Baseline Magnetic Resonance Imaging Research Articles

Explainable machine learning on baseline MRI predicts multiple sclerosis trajectory descriptors.

Multiple sclerosis (MS) is a multifaceted neurological condition characterized by challenges in timely diagnosis and personalized patient management. The application of Artificial Intelligence (AI) to MS holds promises for early detection, accurate diagnosis, and predictive modeling. The objectives of this study are: 1) to propose new MS trajectory descriptors that could be employed in Machine Learning (ML) regressors and classifiers to predict patient evolution; 2) to explore the contribution of ML models in discerning MS trajectory descriptors using only baseline Magnetic Resonance Imaging (MRI) studies. This study involved 446 MS patients who had a baseline MRI, at least two measurements of Expanded Disability Status Scale (EDSS), and a 1-year follow-up. Patients were divided into two groups: 1) for model development and 2) for evaluation. Three descriptors: β1, β2, and EDSS(t), were related to baseline MRI parameters using regression and classification XGBoost models. Shapley Additive Explanations (SHAP) analysis enhanced model transparency by identifying influential features. The results of this study demonstrate the potential of AI in predicting MS progression using the proposed patient trajectories and baseline MRI scans, outperforming classic Multiple Linear Regression (MLR) methods. In conclusion, MS trajectory descriptors are crucial; incorporating AI analysis into MRI assessments presents promising opportunities to advance predictive capabilities. SHAP analysis enhances model interpretation, revealing feature importance for clinical decisions.

Multiparametric quantitative magnetic resonance imaging of uterine fibroids for prediction of growth rate-a pilot study.

Uterine fibroid (UF) growth rate and future morbidity cannot be predicted. This can lead to sub-optimal clinical management, with women being lost to follow-up and later presenting with severe disease that may require hospitalization, transfusions, and urgent surgical interventions. Multi-parametric quantitative magnetic resonance imaging (MRI) could provide a biomarker to predict growth rate facilitating better-informed disease management and better clinical outcomes. We assessed the ability of putative quantitative and qualitative MRI predictive factors to predict UF growth rate. Twenty women with UFs were recruited and completed baseline and follow-up MRI exams, 1-2.5 years apart. The subjects filled out symptom severity and health-related quality of life questionnaires at each visit. A standard clinical pelvic MRI non-contrast exam was performed at each visit, followed by a contrast-enhanced multi-parametric quantitative MRI (mp-qMRI) exam with T2, T2*, and apparent diffusion coefficient (ADC) mapping and dynamic contrast-enhanced MRI. Up to 3 largest fibroids were identified and outlined on the T2-weighted sequence. Fibroid morphology and enhancement patterns were qualitatively assessed on dynamic contrast-enhanced MRI. The UFs' volumes and average T2, T2*, and ADC values were calculated. Pearson correlation coefficients were calculated between UF growth rate and T2, T2*, ADC, and baseline volume. Multiple logistic regression and receiver operating characteristic (ROC) analysis were performed to predict fast-growing UFs using combinations of up to 2 significant predictors. A significance level of alpha =0.05 was used. Forty-four fibroids in 20 women had growth rate measurement available, and 36 fibroids in 16 women had follow-up quantitative MRI available. The distribution of fibroid growth rate was skewed, with approximately 20% of the fibroids exhibiting fast growth (>10 cc/year). However, there were no significant changes in median baseline and follow-up values of symptom severity and health-related quality of life scores. There was no change in average T2, T2*, and ADC at follow-up exams and there was a moderate to strong correlation to the fibroid growth rate in baseline volume and average T2 and ADC in slow-growing fibroids (<10 cc/year). A multiple logistic regression to identify fast growing UFs (>10 cc/year) achieved an area under the curve (AUC) of 0.80 with specificity of 69% at 100% sensitivity. The mp-qMRI parameters T2, ADC, and UF volume obtained at the time of initial fibroid diagnosis may be able to predict UF growth rate. Mp-qMRI could be integrated into the management of UFs, for individualized care and improved clinical outcomes.

Association of Baseline Magnetic Resonance Imaging Prostate Imaging Reporting and Data System Score With Prostate Cancer Active Surveillance Early Biopsy Reclassification: Data From the Michigan Urological Surgery Improvement Collaborative (MUSIC).

The purpose of our study was to evaluate the association of baseline MRI Prostate Imaging Reporting and Data System (PI-RADS) score with biopsy reclassification in a multicenter active surveillance (AS) cohort. We identified men in the Michigan Urological Surgery Improvement Collaborative registry (46 hospital-based/academic/private practice urology groups) with National Comprehensive Cancer Network (NCCN) low-risk and favorable intermediate-risk prostate cancer who underwent MRI within 6 months before or after initial biopsy and enrolled in AS from June 2016 to January 2021. The primary objective was to determine the association of baseline MRI PI-RADS score (≥4 lesion) with reclassification to high-grade prostate cancer (≥grade group 3) on surveillance biopsy. Multivariable Cox proportional hazards regression models were constructed and adjusted for pathologic, MRI, and clinical/biopsy factors, with landmark time of 6 months from diagnostic biopsy. We included an interaction term between PI-RADS score and NCCN group in the Cox model. A total of 1491 men were included with median age 64 years (IQR: 59-69) with median follow-up 11.0 months (IQR: 6.0-23.0) after landmark. Baseline PI-RADS ≥ 4 lesion was associated with an increased hazard of biopsy reclassification (HR: 2.3 [95% CI: 1.6-3.2], P < .001), along with grade group 2 vs 1 (HR: 2.5 [95% CI: 1.7-3.7], P < .001), and increasing age (per 10 years; HR: 1.8 [95% CI: 1.4-2.4], P < .001). The interaction between NCCN risk group with MRI findings was not significant (P = .7). In this multicenter cohort study of real-world data, baseline MRI PI-RADS score was significantly associated with early biopsy reclassification in men undergoing AS with NCCN low- or favorable intermediate-risk prostate cancer.

