Deep Learning Assistance Closes the Accuracy Gap in Fracture Detection Across Clinician Types.

BACKGROUND: Missed fractures are the most common diagnostic errors in musculoskeletal imaging and can result in treatment delays and preventable morbidity. Deep learning, a subfield of artificial intelligence, can be used to accurately detect fractures by training algorithms to emulate the judgments of expert clinicians. Deep learning systems that detect fractures are often limited to specific anatomic regions and require regulatory approval to be used in practice. Once these hurdles are overcome, deep learning systems have the potential to improve clinician diagnostic accuracy and patient care. QUESTIONS/PURPOSES: This study aimed to evaluate whether a Food and Drug Administration-cleared deep learning system that identifies fractures in adult musculoskeletal radiographs would improve diagnostic accuracy for fracture detection across different types of clinicians. Specifically, this study asked: (1) What are the trends in musculoskeletal radiograph interpretation by different clinician types in the publicly available Medicare claims data? (2) Does the deep learning system improve clinician accuracy in diagnosing fractures on radiographs and, if so, is there a greater benefit for clinicians with limited training in musculoskeletal imaging? METHODS: We used the publicly available Medicare Part B Physician/Supplier Procedure Summary data provided by the Centers for Medicare & Medicaid Services to determine the trends in musculoskeletal radiograph interpretation by clinician type. In addition, we conducted a multiple-reader, multiple-case study to assess whether clinician accuracy in diagnosing fractures on radiographs was superior when aided by the deep learning system compared with when unaided. Twenty-four clinicians (radiologists, orthopaedic surgeons, physician assistants, primary care physicians, and emergency medicine physicians) with a median (range) of 16 years (2 to 37) of experience postresidency each assessed 175 unique musculoskeletal radiographic cases under aided and unaided conditions (4200 total case-physician pairs per condition). These cases were comprised of radiographs from 12 different anatomic regions (ankle, clavicle, elbow, femur, forearm, hip, humerus, knee, pelvis, shoulder, tibia and fibula, and wrist) and were randomly selected from 12 hospitals and healthcare centers. The gold standard for fracture diagnosis was the majority opinion of three US board-certified orthopaedic surgeons or radiologists who independently interpreted the case. The clinicians' diagnostic accuracy was determined by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve, sensitivity, and specificity. Secondary analyses evaluated the fracture miss rate (1-sensitivity) by clinicians with and without extensive training in musculoskeletal imaging. RESULTS: Medicare claims data revealed that physician assistants showed the greatest increase in interpretation of musculoskeletal radiographs within the analyzed time period (2012 to 2018), although clinicians with extensive training in imaging (radiologists and orthopaedic surgeons) still interpreted the majority of the musculoskeletal radiographs. Clinicians aided by the deep learning system had higher accuracy diagnosing fractures in radiographs compared with when unaided (unaided AUC: 0.90 [95% CI 0.89 to 0.92]; aided AUC: 0.94 [95% CI 0.93 to 0.95]; difference in least square mean per the Dorfman, Berbaum, Metz model AUC: 0.04 [95% CI 0.01 to 0.07]; p < 0.01). Clinician sensitivity increased when aided compared with when unaided (aided: 90% [95% CI 88% to 92%]; unaided: 82% [95% CI 79% to 84%]), and specificity increased when aided compared with when unaided (aided: 92% [95% CI 91% to 93%]; unaided: 89% [95% CI 88% to 90%]). Clinicians with limited training in musculoskeletal imaging missed a higher percentage of fractures when unaided compared with radiologists (miss rate for clinicians with limited imaging training: 20% [95% CI 17% to 24%]; miss rate for radiologists: 14% [95% CI 9% to 19%]). However, when assisted by the deep learning system, clinicians with limited training in musculoskeletal imaging reduced their fracture miss rate, resulting in a similar miss rate to radiologists (miss rate for clinicians with limited imaging training: 9% [95% CI 7% to 12%]; miss rate for radiologists: 10% [95% CI 6% to 15%]). CONCLUSION: Clinicians were more accurate at diagnosing fractures when aided by the deep learning system, particularly those clinicians with limited training in musculoskeletal image interpretation. Reducing the number of missed fractures may allow for improved patient care and increased patient mobility. LEVEL OF EVIDENCE: Level III, diagnostic study.

