Contrast-enhanced T1-weighted Sequences Research Articles

Assessing the stability and discriminative ability of radiomics features in the tumor microenvironment: Leveraging peri-tumoral regions in vestibular schwannoma

PurposeThe tumor microenvironment (TME) plays a crucial role in tumor progression and treatment response. Radiomics offers a non-invasive approach to studying the TME by extracting quantitative features from medical images. In this study, we present a novel approach to assess the stability and discriminative ability of radiomics features in the TME of vestibular schwannoma (VS). MethodsMagnetic Resonance Imaging (MRI) data from 242 VS patients were analyzed, including contrast-enhanced T1-weighted (ceT1) and high-resolution T2-weighted (hrT2) sequences. Radiomics features were extracted from concentric peri-tumoral regions of varying sizes. The intraclass correlation coefficient (ICC) was used to assess feature stability and discriminative ability, establishing quantile thresholds for ICCmin and ICCmax. ResultsThe identified thresholds for ICCmin and ICCmax were 0.45 and 0.72, respectively. Features were classified into four categories: stable and discriminative (S-D), stable and non-discriminative (S-ND), unstable and discriminative (US-D), and unstable and non-discriminative (US-ND). Different feature groups exhibited varying proportions of S-D features across ceT1 and hrT2 sequences. The similarity of S-D features between ceT1 and hrT2 sequences was evaluated using Jaccard’s index, with a value of 0.78 for all feature groups which is ranging from 0.68 (intensity features) to 1.00 (Neighbouring Gray Tone Difference Matrix (NGTDM) features). ConclusionsThis study provides a framework for identifying stable and discriminative radiomics features in the TME, which could serve as potential biomarkers or predictors of patient outcomes, ultimately improving the management of VS patients.

Focal leptomeningeal vascular anomalies on brain MRI: A mimic of leptomeningeal metastatic disease

Abstract Background The diagnosis of leptomeningeal metastatic disease has major prognostic and therapeutic implications. We report 13 patients with a radiologically distinct kind of focal, enhancing leptomeningeal lesion on brain MRI that mimics leptomeningeal metastatic disease. Methods These patients were assessed at University Health Network (Toronto, Canada) between January 2001 and December 2023. Results Median age was 68 years (range, 55–78 years) and 10 patients were women. All patients had brain magnetic resonance imaging (MRI) including contrast-enhanced T2-weighted fluid attenuation inversion recovery (FLAIR) and T1-weighted spin echo sequences. MRI in all 13 patients showed a focal enhancing lesion located along the leptomeningeal surface of the brain. The MRI exams were reported as possible or likely leptomeningeal metastatic disease for the majority (9/13) of patients. Each lesion was curvilinear rather than sheet-like, and some lesions consisted of multiple connected/branching curvilinear structures with the appearance of abnormal vessels. The lesions were distinct from normal blood vessels. Some lesions had a visible connection with a nearby cortical vein. Follow-up contrast-enhanced brain MRI for 8/13 (62%) patients at a median of 3.9 years (IQR 2.4–6.6 years) showed all lesions were unchanged over time. Another 2/13 (15%) patients had clinical and CT brain follow-up after one year with no evidence of metastatic disease. Conclusions We describe a distinct kind of focal, enhancing leptomeningeal lesion on brain MRI that mimics metastatic disease. These lesions are likely a type of low-flow vascular anomaly. Their curvilinear/branching shape and intense enhancement particularly on T2-weighted FLAIR images distinguish these lesions from tumors.

E.1 Focal leptomeningeal vascular anomalies on brain MRI: a mimic of leptomeningeal metastatic disease

Background: The diagnosis of leptomeningeal metastatic disease has major prognostic implications. We report 13 patients with a radiologically distinct, focal, enhancing leptomeningeal lesion on brain MRI mimicking leptomeningeal metastatic disease. Methods: These patients were assessed at University Health Network between January 2001 and December 2023. Results: Median age was 68 years and 10 patients were women. All patients had brain MRI including contrast-enhanced T2-weighted FLAIR and T1-weighted spin echo sequences. MRI in all patients showed a focal enhancing lesion along the leptomeningeal surface of the brain. The MRI exams were reported as possible metastatic disease for the majority (9/13) of patients. Each lesion was curvilinear rather than sheet-like, and some consisted of multiple connected/branching curvilinear structures with the appearance of abnormal vessels. Some lesions had visible connection with a nearby cortical vein. The lesions were distinct from normal blood vessels. Follow-up contrast-enhanced brain MRI for 8/13 patients at a median of 3.9 years showed all lesions were unchanged over time. Conclusions: We describe a distinct kind of focal, enhancing leptomeningeal lesion on brain MRI that mimics metastatic disease. These lesions are likely a type of low-flow vascular anomaly. Their curvilinear/branching shape and intense enhancement particularly on T2-weighted FLAIR images distinguishes these lesions from tumor.

Preferred Imaging for Target Volume Delineation for Radiotherapy of Recurrent Glioblastoma: A Literature Review of the Available Evidence.