FRONTAL WHITE MATTER CHANGES INDICATE RECOVERY WITH INPATIENT TREATMENT IN HEROIN ADDICTION.

Amidst an unprecedented opioid epidemic, identifying neurobiological correlates of change with medication-assisted treatment of heroin use disorder is imperative. Distributed white matter (WM) impairments in individuals with heroin use disorder (iHUD) have been associated with increased drug craving, a reliable predictor of treatment outcomes. However, little is known about the extent of whole-brain structural connectivity changes with inpatient treatment and abstinence in iHUD. To assess WM microstructure and associations with drug craving changes with inpatient treatment in iHUD (effects of time/re-scan compared to controls; CTL). Longitudinal cohort study (12/2020-09/2022) where iHUD and CTL underwent baseline magnetic resonance imaging (MRI#1) and follow-up (MRI#2) scans, (mean interval of 13.9 weeks in all participants combined). The iHUD and CTL were recruited from urban inpatient treatment facilities and surrounding communities, respectively. Thirty-four iHUD (42.1yo; 7 women), 25 age-/sex-matched CTL (40.5yo; 9 women). Between scans, inpatient iHUD continued their medically-assisted treatment and related clinical interventions. CTL participants were scanned at similar time intervals. Changes in white matter diffusion metrics [fractional anisotropy (FA), mean (MD), axial (AD), and radial diffusivities (RD)] in addition to baseline and cue-induced drug craving, and other clinical outcome variables (mood, sleep, affect, perceived stress, and therapy attendance). Main findings showed HUD-specific WM microstructure changes encompassing mostly frontal major callosal, projection, and association tracts, characterized by increased FA (.949<1- p<.986) and decreased MD (.949<1-p<.997) and RD (.949<1-p<.999). The increased FA (r=- 0.72, p<.00001) and decreased MD (r=0.69, p<.00001) and RD (r=0.67, p<.0001) in the genu and body of the corpus callosum and the left anterior corona radiata in iHUD were correlated with a reduction in baseline craving (.949<1-p<.999). No other WM correlations with outcome variables reached significance. Our findings suggest whole-brain normalization of structural connectivity with inpatient medically-assisted treatment in iHUD encompassing recovery in frontal WM pathways implicated in emotional regulation and top-down executive control. The association with decreases in baseline craving further supports the relevance of these WM markers to a major symptom in drug addiction, with implications for monitoring clinical outcomes. Question: Does white matter (WM) microstructure change with medication-assisted treatment in individuals with heroin use disorder (iHUD)?Findings: In this longitudinal cohort study, diffusion MRI was acquired in 34 inpatient iHUD and 25 healthy controls (CTL) twice, separated by a mean of 13.9 weeks. We found HUD- specific WM microstructure changes with time, characterized by increased anisotropy and decreased diffusivity in fronto-striatal WM pathways. These changes were correlated with decreased baseline drug craving with treatment.Meaning: Frontal WM changes and associated drug craving decreases suggest brain-behavior recovery with inpatient treatment in iHUD, potentially contributing to reduced drug use and sustained abstinence.

Incidence and Risk Factors for New and Recurrent Infarcts in Adults With Sickle Cell Disease.

Most adults with sickle cell disease will experience a silent cerebral infarction (SCI) or overt stroke. Identifying patient subgroups with increased stroke incidence is important for future clinical trials focused on stroke prevention. Our 3-center prospective cohort study tested the primary hypothesis that adults with sickle cell disease and SCIs have a greater incidence of new stroke or SCI compared with those without SCI. A secondary aim focused on identifying additional risk factors for progressive infarcts, particularly traditional risk factors for stroke in adults. This observational study included adults with sickle cell disease and no history of stroke. Magnetic resonance imaging scans of the brain completed at baseline and >1 year later were reviewed by 3 radiologists for baseline SCIs and new or progressive infarcts on follow-up magnetic resonance imaging. Stroke risk factors were abstracted from the medical chart. Time-to-event analysis was utilized for progressive infarcts. Median age was 24.1 years; 45.3% of 95 participants had SCIs on baseline magnetic resonance imaging. Progressive infarcts were present in 17 participants (17.9%), and the median follow-up was 2.1 years. Incidence of new infarcts was 11.95 per 100 patient-years (6.17-20.88) versus 3.74 per 100 patient-years (1.21-8.73) in those with versus without prior SCI. Multivariable Cox regression showed that baseline SCI predicts progressive infarcts (hazard ratio, 3.46 [95% CI, 1.05-11.39]; P=0.041); baseline hypertension was also associated with progressive infarcts (hazard ratio, 3.23 [95% CI, 1.16-9.51]; P=0.025). Selecting individuals with SCIs and hypertension for stroke prevention trials in sickle cell disease may enrich the study population with those at highest risk for infarct recurrence.

Imaging-Based Deep Learning for Predicting Desmoid Tumor Progression.