Is there a role for preoperative liver reducing diet in hepatectomy? A systematic review.

PURPOSE: Preoperative very low-calorie diets (VLCDs) have been shown to reduce liver volume, reduce the risk of liver injury and improve safety during bariatric surgery. Hepatic steatosis (HS) has been associated with poorer outcomes in liver resection. VLCD can be used to improve HS. We aim to explore if preoperative VLCD could improve outcomes for patients with HS undergoing liver resection. METHODS: We performed a systematic review of MEDLINE, EMBASE, PubMed and Cochrane databases. Studies were included if they were full-text articles investigating the effect of a preoperative dietary intervention in patients undergoing liver resection on intra-operative and post-operative outcomes. The last search was performed on 11 Jun 2020. Evidence quality was assessed by "GRADE". A narrative review was undertaken. RESULTS: Five studies were found: one RCT and four cohort studies including 133 patients in intervention groups and 181 controls. Three used diet-only strategies and two diet and exercise strategies with varying time courses and monitoring. The quality of evidence assessed by GRADE was "high" for the RCT and "low" for the four cohort studies. Steatosis objectively improved in three studies, with evidence of reduced liver volume and increased attenuation on imaging in one. All studies showed a reduction in body weight and body mass index (BMI). Intra-operative blood loss was decreased following a diet-only intervention in two studies, and liver mobility improved in one. No difference was found in morbidity, mortality or hospital length of stay between intervention and control groups. CONCLUSIONS: There is evidence of poorer outcomes in liver resection patients with existing HS. There is an expected role for a preoperative VLCD to optimise these patients for surgery. Existing publications support this, but diet interventions and outcome measures are inconsistent, and patient numbers are small. There is scope for a well-designed, multi-centre randomised trial to investigate this further.

Characteristics of respiratory measures in young adults scanned at rest, including systematic changes and "missed" deep breaths.

Breathing rate and depth influence the concentration of carbon dioxide in the blood, altering cerebral blood flow and thus functional magnetic resonance imaging (fMRI) signals. Such respiratory fluctuations can have substantial influence in studies of fMRI signal covariance in subjects at rest, the so-called "resting state functional connectivity" technique. If respiration is monitored during fMRI scanning, it is typically done using a belt about the subject's abdomen to record abdominal circumference. Several measures have been derived from these belt records, including the windowed envelope of the waveform (ENV), the windowed variance in the waveform (respiration variation, RV), and a measure of the amplitude of each breath divided by the cycle time of the breath (respiration volume per time, RVT). Any attempt to gauge respiratory contributions to fMRI signals requires a respiratory measure, but little is known about how these measures compare to each other, or how they perform beyond the small studies in which they were initially proposed. Here, we examine the properties of these measures in hundreds of healthy young adults scanned for an hour each at rest, a subset of the Human Connectome Project chosen for having high-quality physiological records. We find: 1) ENV, RV, and RVT are all correlated, and ENV and RV are more highly correlated to each other than to RVT; 2) respiratory events like deep breaths exhibit characteristic heart rate elevations, fMRI signal changes, head motions, and image quality abnormalities time-locked to large deflections in the belt traces; 3) all measures can "miss" deep breaths; 4) RVT "misses" deep breaths more than ENV or RV; 5) all respiratory measures change systematically over the course of a 14.4-min scan. We discuss the implications of these findings for the literature and ways to move forward in modeling respiratory influences on fMRI scans.

Distinct and similar patterns of emotional development in adolescents and young adults.

