Dynamic Contrast-enhanced Research Articles

MRI-based vector radiomics for predicting breast cancer HER2 status and its changes after neoadjuvant therapy

Purpose: To develop a novel MRI-based vector radiomic approach to predict breast cancer (BC) human epidermal growth factor receptor 2 (HER2) status (zero, low, and positive; task 1) and its changes after neoadjuvant therapy (NAT) (positive-to-positive, positive-to-negative, and positive-to-pathologic complete response; task 2). Materials and Methods: Both dynamic contrast-enhanced (DCE) MRI data and multi-b-value (MBV) diffusion-weighted imaging (DWI) data were acquired in BC patients at two centers. Vector-radiomic and conventional-radiomic features were extracted from both DCE-MRI and MBV-DWI. After feature selection, the following models were built using the retained features and logistic regression: vector model, conventional model, and combined model that integrates the vector-radiomic and conventional-radiomic features. The models’ performances were quantified by the area under the receiver-operating characteristic curve (AUC). Results:The training/external test set (center 1/2) included 483/361 women. For task 1, the vector model (AUCs=0.73∼0.86) was superior to (p<.05) the conventional model (AUCs=0.68∼0.81), and the addition of vector-radiomic features to conventional-radiomic features yielded an incremental predictive value (AUCs=0.80∼0.90, p<.05). For task 2, the combined MBV-DWI model (AUCs=0.85∼0.89) performed better than (p<.05) the conventional MBV-DWI model (AUCs=0.73∼0.82). In addition, for the combined DCE-MRI model and the combined MBV-DWI model, the former (AUCs=0.85∼0.90) outperformed (p<.05) the latter (AUCs=0.80∼0.85) in task 1, whereas the latter (AUCs=0.85∼0.89) outperformed (p<.05) the former (AUCs=0.76∼0.81) in task 2. The above results are true for the training and external test sets. Conclusions:MRI-based vector radiomics may predict BC HER2 status and its changes after NAT and provide significant incremental prediction over and above conventional radiomics.

Added prognostic value of DCE blood volume imaging in patients with suspected recurrent or residual glioblastoma – a hybrid [18F]FET PET/MRI study

Abstract Background MRI cerebral blood volume (CBV) measurements improves diagnosis of recurrent gliomas. The study investigated the prognostic value of dynamic contrast enhanced (DCE) CBV imaging in treated IDH wildtype glioblastoma when added to MRI or amino acid PET. Methods Hybrid [18F]FET PET/MRI with 2CXM (2-compartment exchange model) DCE from 86 adult patients with suspected recurrent or residual glioblastoma were retrospectively analysed. High CBV tumour volume (VOLCBV), and contrast enhancing (VOLCE) and [18F]FET active tumour (VOLFET) volumes were delineated. Absolute and fractional high CBV subvolumes within VOLCE and VOLFET were determined. Associations with overall survival (OS) were assessed by Cox analysis. Results Adjusted for MGMT status and steroid all total tumour volumes were individually associated with shorter OS. Adding VOLCBV to VOLCE, or VOLFET only the effect of VOLCBV was prognostic of OS (HR 1.327, p=0.042 and 1.352, p=0.011, respectively). High CBV subvolume within both VOLCE and VOLFET, were associated with shorter survival (HR 1.448, p=0.042 and 1.416, p=0011, respectively), and the low CBV subvolumes with longer survival (HR 0.504, p=0.002 and 0.365, p=0.001, respectively). The fraction of VOLCE and VOLFET with high CBV was a strong predictor of OS with shorter median OS in upper vs lower tertiles (8.3 vs 14.5 months and 7.1 vs 15.6 months, respectively, both p&amp;lt;0.001). Conclusions The high CBV tumour volume was a strong prognosticator of survival and allowed for separation of high and low risk subvolumes underlining the heterogeneous physiological environment represented in the contrast enhancing or metabolically active tumour volumes of treated glioblastoma.

Enhancing bone metastasis prediction in prostate cancer using quantitative mpMRI features, ISUP grade and PSA density: a machine learning approach.

Bone metastasis is a critical complication in prostate cancer, significantly impacting patient prognosis and quality of life. This study aims to enhance bone metastasis prediction using machine learning (ML) techniques by integrating dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) perfusion features, International Society of Urological Pathology (ISUP) grade, and prostate-specific antigen (PSA) density. A retrospective analysis was conducted on 122 patients with histopathologically confirmed prostate cancer who underwent multiparametric prostate magnetic resonance imaging (mpMRI). Quantitative mpMRI features, PSA density, and ISUP grades were extracted and normalized. The dataset was balanced using oversampling and divided into training (70%) and test (30%) sets. Various ML models were developed and evaluated using area under the curve (AUC) metrics. Bone metastases were present in 26 patients (21.3%) at diagnosis. IAUGC and MaxSlope showed a statistically significant association with bone metastasis (p = 0.035, p = 0.050 respectively). The optimal PSA density cut-off value of 0.24 yielded a sensitivity of 0.88, specificity of 0.60, and AUC of 0.77. Machine learning models were developed using the dataset created with IAUGC, MaxSlope, ISUP grade, and PSA density values. Among the ML models, XGBoost demonstrated superior performance with validation and test AUCs of 91.5% and 92.6%, respectively, along with high precision (93.3%) and recall (93.1%). Integrating quantitative mpMRI features, ISUP grade, and PSA density through machine learning algorithms, particularly XGBoost, significantly improves the accuracy of bone metastasis prediction in prostate cancer patients. This approach can potentially reduce the need for additional imaging modalities and associated radiation exposure.