Desmoid tumors (DTs) are non-metastasizing and locally aggressive soft-tissue mesenchymal neoplasms. Those that become enlarged often become locally invasive and cause significant morbidity. DTs have a varied pattern of clinical presentation, with up to 50-60% not growing after diagnosis and 20-30% shrinking or even disappearing after initial progression. Enlarging tumors are considered unstable and progressive. The management of symptomatic and enlarging DTs is challenging, and primarily consists of chemotherapy. Despite wide surgical resection, DTs carry a rate of local recurrence as high as 50%. There is a consensus that contrast-enhanced magnetic resonance imaging (MRI) or, alternatively, computerized tomography (CT) is the preferred modality for monitoring DTs. Each uses Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1), which measures the largest diameter on axial, sagittal, or coronal series. This approach, however, reportedly lacks accuracy in detecting response to therapy and fails to detect tumor progression, thus calling for more sophisticated methods. The objective of this study was to detect unique features identified by deep learning that correlate with the future clinical course of the disease. Between 2006 and 2019, 51 patients (mean age 41.22 ± 15.5 years) who had a tissue diagnosis of DT were included in this retrospective single-center study. Each had undergone at least three MRI examinations (including a pretreatment baseline study), and each was followed by orthopedic oncology specialists for a median of 38.83 months (IQR 44.38). Tumor segmentations were performed on a T2 fat-suppressed treatment-naive MRI sequence, after which the segmented lesion was extracted to a three-dimensional file together with its DICOM file and run through deep learning software. The results of the algorithm were then compared to clinical data collected from the patients' medical files. There were 28 males (13 stable) and 23 females (15 stable) whose ages ranged from 19.07 to 83.33 years. The model was able to independently predict clinical progression as measured from the baseline MRI with an overall accuracy of 93% (93 ± 0.04) and ROC of 0.89 ± 0.08. Artificial intelligence may contribute to risk stratification and clinical decision-making in patients with DT by predicting which patients are likely to progress.

Abstract PO5-16-10: The efficacy of novel antibody-drug-conjugates on brain metastases in metastatic breast cancer patients – a multicenter real-world analysis

Abstract 15 – 30% of all patients with metastatic breast cancer (MBC) develop brain metastases (BMs). Treatment of BMs include radiation therapy and surgical resection while systemic treatment is challenging and prospective evidence for the treatment of active brain metastases is restricted to the small molecule Tucatinib. Recently, the antibody-drug-conjugates (ADCs) Sacituzumab-Govitecan (SG) and Trastuzumab-Deruxtecan (T-DXd) have shown to be highly effective in the treatment of triple-negative (SG), HER2-positive (T-DXd) and HR+/HER2- (SG and, in case of HER2-low-expression, T-DxD) MBC. However, there is only limited data, whether these macromolecules are also effective in patients with BMs. We therefore aimed to examine the efficacy of SG and T-DXd in patients with stable and active BMs in a multicenter real-world analysis. Female patients with stable or active BMs who were treated with either SG or T-DXd at the breast centers of Tuebingen University, Ulm University or Freiburg University in Germany before June 30, 2023 were included into this analysis. Data cut-off for this analysis was July 10, 2023. BMs were classified as active if they were newly diagnosed and did not require local treatment or if they were progressing and did not require local treatment. Preexisting BMs without intracranial disease progression or preexisting/newly diagnosed BMs that had been treated by surgery and/or radiotherapy directly prior to SG/T-DXd therapy were considered as stable. All patients underwent baseline magnetic resonance imaging (MRI) of the brain and received follow-up MRIs at least every three months. Growth dynamics of BMs were assessed by board certified neuroradiologists. Intracranial disease control rate (icDCR) was defined as the percentage of patients with at least intracranial stable disease at the first follow-up MRI. Intracranial progression-free survival (icPFS) was determined as the period between the first application of SG or T-DXd and intracranial disease progression or treatment cessation due to the patients’ will or the onset of intolerable toxicity. If ADC treatment was discontinued due to extracranial disease progression albeit at least stable intracranial disease patients were censored. Overall survival (OS) was defined as the period between the first application of SG or T-DXd and death. In total, 26 patients were included with a median of two prior therapy lines in the metastatic setting (range 2 – 15). 10 (38%) and 14 (54%) patients received SG and T-DXd, respectively. 2/26 (8%) patients receiving SG were consecutively subjected to T-DXd treatment. SG was applied to 12/12 patients (100%) due to triple-negative MBC and 12/16 patients (75%) treated with T-DXd were HER2-positive. The remaining 4/16 patients (25%) treated with T-DXd showed HER2 low tumor biology. 5/12 (42%) and 8/16 (50%) patients treated with SG or T-DXd had active BMs at treatment initiation. The icDCR was 33% (95% CI: 5% - 61%)/81% (95% CI: 62% - 100%) for all patients treated with SG/T-DXd and 40% (95% CI: 0% - 88%)/75% (95% CI: 43% - 100%) in case of active BMs. After a median follow-up of 7.2 months, 10/12 (83.3%) patients treated with SG and 4/16 (25%) patients subjected to T-DXd had discontinued treatment. Median icPFS was 3.1 months (95% CI: 0.4 – 5.8 months) for SG and not reached for T-DXd. 7/12 (58%) patients treated with SG and 4/16 patients (25%) treated with T-DXd died until data cut-off. Median OS was 7.2 months (95% CI: 3.0 – 11.4 months) for patients treated with SG and not reached for patients treated with T-DXd. SG and T-DXd showed promising clinical activity in both, stable and active cerebral breast cancer metastasis. Updated PFS and OS data will be reported at the meeting. Citation Format: Dominik Dannehl, Henning Schäffler, Tobias Engler, Lea Volmer, Arne Estler, Ilinca Popp, Stefanie Lorenz, Eva-Maria Grischke, Markus Hahn, Ghazaleh Tabatabai, Ingolf Juhasz-Böss, Wolfgang Janni, Sara Brucker, Florin-Andrei Taran, Andreas Hartkopf. The efficacy of novel antibody-drug-conjugates on brain metastases in metastatic breast cancer patients – a multicenter real-world analysis [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO5-16-10.

