A neural network model for detection and classification of lumbar spinal stenosis on MRI.

OBJECTIVES: To develop a three-stage convolutional neural network (CNN) approach to segment anatomical structures, classify the presence of lumbar spinal stenosis (LSS) for all 3 stenosis types: central, lateral recess and foraminal and assess its severity on spine MRI and to demonstrate its efficacy as an accurate and consistent diagnostic tool. METHODS: The three-stage model was trained on 1635 annotated lumbar spine MRI studies consisting of T2-weighted sagittal and axial planes at each vertebral level. Accuracy of the model was evaluated on an external validation set of 150 MRI studies graded on a scale of absent, mild, moderate or severe by a panel of 7 radiologists. The reference standard for all types was determined by majority voting and in case of disagreement, adjudicated by an external radiologist. The radiologists' diagnoses were then compared to the diagnoses of the model. RESULTS: The model showed comparable performance to the radiologist average both in terms of the determination of presence/absence of LSS as well as severity classification, for all 3 stenosis types. In the case of central canal stenosis, the sensitivity, specificity and AUROC of the CNN were (0.971, 0.864, 0.963) for binary (presence/absence) classification compared to the radiologist average of (0.786, 0.899, 0.842). For lateral recess stenosis, the sensitivity, specificity and AUROC of the CNN were (0.853, 0.787, 0.907) compared to the radiologist average of (0.713, 0.898, 805). For foraminal stenosis, the sensitivity, specificity and AUROC of the CNN were (0.942, 0.844, 0.950) compared to the radiologist average of (0.879, 0.877, 0.878). Multi-class severity classifications showed similarly comparable statistics. CONCLUSIONS: The CNN showed comparable performance to radiologist subspecialists for the detection and classification of LSS. The integration of neural network models in the detection of LSS could bring higher accuracy, efficiency, consistency, and post-hoc interpretability in diagnostic practices.

Deep-learning reconstructed lumbar spine 3D MRI for surgical planning: pedicle screw placement and geometric measurements compared to CT.

PURPOSE: To test equivalency of deep-learning 3D lumbar spine MRI with "CT-like" contrast to CT for virtual pedicle screw planning and geometric measurements in robotic-navigated spinal surgery. METHODS: Between December 2021 and June 2022, 16 patients referred for spinal fusion and decompression surgery with pre-operative CT and 3D MRI were retrospectively assessed. Pedicle screws were virtually placed on lumbar (L1-L5) and sacral (S1) vertebrae by three spine surgeons, and metrics (lateral deviation, axial/sagittal angles) were collected. Vertebral body length/width (VL/VW) and pedicle height/width (PH/PW) were measured at L1-L5 by three radiologists. Analysis included equivalency testing using the 95% confidence interval (CI), a margin of ± 1 mm (± 2.08° for angles), and intra-class correlation coefficients (ICCs). RESULTS: Across all vertebral levels, both combined and separately, equivalency between CT and MRI was proven for all pedicle screw metrics and geometric measurements, except for VL at L1 (mean difference: - 0.64 mm; [95%CI - 1.05, - 0.24]), L2 (- 0.65 mm; [95%CI - 1.11, - 0.20]), and L4 (- 0.78 mm; [95%CI - 1.11, - 0.46]). Inter- and intra-rater ICC for screw metrics across all vertebral levels combined ranged from 0.68 to 0.91 and 0.89-0.98 for CT, and from 0.62 to 0.92 and 0.81-0.97 for MRI, respectively. Inter- and intra-rater ICC for geometric measurements ranged from 0.60 to 0.95 and 0.84-0.97 for CT, and 0.61-0.95 and 0.93-0.98 for MRI, respectively. CONCLUSION: Deep-learning 3D MRI facilitates equivalent virtual pedicle screw placements and geometric assessments for most lumbar vertebrae, with the exception of vertebral body length at L1, L2, and L4, compared to CT for pre-operative planning in patients considered for robotic-navigated spine surgery.

MRI after Lumbar Spine Decompression and Fusion Surgery: Technical Considerations, Expected Findings, and Complications.