Exam quality of ultrasound and dynamic contrast-enhanced abbreviated MRI and impact on early-stage HCC detection.

MRI is a potential alternative to ultrasound for hepatocellular carcinoma (HCC) detection. We evaluated the impact of ultrasound and dynamic abbreviated MRI (AMRI) exam quality on early-stage HCC detection. We conducted a multicenter case-control study among patients with cirrhosis (cases with early-stage HCC per Milan Criteria; controls without HCC) who underwent both a liver ultrasound and dynamic contrast-enhanced (DCE) AMRI within 6 months in 2012-2019. Two radiologists performed independent, blinded interpretations of both exams for HCC detection and scored exam quality as no/mild, moderate, or severe limitations. Associations between exam quality, patient characteristics, and HCC detection were assessed by odds ratios (OR). Of 216 cases and 432 controls, severe limitations were reported in 7% and 8% of ultrasounds and DCE-AMRIs, respectively. Severe limitations at ultrasound were associated with obesity (OR 2.08, 95%CI [1.32-3.32]) and metabolic dysfunction-associated steatotic liver disease (MASLD) (OR 1.98 [1.12-3.44]) but not for DCE-AMRI. Decompensated cirrhosis (Child-Pugh C) was associated with severe limitations for both ultrasound (OR 2.54 [1.37-4.58]) and DCE-AMRI (OR 3.96 [2.36-6.58]). Compared to exams with no/mild limitations, exams with severe limitations had lower sensitivity for ultrasound (79.6% vs. 21.7%, P < 0.001) and AMRI (86.1% vs. 50.0%, P = 0.001). In patients in whom ultrasound was severely limited, DCE-AMRI had significantly higher odds of early-stage HCC detection than ultrasound (OR 8.23 [1.25-54.02]). HCC detection by DCE-AMRI may be preferred in patients with severe limitations at ultrasound due to obesity and MASLD. Both modalities remain limited for patients with decompensated cirrhosis, for whom alternative strategies may be needed.

Performance of an AI-powered visualization software platform for precision surgery in breast cancer patients.

Surgery remains the primary treatment modality in the management of early-stage invasive breast cancer. Artificial intelligence (AI)-powered visualization platforms offer the compelling potential to aid surgeons in evaluating the tumor's location and morphology within the breast and accordingly optimize their surgical approach. We sought to validate an AI platform that employs dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to render three-dimensional (3D) representations of the tumor and 5 additional chest tissues, offering clear visualizations as well as functionalities for quantifying tumor morphology, tumor-to-landmark structure distances, excision volumes, and approximate surgical margins. This retrospective study assessed the visualization platform's performance on 100 cases with ground-truth labels vetted by 2 breast-specialized radiologists. We assessed features including automatic AI-generated clinical metrics (e.g., tumor dimensions) as well as visualization tools including convex hulls at desired margins around the tumor to help visualize lumpectomy volume. The statistical performance of the platform's automated features was robust and within the range of inter-radiologist variability. These detailed 3D tumor and surrounding multi-tissue depictions offer both qualitative and quantitative comprehension of cancer topology and may aid in formulating an optimal surgical approach for breast cancer treatment. We further establish the framework for broader data integration into the platform to enhance precision cancer care.

Loss of aquaporin-4 impairs cerebrospinal fluid solute clearance through cerebrospinal fluid drainage pathways.

The aquaporin-4 (AQP4) water channel is essential in neurofluid dynamics. AQP4 loss impairs solute exchange between the cerebrospinal fluid (CSF) and interstitial fluid (ISF). However, whether AQP4 expression affects solute clearance from the CSF space to the extracranial space remains unclear. This study aimed to investigate this using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) following the intrathecal administration of gadolinium-based contrast agents (GBCAs) to AQP4 knockout (KO) rats. AQP4 KO rats showed reduced efflux of intrathecal GBCAs to the extracranial spaces through CSF drainage pathways and increased retention of intrathecal GBCAs in the CSF space compared with the controls. These results suggest that AQP4 loss impairs solute clearance from the CSF space to the extracranial spaces via the CSF drainage pathways. This study revealed a close relationship between AQP4 expression and CSF solute clearance, contributing to a better understanding of the function of AQP4 in neurofluid dynamics.