Adolescence is a developmental period of increased sensitivity to social emotional cues, but it is less known whether young adults demonstrate similar social emotional sensitivity. The current study tested variation in reaction times to emotional face cues during different phases of emotional development. Ex-Gaussian parameters mu, sigma, and tau were computed, in addition to mean, median and standard deviation (SD) in reaction times (RT) during an emotional go/nogo-paradigm with fearful, happy, and calm facial expressions in 377 participants, 6-30 years of age. Across development, mean RT showed slowing to fearful facial expressions relative to both calm and happy facial cues, but mu revealed that this pattern was specific to adolescence. In young adulthood, increased variability to fearful expressions relative to both happy and calm ones was captured by SD and tau. The findings that adolescents had longer response latencies to fearful faces, whereas young adults demonstrated greater response variability to fearful faces, together reflect how social emotional processing continues to evolve from adolescence into early adulthood. The findings suggest that young adulthood is also a vulnerable period for processing social emotional cues that ultimately may be important to better understand why different psychopathologies emerge in early adulthood.

Distinctions among real and apparent respiratory motions in human fMRI data.

Head motion estimates in functional magnetic resonance imaging (fMRI) scans appear qualitatively different with sub-second image sampling rates compared to the multi-second sampling rates common in the past. Whereas formerly the head appeared still for much of a scan with brief excursions from baseline, the head now appears to be in constant motion, and motion estimates often seem to divulge little information about what is happening in a scan. This constant motion has been attributed to respiratory oscillations that do not alias at faster sampling rates, and investigators are divided on the extent to which such motion is "real" motion or only "apparent" pseudomotion. Some investigators have abandoned the use of motion estimates entirely due to these considerations. Here we investigate the properties of motion in several fMRI datasets sampled at rates between 720 and 1160 ms, and describe 5 distinct kinds of respiratory motion: 1) constant real respiratory motion in the form of head nodding most evident in vertical position and pitch, which can be very large; 2) constant pseudomotion at the same respiratory rate as real motion, occurring only in the phase encode direction; 3) punctate real motions occurring at times of very deep breaths; 4) a low-frequency pseudomotion in only the phase encode direction at and after very deep breaths; 5) slow modulation of vertical and anterior-posterior head position by the respiratory envelope. We reformulate motion estimates in light of these considerations and obtain good concordance between motion estimates, physiologic records, image quality measures, and events evident in the fMRI signals. We demonstrate how variables describing respiration or body habitus separately scale with distinct kinds of head motion. We also note heritable aspects of respiration and motion.

Customized head molds reduce motion during resting state fMRI scans.

Head motion causes artifacts in functional magnetic resonance imaging (fMRI) scans, a problem especially relevant for task-free resting state paradigms and for developmental, aging, and clinical populations. In a cohort spanning 7-28 years old (mean age 15) we produced customized head-anatomy-specific Styrofoam molds for each subject that inserted into an MRI head coil. We scanned these subjects under two conditions: using our standard procedure of packing the head coil with foam padding about the head to reduce head motion, and using the customized molds to reduce head motion. In 12 of 13 subjects, the molds reduced head motion throughout the scan and reduced the fraction of a scan with substantial motion (i.e., volumes with motion notably above baseline levels of motion). Motion was reduced in all 6 head position estimates, especially in rotational, left-right, and superior-inferior directions. Motion was reduced throughout the full age range studied, including children, adolescents, and young adults. In terms of the fMRI data itself, quality indices improved with the head mold on, scrubbing analyses detected less distance-dependent artifact in scans with the head mold on, and distant-dependent artifact was less evident in both the entire scan and also during only low-motion volumes. Subjects found the molds comfortable. Head molds are thus effective tools for reducing head motion, and motion artifacts, during fMRI scans.

The Chemistry of Mordançage, a Historic Photographic Process.

Based on a 19th century method, the process known as mordançage has been used to dramatically alter black and white photographs since the 1960s. Mordançage subjects a fully developed photograph to a series of wet chemical reactions, beginning with an acidic copper(II) chloride and hydrogen peroxide bleaching solution. Different types of photographic paper, time in mordançage solution, and the concentration of hydrogen peroxide, copper(II) chloride, and acetic acid are all variables that influence the degree of emulsion lift and corresponding veils observed. We conducted experiments to explore these variables and analyzed processed prints using ImageJ to measure changes in pixel intensity and SEM-EDS for compositional analysis. Fiber-based and cotton rag papers resulted in more emulsion lift than resin-coated paper, and increased time in the mordançage solution had similar results. Quantitative analysis of the bleaching solution reveals how the process variables affect the formation of veils, and a process scheme is presented. This work provides insight into the chemical transformations involved in this artistic process.