Postoperative MRI of the lumbar spine is a mainstay for detailed anatomic assessment and evaluation of complications related to decompression and fusion surgery. Key factors for reliable interpretation include clinical presentation of the patient, operative approach, and time elapsed since surgery. Yet, recent spinal surgery techniques with varying anatomic corridors to approach the intervertebral disc space and implanted materials have expanded the range of normal (expected) and abnormal (unexpected) postoperative changes. Modifications of lumbar spine MRI protocols in the presence of metallic implants, including strategies for metal artifact reduction, provide important diagnostic information. This focused review discusses essential principles for the acquisition and interpretation of MRI after lumbar spinal decompression and fusion surgery, highlights expected postoperative changes, and describes early and delayed postoperative complications with examples.

MRI Advancements in Musculoskeletal Clinical and Research Practice.

Over the past decades, MRI has become increasingly important for diagnosing and longitudinally monitoring musculoskeletal disorders, with ongoing hardware and software improvements aiming to optimize image quality and speed. However, surging demand for musculoskeletal MRI and increased interest to provide more personalized care will necessitate a stronger emphasis on efficiency and specificity. Ongoing hardware developments include more powerful gradients, improvements in wide-bore magnet designs to maintain field homogeneity, and high-channel phased-array coils. There is also interest in low-field-strength magnets with inherently lower magnetic footprints and operational costs to accommodate global demand in middle- and low-income countries. Previous approaches to decrease acquisition times by means of conventional acceleration techniques (eg, parallel imaging or compressed sensing) are now largely overshadowed by deep learning reconstruction algorithms. It is expected that greater emphasis will be placed on improving synthetic MRI and MR fingerprinting approaches to shorten overall acquisition times while also addressing the demand of personalized care by simultaneously capturing microstructural information to provide greater detail of disease severity. Authors also anticipate increased research emphasis on metal artifact reduction techniques, bone imaging, and MR neurography to meet clinical needs.

Deep learning reconstruction for optimized bone assessment in zero echo time MR imaging of the knee.

PURPOSE: To evaluate the impact of deep learning-based reconstruction (DLRecon) on bone assessment in zero echo-time (ZTE) MRI of the knee at 1.5 Tesla. METHODS: This retrospective study included 48 consecutive exams of 46 patients (23 females) who underwent clinically indicated knee MRI at 1.5 Tesla. Standard imaging protocol comprised a sagittal prescribed, isotropic ZTE sequence. ZTE image reconstruction was performed with a standard-of-care (non-DL) and prototype DLRecon method. Exams were divided into subsets with and without osseous pathology based on the radiology report. Using a 4-point scale, two blinded readers qualitatively graded features of bone depiction including artifacts and conspicuity of pathology including diagnostic certainty in the respective subsets. Quantitatively, one reader measured signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of bone. Comparative analyses were conducted to assess the differences between the reconstruction methods. In addition, interreader agreement was calculated for the qualitative gradings. RESULTS: DLRecon significantly improved gradings for bone depiction relative to non-DL reconstruction (all, p < 0.05), while there was no significant difference with regards to artifacts (both, median score of 0; p = 0.058). In the subset with pathologies, conspicuity of pathology and diagnostic confidence were also scored significantly higher in DLRecon compared to non-DL (median 3 vs 2; p ≤ 0.03). Interreader agreement ranged from moderate to almost-perfect (κ = 0.54-0.88). Quantitatively, DLRecon demonstrated significantly enhanced CNR and SNR of bone compared to non-DL (p < 0.001). CONCLUSION: ZTE MRI with DLRecon improved bone depiction in the knee, compared to non-DL. Additionally, DLRecon increased conspicuity of osseous findings together with diagnostic certainty.

Imaging of Discogenic and Vertebrogenic Pain.