Hemorrhagic Transformation in Noncardioembolic Acute Ischemic Stroke: MRI Analysis From PACIFIC-STROKE.

In the phase 2 PACIFIC-STROKE trial (Proper Dosing and Safety of the Oral FXIa Inhibitor BAY 2433334 in Patients Following Acute Noncardioembolic Stroke), asundexian, an oral factor XIa inhibitor, did not increase the risk of hemorrhagic transformation (HT). In this secondary analysis, we aimed to investigate the frequency, types, and risk factors of HT on brain magnetic resonance imaging (MRI). This was a secondary analysis of the PACIFIC-STROKE trial. Patients with mild-to-moderate acute noncardioembolic ischemic stroke were randomly assigned to asundexian or placebo plus guideline-based antiplatelet therapy. Brain MRIs were required at baseline (≤120 hours after stroke onset) and at 26 weeks or end-of-study. HT was defined using the Heidelberg classification and classified as early HT (identified on baseline MRI) or late HT (new HT by 26 weeks) based on iron-sensitive sequences. Multivariable logistic regression models were used to test factors that are associated with early HT and late HT, respectively. Of 1745 patients with adequate baseline brain MRI (mean age, 67 years; mean National Institutes of Health Stroke Scale score, 2.8), early HT at baseline was detected in 497 (28.4%). Most were hemorrhagic infarctions (hemorrhagic infarction type 1: 15.2%; HI2: 12.7%) while a few were parenchymal hematomas (parenchymal hematoma type 1: 0.4%; parenchymal hematoma type 2: 0.2%). Early HT was more frequent with longer symptom onset-to-MRI interval. Male sex, diabetes, higher National Institutes of Health Stroke Scale large (>15 mm) infarct size, cortical involvement by infarct, higher number of acute infarcts, presence of chronic brain infarct, cerebral microbleed, and chronic cortical superficial siderosis were independently associated with early HT in the multivariable logistic regression model. Of 1507 with follow-up MRI, HT was seen in 642 (42.6%) overall, including 361 patients (23.9%) with late HT (new HT: 306; increased grade of baseline HT: 55). Higher National Institutes of Health Stroke Scale, large infarct size, cortical involvement of infarct, and higher number of acute infarcts predicted late HT. About 28% of patients with noncardioembolic stroke had early HT, and 24% had late HT detectable by MRI. Given the high frequency of HT on MRI, more research is needed on how it influences treatment decisions and outcomes.

Estimating Efficacy of Conversion Therapy on Patients with Initially Unresectable Colorectal Cancer Liver Metastases by using MRI: Development of a Predictive Score

The conversion success rate (CSR) has crucial implication for clinical outcomes of initially unresectable colorectal liver metastases (CRLM) following conversion therapy. This study aimed to develop a simple predictive scoring model for identifying CSR according to baseline magnetic resonance imaging (MRI) features, and confirm its performance and prognostic significance in a validation cohort. A total of 155 consecutive patients with initially unresectable CRLM were retrospectively reviewed in the study. A simple MRI-based predictive scoring model for identifying CSR was developed in the development cohort (n=104) by using multivariable logistic regression analyzes. The diagnostic performance was evaluated for the predictive score. Thereafter, patients in the validation cohort (n=51) were stratified into groups with predicted high CSR or low CSR according to the score. The progression-free survival (PFS) and overall survival (OS) were compared between two groups using the log-rank test. The predictive score of CSR, named mrNISE, incorporated the number of CRLM ≥10, the largest size ≥50mm, poorly defined tumor-liver interface, and peritumoral enhancement. The AUC of the mrNISE score was 0.845 for the development cohort and 0.776 for the validation cohort. According to the score, patients with predicted high CSR had better PFS and OS than those with low CSR in both development and validation cohorts. The predictive score demonstrated great performance for identifying CSR of initially unresectable CRLM. Stratifying patients by the score, personalized treatment goals can be formulated before conversion therapy to improve clinical prognosis and reduce adverse events caused by ineffective treatment.

Impact of magnetic resonance imaging-derived skeletal muscle index in locoregionally advanced nasopharyngeal carcinoma.

To evaluate the clinical implication of magnetic resonance imaging (MRI)-derived skeletal muscle index (SMI) in locoregionally advanced nasopharyngeal carcinoma (LANPC) patients undergoing induction chemotherapy (IC) followed by concurrent chemoradiotherapy (CCRT) and further to develop a nomogram for predicting survival prognosis. SMI was determined through baseline MRI at the third cervical level. The nomogram was based on a training cohort involving 409 LANPC patients. We validated the prognostic accuracy of this prognostic model in an internal validation cohort (n = 204) and an external independent cohort (n = 272). SMI was an independent risk factor for OS. A prognostic model comprising age, TNM stage and SMI for individual survival prediction was developed and graphically represented as a nomogram. The model showed favorable discrimination (C-index: 0.686), predictive accuracy [time dependent area under the curve (tAUC) at 5years: 0.70], and calibration, and was further validated in the internal and external validation datasets. A risk stratification derived from the model stratified these patients into three prognostic subgroups with significantly different survival. Low SMI accessed by MRI was significantly associated with poor overall survival in LANPC patients undergoing IC + CCRT. Moreover, we established and validated a novel nomogram involving age, TNM stage and SMI that could provide accurate prognostic stratification among this population.