Prediction of genetic profile of breast carcinoma on MRI using a combination of DCE‐MRI, DWI, and MR spectroscopy: A prospective observational study

AbstractBackgroundClassification of breast cancer based on gene expression has emerged as the standard approach in its management, owing to the distinct prognoses and treatment responses observed among different subtypes. The aim of this study was to prospectively assess the imaging features of the molecular subtypes of breast cancer using multiparametric magnetic resonance imaging (mMRI) with the combined assessment of dynamic contrast‐enhanced magnetic resonance imaging (DCE‐MRI), diffusion‐weighted imaging (DWI), and MR spectroscopy (MRS).MethodsThis was a prospective observational single‐center cohort study, which included women with BI‐RADS 4−5 lesions on mammography/ultrasound (US) who subsequently underwent 1.5 T MRI (encompassing DCE‐MRI, DWI, and MRS). The histological subtypes of breast cancer were assessed. Estrogen receptor (ER), progesterone receptor (PR), Ki‐67 status, and human epidermal growth receptor‐2 (HER2) expression, assessed by immunohistochemistry (IHC), defined four molecular subtypes: luminal A, luminal B, HER2‐enriched (Her2en), and triple‐negative breast carcinoma (TNBC). Statistical associations between the four molecular subtypes and MRI features were investigated.ResultsFifty patients were included in the study. Circumscribed margins were significantly correlated with triple‐negative tumors compared to others (78% versus 6%, p &lt; 0.001). Spiculated margins were observed in non‐triple negative tumors. Rim enhancement was significantly correlated to triple‐negative tumors compared to all other subtypes (71.4% versus 25%, p = 0.035). Mean apparent diffusion coefficient (ADC) values were significantly lower for luminal subtypes compared to non‐luminal subtypes (p &lt; 0.001). The total choline (tCho) signal‐to‐noise ratio (SNR) was higher in triple‐negative tumors. A combined algorithm using DCE‐MRI, DWI, and MRS can predict TNBC and Her2en with specificity of 86.6% and 100%, respectively, and sensitivity of 100% and 85.37%, respectively.ConclusionThe combination of mMRI with DCE‐MRI, DWI, and MRS can accurately differentiate the molecular subtypes of breast carcinoma.

Added-value of dynamic contrast-enhanced MRI to conventional MRI for the differentiation between inflammatory myofibroblastic tumor and squamous cell carcinoma in the sinonasal region.

The purpose of this study was to evaluate the additional value of dynamic contrast-enhanced (DCE) MRI and diffusion weighted MRI (DWI) in differentiation between inflammatory myofibroblastic tumor (IMT) and squamous cell carcinoma (SCC) in the sinonasal cavity. Patients with pathologically proven IMT and SCC in the sinonasal region were enrolled in this retrospective study. All participants underwent conventional MRI and dynamic contrast-enhanced MRI, while a subset of them performed DWI. All the MRI parameters were independently analyzed by two investigators. This retrospective study included 21 patients with IMT and 55 patients with SCC. Significant differences were found in the conventional MR imaging features including mass margin, T2 signal intensity and track sign of maxillary (p < 0.05). For DCE-MRI features, significant differences were found in progressive centripetal continual enhancement and CImax (p < 0.001 and p = 0.026, respectively). A marginal significant difference was found in ADC values between IMT (0.86 ± 0.59) and SCC (1.14 ± 0.25) (p = 0.061). The conventional MRI analysis revealed that the combination of mass margin and track sign of maxillary yielded an accuracy of 81.6%. Using a combination of progressive centripetal continual enhancement on DCE-MRI and track sign of maxillary in multivariate logistic regression analysis, the accuracy was elevated to 92.1%. The incorporation of DCE-MRI features into conventional MRI showed improved diagnostic performance in differentiating IMT from SCC in the sinonasal region. The novel progressive centripetal continual enhancement on DCE-MRI is the most effective feature of IMT.

Abstract 4137891: Treated HIV infection is not associated with carotid vascular inflammation or plaque progression as assessed by dynamic contrast magnetic resonance imaging

Background: Inflammation and immune dysregulation are thought to drive residual cardiovascular disease risk among persons living with HIV (PLWH) despite effective viral suppression with antiretroviral therapy (ART). Question: We investigated differences in carotid vascular inflammation and atherosclerosis in a longitudinal cohort of virally suppressed PLWH (n = 50; on stable ART with CD4 &gt;250 cells/mm 3 , viral load &lt;200 copies/mL for &gt;6 months) and HIV-uninfected controls (n = 51) matched for age, sex, hypertension, diabetes, smoking, hyperlipidemia, and family history of premature coronary artery disease. Methods&amp;Results: Participants were &gt; 40 years old at enrollment, 8% female, and had a high prevalence of cardiovascular risk factors ( Table 1 ). Measures of carotid inflammation and capillary permeability ( K trans ), neovascularization ( V p ), and wall thickness were assessed at baseline, 1 year, and change over 1 year by dynamic contrast-enhanced magnetic resonance imaging. Both PLWH and controls demonstrated a reduction in systolic and diastolic blood pressures and total cholesterol over 1 year; however, the difference was not significant by HIV status. PLWH had a significant reduction in triglycerides compared with controls (-48.8 mg/dL vs 12.8 mg/dL; p = 0.026). HIV was not associated with baseline, follow-up, or change in markers of systemic inflammation assessed by plasma cytokines (C-reactive protein, interleukin-6, interleukin-1ß), nor vascular inflammation or plaque as assessed by K trans , V p , carotid wall thickness, or percent wall volume ( Tables 2 &amp; 3 ). Conclusions: In contrast to other studies of chronically treated and virally suppressed PLWH, HIV infection was not associated with carotid inflammation or plaque.

