Limited capability of MRI radiomics to predict primary tumor histology of brain metastases in external validation.

Background: Growing research demonstrates the ability to predict histology or genetic information of various malignancies using radiomic features extracted from imaging data. This study aimed to investigate MRI-based radiomics in predicting the primary tumor of brain metastases through internal and external validation, using oversampling techniques to address the class imbalance. Methods: This IRB-approved retrospective multicenter study included brain metastases from lung cancer, melanoma, breast cancer, colorectal cancer, and a combined heterogenous group of other primary entities (5-class classification). Local data were acquired between 2003 and 2021 from 231 patients (545 metastases). External validation was performed with 82 patients (280 metastases) and 258 patients (809 metastases) from the publicly available Stanford BrainMetShare and the University of California San Francisco Brain Metastases Stereotactic Radiosurgery datasets, respectively. Preprocessing included brain extraction, bias correction, coregistration, intensity normalization, and semi-manual binary tumor segmentation. Two-thousand five hundred and twenty-eight radiomic features were extracted from T1w (±â€…contrast), fluid-attenuated inversion recovery (FLAIR), and wavelet transforms for each sequence (8 decompositions). Random forest classifiers were trained with selected features on original and oversampled data (5-fold cross-validation) and evaluated on internal/external holdout test sets using accuracy, precision, recall, F1 score, and area under the receiver-operating characteristic curve (AUC). Results: Oversampling did not improve the overall unsatisfactory performance on the internal and external test sets. Incorrect data partitioning (oversampling before train/validation/test split) leads to a massive overestimation of model performance. Conclusions: Radiomics models' capability to predict histologic or genomic data from imaging should be critically assessed; external validation is essential.

Head-to-Head Comparison of Dual-Source and Split-Beam Filter Multi-Energy CT versus SPECT/CT for Assessing Lobar Lung Perfusion in Emphysema.

Purpose: To evaluate dual-source and split-beam filter multi-energy chest CT in assessing pulmonary perfusion on a lobar level in patients with lung emphysema, using perfusion SPECT as the reference standard. Materials and Methods: Patients with emphysema evaluated for lung volume reduction therapy between May 2016 and February 2021 were retrospectively included. All patients underwent SPECT and either dual-source or split-beam filter (SBF) multi-energy CT. To calculate the fractional lobar lung perfusion (FLLP), SPECT acquisitions were co-registered with chest CT scans (hereafter, SPECT/CT) and semi-manually segmented. For multi-energy CT scans, lung lobes were automatically segmented using a U-Net model. Segmentations were manually verified. The FLLP was derived from iodine maps computed from the multi-energy data. Statistical analysis included Pearson and intraclass correlation coefficients and Bland-Altman analysis. Results: Fifty-nine patients (30 male, 29 female; 31 underwent dual-source CT, 28 underwent SBF CT; mean age for all patients, 67 years ± 8 [SD]) were included. Both multi-energy methods significantly correlated with the SPECT/CT acquisitions for all individual lobes (P < .001). Pearson correlation concerning all lobes combined was significantly better for dual-source (r = 0.88) than for SBF multi-energy CT (r = 0.78; P = .006). On the level of single lobes, Pearson correlation coefficient differed for the right upper lobe only (dual-source CT, r = 0.88; SBF CT, r = 0.58; P = .008). Conclusion: Dual-source and SBF multi-energy CT accurately assessed lung perfusion on a lobar level in patients with emphysema compared with SPECT/CT. The overall correlation was higher for dual-source multi-energy CT.Keywords: Chronic Obstructive Pulmonary Disease, Comparative Studies, Computer Applications, CT Spectral Imaging, Image Postprocessing, Lung, Pulmonary Perfusion© RSNA, 2023.

CT findings in "Post-Covid": residua from acute pneumonia or "Post-Covid-ILD"?

The aim of this study was to evaluate if CT findings in patients with pulmonary Post Covid syndrome represent residua after acute pneumonia or if SARS-CoV 2 induces a true ILD. Consecutive patients with status post acute Covid-19 pneumonia and persisting pulmonary symptoms were enrolled. Inclusion criteria were availability of at least one chest CT performed in the acute phase and at least one chest CT performed at least 80 days after symptom onset. In both acute and chronic phase CTs 14 CT features as well as distribution and extent of opacifications were independently determined by two chest radiologists. Evolution of every single CT lesion over time was registered intraindividually for every patient. Moreover, lung abnormalities were automatically segmented using a pre-trained nnU-Net model and volume as well as density of parenchymal lesions were plotted over the entire course of disease including all available CTs. 29 patients (median age 59 years, IQR 8, 22 men) were enrolled. Follow-up period was 80-242 days (mean 134). 152/157 (97 %) lesions in the chronic phase CTs represented residua of lung pathology in the acute phase. Subjective and objective evaluation of serial CTs showed that CT abnormalities were stable in location and continuously decreasing in extent and density. The results of our study support the hypothesis that CT abnormalities in the chronic phase after Covid-19 pneumonia represent residua in terms of prolonged healing of acute infection. We did not find any evidence for a Post Covid ILD.