Post-traumatic and OA-related lesions in the knee at baseline and 2 years after traumatic meniscal injury: Secondary analysis of a randomized controlled trial

ObjectiveTo assess the presence of early degenerative changes on Magnetic Resonance Imaging (MRI) 24 months after a traumatic meniscal tear and to compare these changes in patients treated with arthroscopic partial meniscectomy or physical therapy plus optional delayed arthroscopic partial meniscectomy. DesignWe included patients aged 18-45 years with a recent onset, traumatic, MRI verified, isolated meniscal tear without radiographic osteoarthritis. Patients were randomized to arthroscopic partial meniscectomy or standardized physical therapy with optional delayed arthroscopic partial meniscectomy. MRIs at baseline and 24 months were scored using the MRI Osteoarthritis Knee Score (MOAKS). We compared baseline MRIs to healthy controls aged 18-40 years. The outcome was the progression of bone marrow lesions (BMLs), cartilage defects and osteophytes after 24 months in patients. ResultsWe included 99 patients and 50 controls. At baseline, grade 2 and 3 BMLs were present in 26% of the patients (n=26), compared to 2% of the controls (n=1) (between group difference 24% (95% CI 15% to 34%)). In patients, 35% (n=35) had one or more cartilage defects grade 1 or higher, compared to 2% of controls (n=1) (between group difference 33% (95% CI 23% to 44%)). At 24 months MRI was available for 40 patients randomized to arthroscopic partial meniscectomy and 41 patients randomized to physical therapy. At 24 months 30% (n=12) of the patients randomized to arthroscopic partial meniscectomy showed BML worsening, compared to 22% (n=9) of the patients randomized to physical therapy (between group difference 8% (95% CI -11% to 27%)). Worsening of cartilage defects was present in 40% (n=16) of the arthroscopic partial meniscectomy group, compared to 22% (n=9) of the physical therapy group (between group difference 18% (95% CI -2% to 38%)). Of the patients who had no cartilage defect at baseline, 33% of the arthroscopic partial meniscectomy group had a new cartilage defect at follow-up compared to 14% of the physical therapy group. Osteophyte worsening was present in 18% (n=7) of the arthroscopic partial meniscectomy group and 15% (n=6) of the physical therapy group (between group difference 3% (95% CI -13% to 19%)). ConclusionsOur results might suggest more worsening of BMLs and cartilage defects with arthroscopic partial meniscectomy compared to physical therapy with optional delayed arthroscopic partial meniscectomy at 24-month follow-up in young patients with isolated traumatic meniscal tears without radiographic OA.

Automated graded prognostic assessment for patients with hepatocellular carcinoma using machine learning.

Accurate mortality risk quantification is crucial for the management of hepatocellular carcinoma (HCC); however, most scoring systems are subjective. To develop and independently validate a machine learning mortality risk quantification method for HCC patients using standard-of-care clinical data and liver radiomics on baseline magnetic resonance imaging (MRI). This retrospective study included all patients with multiphasic contrast-enhanced MRI at the time of diagnosis treated at our institution. Patients were censored at their last date of follow-up, end-of-observation, or liver transplantation date. The data were randomly sampled into independent cohorts, with 85% for development and 15% for independent validation. An automated liver segmentation framework was adopted for radiomic feature extraction. A random survival forest combined clinical and radiomic variables to predict overall survival (OS), and performance was evaluated using Harrell's C-index. A total of 555 treatment-naïve HCC patients (mean age, 63.8 years ± 8.9 [standard deviation]; 118 females) with MRI at the time of diagnosis were included, of which 287 (51.7%) died after a median time of 14.40 (interquartile range, 22.23) months, and had median followed up of 32.47 (interquartile range, 61.5) months. The developed risk prediction framework required 1.11 min on average and yielded C-indices of 0.8503 and 0.8234 in the development and independent validation cohorts, respectively, outperforming conventional clinical staging systems. Predicted risk scores were significantly associated with OS (p < .00001 in both cohorts). Machine learning reliably, rapidly, and reproducibly predicts mortality risk in patients with hepatocellular carcinoma from data routinely acquired in clinical practice. Precision mortality risk prediction using routinely available standard-of-care clinical data and automated MRI radiomic features could enable personalized follow-up strategies, guide management decisions, and improve clinical workflow efficiency in tumor boards. • Machine learning enables hepatocellular carcinoma mortality risk prediction using standard-of-care clinical data and automated radiomic features from multiphasic contrast-enhanced MRI. • Automated mortality risk prediction achieved state-of-the-art performances for mortality risk quantification and outperformed conventional clinical staging systems. • Patients were stratified into low, intermediate, and high-risk groups with significantly different survival times, generalizable to an independent evaluation cohort.