Evaluations of a Cutaneous Wound Healing Model Using Oxygen Enhanced - Dynamic Contrast Enhanced Photoacoustic Imaging (OE-DCE PAI).

Preclinical models of cutaneous wound healing can be useful for improving clinical wound care. Oxygen Enhanced Photoacoustic imaging (OE PAI) can measure oxygenation, and Dynamic Contrast Enhanced (DCE) PAI can measure vascular perfusion. We investigated how a combined OE-DCE PAI protocol can measure vascular oxygenation and perfusion to a cutaneous healing model. We developed a cutaneous "punch" wound model and photographed the wounds to track healing for 9 days. We performed OE-DCE PAI on Day 0, 3, 6, and 9. OE PAI was performed with 21% O2 and 100% O2 breathing gases to measure oxyhemoglobin (HbO2), deoxyhemoglobin (Hb), total hemoglobin (HbT), and oxygen saturation (%sO2), along with changes in these parameters caused by a change in breathing gas (ΔHb, ΔHbO2, ΔHbT, ΔsO2). To perform DCE PAI, indocyanine green (ICG) was administered intravenously while monitoring the PAI signal for 10min as the agent washed through the wound area, which was used to evaluate the wash-out rate. The average wound size was significantly smaller only by Day 6. For comparison, OE PAI measured a significant increase in HbO2, Hb, HbT, and %sO2 immediately after creating the wound, which significantly decreased by Day 3 and continued to decrease towards values for normal tissue by Day 9. The vascular wash-out rate significantly increased by Day 3, and continued to increase during the healing process. Notably, the wash-out rate can be assessed at a single PAI absorbance wavelength and by simply comparing signal amplitudes without advanced analysis, which may facilitate clinical translation. OE-DCE PAI can monitor significant changes in vascular perfusion and oxygenation prior to significant changes in cutaneous wound size. These results establish OE-DCE PAI as a tool for future pre-clinical wound healing studies and eventual clinical translation.

Abstract 4134729: Quantitative, time-efficient viability cardiac magnetic resonance with delayed-phase dynamic contrast enhancement model: a pilot study in dogs with myocardial infarction

Introduction: Late gadolinium enhancement (LGE) is the gold standard for myocardial viability imaging, recommended by all major cardiology societies. In this study, we aimed to develop a time-efficient viability imaging technique utilizing a dynamic contrast enhancement(dDCE) model to shorten the LGE wait time, provide quantitative measures of the contrast washout procedure, and potentially enable a fast and quantitative mean for scar imaging. Methods: A canine study(n=10) was conducted on reperfused myocardial infarction(MI). A dDCE model using dynamic post-contrast T1 maps was adopted to depict the contrast washout process. dDCE maps were reconstructed using the whole dataset(dDCE 30min ) and a 5-minute subset of the T1 maps(dDCE 5min ). To test the dDCE models for shortening the LGE wait time, LGE dDCE images were synthesized using the dDCE 5min maps and compared to the clinical LGE images. Remote and MI dDCE parameters on dDCE 30min and dDCE 5min maps were compared to test the model's ability to extract physiological features. Results: dDCE 30min and dDCE 5min map showed a good visuospatial difference between remote and MI(Fig. 1A) and no quantitative difference (Fig.1B). v e and PS showed a significant elevation in MI than in remote (61.12±13.65% vs 13.43±5.00%, P =0.018; 44.62±21.40mL/g/min vs 0.42±0.55mL/g/min, P =0.018, respectively). v p and F p were significantly decreased in MI than in remote (4.17±2.06% vs 10.85±3.77%, P =0.02; 1.11±0.32mL/g/min vs 1.51±0.42mL/g/min, P =0.04, respectively). Representative LGE and LGE dDCE images showed high agreement in the presence of MI (Fig. 2A). Bland-Altman analysis showed good agreement between LGE and LGE dDCE images for the measurement of infarct area (bias, -1.74 ± 6.60 %, Fig. 2B) and transmuraltiy (bias, 1.86 ± 2.73 %, Fig. 2C). The simple liner regression showed strong correlations between LGE and LGE dDCE images for the infarct area (R 2 , 0.95; slope, 0.93, P &lt;0.01; Fig. 2D) and transmurality (R 2 , 0.97; slope, 0.93, P &lt;0.01; Fig. 2E). ROC analysis showed that the AUC was 0.97 (95% CI: 0.94 to 1.00) (Fig. 2F). Increased lesion-blood pool contrast by 10 folds on dDCE images improved subendocardial MI detection (Fig.3). Conclusions: The developed dDCE model provides comparable viability assessment ability to the standard LGE images without the prolonged wait time and reflects MI pathophysiological changes. It shows the potential for a rapid and quantitative myocardial viability imaging protocol using cardiac magnetic resonance.

