FlapMap Visual Language System for Vascular Imaging Prior to Microvascular Free Tissue Transfer.

Preoperative vascular imaging has been shown to be beneficial before free tissue transfer procedures, especially for deep inferior epigastric perforator flap breast reconstruction. Although computerized tomography angiography and magnetic resonance angiogram are increasingly frequently performed, there is no standardized method for recording, analyzing, and communicating the vast amount of clinically relevant information that is obtained from these tomographic imaging studies. Herein, the authors propose a new visual language system for preoperative imaging called "FlapMap," which allows for the creation of a clinically actionable, easily understood, and easily communicated single image that aids in preoperative planning before microvascular free tissue transfer.

Development of a machine learning model using electrocardiogram signals to improve acute pulmonary embolism screening.

Aims: Clinical scoring systems for pulmonary embolism (PE) screening have low specificity and contribute to computed tomography pulmonary angiogram (CTPA) overuse. We assessed whether deep learning models using an existing and routinely collected data modality, electrocardiogram (ECG) waveforms, can increase specificity for PE detection. Methods and results: We create a retrospective cohort of 21 183 patients at moderate- to high suspicion of PE and associate 23 793 CTPAs (10.0% PE-positive) with 320 746 ECGs and encounter-level clinical data (demographics, comorbidities, vital signs, and labs). We develop three machine learning models to predict PE likelihood: an ECG model using only ECG waveform data, an EHR model using tabular clinical data, and a Fusion model integrating clinical data and an embedded representation of the ECG waveform. We find that a Fusion model [area under the receiver-operating characteristic curve (AUROC) 0.81 ± 0.01] outperforms both the ECG model (AUROC 0.59 ± 0.01) and EHR model (AUROC 0.65 ± 0.01). On a sample of 100 patients from the test set, the Fusion model also achieves greater specificity (0.18) and performance (AUROC 0.84 ± 0.01) than four commonly evaluated clinical scores: Wells' Criteria, Revised Geneva Score, Pulmonary Embolism Rule-Out Criteria, and 4-Level Pulmonary Embolism Clinical Probability Score (AUROC 0.50-0.58, specificity 0.00-0.05). The model is superior to these scores on feature sensitivity analyses (AUROC 0.66-0.84) and achieves comparable performance across sex (AUROC 0.81) and racial/ethnic (AUROC 0.77-0.84) subgroups. Conclusion: Synergistic deep learning of ECG waveforms with traditional clinical variables can increase the specificity of PE detection in patients at least at moderate suspicion for PE.

Lymphoepithelioma-like neoplasm of the biliary tract with 'probable low malignant potential'.

AIMS: Lymphoepithelioma-like carcinomas (LELCs) are uncommon epithelial cancers characteristically showing two distinct components consisting of malignant epithelial cells and prominent dense lymphoid infiltrate. Hepatic LELCs consist of two types, the lymphoepithelioma-like hepatocellular carcinoma and lymphoepithelioma-like cholangiocarcinoma (LEL-CCA), with the latter being strongly associated with Epstein-Barr virus (EBV). METHODS AND RESULTS: We present a series of three cases of intrahepatic biliary EBV-associated LEL tumours in which the biliary epithelial component showed a distinctly benign appearance, instead of the usual malignant epithelial features of a typical CCA or EBV-associated LEL-CCA. In the lesions, the biliary epithelium showed interconnecting glands or cords of cells. All had a very low proliferation (Ki-67) index. Immunohistochemistry for IDH1 and TP53 performed on two cases was negative and molecular tests for EGFR and KRAS gene mutations performed on one were negative. Prognosis was very good in all three cases, with patients alive with no evidence of disease 24-62 months after surgery. Intriguingly, all three cases had co-infection of HBV and EBV. These cases are also discussed in the context of the 63 cases of LEL-CCA available in the literature, with a focus on epidemiology, clinicopathological features and potential research interests. CONCLUSIONS: Based on the distinct clinicopathological features and unique survival benefits, we believe these tumours represent the benign end of the spectrum of EBV-associated lymphoepithelial biliary carcinomas. Whether these tumours require a revision of the current nomenclature to 'lymphoepithelioma-like neoplasm of the biliary tract with probable low malignant potential' will require more detailed analysis with larger case-series.

Preoperative Evaluation of a Pancreas Mass: Diagnostic Options.

As modern abdominal imaging equipment advances, pancreatic lesion detection improves. Most of these lesions are incidental, and present a conundrum to the clinician and create great anxiety to the patient until a final diagnosis is made. For the practicing physician, the plethora of diagnostic options is overwhelming. The relevant question is, what is the most efficient algorithm to follow and to arrive at a timely and accurate diagnosis. This article presents a logical approach to the initial evaluation of a pancreatic lesion to get the most information possible with the least amount of testing, and to avoid duplicative measures.

Machine Learning Interface for Medical Image Analysis.

