Text Report Analysis to Identify Opportunities for Optimizing Target Selection for Chest Radiograph Artificial Intelligence Models.

Our goal was to analyze radiology report text for chest radiographs (CXRs) to identify imaging findings that have the most impact on report length and complexity. Identifying these imaging findings can highlight opportunities for designing CXR AI systems which increase radiologist efficiency. We retrospectively analyzed text from 210,025 MIMIC-CXR reports and 168,949 reports from our local institution collected from 2019 to 2022. Fifty-nine categories of imaging finding keywords were extracted from reports using natural language processing (NLP), and their impact on report length was assessed using linear regression with and without LASSO regularization. Regression was also used to assess the impact of additional factors contributing to report length, such as the signing radiologist and use of terms of perception. For modeling CXR report word counts with regression, mean coefficient of determination, R2, was 0.469 ± 0.001 for local reports and 0.354 ± 0.002 for MIMIC-CXR when considering only imaging finding keyword features. Mean R2 was significantly less at 0.067 ± 0.001 for local reports and 0.086 ± 0.002 for MIMIC-CXR, when only considering use of terms of perception. For a combined model for the local report data accounting for the signing radiologist, imaging finding keywords, and terms of perception, the mean R2 was 0.570 ± 0.002. With LASSO, highest value coefficients pertained to endotracheal tubes and pleural drains for local data and masses, nodules, and cavitary and cystic lesions for MIMIC-CXR. Natural language processing and regression analysis of radiology report textual data can highlight imaging targets for AI models which offer opportunities to bolster radiologist efficiency.

Reducing annotation burden in MR: A novel MR-contrast guided contrastive learning approach for image segmentation.

BACKGROUND: Contrastive learning, a successful form of representational learning, has shown promising results in pretraining deep learning (DL) models for downstream tasks. When working with limited annotation data, as in medical image segmentation tasks, learning domain-specific local representations can further improve the performance of DL models. PURPOSE: In this work, we extend the contrastive learning framework to utilize domain-specific contrast information from unlabeled Magnetic Resonance (MR) images to improve the performance of downstream MR image segmentation tasks in the presence of limited labeled data. METHODS: The contrast in MR images is controlled by underlying tissue properties (e.g., T1 or T2) and image acquisition parameters. We hypothesize that learning to discriminate local representations based on underlying tissue properties should improve subsequent segmentation tasks on MR images. We propose a novel constrained contrastive learning (CCL) strategy that uses tissue-specific information via a constraint map to define positive and negative local neighborhoods for contrastive learning, embedding this information in the representational space during pretraining. For a given MR contrast image, the proposed strategy uses local signal characteristics (constraint map) across a set of related multi-contrast MR images as a surrogate for underlying tissue information. We demonstrate the utility of the approach for downstream: (1) multi-organ segmentation tasks in T2-weighted images where a DL model learns T2 information with constraint maps from a set of 2D multi-echo T2-weighted images (n = 101) and (2) tumor segmentation tasks in multi-parametric images from the public brain tumor segmentation (BraTS) (n = 80) dataset where DL models learn T1 and T2 information from multi-parametric BraTS images. Performance is evaluated on downstream multi-label segmentation tasks with limited data in (1) T2-weighted images of the abdomen from an in-house Radial-T2 (Train/Test = 30/20), (2) public Cartesian-T2 (Train/Test = 6/12) dataset, and (3) multi-parametric MR images from the public brain tumor segmentation dataset (BraTS) (Train/Test = 40/50). The performance of the proposed CCL strategy is compared to state-of-the-art self-supervised contrastive learning techniques. In each task, a model is also trained using all available labeled data for supervised baseline performance. RESULTS: The proposed CCL strategy consistently yielded improved Dice scores, Precision, and Recall metrics, and reduced HD95 values across all segmentation tasks. We also observed performance comparable to the baseline with reduced annotation effort. The t-SNE visualization of features for T2-weighted images demonstrates its ability to embed T2 information in the representational space. On the BraTS dataset, we also observed that using an appropriate multi-contrast space to learn T1+T2, T1, or T2 information during pretraining further improved the performance of tumor segmentation tasks. CONCLUSIONS: Learning to embed tissue-specific information that controls MR image contrast with the proposed constrained contrastive learning improved the performance of DL models on subsequent segmentation tasks compared to conventional self-supervised contrastive learning techniques. The use of such domain-specific local representations could help understand, improve performance, and mitigate the scarcity of labeled data in MR image segmentation tasks.