Recurrence in glioblastoma lacks a standardized treatment, prompting an exploration of re-irradiation's efficacy. A comprehensive systematic review from January 2005 to May 2023 assessed the role of MRI sequences in recurrent glioblastoma re-irradiation. The search criteria, employing MeSH terms, targeted English-language, peer-reviewed articles. The inclusion criteria comprised both retrospective and prospective studies, excluding certain types and populations for specificity. The PICO methodology guided data extraction, and the statistical analysis employed Chi-squared tests via MedCalc v22.009. Out of the 355 identified studies, 81 met the criteria, involving 3280 patients across 65 retrospective and 16 prospective studies. The key findings indicate diverse treatment modalities, with linac-based photons predominating. The median age at re-irradiation was 54 years, and the median time interval between radiation courses was 15.5 months. Contrast-enhanced T1-weighted sequences were favored for target delineation, with PET-imaging used in fewer studies. Re-irradiation was generally well tolerated (median G3 adverse events: 3.5%). The clinical outcomes varied, with a median 1-year local control rate of 61% and a median overall survival of 11 months. No significant differences were noted in the G3 toxicity and clinical outcomes based on the MRI sequence preference or PET-based delineation. In the setting of recurrent glioblastoma, contrast-enhanced T1-weighted sequences were preferred for target delineation, allowing clinicians to deliver a safe and effective therapeutic option; amino acid PET imaging may represent a useful device to discriminate radionecrosis from recurrent disease. Future investigations, including the ongoing GLIAA, NOA-10, ARO 2013/1 trial, will aim to refine approaches and standardize methodologies for improved outcomes in recurrent glioblastoma re-irradiation.

MRI-based intratumoral and peritumoral radiomics for preoperative prediction of glioma grade: a multicenter study.

Accurate preoperative prediction of glioma is crucial for developing individualized treatment decisions and assessing prognosis. In this study, we aimed to establish and evaluate the value of integrated models by incorporating the intratumoral and peritumoral features from conventional MRI and clinical characteristics in the prediction of glioma grade. A total of 213 glioma patients from two centers were included in the retrospective analysis, among which, 132 patients were classified as the training cohort and internal validation set, and the remaining 81 patients were zoned as the independent external testing cohort. A total of 7728 features were extracted from MRI sequences and various volumes of interest (VOIs). After feature selection, 30 radiomic models depended on five sets of machine learning classifiers, different MRI sequences, and four different combinations of predictive feature sources, including features from the intratumoral region only, features from the peritumoral edema region only, features from the fusion area including intratumoral and peritumoral edema region (VOI-fusion), and features from the intratumoral region with the addition of features from peritumoral edema region (feature-fusion), were established to select the optimal model. A nomogram based on the clinical parameter and optimal radiomic model was constructed for predicting glioma grade in clinical practice. The intratumoral radiomic models based on contrast-enhanced T1-weighted and T2-flair sequences outperformed those based on a single MRI sequence. Moreover, the internal validation and independent external test underscored that the XGBoost machine learning classifier, incorporating features extracted from VOI-fusion, showed superior predictive efficiency in differentiating between low-grade gliomas (LGG) and high-grade gliomas (HGG), with an AUC of 0.805 in the external test. The radiomic models of VOI-fusion yielded higher prediction efficiency than those of feature-fusion. Additionally, the developed nomogram presented an optimal predictive efficacy with an AUC of 0.825 in the testing cohort. This study systematically investigated the effect of intratumoral and peritumoral radiomics to predict glioma grading with conventional MRI. The optimal model was the XGBoost classifier coupled radiomic model based on VOI-fusion. The radiomic models that depended on VOI-fusion outperformed those that depended on feature-fusion, suggesting that peritumoral features should be rationally utilized in radiomic studies.

Comparison of MRI Sequences to Predict IDH Mutation Status in Gliomas Using Radiomics-Based Machine Learning.

Regarding the 2021 World Health Organization (WHO) classification of central nervous system (CNS) tumors, the isocitrate dehydrogenase (IDH) mutation status is one of the most important factors for CNS tumor classification. The aim of our study is to analyze which of the commonly used magnetic resonance imaging (MRI) sequences is best suited to obtain this information non-invasively using radiomics-based machine learning models. We developed machine learning models based on different MRI sequences and determined which of the MRI sequences analyzed yields the highest discriminatory power in predicting the IDH mutation status. In our retrospective IRB-approved study, we used the MRI images of 106 patients with histologically confirmed gliomas. The MRI images were acquired using the T1 sequence with and without administration of a contrast agent, the T2 sequence, and the Fluid-Attenuated Inversion Recovery (FLAIR) sequence. To objectively compare performance in predicting the IDH mutation status as a function of the MRI sequence used, we included only patients in our study cohort for whom MRI images of all four sequences were available. Seventy-one of the patients had an IDH mutation, and the remaining 35 patients did not have an IDH mutation (IDH wild-type). For each of the four MRI sequences used, 107 radiomic features were extracted from the corresponding MRI images by hand-delineated regions of interest. Data partitioning into training data and independent test data was repeated 100 times to avoid random effects associated with the data partitioning. Feature preselection and subsequent model development were performed using Random Forest, Lasso regression, LDA, and Naïve Bayes. The performance of all models was determined with independent test data. Among the different approaches we examined, the T1-weighted contrast-enhanced sequence was found to be the most suitable for predicting IDH mutations status using radiomics-based machine learning models. Using contrast-enhanced T1-weighted MRI images, our seven-feature model developed with Lasso regression achieved a mean area under the curve (AUC) of 0.846, a mean accuracy of 0.792, a mean sensitivity of 0.847, and a mean specificity of 0.681. The administration of contrast agents resulted in a significant increase in the achieved discriminatory power. Our analyses show that for the prediction of the IDH mutation status using radiomics-based machine learning models, among the MRI images acquired with the commonly used MRI sequences, the contrast-enhanced T1-weighted images are the most suitable.