Abstract 894: Radiomics at baseline MRI are associated with tumor shrinkage in patients with unresectable renal cell carcinoma treated with neo-adjuvant Cabozantinib

Abstract Background: Neoadjuvant Cabozantinib targeted therapy (nCabo) has been shown to facilitate tumor size reduction in patients with renal cell carcinoma (RCC). However, there is unmet clinical need to identify pre-treatment markers predictive of response to nCabo to limit financial, dose-related toxicity. We investigated computationally derived radiomic signatures from baseline magnetic resonance imaging (MRI) for their association with response to tumor shrinkage following nCabo.Methods Seventeen patients with non-metastatic, clinical stage ≥T3, biopsy-confirmed clear cell RCC were enrolled in a phase II clinical trial (NCT04022343) examining primary tumor response to nCabo. Patients underwent 3 Tesla MRI at baseline, 6 and 12 weeks. Partial response (PR) or stable disease (SD) was determined per Response Evaluation Criteria in Solid Tumors at 12 weeks. Tumor regions of interest (ROI) were delineated on axial, arterial phase T1-weighted (T1W) series under the guidance of expert radiologist. T1W were normalized with respect to enhancement in cortex region of kidney. 75 radiomic features quantifying texture heterogeneity were derived from tumor ROI on a per-voxel basis at baseline. Wilcoxon ranksum test was used to evaluate significant differences in radiomics between patient groups PR and SD. Results: Following nCabo, 6/17 patients experienced PR. 22 radiomic features from T1W quantifying intensity based heterogeneity showed significant differences between patients with SD and PR after nCabo (p&amp;lt;0.05; top 3 in Table). None of the clinical variables showed significant differences between SD and PR. Conclusion: In this proof-of-concept preliminary study, we observed that radiomics at baseline MRI may allow for identifying RCC patients who would favorably respond to nCabo in terms of tumor shrinkage. Future studies will include analysis on longitudinal data, correlation against survival on large-scale datasets. Association between tumor shrinkage following nCabo and radiomic features, clinical variables Stable Disease (N=6) Partial Response (N=11) Category Feature P-value median IQR median IQR Radiomic Haralick 0.01 1.50 0.51 2.31 0.51 Gradient Filter 0.02 88.06 107.28 289.58 107.28 CoLlAGe 0.03 224.28 10.73 210.50 10.73 Clinical Age 0.12 56.00 14.00 52.50 14.00 BMI 0.48 29.90 8.50 27.30 8.50 modified Glasgow Prognostic Score 0.21 1.00 1.00 0.00 1.00 Baseline Tumor Size (cm) 0.48 9.50 5.15 9.50 5.15 ECOG PS at presentation 0.23 0.00 1.00 0.00 1.00 Charlson Comorbidity Index 0.72 6.00 6.00 4.00 6.00 Citation Format: Liping LI, Edouard Nicaise, Benjamin Schmeusser, Abhishek Midya, Amir Davarpanahfakhr, Anant Madabhushi, Mehmet Asim Bilen, Viraj A. Master, Rakesh Shiradkar. Radiomics at baseline MRI are associated with tumor shrinkage in patients with unresectable renal cell carcinoma treated with neo-adjuvant Cabozantinib [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 894.

Morphological and Quantitative Parametric MRI Follow-up of Cartilage Changes Before and After Intra-articular Injection Therapy in Patients With Mild to Moderate Knee Osteoarthritis: A Randomized, Placebo-Controlled Trial.

Intra-articular injections are routinely used for conservative treatment of knee osteoarthritis (OA). The detailed comparative therapeutic effects of these injections on cartilage tissue are still unclear. The aim of this study was to detect and compare knee cartilage changes after intra-articular injection of glucocorticoid, hyaluronic acid, or platelet-rich plasma (PRP) to placebo using quantitative (T2 and T2* mapping) and morphological magnetic resonance imaging parameters in patients with mild or moderate osteoarthritis. In a double-blinded, placebo-controlled, single-center trial, knees with mild or moderate osteoarthritis (Kellgren-Lawrence grade 1-3) were randomly assigned to an intra-articular injection with 1 of these substances: glucocorticoid, hyaluronic acid, PRP, or placebo. Cartilage degeneration on baseline and follow-up magnetic resonance imaging scans (after 3 and 12 months) was assessed by 2 readers using quantitative T2 and T2* times (milliseconds) and morphological parameters (modified Outerbridge grading, subchondral bone marrow edema, subchondral cysts, osteophytes). One hundred twenty knees (30 knees per treatment group) were analyzed with a median patient age of 60 years (interquartile range, 54.0-68.0 years). Interreader reliability was good for T2 (ICC, 0.76; IQR, 0.68-0.83) and T2* (ICC, 0.83; IQR, 0.76-0.88) measurements. Morphological parameters showed no significant changes between all groups after 3 and 12 months. T2 mapping after 12 months showed the following significant ( P = 0.001-0.03) changes between groups in 6 of 14 compartments: values after PRP injection decreased compared with glucocorticoid in 4 compartments (complete medial femoral condyle and central part of lateral condyle) and compared with placebo in 2 compartments (anterior and central part of medial tibial plateau); values after glucocorticoid injection decreased compared with placebo in 1 compartment (central part of medial tibial plateau). No significant changes were seen for T2 and T2* times after 3 months and T2* times after 12 months. No correlation was found between T2/T2* times and Kellgren-Lawrence grade, age, body mass index, or pain (Spearman ρ, -0.23 to 0.18). Platelet-rich plasma injection has a positive long-term effect on cartilage quality in the medial femoral compartment compared to glucocorticoid, resulting in significantly improved T2 values after 12 months. For morphological cartilage parameters, injections with glucocorticoid, PRP, or hyaluronic acid showed no better effect in the short or long term compared with placebo.

Pancreatic iron in pediatric transfusion-dependent beta-thalassemia patients: A longitudinal MRI study.