Diagnostic Value of Diffusion-Weighted Imaging with Background Body Signal Suppression (DWIBS) for the Pre-Therapeutic Loco-Regional Staging of Cervical Cancer: A Feasibility and Interobserver Reliability Study.

(1) Background: cervical cancer is one of the leading causes of cancer-related deaths and the fourth most common cancer among women worldwide. Magnetic resonance imaging (MRI) is the modality of choice for loco-regional staging of cervical cancer in the primary diagnostic workup beginning with at least stage IB. (2) Methods: we retrospectively analyzed 16 patients with histopathological proven cervical cancer (FIGO IB1−IVA) for the diagnostic accuracy of standard MRI and standard MRI with diffusion-weighted imaging with background body signal suppression (DWIBS) for the correct pre-therapeutic assessment of the definite FIGO category. (3) Results: In 7 out of 32 readings (22%), DWIBS improved diagnostic accuracy. With DWIBS, four (13%) additional readings were assigned the correct major (I−IV) FIGO stages pre-therapeutically. Interobserver reliability of DWIBS was weakest for parametrial infiltration (k = 0.43; CI-95% 0.00−1.00) and perfect for tumor size <2 cm, infiltration of the vaginal lower third, infiltration of adjacent organs and loco-regional nodal metastases (k = 1.000; CI-95% 1.00−1.00). (4) Conclusions: the pre-therapeutic staging of cervical cancer has a high diagnostic accuracy and interobserver reliability when using standard MRI but can be further optimized with the addition of DWIBS sequences when reporting is performed by an experienced radiologist.

Clinical Value of Diffusion-Weighted Whole-Body Imaging with Background Body Signal Suppression (DWIBS) for Staging of Patients with Suspected Head and Neck Cancer.

(1) Background: To determine the importance of diffusion-weighted whole-body MRI with background body signal suppression (DWIBS) in the staging process of patients with suspected head and neck carcinomas. (2) Methods: A total of 30 patients (24 male, 6 female) with a median age of 67 years with clinically suspected head and neck carcinoma with pathologic cervical nodal swelling in ultrasound underwent the staging procedure with computed tomography (CT) and whole-body MRI including DWIBS. (3) Results: In a total of 9 patients, abnormalities in the routine work-up of pretherapeutic staging were found. Five cases of either secondary cancer or distant metastases were only visible in DWIBS, while being missed on CT. One diagnosis was only detectable in CT and not in DWIBS, whereas three diagnoses were recognizable in both modalities. (4) Conclusions: DWIBS in addition to a standard neck MRI in cervical lymphadenopathy suspicious for head and neck cancer yielded additional clinically relevant diagnoses in 17% of cases that would have been missed by current staging routine procedures. DWIBS offered a negative predictive value of 98.78% for ruling out distant metastases or secondary malignancies.

Application of A U-Net for Map-like Segmentation and Classification of Discontinuous Fibrosis Distribution in Gd-EOB-DTPA-Enhanced Liver MRI.

We aimed to evaluate whether U-shaped convolutional neuronal networks can be used to segment liver parenchyma and indicate the degree of liver fibrosis/cirrhosis at the voxel level using contrast-enhanced magnetic resonance imaging. This retrospective study included 112 examinations with histologically determined liver fibrosis/cirrhosis grade (Ishak score) as the ground truth. The T1-weighted volume-interpolated breath-hold examination sequences of native, arterial, late arterial, portal venous, and hepatobiliary phases were semi-automatically segmented and co-registered. The segmentations were assigned the corresponding Ishak score. In a nested cross-validation procedure, five models of a convolutional neural network with U-Net architecture (nnU-Net) were trained, with the dataset being divided into stratified training/validation (n = 89/90) and holdout test datasets (n = 23/22). The trained models precisely segmented the test data (mean dice similarity coefficient = 0.938) and assigned separate fibrosis scores to each voxel, allowing localization-dependent determination of the degree of fibrosis. The per voxel results were evaluated by the histologically determined fibrosis score. The micro-average area under the receiver operating characteristic curve of this seven-class classification problem (Ishak score 0 to 6) was 0.752 for the test data. The top-three-accuracy-score was 0.750. We conclude that determining fibrosis grade or cirrhosis based on multiphase Gd-EOB-DTPA-enhanced liver MRI seems feasible using a 2D U-Net. Prospective studies with localized biopsies are needed to evaluate the reliability of this model in a clinical setting.