Multiphase MRI-Based Radiomics for Predicting Histological Grade of Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) is a highly heterogeneous cancer. Accurate preoperative prediction of histological grade holds potential for improving clinical management and disease prognostication. To evaluate the performance of a radiomics signature based on multiphase MRI in assessing histological grade in solitary HCC. Retrospective. A total of 405 patients with histopathologically confirmed solitary HCC and with liver gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced MRI within 1 month of surgery. Contrast-enhanced T1-weighted spoiled gradient echo sequence (LAVA) at 1.5 or 3.0 T. Tumors were graded (low/high) according to results of histopathology. Basic clinical characteristics (including age, gender, serum alpha-fetoprotein (AFP) level, history of hepatitis B, and cirrhosis) were collected and tumor size measured. Radiomics features were extracted from Gd-EOB-DTPA-enhanced MRI data. Three feature selection strategies were employed sequentially to identify the optimal features: SelectFromModel (SFM), SelectPercentile (SP), and recursive feature elimination with cross-validation (RFECV). Probabilities of five single-phase radiomics-based models were averaged to generate a radiomics signature. A combined model was built by combining the radiomics signature and clinical predictors. Pearson χ2 test/Fisher exact test, Wilcoxon rank sum test, interclass correlation coefficient (ICC), univariable/multivariable logistic regression analysis, area under the receiver operating characteristic (ROC) curve (AUC), DeLong test, calibration curve, Brier score, decision curve, Kaplan-Meier curve, and log-rank test. A P-value <0.05 was considered statistically significant. High-grade HCCs were present in 33.8% of cases. AFP levels (odds ratio [OR] 1.89) and tumor size (>5 cm; OR 2.33) were significantly associated with HCC grade. The combined model had excellent performance in assessing HCC grade in the test dataset (AUC: 0.801), and demonstrated satisfactory calibration and clinical utility. A model that combined a radiomics signature derived from preoperative multiphase Gd-EOB-DTPA-enhanced MRI and clinical predictors showed good performance in assessing HCC grade. 3 TECHNICAL EFFICACY: Stage 5.

Inflammatory Knee Synovitis: Evaluation of an Accelerated FLAIR Sequence Compared With Standard Contrast-Enhanced Imaging.

The aim of this study was to assess the diagnostic value and accuracy of a deep learning (DL)-accelerated fluid attenuated inversion recovery (FLAIR) sequence with fat saturation (FS) in patients with inflammatory synovitis of the knee. Patients with suspected knee synovitis were retrospectively included between January and September 2023. All patients underwent a 3 T knee magnetic resonance imaging including a DL-accelerated noncontrast FLAIR FS sequence (acquisition time: 1 minute 38 seconds) and a contrast-enhanced (CE) T1-weighted FS sequence (acquisition time: 4 minutes 50 seconds), which served as reference standard. All knees were scored by 2 radiologists using the semiquantitative modified knee synovitis score, effusion synovitis score, and Hoffa inflammation score. Diagnostic confidence, image quality, and image artifacts were rated on separate Likert scales. Wilcoxon signed rank test was used to compare the semiquantitative scores. Interreader and intrareader reproducibility were calculated using Cohen κ. Fifty-five patients (mean age, 52 ± 17 years; 28 females) were included in the study. Twenty-seven patients (49%) had mild to moderate synovitis (synovitis score 6-13), and 17 patients (31%) had severe synovitis (synovitis score >14). No signs of synovitis were detected in 11 patients (20%) (synovitis score <5). Semiquantitative assessment of the whole knee synovitis score showed no significant difference between the DL-accelerated FLAIR sequence and the CE T1-weighted sequence (mean FLAIR score: 10.69 ± 8.83, T1 turbo spin-echo FS: 10.74 ± 10.32; P = 0.521). Both interreader and intrareader reproducibility were excellent (range Cohen κ [0.82-0.96]). Assessment of inflammatory knee synovitis using a DL-accelerated noncontrast FLAIR FS sequence was feasible and equivalent to CE T1-weighted FS imaging.

Utility of Prostate Health Index Density for Biopsy Strategy in Biopsy-Naïve Patients With PI-RADS v2.1 Category 3 Lesions.