NIMG-36. PERFUSION, DIFFUSION, AND ANATOMICAL MRI BIOMARKERS FOR ASSESSING HISTOLOGICALLY-CONFIRMED MALIGNANT TRANSFORMATION IN MOLECULAR SUBTYPES OF IDH-MUTANT GLIOMAS

Abstract BACKGROUND Isocitrate dehydrogenase-mutant (IDHm) gliomas often initially present as low-grade (grade 2) gliomas but are expected to later progress into more aggressive higher-grade (grade 3-4) gliomas through malignant transformation (MT). MT is often determined non-invasively by emergence of contrast-enhancement on post-contrast T1-weighted magnetic resonance imaging (MRI), but contrast-enhancement is an imperfect surrogate for histologically-confirmed MT. This study explored perfusion, diffusion, and anatomical MRI biomarkers besides contrast-enhancement to study histologically-confirmed MT in IDHm 1p/19q-intact astrocytomas (IDHm-A) and IDHm 1p/19q-co-deleted oligodendrogliomas (IDHm-O). METHODS N=64 patients with initial histopathological grade 2 IDH-mutant glioma diagnosis who underwent repeated tissue sampling and were classified as histologically-confirmed MT (n=35) or non-MT (n=29) were retrospectively studied. Pre-surgical anatomical MRI (n=64) and, if available, diffusion-weighted imaging (n=61) and dynamic susceptibility contrast perfusion MRI (n=53) were analyzed. Tumor volumes of interest were segmented on the whole T2/FLAIR-hyperintense tumor. Measurable contrast enhancement (&amp;gt;1000mm3), tumor volume, sphericity, median apparent diffusion coefficient (ADC), and normalized relative cerebral blood volume (nrCBV) were compared between MT vs. non-MT IDHm-A and IDHm-O. RESULTS 81% of contrast-enhancing IDHm-A and 100% of contrast-enhancing IDHm-O underwent MT, while 41% of MT IDHm-A and 62% of MT IDHm-O were non-enhancing. Tumor volumes were significantly larger in the MT vs. non-MT groups for both IDHm-A (P=0.02) and IDHm-O (P=0.04). Whole tumor nrCBV was significantly higher (P=0.002), whole tumor ADC trended lower (P=0.06), and non-enhancing tumor sphericity was significantly lower (P=0.03) in MT vs. non-MT IDHm-A. There were no significant differences in ADC, nrCBV, nor sphericity when comparing MT vs. non-MT IDHm-O (P&amp;gt;0.05). CONCLUSIONS Many MT IDHm-gliomas remain non-enhancing. Tumor volumes can determine MT for both IDHm-A and IDHm-O. Diffusion and perfusion MRI show differences for MT in IDHm-A but not IDHm-O, which may reflect the different tumor biology of these IDHm molecular subtypes and their need for separate imaging biomarkers.

Comparing DCE-MRI and DSA: Understanding the embolization of hypervascular spinal metastases.

This study aims to examine and compare the effectiveness of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and Digital Subtraction Angiography (DSA) in evaluating hypervascular spinal metastases. A comprehensive literature review was conducted, utilizing top-tier databases such as PubMed, Scopus and Google Scholar, to compile an authoritative and up-to-date overview of the current advancements in the field. We synthesized key studies focusing on the advantages, limitations and efficacy of both imaging techniques. DCE-MRI provides a non-invasive method for evaluating tissue morphology, perfusion and vascularity, offering valuable information for cancer diagnosis and treatment monitoring. In contrast, DSA is an invasive procedure primarily used for embolization and diagnosing cerebrovascular events. Both modalities have distinct features regarding image acquisition, contrast agents, resolution and accessibility. DCE-MRI shows promise for cancer-related applications, offering advantages over conventional MRI by incorporating anatomical and hemodynamic parameters. While DSA remains important for cases requiring critical vascular information, further research is necessary to explore its potential therapeutic benefits in assessing vessel patency. Continued investigations are crucial to uncover additional insights and therapeutic applications for both DCE-MRI and DSA in medical imaging.