How interview questions are placed in time influences caregiver description of social communication symptoms on the ADI-R.

BACKGROUND: Caregiver report is crucial for the diagnosis of childhood onset psychiatric disorders, particularly autism. Three experiments were conducted to determine whether caregiver reports of past and current behaviors are affected by question timing and ordering. METHODS: Using the Autism Diagnostic Interview - Revised (ADI-R), two studies systematically varied the order in which caregivers were asked about behaviors. In a third study, descriptions of children's current behaviors at age 5 were compared to retrospective descriptions of behaviors at age 5 collected at age 10. RESULTS: Caregivers, who were first asked about a history of symptoms, described less severe past and present behavior than caregivers reporting current behaviors as well as caregivers reporting current and history of symptoms together. Caregivers retrospectively reported more severe behaviors for age 5 when their children were age 10 than they had when their children were age 5. CONCLUSIONS: Caregivers describe past behaviors differently depending on whether they are asked about current symptoms first. Methods of caregiver reporting can influence interpretations of symptom severity with effects on diagnoses and research findings.

Adolescent-specific patterns of behavior and neural activity during social reinforcement learning.

Humans are sophisticated social beings. Social cues from others are exceptionally salient, particularly during adolescence. Understanding how adolescents interpret and learn from variable social signals can provide insight into the observed shift in social sensitivity during this period. The present study tested 120 participants between the ages of 8 and 25 years on a social reinforcement learning task where the probability of receiving positive social feedback was parametrically manipulated. Seventy-eight of these participants completed the task during fMRI scanning. Modeling trial-by-trial learning, children and adults showed higher positive learning rates than did adolescents, suggesting that adolescents demonstrated less differentiation in their reaction times for peers who provided more positive feedback. Forming expectations about receiving positive social reinforcement correlated with neural activity within the medial prefrontal cortex and ventral striatum across age. Adolescents, unlike children and adults, showed greater insular activity during positive prediction error learning and increased activity in the supplementary motor cortex and the putamen when receiving positive social feedback regardless of the expected outcome, suggesting that peer approval may motivate adolescents toward action. While different amounts of positive social reinforcement enhanced learning in children and adults, all positive social reinforcement equally motivated adolescents. Together, these findings indicate that sensitivity to peer approval during adolescence goes beyond simple reinforcement theory accounts and suggest possible explanations for how peers may motivate adolescent behavior.

A Decade of Discovery-Eukaryotic Replisome Disassembly at Replication Termination.

The eukaryotic replicative helicase (CMG complex) is assembled during DNA replication initiation in a highly regulated manner, which is described in depth by other manuscripts in this Issue. During DNA replication, the replicative helicase moves through the chromatin, unwinding DNA and facilitating nascent DNA synthesis by polymerases. Once the duplication of a replicon is complete, the CMG helicase and the remaining components of the replisome need to be removed from the chromatin. Research carried out over the last ten years has produced a breakthrough in our understanding, revealing that replication termination, and more specifically replisome disassembly, is indeed a highly regulated process. This review brings together our current understanding of these processes and highlights elements of the mechanism that are conserved or have undergone divergence throughout evolution. Finally, we discuss events beyond the classic termination of DNA replication in S-phase and go over the known mechanisms of replicative helicase removal from chromatin in these particular situations.

Adaptive thresholding increases sensitivity to detect changes in the rate of skin conductance responses to psychologically arousing stimuli in both laboratory and ambulatory settings.

