Surgical and endovascular cerebral revascularization for cerebral vasculitis with inflammatory vessel stenosis: a case series.

Autoimmune vasculitides affect the cerebral vasculature significantly in a considerable number of cases. When immunosuppressive treatments fail to prevent stenosis in cerebral vessels, treatment options for affected patients become limited. In this case series, we present four cases of pharmacoresistant vasculitis with recurrent transient ischemic attacks (TIAs) or stroke successfully treated with either extracranial-intracranial (EC-IC) bypass surgery or endovascular stenting. Both rescue treatments were effective and safe in the selected cases. Our experience suggests that cases of pharmacoresistant cerebral vasculitis with recurrent stroke may benefit from rescue revascularization in combination with maximum medical management.

Prediction of therapy response of breast cancer patients with machine learning based on clinical data and imaging data derived from breast [18F]FDG-PET/MRI.

PURPOSE: To evaluate if a machine learning prediction model based on clinical and easily assessable imaging features derived from baseline breast [18F]FDG-PET/MRI staging can predict pathologic complete response (pCR) in patients with newly diagnosed breast cancer prior to neoadjuvant system therapy (NAST). METHODS: Altogether 143 women with newly diagnosed breast cancer (54 ± 12 years) were retrospectively enrolled. All women underwent a breast [18F]FDG-PET/MRI, a histopathological workup of their breast cancer lesions and evaluation of clinical data. Fifty-six features derived from positron emission tomography (PET), magnetic resonance imaging (MRI), sociodemographic / anthropometric, histopathologic as well as clinical data were generated and used as input for an extreme Gradient Boosting model (XGBoost) to predict pCR. The model was evaluated in a five-fold nested-cross-validation incorporating independent hyper-parameter tuning within the inner loops to reduce the risk of overoptimistic estimations. Diagnostic model-performance was assessed by determining the area under the curve of the receiver operating characteristics curve (ROC-AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy. Furthermore, feature importances of the XGBoost model were evaluated to assess which features contributed most to distinguish between pCR and non-pCR. RESULTS: Nested-cross-validation yielded a mean ROC-AUC of 80.4 ± 6.0% for prediction of pCR. Mean sensitivity, specificity, PPV, and NPV of 54.5 ± 21.3%, 83.6 ± 4.2%, 63.6 ± 8.5%, and 77.6 ± 8.1% could be achieved. Histopathological data were the most important features for classification of the XGBoost model followed by PET, MRI, and sociodemographic/anthropometric features. CONCLUSION: The evaluated multi-source XGBoost model shows promising results for reliably predicting pathological complete response in breast cancer patients prior to NAST. However, yielded performance is yet insufficient to be implemented in the clinical decision-making process.

Clinical Decision Support for Axillary Lymph Node Staging in Newly Diagnosed Breast Cancer Patients Based on 18F-FDG PET/MRI and Machine Learning.

In addition to its high prognostic value, the involvement of axillary lymph nodes in breast cancer patients also plays an important role in therapy planning. Therefore, an imaging modality that can determine nodal status with high accuracy in patients with primary breast cancer is desirable. Our purpose was to investigate whether, in newly diagnosed breast cancer patients, machine-learning prediction models based on simple assessable imaging features on MRI or PET/MRI are able to determine nodal status with performance comparable to that of experienced radiologists; whether such models can be adjusted to achieve low rates of false-negatives such that invasive procedures might potentially be omitted; and whether a clinical framework for decision support based on simple imaging features can be derived from these models. Methods: Between August 2017 and September 2020, 303 participants from 3 centers prospectively underwent dedicated whole-body 18F-FDG PET/MRI. Imaging datasets were evaluated for axillary lymph node metastases based on morphologic and metabolic features. Predictive models were developed for MRI and PET/MRI separately using random forest classifiers on data from 2 centers and were tested on data from the third center. Results: The diagnostic accuracy for MRI features was 87.5% both for radiologists and for the machine-learning algorithm. For PET/MRI, the diagnostic accuracy was 89.3% for the radiologists and 91.2% for the machine-learning algorithm, with no significant differences in diagnostic performance between radiologists and the machine-learning algorithm for MRI (P = 0.671) or PET/MRI (P = 0.683). The most important lymph node feature was tracer uptake, followed by lymph node size. With an adjusted threshold, a sensitivity of 96.2% was achieved by the random forest classifier, whereas specificity, positive predictive value, negative predictive value, and accuracy were 68.2%, 78.1%, 93.8%, and 83.3%, respectively. A decision tree based on 3 simple imaging features could be established for MRI and PET/MRI. Conclusion: Applying a high-sensitivity threshold to the random forest results might potentially avoid invasive procedures such as sentinel lymph node biopsy in 68.2% of the patients.

Potential of a machine-learning model for dose optimization in CT quality assurance.