NIMG-37. “SYNTHETIC” PERFUSION MRI WITH ADJUSTABLE ACQUISITION PARAMETERS IN BRAIN TUMORS USING DYNAMIC SPIN-AND-GRADIENT-ECHO ECHOPLANAR IMAGING

Abstract INTRODUCTION Dynamic susceptibility contrast (DSC) perfusion MRI can assess normalized relative cerebral blood volume (nrCBV) and percentage of signal recovery (PSR) in brain tumors, which can aid in differential diagnosis, molecular profiling, and identification of disease progression. However, CBV-optimized and PSR-optimized protocols differ in terms of acquisition parameters. Additionally, DSC dependency on acquisition parameters limits the universalizability of diagnostic cutoffs for nrCBV and PSR across institutions and scanners. This study aims to generate “synthetic” DSC datasets with adjustable synthetic acquisition parameters using dynamic spin-and-gradient-echo echoplanar imaging (dynamic SAGE-EPI) to: 1) simultaneously generate nrCBV and PSR with optimal acquisition parameters, 2) compare DSC datasets with heterogeneous external cohorts. METHODS Thirty-eight patients with contrast-enhancing brain tumors were prospectively imaged with dynamic SAGE-EPI during a non-preloaded single-dose contrast injection. Multiple synthetic DSC datasets with desired pulse sequence parameters were generated using the Bloch equations applied to the dual-echo gradient-echo (GE) data extracted from dynamic SAGE-EPI datasets, with or without optional preload simulation. For each patient, synthetic protocols were set to match guideline-compliant CBV-optimized protocols, PSR-optimized protocols, and protocols used in external cohorts from the literature, for comparison. RESULTS Dynamic SAGE-EPI allowed for simultaneous generation of CBV-optimized and PSR-optimized DSC datasets with a single contrast injection. Conversely, suboptimal PSR computation from guideline-compliant CBV-optimized protocols resulted in rank variations within the cohort (Spearman’s ρ=0.83-0.89, i.e. 31%-21% rank variation). Treatment-naïve glioblastoma exhibited lower parameter-matched PSR compared to the external cohorts of treatment-naïve primary CNS lymphomas (PCNSL) (p&amp;lt;0.0001), supporting a role of synthetic DSC for multicenter comparisons. CONCLUSIONS Dynamic SAGE-EPI allows for simultaneous generation of CBV-optimized and PSR-optimized DSC data with a single injection, facilitating use of a single perfusion protocol for all DSC applications. This approach may also be useful for comparisons of perfusion parameters across heterogeneous multicenter datasets.

BIOM-54. AI-DRIVEN RISK-OF-PROGRESSION (AIRIP) CLASSIFIER FOR DISTINGUISHING RECURRENT BRAIN METASTASES FROM RADIATION TREATMENT EFFECT: A MULTI-INSTITUTIONAL COMPARATIVE STUDY WITH ADVANCED MULTIMODAL IMAGING

Abstract Following radiation therapy, a significant challenge in brain metastases (BM) management is differentiating radiation-induced-treatment effect (TrE) from tumor recurrence (TuR). TrE can be indistinguishable from TuR using conventional MRI. Advanced imaging techniques (e.g., perfusion MRI, PET/MRI) are not consistently used, and the standardized Response Assessment in Neuro-Oncology for brain metastases (RANO-BM) is sensitive to inter-reader variability. The performance of an artificial intelligence (AI)-driven risk-of-progression (AiRiP) classifier, which has been shown to capture pathophysiologic differences between TrE and TuR on routine MRI, was compared to that of clinical assessments and advanced imaging methods, in a multi-institutional setting. A total of n=261 lesions with pathologically-confirmed diagnoses in n=189 patients were analyzed. 115 lesions (73 TuR, 42 TrE) from site 1, 86 lesions (38 TuR, 48 TrE) from site 2, and 60 lesions (33 TuR, 27 TrE) from site 3 were used for training and testing the AiRiP-model. Gd-T1w, T2w, FLAIR MRI were preprocessed, and lesions were segmented by experts. Texture features (n=856) were extracted from each lesion. Random-forest classifier was employed for 3-fold cross-validation. Top-performing AiRiP-features, RANO-BM criteria, perfusion MRI and PET/MRI were evaluated in a sub-group analysis. For n=51 lesions on the test-set (site 3), 14 were classified as stable disease and 37 as TuR using RANO-BM (accuracy=54.1%). AiRiP-model achieved an accuracy of 76.5% on the same test-set and accurately classified 78.6% of the stable lesions as TrE or TuR. For another subset of lesions (n=27) on the same test-set, perfusion MRI and AiRiP-model achieved an accuracy of 59.3% and 70.4%, respectively. Lastly, for a subset of lesions (n=35) on the test-set (site 2), multimodal (perfusion, PET) imaging and AiRiP-model accurately classified 60% and 74.3% of lesions, respectively. 15 lesions were considered indeterminate via multimodal imaging, 73.3% of which AiRiP-model accurately classified as TrE or TuR. Our results suggest AI-driven models on conventional MRI may reliably distinguish TuR from TrE.