Psychophysiologists recording electrodermal activity (EDA) often derive measures of slow, tonic activity-skin conductance level (SCL)-and faster, more punctate changes-skin conductance responses (SCRs). A SCR is conventionally considered to have occurred when the local amplitude of the EDA signal exceeds a researcher-determined threshold (e.g., 0.05 µS), typically fixed across study participants and conditions. However, fixed SCR thresholds can preferentially exclude data from individuals with low SCL because their SCRs are smaller on average, thereby reducing statistical power for group-level analyses. Thus, we developed a fixed plus adaptive (FA) thresholding method that adjusts identification of SCRs based on an individual's SC at the onset of the SCR to increase statistical power and include data from more participants. We assess the utility of applying FA thresholding across two independent samples and explore age and race-related associations with EDA outcomes. Study 1 uses wired EDA measurements from 254 healthy adults responding to evocative images and sounds in a laboratory setting. Study 2 uses wireless EDA measurements from 20 children with autism in a clinical environment while they completed behavioral tasks. Compared to a 0.01, 0.03, and 0.05 µS fixed threshold, FA thresholding at 1.9% modestly increases statistical power to detect a difference in SCR rate between tasks with higher vs. lower subjective arousal and reduces exclusion of participants by up to 5% across both samples. This novel method expands the EDA analytical toolbox and may be useful in populations with highly variable basal SCL or when comparing groups with different basal SCL. Future research should test for reproducibility and generalizability in other tasks, samples, and contexts. IMPACT STATEMENTS: This article is important because it introduces a novel method to enhance sensitivity and statistical power in analyses of skin conductance responses from electrodermal data.

Characterizing replisome disassembly in human cells.

To ensure timely duplication of the entire eukaryotic genome, thousands of replication machineries (replisomes) act on genomic DNA at any time during S phase. In the final stages of this process, replisomes are unloaded from chromatin. Unloading is driven by polyubiquitylation of MCM7, a subunit of the terminated replicative helicase, and processed by p97/VCP segregase. Most of our knowledge of replication termination comes from model organisms, and little is known about how this process is executed and regulated in human somatic cells. Here we show that replisome disassembly in this system requires CUL2LRR1-driven MCM7 ubiquitylation, p97, and UBXN7 for unloading and provide evidence for "backup" mitotic replisome disassembly, demonstrating conservation of such mechanisms. Finally, we find that small-molecule inhibitors against Cullin ubiquitin ligases (CULi) and p97 (p97i) affect replisome unloading but also lead to induction of replication stress in cells, which limits their usefulness to specifically target replisome disassembly processes.

The medial prefrontal cortex and the emergence of self-conscious emotion in adolescence.

In the present study, we examined the relationship between developmental modulation of socioaffective brain systems and adolescents' preoccupation with social evaluation. Child, adolescent, and adult participants viewed cues indicating that a camera was alternately off, warming up, or projecting their image to a peer during the acquisition of behavioral-, autonomic-, and neural-response (functional MRI) data. Believing that a peer was actively watching them was sufficient to induce self-conscious emotion that rose in magnitude from childhood to adolescence and partially subsided into adulthood. Autonomic arousal was uniquely heightened in adolescents. These behavioral patterns were paralleled by emergent engagement of the medial prefrontal cortex (MPFC) and striatum-MPFC connectivity during adolescence, which are thought to promote motivated social behavior in adolescence. These findings demonstrate that adolescents' self-consciousness is related to age-dependent sensitivity of brain systems critical to socioaffective processes. Further, unique interactions between the MPFC and striatum may provide a mechanism by which social-evaluation contexts influence adolescent behavior.

Replication fork dynamics and the DNA damage response.

Prevention and repair of DNA damage is essential for maintenance of genomic stability and cell survival. DNA replication during S-phase can be a source of DNA damage if endogenous or exogenous stresses impair the progression of replication forks. It has become increasingly clear that DNA-damage-response pathways do not only respond to the presence of damaged DNA, but also modulate DNA replication dynamics to prevent DNA damage formation during S-phase. Such observations may help explain the developmental defects or cancer predisposition caused by mutations in DNA-damage-response genes. The present review focuses on molecular mechanisms by which DNA-damage-response pathways control and promote replication dynamics in vertebrate cells. In particular, DNA damage pathways contribute to proper replication by regulating replication initiation, stabilizing transiently stalled forks, promoting replication restart and facilitating fork movement on difficult-to-replicate templates. If replication fork progression fails to be rescued, this may lead to DNA damage and genomic instability via nuclease processing of aberrant fork structures or incomplete sister chromatid separation during mitosis.