Chronic low back pain is a major source of pain and disability globally involving multifactorial causes. Historically, intervertebral disc degeneration and disruption have been associated as primary back pain triggers of the anterior column, termed "discogenic pain." Recently, the vertebral endplates have been identified as another possible pain trigger of the anterior column. This "endplate-driven" model, defined "vertebrogenic pain," is often interconnected with disc degeneration. Diagnosis of vertebrogenic and discogenic pain relies on imaging techniques that isolate pain generators and exclude comorbid conditions. Traditional methods, like radiographs and discography, are augmented by more sensitive methods, including SPECT, CT, and MRI. Morphologic MRI is pivotal in revealing indicators of vertebrogenic (eg, Modic endplate changes) and discogenic pain (eg, disc degeneration and annular fissures). More advanced methods, like ultra-short-echo time imaging, and quantitative MRI further amplify MRI's accuracy in the detection of painful endplate and disc pathology. This review explores the pathophysiology of vertebrogenic and discogenic pain as well as the impact of different imaging modalities in the diagnosis of low back pain. We hope this information can help identify patients who may benefit from personalized clinical treatment and image-guided therapies.

Titanium-Coated Polyetheretherketone Cages Vs Full Titanium Cages for Stand-Alone 1- or 2-Level Anterior Cervical Discectomy and Fusion: A Comparative Study.

BACKGROUND: Anterior cervical discectomy and fusion (ACDF) for the surgical treatment of cervical degenerative disease often includes an intervertebral cage, which restores disc height and lordosis while promoting fusion . Cage materials include titanium (TTN) or polyetheretherketone (PEEK). Controversy in material selection stems from higher fusion rates with TNN, despite a higher subsidence rate, while PEEK cages demonstrate superior preservation of interspace height. Combining the advantages of both materials, TTN-coated PEEK (TCPEEK) cages were developed, featuring a PEEK core with similar stiffness to the bone, enveloped with a TTN coat, improving osteointegration. However, the potential superiority of TCPEEK over TTN cages has not been investigated. This study aimed to compare clinical and radiographic outcomes following single- or double-level ACDF using either TTN or TCPEEK cages. METHODS: This retrospective single-center study included patients undergoing single- or double-level ACDF between 2017 and 2019. Clinical outcomes included the Neck Disability Index and revision surgery incidence. Radiographic parameters included cervical and segmental lordosis, C2 to C7 sagittal vertical axis, fusion, subsidence, and adjacent segment degeneration at a minimum 12-month follow-up. RESULTS: A total of 45 patients (16 TTN; 29 TCPEEK) and 58 cervical levels (21 TTN; 37 TCPEEK) were included. Both cages significantly improved Neck Disability Index scores (TTN -10.0; TCPEEK -14.1) without significant differences. Two single-level TCPEEK patients required revision surgery due to non-union. In the radiological assessments, no significant difference was found for subsidence rates (TTN 52.4%; TCPEEK 56.8%), adjacent segment degeneration, cervical and segmental lordosis, and changes in C2 to C7 sagittal vertical axis. Though not statistically significant, fusion rates trended slightly higher with TTN (90.5%) vs TCPEEK cages (86.5%). CONCLUSION: TTN and TCPEEK cages achieve satisfactory clinical and radiological outcomes in single- or double-level ACDF. This finding suggests that the choice between them can be based on other factors, such as surgeon preference or availability, rather than specific material properties. CLINICAL RELEVANCE: This study found that the selection of ACDF cage material did not affect clinical outcomes.

An Artificial Intelligence-Based Support Tool for Lumbar Spinal Stenosis Diagnosis from Self-Reported History Questionnaire.

OBJECTIVES: Symptomatic lumbar spinal stenosis (LSS) leads to functional impairment and pain. While radiologic characterization of the morphological stenosis grade can aid in the diagnosis, it may not always correlate with patient symptoms. Artificial intelligence (AI) may diagnose symptomatic LSS in patients solely based on self-reported history questionnaires. METHODS: We evaluated multiple machine learning (ML) models to determine the likelihood of LSS using a self-reported questionnaire in patients experiencing low back pain and/or numbness in the legs. The questionnaire was built from peer-reviewed literature and a multidisciplinary panel of experts. Random forest, lasso logistic regression, support vector machine, gradient boosting trees, deep neural networks, and automated machine learning models were trained and performance metrics were compared. RESULTS: Data from 4827 patients (4690 patients without LSS: mean age 62.44, range 27-84 years, 62.8% females, and 137 patients with LSS: mean age 50.59, range 30-71 years, 59.9% females) were retrospectively collected. Among the evaluated models, the random forest model demonstrated the highest predictive accuracy with an area under the receiver operating characteristic curve (AUROC) between model prediction and LSS diagnosis of 0.96, a sensitivity of 0.94, a specificity of 0.88, a balanced accuracy of 0.91, and a Cohen's kappa of 0.85. CONCLUSIONS: Our results indicate that ML can automate the diagnosis of LSS based on self-reported questionnaires with high accuracy. Implementation of standardized and intelligence-automated workflow may serve as a supportive diagnostic tool to streamline patient management and potentially lower health care costs.