Structural Connectivity Patterns of Side Effects Induced by Subthalamic Deep Brain Stimulation for Parkinson's Disease.

Background: Tractography based on diffusion-weighted magnetic resonance imaging (DWI) models the structural connectivity of the human brain. Deep brain stimulation (DBS) targeting the subthalamic nucleus is an effective treatment for advanced Parkinson's disease, but may induce adverse effects. This study investigated the relationship between structural connectivity patterns of DBS electrodes and stimulation-induced side effects. Materials and Methods: Twenty-one patients with Parkinson's disease treated with bilateral subthalamic DBS were examined. Overall, 168 electrode contacts were categorized as inducing or noninducing depending on their capability for inducing side effects such as motor effects, paresthesia, dysarthria, oculomotor effects, hyperkinesia, and other complications as assessed during the initial programming session. Furthermore, the connectivity of each contact with target regions was evaluated by probabilistic tractography based on DWI. Finally, stimulation sites and structural connectivity patterns of inducing and noninducing contacts were compared. Results: Inducing contacts differed across the various side effects and from those mitigating Parkinson's symptoms. Although contacts showed a largely overlapping spatial distribution within the subthalamic region, they could be distinguished by their connectivity patterns. In particular, inducing contacts were more likely connected with supplementary motor areas (hyperkinesia, dysarthria), frontal cortex (oculomotor), fibers of the internal capsule (paresthesia), and the basal ganglia-thalamo-cortical circuitry (dysarthria). Discussion: Side effects induced by DBS seem to be associated with distinct connectivity patterns. Cerebellar connections are hardly associated with side effects, although they seem relevant for mitigating motor symptoms in Parkinson's disease. A symptom-specific, connectivity-based approach for target planning in DBS may enhance treatment outcomes and reduce adverse effects. Impact statement Tractography based on diffusion-weighted magnetic resonance imaging has become a prominent technique for investigating the connectivity of human brain networks in vivo. However, the relationship between structural connections and brain function is still hardly known. The present study examined the relationship between adverse behavioral effects induced by deep brain stimulation (DBS) and tractography patterns in individual brains. The results suggest that DBS-based side effects depend on the structural connections of electrode contacts rather than their location. Network-based target planning in DBS may improve treatment by avoiding side effects. Moreover, the adopted approach may serve as a paragon for investigating structure/function relationships.

Deep brain stimulation: Connectivity profile for bradykinesia alleviation.

OBJECTIVE: Subthalamic deep brain stimulation may alleviate bradykinesia in Parkinson patients. Research suggests that this stimulation effect may be mediated by brain networks like the corticocerebellar loop. This study investigated the connectivity between stimulation sites and cortical and subcortical structures to identify connections for effective stimulation. METHODS: We retrospectively investigated 21 patients with Parkinson disease with bilateral subthalamic deep brain stimulation. Stimulation effectiveness in reducing bradykinesia, tremor, and rigidity was evaluated for each electrode contact in brain hemispheres contralateral to the affected hemibody. Dysarthric side effects were also examined. Probabilistic tractography based on diffusion-weighted imaging was performed in individual patient-specific brains using electrode contacts as seeds. Connectivity profiles of contacts with effective and noneffective stimulation were compared. RESULTS: Connectivity profiles of effective and noneffective contacts differed. Moreover, the connectivity profile for bradykinesia differed from that for rigidity, tremor, or dysarthria. Regarding bradykinesia, effective contacts were significantly more often connected with the ipsilateral superior cerebellar peduncle and the ipsilateral dentate nucleus, which correspond to the ipsilateral portion of the cerebellothalamocortical pathway. Rigidity was mitigated by stimulation of ascending brainstem and intralaminar thalamic connections. Tremor alleviation was related to connections with the internal capsule (anterior limb) and the pallidum. Dysarthric side effects were associated with connections to the supplementary motor area and the decussating cerebellothalamocortical pathway. INTERPRETATION: Whereas bradykinesia seems to be mitigated by stimulation of the ascending, ipsilateral cerebellothalamocortical pathway, stimulation of the descending corticopontocerebellar pathway may be ineffective. Rigidity, tremor, and dysarthric side effects seem to be influenced by different neural networks. ANN NEUROL 2019;85:852-864.