TRAIP resolves DNA replication-transcription conflicts during the S-phase of unperturbed cells.

Cell division is the basis for the propagation of life and requires accurate duplication of all genetic information. DNA damage created during replication (replication stress) is a major cause of cancer, premature aging and a spectrum of other human disorders. Over the years, TRAIP E3 ubiquitin ligase has been shown to play a role in various cellular processes that govern genome integrity and faultless segregation. TRAIP is essential for cell viability, and mutations in TRAIP ubiquitin ligase activity lead to primordial dwarfism in patients. Here, we have determined the mechanism of inhibition of cell proliferation in TRAIP-depleted cells. We have taken advantage of the auxin induced degron system to rapidly degrade TRAIP within cells and to dissect the importance of various functions of TRAIP in different stages of the cell cycle. We conclude that upon rapid TRAIP degradation, specifically in S-phase, cells cease to proliferate, arrest in G2 stage of the cell cycle and undergo senescence. Our findings reveal that TRAIP works in S-phase to prevent DNA damage at transcription start sites, caused by replication-transcription conflicts.

Behavioral and neural properties of social reinforcement learning.

Social learning is critical for engaging in complex interactions with other individuals. Learning from positive social exchanges, such as acceptance from peers, may be similar to basic reinforcement learning. We formally test this hypothesis by developing a novel paradigm that is based on work in nonhuman primates and human imaging studies of reinforcement learning. The probability of receiving positive social reinforcement from three distinct peers was parametrically manipulated while brain activity was recorded in healthy adults using event-related functional magnetic resonance imaging. Over the course of the experiment, participants responded more quickly to faces of peers who provided more frequent positive social reinforcement, and rated them as more likeable. Modeling trial-by-trial learning showed ventral striatum and orbital frontal cortex activity correlated positively with forming expectations about receiving social reinforcement. Rostral anterior cingulate cortex activity tracked positively with modulations of expected value of the cues (peers). Together, the findings across three levels of analysis--social preferences, response latencies, and modeling neural responses--are consistent with reinforcement learning theory and nonhuman primate electrophysiological studies of reward. This work highlights the fundamental influence of acceptance by one's peers in altering subsequent behavior.

The C-terminal region of the exosome-associated protein Rrp47 is specifically required for box C/D small nucleolar RNA 3'-maturation.

Cells lacking the exosome-associated protein Rrp47 show similar defects in stable RNA processing to those observed in the absence of the catalytic subunit Rrp6, but the precise mechanism(s) by which Rrp47 functions together with Rrp6 remains unclear. Deletion complementation analyses defined an N-terminal region of Rrp47, largely coincident with the bioinformatically defined Sas10/C1D domain, which was sufficient for protein function in vivo. In vitro protein interaction studies demonstrated that this domain of Rrp47 binds the PMC2NT domain of Rrp6. Expression of the N-terminal domain of Rrp47 in yeast complemented most RNA-processing defects associated with the rrp47Δ mutant but failed to complement the defect observed in 3'-end maturation of box C/D small nucleolar RNAs. Consistent with these results, protein capture assays revealed an interaction between the C-terminal region of Rrp47 and the small nucleolar ribonucleoproteins Nop56 and Nop58. Filter binding assays demonstrated that deletion of the lysine-rich sequence at the C terminus of Rrp47 blocked RNA binding in vitro. Furthermore, a protein mutated both at the C terminus and within the N-terminal domain showed a synergistic defect in RNA binding without impacting on its ability to interact with Rrp6. These studies provide evidence for a role of Rrp47 in registering a small nucleolar ribonucleoprotein particle assembly, functionally characterize the Sas10/C1D domain of Rrp47, and show that both the C terminus of Rrp47 and the N-terminal domain contribute to its RNA-binding activity.

Annual research review: re-thinking the classification of autism spectrum disorders.