Advanced Neuroimaging With Photon-Counting Detector CT.

ABSTRACT: Photon-counting detector computed tomography (PCD-CT) is an emerging technology and promises the next step in CT evolution. Photon-counting detectors count the number of individual incoming photons and assess the energy level of each of them. These mechanisms differ substantially from conventional energy-integrating detectors. The new technique has several advantages, including lower radiation exposure, higher spatial resolution, reconstruction of images with less beam-hardening artifacts, and advanced opportunities for spectral imaging. Research PCD-CT systems have already demonstrated promising results, and recently, the first whole-body full field-of-view PCD-CT scanners became clinically available. Based on published studies of preclinical systems and the first experience with clinically approved scanners, the performance can be translated to valuable neuroimaging applications, including brain imaging, intracranial and extracranial CT angiographies, or head and neck imaging with detailed assessment of the temporal bone. In this review, we will provide an overview of the current status in neuroimaging with upcoming and potential clinical applications.

Robotic-navigated assistance in spine surgery.

The aim of this study was to assess the accuracy of pedicle screw placement, as well as intraoperative factors, radiation exposure, and complication rates in adult patients with degenerative disorders of the thoracic and lumbar spines who have undergone robotic-navigated spinal surgery using a contemporary system. The authors reviewed the prospectively collected data on 196 adult patients who had pedicle screws implanted with robot-navigated assistance (RNA) using the Mazor X Stealth system between June 2019 and March 2022. Pedicle screws were implanted by one experienced spinal surgeon after completion of a learning period. The accuracy of pedicle screw placement was determined using intraoperative 3D fluoroscopy. A total of 1,123 pedicle screws were implanted: 1,001 screws (89%) were placed robotically, 63 (6%) were converted from robotic placement to a freehand technique, and 59 (5%) were planned to be implanted freehand. Of the robotically placed screws, 942 screws (94%) were determined to be Gertzbein and Robbins grade A with median deviation of 0.8 mm (interquartile range 0.4 to 1.6). Skive events were noted with 20 pedicle screws (1.8%). No adverse clinical sequelae were noted in the 90-day follow-up. The mean fluoroscopic exposure per screw was 4.9 seconds (SD 3.8). RNA is highly accurate and reliable, with a low rate of abandonment once mastered. No adverse clinical sequelae occurred after implanting a large series of pedicle screws using the latest generation of RNA. Understanding of patient-specific anatomical features and the real-time intraoperative identification of risk factors for suboptimal screw placement have the potential to improve accuracy further.

Lumbar dorsal root ganglion displacement between supine and prone positions evaluated with 3D MRI.

OBJECTIVE: Pre-operative lumbar spine MRI is usually acquired with the patient supine, whereas lumbar spine surgery is most commonly performed prone. For MRI to be used reliably and safely for intra-operative navigation for foraminal and extraforaminal decompression, the magnitude of dorsal root ganglion (DRG) displacement between supine and prone positions needs to be understood. METHODS: A prospective study of a degenerative lumbar spine cohort of 18 subjects indicated for lumbar spine surgery. Three-dimensional T2-weighted fast spin echo and T1-weighted spoiled gradient echo sequences were acquired at 3 T. Displacement and cross-sectional area (CSA) of the bilateral DRGs at 5 motion levels (L1-2 to L5-S1) were determined via 3D segmentation by 2 independent evaluators. Wilcoxon rank-sum tests without correction for multiple comparison were performed against hypothesized 1-mm absolute displacement and corresponding 24% CSA change. RESULTS: DRG mean absolute displacement was <1 mm (p > 0.99, mean = 0.707 mm, 95% confidence interval (CI) = 0.659 to 0.755 mm), with the largest directional displacement in the dorsal-to-ventral direction from supine to prone (mean = 0.141 mm, 95% CI = 0.082 to 0.200 mm). Directional displacements caudal-to-cephalad were 0.087 mm (95% CI = 0.022 to 0.151 mm), and left-right were -0.030 mm (95%CI = -0.059 to -0.001 mm). Mean CSA change was within 24% (p > 0.99, mean = -8.30%, 95% CI = -10.5 to -6.09%). Mean absolute displacement was largest for the L1 (mean = 0.811 mm) and L2 (mean = 0.829 mm) DRGs. CONCLUSIONS: Minimal, non-statistically significant soft tissue displacement and morphological area differences were demonstrated between supine and prone positions during 3D lumbar spine MRI.