Category 3 lesions in PI-RADSv2.1 pose diagnostic challenges, complicating biopsy decisions. Recent biomarkers like prostate health index (PHI) have shown higher specificity in detecting clinically significant prostate cancer (csPCa) than prostate-specific antigen (PSA). Yet their integration with MRI remains understudied. To evaluate the utility of PSA and PHI with its derivatives for detecting csPCa in biopsy-naïve patients with category 3 lesion on initial prostate MRI scan. Retrospective. One hundred ninety-three biopsy-naïve patients who underwent MRI, PSA, and PHI testing, followed by both targeted and systematic biopsies. Turbo spin-echo T2-weighted imaging, diffusion-weighted single-shot echo-planar imaging, and dynamic contrast-enhanced T1-weighted fast field echo sequence imaging in 3 T. PHI density (PHID) and PSA density (PSAD) derived by dividing serum PHI and PSA with prostate volume (MRI based methodology suggested by PI-RADSv2.1). Risk-stratified models to evaluate the utility of markers in triaging patients for biopsy, including low-, intermediate-, and high-risk groups. Independent t-test, Mann-Whitney U test, Mantel-Haenszel test, generalized estimating equation, and receiver operating characteristic (ROC) curve analysis were used. Statistical significance defined as P < 0.05. CsPCa was found in 16.6% (32/193) of patients. PHID had the highest area under the ROC curve (AUROC) of 0.793, followed by PHI of 0.752, PSAD of 0.750, and PSA of 0.654. PHID with two cut-off points (0.88/mL and 1.82/mL) showed the highest potential biopsy avoidance of 47.7% (92/193) with 5% missing csPCa, and the lowest intermediate-risk group (borderline decision group) at 38.9% (75/193), compared to PSA and PHI. PHID demonstrated better potential in triaging patients with category 3 lesions, possibly aiding more selective and confident biopsy decisions for csPCa detection, than traditional markers. 4 TECHNICAL EFFICACY: Stage 5.

MRI Treatment Response Assessment Maps (TRAMs) for differentiating recurrent glioblastoma from radiation necrosis.

MRI treatment response assessment maps (TRAMs) were introduced to distinguish recurrent malignant glioma from therapy related changes. TRAMs are calculated with two contrast-enhanced T1-weighted sequences and reflect the "late" wash-out (or contrast clearance) and wash-in of gadolinium. Vital tumor cells are assumed to produce a wash-out because of their high turnover rate and the associated hypervascularization, whereas contrast medium slowly accumulates in scar tissue. To examine the real value of this method, we compared TRAMs with the pathology findings obtained after a second biopsy or surgery when recurrence was suspected. We retrospectively evaluated TRAMs in adult patients with histologically demonstrated glioblastoma, contrast-enhancing tissue and a pre-operative MRI between January 1, 2017, and December 31, 2022. Only patients with a second biopsy or surgery were evaluated. Volumes of the residual tumor, contrast clearance and contrast accumulation before the second surgery were analyzed. Among 339 patients with mGBM who underwent MRI, we identified 29 repeated surgeries/biopsies in 27 patients 59 ± 12 (mean ± standard deviation) years of age. Twenty-eight biopsies were from patients with recurrent glioblastoma histology, and only one was from a patient with radiation necrosis. We volumetrically evaluated the 29 pre-surgery TRAMs. In recurrent glioblastoma, the ratio of wash-out volume to tumor volume was 36 ± 17% (range 1-73%), and the ratio of the wash-out volume to the sum of wash-out and wash-in volumes was 48 ± 21% (range 22-92%). For the one biopsy with radiation necrosis, the ratios were 42% and 54%, respectively. Typical recurrent glioblastoma shows a > 20%ratio of the wash-out volume to the sum of wash-out and wash-in volumes. The one biopsy with radiation necrosis indicated that such necrosis can also produce high wash-out in individual cases. Nevertheless, the additional information provided by TRAMs increases the reliability of diagnosis.

Presentation and Management of Cerebral Venous Sinus Thrombosis After Supratentorial Craniotomy.

Cerebral venous sinus thrombosis (CVST) after supratentorial craniotomy is a poorly studied complication, for which there are no management guidelines. This study assessed the incidence, associated risk factors, and management of postoperative CVST after awake craniotomy. This is an observational, retrospective, monocentric analysis of patients who underwent a supratentorial awake craniotomy. Postoperative CVST was defined as a flow defect on the postoperative contrast-enhanced 3D T1-weighted sequence and/or as a T2* hypointensity within the sinus. In 401 supratentorial awake craniotomies (87.3% of diffuse glioma), the incidence of postoperative CVST was 4.0% (95% CI 2.5-6.4): 14/16 thromboses located in the superior sagittal sinus and 12/16 located in the transverse sinus. A venous sinus was exposed during craniotomy in 45.4% of cases, and no intraoperative injury to a cerebral venous sinus was reported. All thromboses were asymptomatic, and only two cases were diagnosed at the time of the first postoperative imaging (0.5%). Postoperative complications, early postoperative Karnofsky Performance Status score, and duration of hospital stay did not significantly differ between patients with and without postoperative CVST. Adjusted independent risk factors of postoperative CVST were female sex (adjusted Odds Ratio 4.00, 95% CI 1.24-12.91, P = .021) and a lesion ≤1 cm to a venous sinus (adjusted Odds Ratio 10.58, 95% CI 2.93-38.20, P < .001). All patients received standard prophylactic-dose anticoagulant therapy, and none received treatment-dose anticoagulant therapy. No thrombosis-related adverse event was reported. All thromboses presented spontaneous sinus recanalization radiologically at a mean of 89 ± 41 days (range, 7-171). CVST after supratentorial awake craniotomy is a rare event with satisfactory clinical outcomes and spontaneous sinus recanalization under conservative management without treatment-dose anticoagulant therapy. These findings are comforting to neurosurgeons confronted with postoperative MRI reports suggesting CVST.