BACKGROUND: The nosology of autism spectrum disorders (ASD) is at a critical point in history as the field seeks to better define dimensions of social-communication deficits and restricted/repetitive behaviors on an individual level for both clinical and neurobiological purposes. These different dimensions also suggest an increasing need for quantitative measures that accurately map their differences, independent of developmental factors such as age, language level and IQ. METHOD: Psychometric measures, clinical observation as well as genetic, neurobiological and physiological research from toddlers, children and adults with ASD are reviewed. RESULTS: The question of how to conceptualize ASDs along dimensions versus categories is discussed within the nosology of autism and the proposed changes to the DSM-5 and ICD-11. Differences across development are incorporated into the new classification frameworks. CONCLUSIONS: It is crucial to balance the needs of clinical practice in ASD diagnostic systems, with neurobiologically based theories that address the associations between social-communication and restricted/repetitive dimensions in individuals. Clarifying terminology, improving description of the core features of ASD and other dimensions that interact with them and providing more valid and reliable ways to quantify them, both for research and clinical purposes, will move forward both practice and science.

Transcranial Neuromodulation Array With Imaging Aperture for Simultaneous Multifocus Stimulation in Nonhuman Primates.

Even simple behaviors arise from the simultaneous activation of multiple regions in the brain. Thus, the ability to simultaneously stimulate multiple regions within a brain circuit should allow for better modulation of function. However, performing simultaneous multifocus ultrasound neuromodulation introduces challenges to transducer design. Using 3-D Fullwave simulations, we have designed an ultrasound neuromodulation array for nonhuman primates that: 1) can simultaneously focus on multiple targets and 2) include an imaging aperture for additional functional imaging. This design is based on a spherical array, with 128 15-mm elements distributed in a spherical helix pattern. It is shown that clustering the elements tightly around the 65-mm imaging aperture located at the top of the array improves targeting at shallow depths, near the skull surface. Spherical arrays have good focusing capabilities through the skull at the center of the array, but focusing on off-center locations is more challenging due to the natural geometric configuration and the angle of incidence with the skull. In order to mitigate this, the 64 elements closest to the aperture were rotated toward and focusing on a shallow target, and the 64 elements farthest from the aperture were rotated toward and focusing on a deeper target. Data illustrated that this array produced focusing on the somatosensory cortex with a gain of 4.38 and to the thalamus with a gain of 3.82. To improve upon this, the array placement was optimized based on phase aberration simulations, allowing for the elements with the largest impact on the gain at each focal point to be found. This optimization resulted in an array design that can focus on the somatosensory cortex with a gain of 5.19 and the thalamus with a gain of 4.45. Simulations were also performed to evaluate the ability of the array to focus on 28 additional brain regions, showing that off-center target regions can be stimulated, but those closer to the skull will require corrective steps to deliver the same amount of energy to those locations. This simulation and design process can be adapted to an individual monkey or human skull morphologies and specific target locations within individuals by using orientable 3-D printing of the transducer case and by electronic phase aberration correction.

Reevaluation of missed lung cancer with artificial intelligence.

Lung cancer is often missed on chest radiographs, despite chest radiography typically being the first imaging modality in the diagnosis pathway. We present a 46 year-old male with chest pain referred for chest X-ray, and initial interpretation reported no abnormality within the patient's lungs. The patient was discharged but returned 4 months later with persistent and worsening symptoms. At this time, chest X-ray was again performed and revealed an enlarging left perihilar mass with post-obstructive atelectasis in the left lower lobe. Follow-up chest computerized tomography scan confirmed lung cancer with post-obstructive atelectasis, and subsequent bronchoscopy-assisted biopsy confirmed squamous cell carcinoma. Retrospective analysis of the initial chest radiograph, which had reported normal findings, was performed with Chest-CAD, a Food and Drug Administration (FDA) cleared computer-assisted detection (CAD) software device that analyzes chest radiograph studies using artificial intelligence. The device highlighted the perihilar region of the left lung as suspicious. Additional information provided by artificial intelligence software holds promise to prevent missed detection of lung cancer on chest radiographs.