Impact of photon counting detector CT derived virtual monoenergetic images and iodine maps on the diagnosis of pleural empyema.

PURPOSE: The purpose of this study was to evaluate the impact of virtual monoenergetic image (VMI) energies and iodine maps on the diagnosis of pleural empyema with photon counting detector computed tomography (PCD-CT). MATERIALS AND METHODS: In this IRB-approved retrospective study, consecutive patients with non-infectious pleural effusion or histopathology-proven empyema were included. PCD-CT examinations were performed on a dual-source PCD-CT in the multi-energy (QuantumPlus) mode at 120 kV with weight-adjusted intravenous contrast-agent. VMIs from 40-70 keV obtained in 10 keV intervals and an iodine map was reconstructed for each scan. CT attenuation was measured in the aorta, the pleura and the peripleural fat (between autochthonous dorsal muscles and dorsal ribs). Contrast-to-noise (CNR) and signal-to-noise (SNR) ratios were calculated. Two blinded radiologists evaluated if empyema was present (yes/no), and rated diagnostic confidence (1 to 4; not confident to fully confident, respectively) with and without using the iodine map. Sensitivity, specificity and diagnostic confidence were estimated. Interobserver agreement was estimated using an unweighted Cohen kappa test. A one-way ANOVA was used to compare variables. Differences in sensitivity and specificity between the different levels of energy were searched using McNemar test. RESULTS: Sixty patients (median age, 60 years; 26 women) were included. A strong negative correlation was found between image noise and VMI energies (r = -0.98; P = 0.001) and CNR increased with lower VMI energies (r = -0.98; P = 0.002). Diagnostic accuracy (96%; 95% CI: 82-100) as well as diagnostic confidence (3.4 ± 0.75 [SD]) were highest at 40 keV. Diagnostic accuracy and confidence at higher VMI energies improved with the addition of iodine maps (P ≤0.001). Overall, no difference in CT attenuation of peripleural fat between patients with empyema and those with pleural effusion was found (P = 0.07). CONCLUSION: Low VMI energies lead to a higher diagnostic accuracy and diagnostic confidence in the diagnosis of pleural empyema. Iodine maps help in diagnosing empyema only at high VMI energies.

Prospective Comparison of Two Robotically Navigated Pedicle Screw Instrumentation Techniques.

This study aimed to compare screw accuracy and incidence of skive between two robotically navigated instrumented techniques in posterior spine fusion surgery: manual anti-skive instrumentation with an anti-skive cannula (ASC) and the use of a navigated, high-speed drill (HSD). Over a 3-year period, consecutive patients are undergoing RNA posterior fusion surgery with either ASC (n = 53) or HSD (n = 63). Both groups met a value of approximately 292 screws in our analysis (296 ASC, 294 HSD), which was determined by a biostatistician at an academic institution. Screw accuracy and skive was analyzed using preoperative CT and intraoperative three-dimensional (3D) fluoroscopy. Among 590 planned robotically inserted pedicle screws (296 ASC, 294 HSD), 245 ASC screws (82.8%) and 283 HSD screws (96.3%) were successfully inserted (p < 0.05). Skive events occurred in 4/283 (1.4%) HSD screws and 15/245 (6.2%) ASC screws (p < 0.05). HSD screws showed better accuracy in the axial and sagittal planes, being closer to planned trajectories in all directions except cranial deviation (p < 0.05). Additionally, HSD had a significantly lower time per screw (1.9 ± 1.0 min) compared to ASC (3.2 ± 2.0 min, p < 0.001). No adverse clinical effects were observed. The HSD technique showed significant improvements in time and screw accuracy compared to ASC. Biplanar fluoroscopy and 3D imaging resulted in significantly lower radiation exposure and time compared to ASC. These significant findings in the HSD group may be attributed to the lower occurrence of malpositioned screws, leading to a decrease in the need for second authentication. This represents a notable iterative improvement of the RNA platform.

