates CRS and ICANS relative to the MajesTEC-1 study, and these were mostly limited to grade 1/2. MVA demonstrated a significantly higher risk of CRS grade ≥2 and/or any-grade ICANS in pts with baseline platelet &lt;50 and/or ECOG ≥2, the former may reflect BM disease burden. These results may help identify pts who are at higher risk for IEC related toxicities with Tec that would warrant inpatient monitoring during initial step-up dosing. OP, AA, & LF: co-first authors. AG, HCL & JK: co-senior authors

Background: Hypomethylating agents (HMA) with venetoclax (VEN) have shown encouraging clinical efficacy in single arm clinical trials for patients with myelodysplastic syndrome (MDS) (Garcia JS et al. ASH 2023, Zeidan AM et al. Am J Hematol. 2023). Many centers have started utilizing HMA-VEN in the clinical management of MDS due to unmet need, pending results of the ongoing phase 3 randomized clinical trial (VERONA) using this combination. However, real world studies with large sample size investigating the use of these agents in MDS, practice patterns and clinical outcomes are lacking. Methods: This is a multicenter retrospective study of 13 US academic medical centers. We included contemporary cohorts of adults (age≥18 years) with MDS who were treated with upfront combination of HMA-VEN, upfront HMA monotherapy, or VEN added after inadequate response or progression on HMA (post-HMA failure VEN). Outcomes analyzed includes clinical response, adverse events (AE) of interest, dose administration and survival. Descriptive statistics were used for baseline characteristics. Overall survival and event free survival (EFS: progression, leukemic transformation, death as events) analysis was conducted using Kaplan-Meier method. Determinants of outcomes were analyzed using multivariable regression models. All analyses were conducted at a significance level of p&lt;0.05. Results: Among 454 patients included, 258 patients received VEN in combination with HMA [upfront HMA-VEN (n=175), post-HMA failure VEN (n=83)] and 196 patients received HMA monotherapy. Median age was 69 (59-77) years and majority of patients had MDS with excess blasts (75%), high risk (31.9%) or very-high (41.8%) risk disease by IPSS-R, 23.7% had therapy related disease, 33% had deletion17p or TP53 mutation and 40.9% underwent allogeneic stem cell transplantation. VEN was predominantly started in outpatient setting (78.6%) with tumor lysis syndrome (TLS) prophylaxis and VEN doses ranged from 70-400mg, adjusted for concomitant antifungal drugs. Incidence of TLS was minimal (No TLS- 92.9%, lab-only TLS 6.2%, 1 pt with clinical TLS). Neutropenic fever was the main AE noted [upfront VEN-HMA (34.2%), HMA monotherapy (22.5%), post-HMA VEN (44.5%)]. In the upfront setting, response rate was significantly higher with HMA-VEN than HMA monotherapy (CR: 33% vs 12%; marrow CR: 40% vs 27%, p&lt;.001). VEN given in the post HMA failure setting also resulted in encouraging response rates (CR 10%, marrow CR 32%). After adjusting for covariates in multivariable analysis, patients treated with HMA-VEN in the upfront setting had significantly improved event free survival as compared to HMA monotherapy (HR 0.59, HR 0.44-0.78, p&lt;0.01), but overall survival did not reach statistical significance (overall survival HR 0.77, 95%CI 0.57-1.04, p=0.08). Conclusions: In this large multicenter retrospective analysis from US academic medical centers, we found that VEN in combination with HMA resulted in encouraging clinical outcomes and was effectively administered in the outpatient setting with minimal TLS. Infection risk management and supportive measures would be important to improve the ability to use this treatment widely, pending results of the ongoing phase 3 randomized clinical trial. Updated results with additional variables will be available at the main meeting.