In pediatric transfusion-dependent thalassemia (TDT) patients, we evaluated the prevalence, pattern, and clinical associations of pancreatic siderosis and the changes in pancreatic iron levels and their association with baseline and changes in total body iron balance. We considered 86 pediatric TDT patients consecutively enrolled in the Extension-Myocardial Iron Overload in Thalassemia Network. Iron overload (IO) was quantified by R2* magnetic resonance imaging (MRI). Sixty-three (73%) patients had pancreatic IO (R2*>38Hz). Global pancreas R2* values were significantly correlated with mean serum ferritin levels, MRI liver iron concentration (LIC) values, and global heart R2* values. Global pancreas R2* values were significantly higher in patients with altered versus normal glucose metabolism. Thirty-one patients also performed the follow-up MRI at 18±3months. Higher pancreatic R2* values were detected at the follow-up, but the difference versus the baseline MRI was not significant. The 20% of patients with baseline pancreatic IO showed no pancreatic IO at the follow-up. The 46% of patients without baseline pancreatic IO developed pancreatic siderosis. The changes in global pancreas R2* between the two MRIs were not correlated with baseline serum ferritin levels, baseline, final, and changes in MRI LIC values, or baseline pancreatic iron levels. In children with TDT, pancreatic siderosis is a frequent finding associated with hepatic siderosis and represents a risk factor for myocardial siderosis and alterations of glucose metabolism. Iron removal from the pancreas is exceptionally challenging and independent from hepatic iron status.

Abstract TP221: Progression of Intracranial Artery Stenosis and Cerebral Small Vessel Disease and Associations Between the Two Entities: A Prospective Community-Based Cohort Study

Background: Both intracranial artery stenosis (ICAS) and cerebral small vessel disease (CSVD) can cause heavy socioeconomic burdens. However, their longitudinal changes remain controversial. To investigate the risk factors for ICAS and CSVD progression and their associations in a community population. Methods: This longitudinal analysis included 756 participants who underwent both baseline and follow-up brain magnetic resonance imaging (MRI) and magnetic resonance angiography from Shunyi cohort. The follow-up duration was 5.58±0.49 years. Incident ICAS was defined as new stenosis occurring in at least one artery or increased severity of the original artery stenosis. CSVD markers included lacunes, cerebral microbleeds (CMB), and white matter hyperintensities (WMH). Results: Of the 756 participants (aged 53.7±8.0 years, 34.9% men), 8.5% had incident ICAS. Body mass index (BMI) and diabetes mellitus are independent risk factors for ICAS. Hypertension is an independent risk factor for incident lacunes and CMB, whereas WMH progression is primarily affected by BMI. Higher LDL cholesterol level independently protects against WMH progression. Baseline ICAS burden predicted an increased risk of incident lacunes (odds ratio [OR] = 2.17, 95% confidence interval [CI] = 1.10-4.29, p = 0.025) after adjusting for age, sex, follow-up time, vascular risk factors, and baseline lacunes burden. Incident ICAS was also significantly associated with incident lacunes (OR = 3.29, 95% CI = 1.61-6.72, p = 0.001). No significant correlation was identified between ICAS and CMB or WMH. Conclusions: Patients with ICAS and CSVD progression had different modifiable risk factor profiles. The disparate association between ICAS and progressions of different MRI markers of CSVD imply vital differences of their underline mechanism.

Magnetic Resonance Imaging Tissue Signatures Associated With White Matter Changes Due to Sporadic Cerebral Small Vessel Disease Indicate That White Matter Hyperintensities Can Regress.

White matter hyperintensities (WMHs) might regress and progress contemporaneously, but we know little about underlying mechanisms. We examined WMH change and underlying quantitative magnetic resonance imaging tissue measures over 1 year in patients with minor ischemic stroke with sporadic cerebral small vessel disease. We defined areas of stable normal-appearing white matter, stable WMHs, progressing and regressing WMHs based on baseline and 1-year brain magnetic resonance imaging. In these areas we assessed tissue characteristics with quantitative T1, fractional anisotropy (FA), mean diffusivity (MD), and neurite orientation dispersion and density imaging (baseline only). We compared tissue signatures cross-sectionally between areas, and longitudinally within each area. WMH change masks were available for N=197. Participants' mean age was 65.61 years (SD, 11.10), 59% had a lacunar infarct, and 68% were men. FA and MD were available for N=195, quantitative T1 for N=182, and neurite orientation dispersion and density imaging for N=174. Cross-sectionally, all 4 tissue classes differed for FA, MD, T1, and Neurite Density Index. Longitudinally, in regressing WMHs, FA increased with little change in MD and T1 (difference estimate, 0.011 [95% CI, 0.006-0.017]; -0.002 [95% CI, -0.008 to 0.003] and -0.003 [95% CI, -0.009 to 0.004]); in progressing and stable WMHs, FA decreased (-0.022 [95% CI, -0.027 to -0.017] and -0.009 [95% CI, -0.011 to -0.006]), whereas MD and T1 increased (progressing WMHs, 0.057 [95% CI, 0.050-0.063], 0.058 [95% CI, 0.050 -0.066]; stable WMHs, 0.054 [95% CI, 0.045-0.063], 0.049 [95% CI, 0.039-0.058]); and in stable normal-appearing white matter, MD increased (0.004 [95% CI, 0.003-0.005]), whereas FA and T1 slightly decreased and increased (-0.002 [95% CI, -0.004 to -0.000] and 0.005 [95% CI, 0.001-0.009]). Quantitative magnetic resonance imaging shows that WMHs that regress have less abnormal microstructure at baseline than stable WMHs and follow trajectories indicating tissue improvement compared with stable and progressing WMHs.