Determining Prior Authorization Approval for Lumbar Stenosis Surgery With Machine Learning.

STUDY DESIGN: Medical vignettes. OBJECTIVES: Lumbar spinal stenosis (LSS) is a degenerative condition with a high prevalence in the elderly population, that is associated with a significant economic burden and often requires spinal surgery. Prior authorization of surgical candidates is required before patients can be covered by a health plan and must be approved by medical directors (MDs), which is often subjective and clinician specific. In this study, we hypothesized that the prediction accuracy of machine learning (ML) methods regarding surgical candidates is comparable to that of a panel of MDs. METHODS: Based on patient demographic factors, previous therapeutic history, symptoms and physical examinations and imaging findings, we propose an ML which computes the probability of spinal surgical recommendations for LSS. The model implements a random forest model trained from medical vignette data reviewed by MDs. Sets of 400 and 100 medical vignettes reviewed by MDs were used for training and testing. RESULTS: The predictive accuracy of the machine learning model was with a root mean square error (RMSE) between model predictions and ground truth of .1123, while the average RMSE between individual MD's recommendations and ground truth was .2661. For binary classification, the AUROC and Cohen's kappa were .959 and .801, while the corresponding average metrics based on individual MD's recommendations were .844 and .564, respectively. CONCLUSIONS: Our results suggest that ML can be used to automate prior authorization approval of surgery for LSS with performance comparable to a panel of MDs.

Flexible array coil for cervical and extraspinal (FACE) MRI at 3.0 Tesla.

Objective.High-resolution MRI of the cervical spine (c-spine) and extraspinal neck region requires close-fitting receiver coils to maximize the signal-to-noise ratio (SNR). Conventional, rigid C-spine receiver coils do not adequately contour to the neck to accommodate varying body shapes, resulting in suboptimal SNR. Recent innovations in flexible surface coil array designs may provide three-dimensional (3D) bendability and conformability to optimize SNR, while improving capabilities for higher acceleration factors.Approach.This work describes the design, implementation, and preliminaryin vivotesting of a novel, conformal 23-channel receive-only flexible array for cervical and extraspinal (FACE) MRI at 3-Tesla (T), with use of high-impedance elements to enhance the coil's flexibility. Coil performance was tested by assessing SNR and geometry factors (g-factors) in a phantom compared to a conventional 21-channel head-neck-unit (HNU).In vivoimaging was performed in healthy human volunteers and patients using high-resolution c-spine and neck MRI protocols at 3T, including MR neurography (MRN).Main results.Mean SNR with the FACE was 141%-161% higher at left, right, and posterior off-isocenter positions and 4% higher at the isocenter of the phantom compared to the HNU. Parallel imaging performance was comparable for an acceleration factor (R) = 2 × 2 between the two coils, but improved forR= 3 × 3 with meang-factors ranging from 1.46-2.15 with the FACE compared to 2.36-3.62 obtained with the HNU. Preliminary human volunteer and patient testing confirmed that equivalent or superior image quality could be obtained for evaluation of osseous and soft tissue structures of the cervical region with the FACE.Significance.A conformal and highly flexible cervical array with high-impedance coil elements can potentially enable higher-resolution imaging for cervical imaging.

Detecting Abnormal Axillary Lymph Nodes on Mammograms Using a Deep Convolutional Neural Network.