A deep learning model integrating multisequence MRI to predict EGFR mutation subtype in brain metastases from non-small cell lung cancer.

To establish a predictive model based on multisequence magnetic resonance imaging (MRI) using deep learning to identify wild-type (WT) epidermal growth factor receptor (EGFR), EGFR exon 19 deletion (19Del), and EGFR exon 21-point mutation (21L858R) simultaneously. A total of 399 patients with proven brain metastases of non-small cell lung cancer (NSCLC) were retrospectively enrolled and divided into training (n = 306) and testing (n = 93) cohorts separately based on two timepoints. All patients underwent 3.0-T brain MRI including T2-weighted, T2-weighted fluid-attenuated inversion recovery, diffusion-weighted imaging, and contrast-enhanced T1-weighted sequences. Radiomics features were extracted from each lesion based on four sequences. An algorithm combining radiomics approach with graph convolutional networks architecture (Radio-GCN) was designed for the prediction of EGFR mutation status and subtype. The area under the curve (AUC) at receiver operating characteristic analysis was used to evaluate the predication capabilities of each model. We extracted 1,290 radiomics features from each MRI sequence. The AUCs of the Radio-GCN model for identifying EGFR 19Del, 21L858R, and WT for the lesion-wise analysis were 0.996 ± 0.004, 0.971 ± 0.013, and 1.000 ± 0.000 on the independent testing cohort separately. It also yielded AUCs of 1.000 ± 0.000, 0.991 ± 0.009, and 1.000 ± 0.000for predicting EGFR mutations respectively for the patient-wise analysis. The κ coefficients were 0.735 and 0.812, respectively. The constructed Radio-GCN model is a new potential tool to predict the EGFR mutation status and subtype in NSCLC patients with brain metastases. The study demonstrated that a deep learning approach based on multisequence MRI can help to predict the EGFR mutation status in NSCLC patients with brain metastases, which is beneficial to guide a personalized treatment. • This is the first study to predict the EGFR mutation subtype simultaneously. • The Radio-GCN model holds the potential to be used as a diagnostic tool. • This study provides an imaging surrogate for identifying the EGFR mutation subtype.

Fully automated MR-based virtual biopsy of primary CNS lymphomas.

Primary central nervous system lymphomas (PCNSL) pose a challenge as they may mimic gliomas on magnetic resonance imaging (MRI) imaging, compelling precise differentiation for appropriate treatment. This study focuses on developing an automated MRI-based workflow to distinguish between PCNSL and gliomas. MRI examinations of 240 therapy-naive patients (141 males and 99 females, mean age: 55.16 years) with cerebral gliomas and PCNSLs (216 gliomas and 24 PCNSLs), each comprising a non-contrast T1-weighted, fluid-attenuated inversion recovery (FLAIR), and contrast-enhanced T1-weighted sequence were included in the study. HD-GLIO, a pre-trained segmentation network, was used to generate segmentations automatically. To validate the segmentation efficiency, 237 manual segmentations were prepared (213 gliomas and 24 PCNSLs). Subsequently, radiomics features were extracted following feature selection and training of an XGBoost algorithm for classification. The segmentation models for gliomas and PCNSLs achieved a mean Sørensen-Dice coefficient of 0.82 and 0.80 for whole tumors, respectively. Three classification models were developed in this study to differentiate gliomas from PCNSLs. The first model differentiated PCNSLs from gliomas, with an area under the curve (AUC) of 0.99 (F1-score: 0.75). The second model discriminated between high-grade gliomas and PCNSLs with an AUC of 0.91 (F1-score: 0.6), and the third model differentiated between low-grade gliomas and PCNSLs with an AUC of 0.95 (F1-score: 0.89). This study serves as a pilot investigation presenting an automated virtual biopsy workflow that distinguishes PCNSLs from cerebral gliomas. Prior to clinical use, it is necessary to validate the results in a prospective multicenter setting with a larger number of PCNSL patients.

Prediction of diabetes insipidus occurrence after endoscopic endonasal removal of sellar lesions using MRI-based radiomics and machine learning.