An automated deep learning pipeline for EMVI classification and response prediction of rectal cancer using baseline MRI: a multi-centre study

ABSTRACTThe classification of extramural vascular invasion status using baseline magnetic resonance imaging in rectal cancer has gained significant attention as it is an important prognostic marker. Also, the accurate prediction of patients achieving complete response with primary staging MRI assists clinicians in determining subsequent treatment plans. Most studies utilised radiomics-based methods, requiring manually annotated segmentation and handcrafted features, which tend to generalise poorly. We retrospectively collected 509 patients from 9 centres, and proposed a fully automated pipeline for EMVI status classification and CR prediction with diffusion weighted imaging and T2-weighted imaging. We applied nnUNet, a self-configuring deep learning model, for tumour segmentation and employed learned multiple-level image features to train classification models, named MLNet. This ensures a more comprehensive representation of the tumour features, in terms of both fine-grained detail and global context. On external validation, MLNet, yielding similar AUCs as internal validation, outperformed 3D ResNet10, a deep neural network with ten layers designed for analysing spatiotemporal data, in both CR and EMVI tasks. For CR prediction, MLNet showed better results than the current state-of-the-art model using imaging and clinical features in the same external cohort. Our study demonstrated that incorporating multi-level image representations learned by a deep learning based tumour segmentation model on primary MRI improves the results of EMVI classification and CR prediction with good generalisation to external data. We observed variations in the contributions of individual feature maps to different classification tasks. This pipeline has the potential to be applied in clinical settings, particularly for EMVI classification.

Psychometric evaluation of the 5-item Medication Adherence Report Scale questionnaire in persons with multiple sclerosis.

The 5-item Medication Adherence Report Scale (MARS-5) is a reliable and valid questionnaire for evaluating adherence in patients with asthma, hypertension, and diabetes. Validity has not been determined in multiple sclerosis (MS). We aimed to establish criterion validity and reliability of the MARS-5 in persons with MS (PwMS). Our prospective study included PwMS on dimethyl fumarate (DMF). PwMS self-completed the MARS-5 on the same day before baseline and follow-up brain magnetic resonance imaging (MRI) 3 and 9 months after treatment initiation and were graded as highly and medium adherent upon the 24-cut-off score, established by receiver operator curve analysis. Health outcomes were represented by relapse occurrence from the 1st DMF dispense till follow-up brain MRI and radiological progression (new T2 MRI lesions and quantitative analysis) between baseline and follow-up MRI. Criterion validity was established by association with the Proportion of Days Covered (PDC), new T2 MRI lesions, and Beliefs in Medicines questionnaire (BMQ). The reliability evaluation included internal consistency and the test-retest method. We included 40 PwMS (age 37.6 ± 9.9 years, 75% women), 34 were treatment-naive. No relapses were seen during the follow-up period but quantitative MRI analysis showed new T2 lesions in 6 PwMS. The mean (SD) MARS-5 score was 23.1 (2.5), with 24 PwMS graded as highly adherent. The higher MARS-5 score was associated with higher PDC (b = 0.027, P<0.001, 95% CI: (0.0134-0.0403)) and lower medication concerns (b = -1.25, P<0.001, 95% CI: (-1.93-(-0,579)). Lower adherence was associated with increased number (P = 0.00148) and total volume of new T2 MRI lesions (P = 0.00149). The questionnaire showed acceptable internal consistency (Cronbach α = 0.72) and moderate test-retest reliability (r = 0.62, P < 0.0001, 95% CI: 0.33-0.79). The MARS-5 was found to be valid and reliable for estimating medication adherence and predicting medication concerns in persons with MS.

Impact of APOE-ε alleles on brain structure and cognitive function in healthy older adults: A VBM and DTI replication study.

The Apolipoprotein E (APOE) gene has been established in the Alzheimer's disease (AD) literature to impact brain structure and function and may also show congruent effects in healthy older adults, although findings in this population are much less consistent. The current study aimed to replicate and expand the multimodal approach employed by Honea et al. Structural magnetic resonance imaging (MRI), diffusion tensor imaging (DTI), and neuropsychological measures were used to investigate the impact of APOE-ε status on grey matter structure, white matter integrity, and cognitive functioning. Data were obtained from the Alzheimer's Disease Initiative Phase 3 (ADNI3) database. Baseline MRI, DTI and cognitive composite scores for memory (ADNI-Mem) and executive function (ADNI-EF) were acquired from 116 healthy controls. Participants were grouped according to APOE allele presence (APOE-ε2+ N = 17, APOE-ε3ε3 N = 64, APOE-ε4+ N = 35). Voxel-based morphometry (VBM) and tract-based spatial statistics (TBSS) were used to compare grey matter volume (GMV) and white matter integrity, respectively, between APOE-ε2+ and APOE-ε3ε3 controls, and again between APOE-ε4+ and APOE-ε3ε3 controls. Multivariate analysis of covariance (MANCOVA) was used to examine the effects of APOE polymorphism on memory and EF across all APOE groups with age, sex and education as regressors of no interest. Cognitive scores were correlated (Pearson r) with imaging metrics within groups. No significant differences were seen across groups, within groups in MRI metrics, or cognitive performance (p>0.05, corrected for multiple comparisons). The current study partially replicated and extended previous findings from an earlier multimodal study (Honea 2009). Future studies should clarify APOE mechanisms in healthy ageing by adding other imaging, cognitive, and lifestyle metrics and longitudinal design in larger sample sizes.