TensorFlow is a second-generation open-source machine learning software library with a built-in framework for implementing neural networks in wide variety of perceptual tasks. Although TensorFlow usage is well established with computer vision datasets, the TensorFlow interface with DICOM formats for medical imaging remains to be established. Our goal is to extend the TensorFlow API to accept raw DICOM images as input; 1513 DaTscan DICOM images were obtained from the Parkinson's Progression Markers Initiative (PPMI) database. DICOM pixel intensities were extracted and shaped into tensors, or n-dimensional arrays, to populate the training, validation, and test input datasets for machine learning. A simple neural network was constructed in TensorFlow to classify images into normal or Parkinson's disease groups. Training was executed over 1000 iterations for each cross-validation set. The gradient descent optimization and Adagrad optimization algorithms were used to minimize cross-entropy between the predicted and ground-truth labels. Cross-validation was performed ten times to produce a mean accuracy of 0.938 ± 0.047 (95 % CI 0.908-0.967). The mean sensitivity was 0.974 ± 0.043 (95 % CI 0.947-1.00) and mean specificity was 0.822 ± 0.207 (95 % CI 0.694-0.950). We extended the TensorFlow API to enable DICOM compatibility in the context of DaTscan image analysis. We implemented a neural network classifier that produces diagnostic accuracies on par with excellent results from previous machine learning models. These results indicate the potential role of TensorFlow as a useful adjunct diagnostic tool in the clinical setting.

Non-invasive prediction of portal pressures using CT and MRI in chronic liver disease.

PURPOSE: To assess the diagnostic value of a fast scoring system based on non-invasive cross-sectional imaging to predict portal hypertension (PH) in patients with liver disease. METHODS: In this retrospective study, we included patients who underwent contrast-enhanced CT or MRI within 3 months of hepatic venous pressure gradient (HVPG) measurements. Two independent observers provided an imaging-based scoring system (max of 9): number of variceal sites, volume of ascites, and spleen size. ROC analysis was performed to predict the presence of PH (HVPG ≥ 5 mmHg) and clinically significant PH (HVPG ≥ 10 mmHg). RESULTS: Our cohort consists of 143 patients with mean HVPG of 13.1 ± 2.0 mmHg. Mean PH scores from the two observers were 3.9 ± 2.7 and 3.2 ± 2.5. There was a significant correlation between PH score and HVPG (r = 0.58, p < 0.001 for both observers) with high inter-observer agreement (kappa 0.71). AUCs of 0.78-0.76 and 0.83-0.81 were observed for diagnosing HVPG ≥ 5 mmHg and HVPG ≥ 10 mmHg, respectively, for observers 1 and 2. CONCLUSIONS: We have developed a fast PH imaging-based composite score, which could be used for non-invasive detection of clinically significant PH.

Modern Perforator Flap Imaging with High-Resolution Blood Pool MR Angiography.

Advances in microsurgical techniques have improved autologous reconstructions by providing new donor site options while decreasing donor site morbidity. Various preoperative imaging modalities have been studied to assess the relevant vascular anatomic structures, with magnetic resonance (MR) angiography traditionally lagging behind computed tomography (CT) with respect to spatial resolution. Blood pool MR angiography with gadofosveset trisodium, a gadolinium-based contrast agent with extended intravascular retention, has allowed longer multiplanar acquisitions with resultant voxel sizes similar to or smaller than those of CT and with improved signal-to-noise ratio and soft-tissue contrast while maintaining the ability to depict flow with time-resolved imaging. The resultant vascular detail enables precise evaluation of the relevant vascular anatomic structures, including the vessel course, size, and branching pattern, as well as the venous arborization pattern. In addition, any architectural distortion, vessel alteration, or injury from prior surgery can be depicted. The reporting radiologist should be aware of pertinent and incidental findings relevant to the planned surgery and the patient's disease so that he or she can assist the microsurgeon in flap design as a member of the multidisciplinary team. Given the lack of ionizing radiation exposure in patients who often have an elevated body mass index, high-spatial-resolution blood pool MR angiography has become the imaging reference standard for the preoperative assessment of perforator flap vascular and soft-tissue morphology in our practice.

Comparison of comfort and image quality with two endorectal coils in MRI of the prostate.

PURPOSE: To prospectively evaluate comfort and image quality of prostate MRI using two different endorectal (ER) coils. MATERIALS AND METHODS: Thirty consecutive patients were prospectively randomized to receive prostate MRI using either a prostate endocoil (PEC) or colorectal endocoil (CEC). Patients and operators were surveyed with regard to endocoil placement. Four Body MRI trained radiologists rated image quality for each examination and additional selected blinded coronal T2 weighted images. RESULTS: Average patient discomfort (on a 0-10 pain scale) was greater with the PEC (5.0 for PEC and 2.7 for CEC) with a statistically significant difference (P < 0.03). Ability to identify the neurovascular bundle (NVB) was 5.8 times more likely to be rated excellent with the CEC compared with the PEC (P < 0.003). Image quality with CEC was 3.5 times more highly rated (P < 0.04). In particular, signal-to-noise ratio (SNR) with the CEC was 3.0 times more highly rated than with the PEC (P < 0.05). CONCLUSION: The smaller CEC was better tolerated by patients than the traditional PEC, and resulted in at least equivalent, and in some instances improved image quality. This may result in fewer aborted cases and lead to decreased procedural intolerance to endorectal coil MRI.

Aggressive angiomyxoma of the scrotum.

INTRODUCTION: Aggressive angiomyxoma (AAM) is a rare, benign mass with propensity for local invasion and recurrence after resection. Infrequently, this tumor can be found arising from the scrotum or cord structures in males. AIM/METHODS: A case report is presented followed by a review of relevant literature addressing the diagnosis, imaging, management and follow-up for aggressive angiomyxoma of the scrotum. RESULTS: Imaging can assist in further characterization of masses noted on physical exam. Scrotal sonography is typically the primary imaging modality utilized and magnetic resonance imaging is able to provide further anatomic detail. Treatment mainstay is surgical resection with necessary long term surveillance.