Pituitary adenomas and craniopharyngiomas are the most common lesions of the sellar region. These tumors are responsible for invasion or compression of crucial neurovascular structures. The involvement of the pituitary stalk warrants high rates of both pre- and post- operative diabetes insipidus. The aim of our study was to assess the accuracy of machine learning analysis from preoperative MRI of pituitary adenomas and craniopharyngiomas for the prediction of DI occurrence. All patients underwent MRI exams either on a 1.5- or 3-T MR scanner from two Institutions, including coronal T2-weighted (T2-w) and contrast-enhanced T1-weighted (CE T1-w) Turbo Spin Echo sequences. Feature selection was carried out as a multi-step process, with a threshold of 0.75 to identify robust features. Further feature selection steps included filtering based on feature variance (threshold >0.01) and pairwise correlation (threshold <0.80). A Bayesian Network model was trained with 10-fold cross validation employing SMOTE to balance classes exclusively within the training folds. Thirty patients were included in this study. In total 2394 features were extracted and 1791 (75%) resulted stable after ICC analysis. The number of variant features was 1351 and of non-colinear features was 125. Finally, 10 features were selected by oneR ranking. The Bayesian Network model showed an accuracy of 63% with a precision of 77% for DI prediction (0.68 area under the precision-recall curve). We assessed the accuracy of machine learning analysis of texture-derived parameters from preoperative MRI of pituitary adenomas and craniopharyngiomas for the prediction of DI occurrence.

Sudden Sensorineural Hearing Loss from a Jugular Bulb Diverticulum

A 19-year-old woman presented with an 11-month history of sudden-onset left sided hearing loss accompanied by vertigo and headache. Audiometric testing revealed profound left- sided hearing loss. A contrast-enhanced MRI of the internal auditory canal performed 5 months after symptom onset was interpreted as showing a vascular loop, probably the anterior inferior cerebellar artery, abutting and indenting on the left vestibulocochlear nerve; and a prominent and high-riding left jugular bulb. In this study, the internal auditory canals were assessed to be of normal width, with walls that were smooth and sharply defined. A cerebral CT angiogram subsequently performed did not show any abnormal findings related to the previously identified vascular loop. On the basis of these radiologic findings, the patient was advised surgery by physicians at a tertiary- care institution, presumably to address the identified vascular loop. A second opinion was sought by the patient.&#x0D; Review of the MRI initially focused on the axial high-resolution T2-weighted sequence (T2-DRIVE), as the fast spin-echo T2-weighted sequence has been recommended as a reliable and cost-effective MR screening protocol for the detection of masses in the IAC.1 In contrast to the official radiology report, stenosis of the left internal auditory canal by a protrusion (Figure 1, white asterisk) originating from the posteromedial wall of the internal auditory canal was noted. This protrusion, which had no MR signal intensity, appeared to abut and compress the cranial nerves within the IAC. Reconstruction of the images in non-orthogonal planes aligned with the orientation and direction of the left 8th cranial nerve showed the protrusion causing upward compression on and distortion of the nerve. (Figure 2, white arrows)&#x0D; Attention was directed to the axial high-resolution T1-weighted sequence (T1W-3D FFE), which revealed that the protrusion contained an isointense soft tissue structure (Figure 3, white asterisk) located within the petrous bone medial to the posterior semicircular canal. This structure appeared to be an upward extension of the jugular bulb.&#x0D; The axial high-resolution contrast-enhanced T1-weighted sequence (T1W-3D TFE Gd) showed smooth, vivid enhancement of the identified structure (Figure 4, black asterisk), which connected with the sigmoid sinus in lower cuts. This confirmed the presence of a high-riding jugular bulb that encroached on the internal auditory canal.&#x0D; Any doubt as to its true nature was dispelled by a review of the temporal bone structures on high-resolution CT which was fortunately available in the cerebral CT angiogram. This revealed a protrusion of the high-riding jugular bulb with a waist-like margin (Figure 5A and B, black arrows), allowing further characterization of the lesion as a jugular bulb diverticulum.2&#x0D; A high-riding jugular bulb that projects into the middle ear is not an uncommon anatomic variation. On the other hand, a jugular bulb diverticulum, which is an outpouching of the jugular bulb that can extend superiorly, medially, and posteriorly in the petrous bone, is a true venous anomaly that has been described rarely in the medical literature.3,4 When symptomatic, patients with this anomaly can present with sensorineural hearing loss, tinnitus, vertigo and auricular pain.3-5 Proper identification of a jugular bulb diverticulum in the evaluation of a patient with neurotologic symptoms is necessary to avoid inappropriate medical and surgical intervention.&#x0D; As demonstrated in this patient, a jugular bulb diverticulum may not be identified by a screening MRI that utilizes only a T2-weighted sequence. T1-weighted MRI sequences with and without contrast are necessary to demonstrate its soft tissue imaging characteristics. Although not the initial imaging study of choice for sudden sensorineural hearing loss, high-resolution bone-window CT may be necessary to delineate the bony anatomy of the jugular foramen and confirm the presence of this anomaly.

Glioblastoma and Solitary Brain Metastasis: Differentiation by Integrating Demographic-MRI and Deep-Learning Radiomics Signatures.