A scoping review of magnetic resonance angiography and perfusion image synthesis

IntroductionDeregulation of the cerebrovascular system has been linked to neurodegeneration, part of a putative causal pathway into etiologies such as Alzheimer's disease (AD). In medical imaging, time-of-flight magnetic resonance angiography (TOF-MRA) and perfusion MRI are the most common modalities used to study this system. However, due to lack of resources, many large-scale studies of AD are not acquiring these images; this creates a conundrum, as the lack of evidence limits our knowledge of the interaction between the cerebrovascular system and AD. Deep learning approaches have been used in recent developments to generate synthetic medical images from existing contrasts. In this review, we study the use of artificial intelligence in the generation of synthetic TOF-MRA and perfusion-related images from existing neuroanatomical and neurovascular acquisitions for the study of the cerebrovascular system.MethodFollowing the PRISMA reporting guidelines we conducted a scoping review of 729 studies relating to image synthesis of TOF-MRA or perfusion imaging, from which 13 met our criteria.ResultsStudies showed that T1-w, T2-w, and FLAIR can be used to synthesize perfusion map and TOF-MRA. Other studies demonstrated that synthetic images could have a greater signal-to-noise ratio compared to real images and that some models trained on healthy subjects could generalize their outputs to an unseen population, such as stroke patients.DiscussionThese findings suggest that generating TOF-MRA and perfusion MRI images holds significant potential for enhancing neurovascular studies, particularly in cases where direct acquisition is not feasible. This approach could provide valuable insights for retrospective studies of several cerebrovascular related diseases such as stroke and AD. While promising, further research is needed to assess their sensitivity and specificity, and ensure their applicability across diverse populations. The use of models to generate TOF-MRA and perfusion MRI using commonly acquired data could be the key for the retrospective study of the cerebrovascular system and elucidate its role in the development of dementia.

NIMG-57. EMBEDDED GRAPH DERIVED FROM PSEUDO-RESTING-STATE FUNCTIONAL MRI CAN PREDICT COGNITIVE IMPAIRMENT IN GLIOMA

Abstract Resting-state brain network analyses are of great interest in studying the neurocognition and measuring the functional connectivity (FC) patterns of glioma patients. However, resting-state functional MRI (rs-fMRI) is not assessed in clinical routine due to the scan time and cost. The current study used pseudo-resting-state functional MRI (pseudo-rs-fMRI) derived from dynamic susceptibility contrast (DSC) perfusion MRI to predict cognitive impairment in glioma. 37 glioma patients were enrolled consecutively in the current study with DSC perfusion MRI acquired and neurocognition assessed, where 24 glioma patients also got rs-fMRI collected. The pseudo-rs-fMRI was created by voxel-wise subtracting the Gamma-variate modeled signal of contrast agent bolus from the original DSC perfusion signal. Following the pre-processing of pseudo-rs-fMRI and full rs-fMRI, the functional connectivity network (FCN) was established for each patient. Graph embedding was implemented to learn and extract features of each node in the binarized FCN, and the decision tree classification algorithm was applied to distinguish cognitive impairment. FCNs of 24 patients with DSC perfusion and rs-fMRI acquired were used to train the model, and FCNs of 13 patients with only DSC perfusion MRI acquired were used to test the model. The AUC (area under the receiver operating characteristic curve) was used to evaluate the model’s performance. The similarity and mean-square-difference between average FCNs extracted from rs-fMRI and pseudo-rs-fMRI were 0.6769 and 0.0263, respectively. Classifier trained using 24 FCNs of pseudo-rs-fMRI has an AUC of 0.750 in identifying cognitively impaired patients in the testing cohort. Combining FCNs of both rs-fMRI and pseudo-rs-fMRI to train the classifier resulted in a higher AUC of 0.833 in the testing cohort. To summary, DSC perfusion MRI-derived pseudo-rs-fMRI can be used to predict cognitive impairment in patients with gliomas. External validation would be required.

NIMG-88. EVALUATION OF DIFFUSION AND PERFUSION MRI FEATURES FOR DIFFERENTIATING RECURRENT BRAIN METASTASES AND RADIATION NECROSIS

Abstract BACKGROUND Brain metastases (BM) are a frequent and serious complication of cancer. Differentiating between BM and radiation-induced necrosis (RN) after radiotherapy remains challenging. Diffusion-weighted imaging (DWI) derived apparent diffusion coefficient (ADC) maps and perfusion-weighted imaging (PWI) derived relative cerebral blood volume (rCBV) have the potential to distinguish between BM and RN. Despite the potential of these imaging modalities, previous studies have been limited by the absence of histopathological confirmation. In this study, we utilized a patient cohort that underwent stereotactic biopsy as part of their Laser Interstitial Thermal Therapy (LITT) following suspicious radiographic progression. This allowed us to evaluate the efficacy of ADC and rCBV imaging features with histopathological validation. METHODS In this study, 29 patients underwent MRI scans (3 with 1.5T, 26 with 3T) prior to LITT. Pathology confirmed 11 patients with BM and 18 with RN. Regions of interest (ROIs) in the contrast-enhanced lesion and T2 hyperintensity lesion were identified using post-contrast T1-weighted images and fluid-attenuated inversion recovery (FLAIR) images. Mean values and histogram-based features (standard deviation, 10%, 90%) of lesion volume, ADC, and rCBV within tumor ROIs were compared between the BM and RN using the Wilcoxon rank-sum test. RESULTS Our findings indicated that the mean values of ADC in the enhancing and non-enhancing lesion ROIs did not significantly differ between BM and RN patients. However, in the contrast-enhanced lesion regions, we observed a significant difference in the standard deviation of rCBV (p&amp;lt;0.05), and the 90th percentile value of rCBV also suggested potential differences between BM and RN (p=0.069). In the T2 hyperintensity lesion regions, lesion volume (p&amp;lt;0.05) and the mean of rCBV (p&amp;lt;0.05) show significant differences between BM and RN. CONCLUSION This study demonstrates the potential of using rCBV to aid in the clinical diagnosis of BM. Additionally, our findings indicate that T2 hyperintensity regions, beyond just contrast-enhancing lesions, may provide valuable information for differential diagnosis.