MRI Imaging Appearance of Hyperostosis Frontalis Interna (HFI): A Case Report of Focal Benign Enhancement.

Hyperostosis frontalis interna (HFI) is a common and often incidental finding seen on imaging. There is a significant paucity of radiology literature, particularly regarding the MRI imaging appearance of HFI. We reported two cases of HFI on MRI, which showed focal enhancement. These were stable on long-term follow-up studies and thought to be most consistent with benign enhancement. Further studies are needed to elucidate the underlying pathogenesis; however, it is important to be aware that regions of HFI may demonstrate variable enhancement and are sometimes mistaken for osseous metastatic disease.

Thromboembolic Disorder in COVID-19 Infection.

Coronavirus (COVID-19) is a global pandemic with over 600 million cases identified. In addition to extensive pulmonary complications of COVID-19, one feature unique to many patients with severe COVID-19 infections is coagulopathy with a rising prevalence of multi-systemic thromboembolic manifestations. Global data suggests a relationship between coagulopathy and mortality. In this review, we highlight multiple COVID-19 thromboembolic complications with emphasis on pathophysiology, clinical management, and radiological manifestations.

Spectrum of injuries with "less-lethal" beanbag weapons: pictorial essay.

Beanbag weapons are gaining popularity with increasing daily use as a non-lethal or less-lethal alternative to traditional firearms. While these are considered "less-lethal," these are associated with a spectrum of serious injuries. We present a pictorial essay of these injuries ranging from mild skin contusions to more severe solid organ injuries.

Performance of overnight on-call radiology residents in interpreting unenhanced abdominopelvic magnetic resonance imaging studies performed for pediatric right lower quadrant abdominal pain.

BACKGROUND: Abdominopelvic magnetic resonance imaging (MRI) is increasingly being used to evaluate children with abdominal pain suspected of having acute appendicitis. At our institution, these examinations are preliminarily interpreted by radiology residents, especially when performed after hours. OBJECTIVE: To determine the accuracy of preliminary reports rendered by radiology residents in this setting. MATERIALS AND METHODS: Three hundred seventy-seven pediatric abdominopelvic MRI examinations were included. The preliminary (resident) and final (attending) radiology reports were coded as diagnosing acute appendicitis or no acute appendicitis. The concordance between resident and attending radiologist interpretations was calculated. Additionally, both resident and attending reports were compared to available surgical pathology or clinical follow-up data. RESULTS: Overall concordance rate for the diagnosis of acute appendicitis was 97.1%. Concordance for verified cases of acute appendicitis was 93.4%. Concordance rates did not differ by residents' postgraduate year levels. When compared against surgical pathology or clinical follow-up data, residents demonstrated 91.2% sensitivity and 97.6% specificity. There was no statistically significant difference in the sensitivity or specificity of resident or attending radiologist interpretations. CONCLUSION: Radiology residents demonstrate high concordance with attending pediatric radiologists in their interpretations of pediatric abdominopelvic MRI for acute appendicitis. The diagnostic performances of residents and attendings were comparable.

Role of magnetic resonance imaging in Bouveret's syndrome: A case report with review of the literature.

Bouveret's syndrome is a rare form of gallstone ileus occurring due to obstructing gallstone into the proximal duodenum through a cholecystoduodenal fistula. We report the case of a 72-year-old female presenting with abdominal pain secondary to a large gallstone in the region of the duodenal bulb, causing the upstream gastric obstruction. Here we discuss the clinical features, imaging technologies, and surgical management of Bouveret's syndrome.

Pearls and Pitfalls of Quantitative Software Analysis of Dopamine Transporter SPECT.

Dopamine transporter SPECT with I-N-ω-fluoropropyl-2ß-carbomethoxy-3ß-(4-iodophenyl) nortropane (I-ioflupane) aids in the diagnosis of parkinsonian syndrome by demonstrating loss of striatal dopamine transporters, proportional to nigrostriatal dopaminergic neuronal loss. Quantitative software analysis (QSA) is a helpful adjunct to visual interpretation. An atlas of pearls and pitfalls of QSA is presented. Examples include correction for head tilt/orientation, scaling artifacts, and detection of balanced loss of activity. Additional examples are provided where QSA can potentially fail such as patient variation and vascular disease.