OBJECTIVES: To evaluate machine learning (ML) to detect chest CT examinations with dose optimization potential for quality assurance in a retrospective, cross-sectional study. METHODS: Three thousand one hundred ninety-nine CT chest examinations were used for training and testing of the feed-forward, single hidden layer neural network (January 2016-December 2017, 60% male, 62 ± 15 years, 80/20 split). The model was optimized and trained to predict the volumetric computed tomography dose index (CTDIvol) based on scan patient metrics (scanner, study description, protocol, patient age, sex, and water-equivalent diameter (DW)). The root mean-squared error (RMSE) was calculated as performance measurement. One hundred separate, consecutive chest CTs were used for validation (January 2018, 60% male, 63 ± 16 years), independently reviewed by two blinded radiologists with regard to dose optimization, and used to define an optimal cutoff for the model. RESULTS: RMSE was 1.71, 1.45, and 1.52 for the training, test, and validation dataset, respectively. The scanner and DW were the most important features. The radiologists found dose optimization potential in 7/100 of the validation cases. A percentage deviation of 18.3% between predicted and actual CTDIvol was found to be the optimal cutoff: 8/100 cases were flagged as suboptimal by the model (range 18.3-53.2%). All of the cases found by the radiologists were identified. One examination was flagged only by the model. CONCLUSIONS: ML can comprehensively detect CT examinations with dose optimization potential. It may be a helpful tool to simplify CT quality assurance. CT scanner and DW were most important. Final human review remains necessary. A threshold of 18.3% between the predicted and actual CTDIvol seems adequate for CT quality assurance. KEY POINTS: • Machine learning can be integrated into CT quality assurance to improve retrospective analysis of CT dose data. • Machine learning may help to comprehensively detect dose optimization potential in chest CT, but an individual review of the results by an experienced radiologist or radiation physicist is required to exclude false-positive findings.

Gadolinium Brain Deposition after Macrocyclic Gadolinium Administration: A Pediatric Case-Control Study.

Purpose To determine whether signal intensity (SI) in T1 sequences as a potential indicator of gadolinium deposition increases after repeated administration of the macrocyclic gadolinium-based contrast agents (GBCAs) gadoteridol and gadoterate meglumine in a pediatric cohort. Materials and Methods This retrospective case-control study of children with brain tumors who underwent nine or more contrast material-enhanced brain magnetic resonance (MR) imaging studies from 2008 to 2015 was approved by the local ethics board. Informed consent was obtained for MR imaging. Twenty-four case patients aged 5-18 years and appropriate control patients with nonpathologic MR neuroimaging findings (and no GBCA administration), matched for age and sex, were inculded. SI was measured on unenhanced T1-weighted MR images for the following five regions of interest (ROIs): the dentate nucleus (DN), pons, substantia nigra (SN), pulvinar thalami, and globus pallidus (GP). Paired t tests were used to compare SI and SI ratios (DN to pons, GP to thalamus) between case patients and control patients. Pearson correlations between relative signal changes and the number of GBCA administrations and total GBCA dose were calculated. Results The mean number of GBCA administrations was 14.2. No significant differences in mean SI for any ROI and no group differences were found when DN-to-pons and GP-to-pulvinar ratios were compared (DN-to-pons ratio in case patients: mean, 1.0083 ± 0.0373 [standard deviation]; DN-to-pons ratio in control patients: mean, 1.0183 ± 0.01917; P = .37; GP-to-pulvinar ratio in case patients: mean, 1.1335 ± 0.04528; and GP-to-pulvinar ratio in control patients: mean, 1.1141 ± 0.07058; P = .29). No correlation was found between the number of GBCA administrations or the total amount of GBCA administered and signal change for any ROI. (Number of GBCA applications: DN: r = -0.254, P = .31; pons: r = -0.097, P = .65; SN: r = -0.194, P = .38; GP: r = -0.175, P = .41; pulvinar: r = -0.067, P = .75; total amount of administered GBCA: DN: r = 0.091, P = .72; pons: r = 0.106, P = .62; SN: r = -0.165, P = .45; GP: r = 0.111, P = .61; pulvinar: r = 0.173, P = .42.) Conclusion Multiple intravenous administrations of these macrocyclic GBCAs in children were not associated with a measurable increase in SI in T1 sequences as an indicator of brain gadolinium deposition detectable by using MR imaging. Additional imaging and pathologic studies are needed to confirm these findings. © RSNA, 2017 Online supplemental material is available for this article.

Optimizing radiation exposure in screening of body packing: image quality and diagnostic acceptability of an 80 kVp protocol with automated tube current modulation.

The aim of this study was to evaluate the objective and subjective image quality of a novel computed tomography (CT) protocol with reduced radiation dose for body packing with 80 kVp and automated tube current modulation (ATCM) compared to a standard body packing CT protocol. 80 individuals who were examined between March 2012 and July 2015 in suspicion of ingested drug packets were retrospectively included in this study. Thirty-one CT examinations were performed using ATCM and a fixed tube voltage of 80 kVp (group A). Forty-nine CT examinations were performed using a standard protocol with a tube voltage of 120 kVp and a fixed tube current time product of 40 mAs (group B). Subjective and objective image quality and visibility of drug packets were assessed. Radiation exposure of both protocols was compared. Contrast-to-noise ratio (group A: 0.56 ± 0.36; group B: 1.13 ± 0.91) and Signal-to-noise ratio (group A: 3.69 ± 0.98; group B: 7.08 ± 2.67) were significantly lower for group A compared to group B (p < 0.001). Subjectively, image quality was decreased for group A compared to group B (2.5 ± 0.8 vs. 1.2 ± 0.4; p < 0.001). Attenuation of body packets was higher with the new protocol (group A: 362.2 ± 70.3 Hounsfield Units (HU); group B: 210.6 ± 60.2 HU; p = 0.005). Volumetric Computed Tomography Dose Index (CTDIvol) and Dose Length Product (DLP) were significantly lower in group A (CTDIvol 2.2 ± 0.9 mGy, DLP 105.7 ± 52.3 mGycm) as compared to group B (CTDIvol 2.7 ± 0.1 mGy, DLP 126.0 ± 9.7 mGycm, p = 0.002 and p = 0.01). The novel 80 kVp CT protocol with ATCM leads to a significant dose reduction compared to a standard CT body packing protocol. The novel protocol led to a diagnostic image quality and cocaine body packets were reliably detected due to the high attenuation.