Quantitative Estimation of Iron and Fat Content in Prostate Cancer by Multiparametric MRI and Its Application in Optimizing D'Amico Score.

The risk of biochemical recurrence (BCR) in prostate cancer (PCa) is typically assessed using D'Amico score. However, iron and fat content in PCa are closely related to tumor cell proliferation and the risk of BCR may be estimated using multiparametric MRI (mpMRI). To noninvasively estimate fat and iron content in PCa and to evaluate their utility in enhancing D'Amico scores for predicting BCR in PCa patients. Prospective. Forty-eight male patients in the BCR group (age 71.31 ± 5.74 years) and 27 male patients in the non-BCR group (age 70.3 ± 6.04 years). 3.0 T, Turbo-spin echo T2-weighted imaging, diffusion-weighted imaging (DWI), dynamic contrast-enhanced (DCE) imaging, Gradient echo Q-Dixon sequence. The mean fat fraction (FF) and T2* values of lesions were extracted from the FF map and the T2* map. Additionally, prostate volume, mean apparent diffusion coefficient (ADC) value, periprostatic fat thickness (PPFT), subcutaneous fat thickness (SFT), blood lipid content, pre- and post-operative prostate-specific antigen (PSA) values were collected. Stepwise-COX regression analysis was employed to identify the significant predictors of BCR, which led to the construction of an improvement-adjusted (IA) model. Then the IA model as well as the D'Amico score were evaluated using C-index and time-dependent AUC, decision-curve analysis, and Kaplan-Meier curve. P < 0.05 was statistically significant. Significant differences were observed in PSA, D'Amico score, ISUP grade, T2*, FF, and ADC values of the lesions in the BCR group compared with the non-BCR group. Mean T2*, FF, and ADC values of the lesions were screened to construct the IA model incorporated into the D'Amico score (IA Model: C-index = 0.749; AUC = 0.812; D'Amico score: C-index = 0.672; AUC = 0.723). This study demonstrated that mpMRI can quantitatively estimate fat and iron within PCa lesions. By integrating ADC, FF, and T2* values into the D'Amico score, the preoperative-risk assessment for BCR can be improved. 2 TECHNICAL EFFICACY: Stage 2.

NIMG-84. BELZUTIFAN REDUCTION OF CEREBRAL BLOOD FLOW IN VHL-ASSOCIATED HEMANGIOBLASTOMA PREDICTS RESPONSE TO TREATMENT

Abstract Hemangioblastoma (HMB) is a vascular-origin tumor occurring primarily in the cerebellum. Treatment typically involves surgical resection of the solid part of the tumor or stereotactic radiosurgery. An emerging treatment for hemangioblastoma is belzutifan (BZF), a hypoxia-inducible factor 2a (HIF2a) inhibitor. A recent trial found BZF to control hemangioblastoma in patients with von Hippel Lindau (vHL) syndrome. Perfusion MRI techniques including arterial spin labelling (ASL) and can be used to determine relative cerebral blood flow (rCBF). No prior studies have investigated rCBF in hemangioblastoma with relation to belzutifan, and the mechanism of action of belzutifan on hemangioblastoma is not well understood. Here, we aim to measure the response in rCBF in patients with vHL-associated hemangioblastoma treated with belzutifan. This retrospective cohort study constituted a group of seven patients with vHL-associated hemangioblastoma treated with Belzutifan at a daily oral dose of 120mg. Perfusion MRI with ASL sequences were acquired as part of the tumor imaging protocol. We quantified rCBF of the tumors prior to treatment, following intiation of BZF, and at most recent follow-up. Statistics were performed using R studio. The long axis diameter, volume of the solid tumor, rCBF mean, rCBF max, N mean, N max of lesions all decreased after Belzutifan treatment (p&amp;lt;0.001). While the long axis diameter of the cyst in cystic lesions did not show statistically significant change, the volume of cysts decreased significantly on Belzutifan (p=0.016). Spearman’s rank correlation analysis demonstrated association between change in rCBF mean, rCBF max, N mean and N max to % reduction in tumor volume. In addition, change in rCBF mean, change in rCBF max and change in N max all showed statistically significant association on multiple linear regression. Finally, early change in CBF and N predicted further volumetric response to treatment with belzutifan. We demonstrate that vHL-associated hemangioblastoma demonstrates decrease in rCBF in conjunction with and predictive of decreased volume following treatment with belzutifan.