Studies have shown that deep-learning radiomics (DLR) could help differentiate glioblastoma (GBM) from solitary brain metastasis (SBM), but whether integrating demographic-MRI and DLR features can more accurately distinguish GBM from SBM remains uncertain. To construct and validate a demographic-MRI deep-learning radiomics nomogram (DDLRN) integrating demographic-MRI and DLR signatures to differentiate GBM from SBM preoperatively. Retrospective. Two hundred and thirty-five patients with GBM (N = 115) or SBM (N = 120), randomly divided into a training cohort (90 GBM and 98 SBM) and a validation cohort (25 GBM and 22 SBM). Axial T2-weighted fast spin-echo sequence (T2WI), T2-weighted fluid-attenuated inversion recovery sequence (T2-FLAIR), and contrast-enhanced T1-weighted spin-echo sequence (CE-T1WI) using 1.5-T and 3.0-T scanners. The demographic-MRI signature was constructed with seven imaging features ("pool sign," "irregular ring sign," "regular ring sign," "intratumoral vessel sign," the ratio of the area of peritumoral edema to the enhanced tumor, the ratio of the lesion area on T2-FLAIR to CE-T1WI, and the tumor location) and demographic factors (age and sex). Based on multiparametric MRI, radiomics and deep-learning (DL) models, DLR signature, and DDLRN were developed and validated. The Mann-Whitney U test, Pearson test, least absolute shrinkage and selection operator, and support vector machine algorithm were applied for feature selection and construction of radiomics and DL models. DDLRN showed the best performance in differentiating GBM from SBM with area under the curves (AUCs) of 0.999 and 0.947 in the training and validation cohorts, respectively. Additionally, the DLR signature (AUC = 0.938) outperformed the radiomics and DL models, and the demographic-MRI signature (AUC = 0.775) was comparable to the T2-FLAIR radiomics and DL models in the validation cohort (AUC = 0.762 and 0.749, respectively). DDLRN integrating demographic-MRI and DLR signatures showed excellent performance in differentiating GBM from SBM. 3 TECHNICAL EFFICACY: Stage 2.

A Machine Learning-Based Unenhanced Radiomics Approach to Distinguishing Between Benign and Malignant Breast Lesions Using T2-Weighted and Diffusion-Weighted MRI.

Breast MRI has been recommended as supplemental screening tool to mammography and breast ultrasound of breast cancer by international guidelines, but its long examination time and use of contrast material remains concerning. To develop an unenhanced radiomics model with using non-gadolinium based sequences for detecting breast cancer based on T2-weighted (T2W) and diffusion-weighted (DW) MRI. Retrospective analysis followed by retrospective and prospective cohorts study. 1760 patients: Of these, 1293 for model construction (n = 775 for training and 518 for validation). The remaining patients for model testing in internal retrospective (n = 167), internal prospective (n = 188), and external retrospective (n = 112) cohorts. 3.0T MR scanners from two institution. T2WI, DWI, and first contrast-enhanced T1-weighted sequence. AUCs in distinguishing breast cancer were compared between combined model with gadolinium agent sequence and unenhanced model. Subsequently, the AUCs in testing cohorts of unenhanced model was compared with two radiologists' diagnosis for this research. Finally, patient subgroup analysis in testing cohorts was performed based on clinical subgroups and different types of malignancies. Mann-Whitney U test, Kruskal-Wallis H test, chi-square test, weighted kappa test, and DeLong's test. The unenhanced radiomics model performed best under Gaussian process (GP) classifiers (AUC: training, 0.893; validation, 0.848) compared to support vector machine (SVM) and logistic, showing favorable prediction in testing cohorts (AUCs, 0.818-0.840). The AUCs for the unenhanced radiomics model were not statistically different in five cohorts from those of the combined radiomics model (P, 0.317-0.816), as well as the two radiologists (P, 0.181-0.918). The unenhanced radiomics model was least successful in identifying ductal carcinoma in situ, whereas did not show statistical significance in other subgroups. An unenhanced radiomics model based on T2WI and DWI has comparable diagnostic accuracy to the combined model using the gadolinium agent. 4 TECHNICAL EFFICACY: Stage 2.

Deep Learning-Based Synthesis of Contrast-Enhanced MRI for Automated Delineation of Primary Gross Tumor Volume in Radiotherapy of Nasopharyngeal Carcinoma.

Contrast-enhanced MRIs are necessary to delineate the primary gross tumor volume (GTVp) in radiotherapy of nasopharyngeal carcinoma (NPC). However, using contrast agents to scan contrast-enhanced MRIs is not applicable to some patients due to metal implants or their allergy, and it increases the treatment cost of patients. To address these problems, this work aims at synthesizing contrast-enhance MRIs from unenhanced MRIs by implementing generative adversarial network (GAN). In this work, 324 MRI datasets of patients with NPC were retrospectively collected between September 2016 and September 2017 from a single institute. MRI examinations were performed with un-enhanced T1-weighted (T1) and T2-weighted (T2) sequences, and contrast-enhanced T1-weighted (T1C) and fat-suppressed T1-weighted (T1FSC) sequences. We designed and developed a modified pix2pix network to synthesize T1C (sT1C) and T1FSC (sT1FSC) from real T1. The end of the generator in this network was assembled with multiple heads (the classification head and gradient head) to learn more representation information and features from real images, the discriminator in this network distinguished whether the synthesized image is real and fake and supervised that the generator outputs more realistic synthesized image. We verified the performance of the synthesized images for automated delineation of GTVp. In an independent testing set of 11 patients, the synthesized sT1C and sT1FSC were inputted into the segmentation deep learning network along with their corresponding T1 and T2 sequences to generate GTVp contours. Delineation performance of the synthesized images and real images for automated delineation were evaluated by dice similarity coefficient (DSC), and average surface distance (ASD), using human expert contours as the ground truth. In automated contouring of GTVp for NPC, the segmentation deep learning network using one or two synthesized MRIs showed equivalent performance when compared with the automated contours which generated from four real MRI sequences. Mean DSCs between automated contours by sT1C-replaced or sT1C and sT1FSC-replaced network and ground truth contours were 0.726 ± 0.143 and 0.711 ± 0.157, respectively, slightly inferior to that of contours generated from four real MRI sequences (0.740 ± 0.154, both P >0.05). In terms of mean ASD, there was also no significant difference between automated contours generated from synthesized images and real images (3.056 ± 4.216 mm and 3.537 ± 4.793 mm vs. 3.124 ± 4.637 mm; both P > 0.05). We proposed an MRI-synthesis method based on GAN and the synthesized contrast-enhanced MRIs performed equivalent as the real contrast-enhanced MRIs in the automated delineation of gross tumor volume for radiotherapy of NPC.