First-Line Diagnostic Evaluation with MRI of Children Suspected of Having Acute Appendicitis.

Background Advances in abdominal MRI have enabled rapid, free-breathing imaging without the need for intravenous or oral contrast material. The use of MRI as the primary imaging modality for suspected appendicitis has not been previously studied. Purpose To determine the diagnostic performance of MRI as the initial imaging modality in children suspected of having acute appendicitis. Materials and Methods The study included consecutive patients 18 years of age and younger presenting with acute abdominal pain at a tertiary care institution from January 2013 through June 2016 who subsequently underwent an unenhanced MRI examination as the primary diagnostic imaging modality. Electronic medical records and radiology reports were retrospectively evaluated for the feasibility and diagnostic performance of MRI, with surgical pathology and follow-up electronic records as reference standards. Statistical analyses were performed by using simple binomial proportions to quantify sensitivity, specificity, and accuracy, and exact 95% confidence intervals (CIs) were obtained. Results After exclusions, 402 patients (median age: 13 years; interquartile range [IQR], 9-15 years; 235 female patients; 167 male patients) were included. Sedation for MRI was required in 13 of 402 patients (3.2%; 95% CI: 1.7%, 5.5%). The appendix was visualized in 349 of 402 patients (86.8%; 95% CI: 83.1%, 90%); for the remaining patients, a diagnosis was provided on the basis of secondary signs of appendicitis. The sensitivity, specificity, and accuracy of MRI as the primary diagnostic imaging modality for the evaluation of acute appendicitis were 97.9% (95 of 97; 95% CI: 92.8%, 99.8%), 99% (302 of 305; 95% CI: 97.2%, 99.8%), and 98.8% (397 of 402; 97.1%, 99.6%), respectively. Among patients with negative findings for appendicitis at MRI, an alternate diagnosis was provided in 113 of 304 patients (37.2%; 95% CI: 31.7%, 42.9%). Conclusion When performed as the initial imaging modality in children suspected of having acute appendicitis, MRI examinations had high diagnostic performance for the diagnosis of acute appendicitis and in providing alternative diagnoses. © RSNA, 2019 See also the editorial by Dillman and Trout in this issue.

Diffuse periostitis as the primary presenting radiological finding in an AML patient with disease relapse.

Acute myelogenous leukemia is a hematologic malignancy defined by the presence of myeloid blasts causing bone marrow infiltration. Evaluation and workup of acute myelogenous leukemia is based on comprehensive medical history, physical examination, laboratory evaluation, and bone marrow sampling. Magnetic resonance (MR) imaging is the study of choice in the evaluation of this disease including the initial evaluation, treatment follow-up, and complications. Herein, we report a case of relapse of the acute myelogenic leukemia in an adult patient who presented with diffuse periostitis in his lower extremities diagnosed on MR imaging and confirmed on Technetium bone scan, which also showed periostitis along the bilateral humeri. To our knowledge, this was not previously reported in the English literature.

Multimodality management of vertebral artery injury sustained during cervical or craniocervical surgery.

BACKGROUND: Iatrogenic vertebral artery (VA) injury is a rare but potentially devastating complication associated with cervical and craniocervical surgery. OBJECTIVE: To retrospectively evaluate treatment modalities and outcomes associated with iatrogenic VA injury. METHODS: Our institutional surgical database was queried for patients who underwent cervical or craniocervical surgery from January 1997 to August 2012. RESULTS: During this time period, 8213 patients underwent cervical or craniocervical surgery, and 17 (0.2%) cases of VA injury were identified. Eight (47%) of these injuries occurred during C1-2 instrumentation procedures. Primary microsurgical repair of the VA was performed in 5 patients. Other cases were managed by either surgical or endovascular VA occlusion. Of the 17 patients, 15 underwent immediate angiography, 9 of whom were ultimately treated by the use of endovascular techniques. CONCLUSION: VA injury is an uncommon complication of cervical and/or skull base surgery. Standardized management recommendations may help reduce complications associated with these rare but potentially devastating injuries.