The purpose of this study was to determine the feasibility of a deep convolutional neural network (dCNN) to accurately detect abnormal axillary lymph nodes on mammograms. In this retrospective study, 107 mammographic images in mediolateral oblique projection from 74 patients were labeled to three classes: (1) "breast tissue", (2) "benign lymph nodes", and (3) "suspicious lymph nodes". Following data preprocessing, a dCNN model was trained and validated with 5385 images. Subsequently, the trained dCNN was tested on a "real-world" dataset and the performance compared to human readers. For visualization, colored probability maps of the classification were calculated using a sliding window approach. The accuracy was 98% for the training and 99% for the validation set. Confusion matrices of the "real-world" dataset for the three classes with radiological reports as ground truth yielded an accuracy of 98.51% for breast tissue, 98.63% for benign lymph nodes, and 95.96% for suspicious lymph nodes. Intraclass correlation of the dCNN and the readers was excellent (0.98), and Kappa values were nearly perfect (0.93-0.97). The colormaps successfully detected abnormal lymph nodes with excellent image quality. In this proof-of-principle study in a small patient cohort from a single institution, we found that deep convolutional networks can be trained with high accuracy and reliability to detect abnormal axillary lymph nodes on mammograms.

Agony of choice: How anesthetics affect the composition and function of extracellular vesicles.

The choice of the anesthetic regime is suggested to affect clinical outcomes following major surgery. Propofol was shown to exert beneficial effects on different cancer outcomes, while volatile anesthetics may be favorable in cardiac surgery. Recently, extracellular vesicles (EVs) were discovered as essential signal mediators in physiological and pathophysiological processes including carcinogenesis and metastasis. Furthermore, depending on their cell source, EVs fulfill therapeutic functions. In addition to extracorporally produced EVs, appropriate systemic intervention such as remote ischemic preconditioning (RIPC) is considered to promote endogenous release of therapeutically active EVs to mediate cardioprotective effects. EVs are assembled in cell-type specific manners and the composition of EVs is not only affected by the disease, but also by the applied anesthetic of anesthetized patients. Here, we compare known impacts of anesthetic agents on outcomes in cancer surgery and cardioprotection and link these effects to the composition and therapeutic potential of EVs.

Extracellular vesicles isolated from patients undergoing remote ischemic preconditioning decrease hypoxia-evoked apoptosis of cardiomyoblasts after isoflurane but not propofol exposure.

Remote ischemic preconditioning (RIPC) can evoke cardioprotection following ischemia/reperfusion and this may depend on the anesthetic used. We tested whether 1) extracellular vesicles (EVs) isolated from humans undergoing RIPC protect cardiomyoblasts against hypoxia-induced apoptosis and 2) this effect is altered by cardiomyoblast exposure to isoflurane or propofol. EVs were isolated before and 60 min after RIPC or Sham from ten patients undergoing coronary artery bypass graft surgery with isoflurane anesthesia and quantified by Nanoparticle Tracking Analysis. Following EV-treatment for 6 hours under exposure of isoflurane or propofol, rat H9c2 cardiomyoblasts were cultured for 18 hours in normoxic or hypoxic atmospheres. Apoptosis was detected by flow cytometry. Serum nanoparticle concentrations in patients had increased sixty minutes after RIPC compared to Sham (2.5x1011±4.9x1010 nanoparticles/ml; Sham: 1.2x1011±2.0x1010; p = 0.04). Hypoxia increased apoptosis of H9c2 cells (hypoxia: 8.4%±0.6; normoxia: 2.5%±0.1; p<0.0001). RIPC-EVs decreased H9c2 cell apoptosis compared to control (apoptotic ratio: 0.83; p = 0.0429) while Sham-EVs showed no protection (apoptotic ratio: 0.97). Prior isoflurane exposure in vitro even increased protection (RIPC-EVs/control, apoptotic ratio: 0.79; p = 0.0035; Sham-EVs/control, apoptotic ratio:1.04) while propofol (50µM) abrogated protection by RIPC-EVs (RIPC-EVs/control, Apoptotic ratio: 1.01; Sham-EVs/control, apoptotic ratio: 0.94; p = 0.602). Thus, EVs isolated from patients undergoing RIPC under isoflurane anesthesia protect H9c2 cardiomyoblasts against hypoxia-evoked apoptosis and this effect is abrogated by propofol. This supports a role of human RIPC-generated EVs in cardioprotection and underlines propofol as a possible confounder in RIPC-signaling mediated by EVs.