A Multicenter Study on Deep Learning for Glioblastoma Auto-Segmentation with Prior Knowledge in Multimodal Imaging

A precise radiotherapy plan is required to ensure accurate delineation of gross tumor volumes (GTV) and clinical target volumes (CTV1 and CTV2) of glioblastomas (GBMs). However, traditional manual delineation is labor intensive and highly dependent on oncologists' experience. To construct and evaluate a deep learning-based automatic delineation method using prior knowledge in multimodal medical imaging to automate precise GTV, CTV1 and CTV2 contouring in GBM patients. We retrospectively collected the CT and MRI scans of 55 eligible patients with histologically proven high-grade glioma (HGG) from an institute, these scans were performed with non-enhanced CT (CT), contrast-enhanced T1-weighted (T1C) and T2-FLAIR (T2F) sequences. We proposed a two-stage automatic segmentation framework (PKMI-Net) for GTV, CTV1 and CTV2 based on deep learning using prior knowledge in multimodal medical imaging, and its segmentation performance was evaluated with dice similarity coefficient (DSC), 95% Harsdorff distance (HD95), average surface distance (ASD) and relative volume difference (RVD). To further investigate the generalizability of our method, we designed and conducted two evaluation strategies (Mix and Cross) on four multicenter datasets (including 55 patients, 37 patients, 21 patients and 35 patients). The evaluation results with an 11-patient test set from the single institute were summarized in Table 1, the proposed method demonstrated the best accuracy in segmenting, respectively, GTV, CTV1 and CTV, achieving a DSC of 0.94, 0.95 and 0.92; HD95 of 2.07 mm, 1.18 mm and 3.80 mm; ASD of 0.69 mm, 0.39 mm and 1.13 mm and RVE of 5.50%, 3.97% and 9.68%. In the multicenter evaluation, the segmentation performance of our method implemented with the Cross strategy was comparable to that with the Mix strategy, demonstrating that our method had high and stable generalizability across multicenter datasets in automatically segmenting GTV, CTV1 and CTV2. Our proposed method achieved promising results in automatically segmenting gliomas across various datasets, which could improve the quality and efficiency of glioblastoma radiotherapy.

AI Evaluation of Imaging Factors in the Evolution of Stage-Treated Metastases Using Gamma Knife.

The study investigated whether three deep-learning models, namely, the CNN_model (trained from scratch), the TL_model (transfer learning), and the FT_model (fine-tuning), could predict the early response of brain metastases (BM) to radiosurgery using a minimal pre-processing of the MRI images. The dataset consisted of 19 BM patients who underwent stereotactic-radiosurgery (SRS) within 3 months. The images used included axial fluid-attenuated inversion recovery (FLAIR) sequences and high-resolution contrast-enhanced T1-weighted (CE T1w) sequences from the tumor center. The patients were classified as responders (complete or partial response) or non-responders (stable or progressive disease). A total of 2320 images from the regression class and 874 from the progression class were randomly assigned to training, testing, and validation groups. The DL models were trained using the training-group images and labels, and the validation dataset was used to select the best model for classifying the evaluation images as showing regression or progression. Among the 19 patients, 15 were classified as "responders" and 4 as "non-responders". The CNN_model achieved good performance for both classes, showing high precision, recall, and F1-scores. The overall accuracy was 0.98, with an AUC of 0.989. The TL_model performed well in identifying the "progression" class, but could benefit from improved precision, while the "regression" class exhibited high precision, but lower recall. The overall accuracy of the TL_model was 0.92, and the AUC was 0.936. The FT_model showed high recall for "progression", but low precision, and for the "regression" class, it exhibited a high precision, but lower recall. The overall accuracy for the FT_model was 0.83, with an AUC of 0.885. Among the three models analyzed, the CNN_model, trained from scratch, provided the most accurate predictions of SRS responses for unlearned BM images. This suggests that CNN models could potentially predict SRS prognoses from small datasets. However, further analysis is needed, especially in cases where class imbalances exist.