Hospital Picture Archiving Research Articles

Long-Term Clinical Outcomes Following Nonsurgical Management of Salter-Harris Type II Fractures of the Proximal Phalanx of the Small Finger: A Prospective Cohort Study

Juxta-physeal fractures at the base of the proximal phalanx (FBPP) of the small finger are one of the most common hand fractures in children. Although many of these fractures are treated nonsurgically, it is unclear which fractures benefit from surgery or the degree of acceptable angulation appropriate for nonsurgical management. Our study aimed to assess long-term, patient-reported outcomes regarding function, appearance, and pain after nonsurgical management of FBPP of the small finger in a pediatric population. Our hospital Picture Archiving and Communication Systems database was queried to identify radiographs of the small finger of children between the ages of 8 and 16 years old taken from 2011 to 2021. Displacement on initial injury radiographs was calculated using the diaphyseal-metacarpal head angle. Patient-reported function, appearance, and pain were measured using standardized assessment tools sent to patients and parents electronically. Optional clinical photographs were uploaded by parents and assessed for residual clinical deformity. One hundred eighty-one eligible subjects were identified. Eighty (44%) agreed to participate, and 40 (22%) parent and 38 patient surveys were completed. The mean age at the time of injury was 11 years old (8-14 years), and the mean age at the time of survey completion was 17 years old (11-23 years). Patient T-scores were higher than the average reference T-score on the Patient-Reported Outcomes Measurement Information System Short Form v2.0-Upper Extremity. Overall, greater than 78% of patients and parents reported appearance as 8/10 or above on a VAS. Ninety-five percent (108/114) of patients reported no pain (0/10) for pain during activities and at rest. Children with up to 26° of initial coronal plane angulation reported better function than a reference population, good appearance, and no pain, at a mean of 6 years after injury. Our findings support nonsurgical management of most FBPP of the small finger in children. Therapeutic IV.

Establishment of local diagnostic reference levels for CT colonography at a tertiary hospital.

Diagnostic reference levels (DRLs) are an important metric in identifying abnormally high radiation doses in diagnostic examinations. National DRLs for CT colonography do not currently exist in South Africa, but there are efforts to collect data for a national DRL project. This study investigated radiation doses for CT colonography in adult patients at a large tertiary hospital in South Africa with the aim of setting local DRLs. Patient data from two CT scanners (Philips Ingenuity and Siemens Somatom go.Top) in the period March 2020 - March 2023 were obtained from the hospital's picture archiving and communication system (PACS) (n = 115). Analysis involved determining the median computed tomography dose index-volume (CTDIvol) and dose-length product (DLP) values. The findings were compared with DRLs established internationally. Ingenuity median CTDIvol was 20 mGy and DLP was 2169 mGy*cm; Somatom median CTDIvol was 6 mGy and DLP was 557 mGy*cm. Ingenuity exceeded the United Kingdom's (UK) recommended DRLs by 82% and 214%, respectively. Somatom median CTDIvol and DLP were 45% and 19% lower than UK NDRLs. Somatom's tin filter and other dose reduction features provided significant dose reduction. These data were used to set DRLs for CT colonography at the hospital; CTDIvol: 6 mGy and DLP: 557 mGy*cm. In addition to informing radiation protection practices at the level of the institution, the established local DRLs contribute towards implementing regional and national DRLs.

RECIST-Induced Reliable Learning: Geometry-Driven Label Propagation for Universal Lesion Segmentation.

Automatic universal lesion segmentation (ULS) from Computed Tomography (CT) images can ease the burden of radiologists and provide a more accurate assessment than the current Response Evaluation Criteria In Solid Tumors (RECIST) guideline measurement. However, this task is underdeveloped due to the absence of large-scale pixel-wise labeled data. This paper presents a weakly-supervised learning framework to utilize the large-scale existing lesion databases in hospital Picture Archiving and Communication Systems (PACS) for ULS. Unlike previous methods to construct pseudo surrogate masks for fully supervised training through shallow interactive segmentation techniques, we propose to unearth the implicit information from RECIST annotations and thus design a unified RECIST-induced reliable learning (RiRL) framework. Particularly, we introduce a novel label generation procedure and an on-the-fly soft label propagation strategy to avoid noisy training and poor generalization problems. The former, named RECIST-induced geometric labeling, uses clinical characteristics of RECIST to preliminarily and reliably propagate the label. With the labeling process, a trimap divides the lesion slices into three regions, including certain foreground, background, and unclear regions, which consequently enables a strong and reliable supervision signal on a wide region. A topological knowledge-driven graph is built to conduct the on-the-fly label propagation for the optimal segmentation boundary to further optimize the segmentation boundary. Experimental results on a public benchmark dataset demonstrate that the proposed method surpasses the SOTA RECIST-based ULS methods by a large margin. Our approach surpasses SOTA approaches over 2.0%, 1.5%, 1.4%, and 1.6% Dice with ResNet101, ResNet50, HRNet, and ResNest50 backbones.

The Incidence of the Associated Meniscal Ligament (ML) With the Anterior Cruciate Ligament (ACL) Tears

Objective: To assess the incidence of the associated meniscal ligament (ML) with the anterior cruciate ligament (ACL) tears in hospital A in Medina-KSA. And to see the effect of sports activities in comparison between males and females. 
 Materials and Methods: In a retrospective study, ninety-nine patients' knee magnetic resonance imaging (MRI) records were selected conveniently from the (A) hospital's picture archiving and communication system (PACS). Patients with suspected anterior cruciate and meniscal ligament' tears were included. Patients with posterior cruciate ligament (PCL) tear, medial collateral ligament (MCL) or lateral collateral ligament (LCL) injury, and previously operated index knees were excluded. All patients were scanned by a 1.5T MRI machine using standard MRI knee scanning protocol. The categorical and continuous variables were presented as percentages, frequencies, and descriptive statistics. The cross-tabulations and chi-square test were applied to obtain the statistical associations between the study variables. A statistical association was considered significant if the P-value was < 0.05. 
 Results: 72 males and 27 females were included, with a mean age of 33 ± 2.43. 42.4% of the knee tears were ML tears, followed by associated (ACL with ML) tears, noted in 31.3%, and 26.3% were ACL tears. The tears were more frequent in the age group (31-40), reported in 35.3% (out of 99). There was no statistical association between the tears and the participants' age, gender, or sports activities; P value = 0.3, 0.5, and 0.9, respectively. 
 Conclusion: The study concluded that the isolated ML tears were more frequent than the associated (ACL with ML) tears—moreover, there was no significant statistical association between the participants' age, gender, or sports activities.

Celiac Disease and Intussusception: A Common Association in Children.

In young childhood, intestinal intussusception (IS) is the most common cause of small bowel obstruction. A lead point such as Meckel diverticulum, polyps, tumors, enlarged lymph nodes, cystic fibrosis, and Schoenlein-Henoch purpura are recognized causes. Association between celiac disease (CD) and IS has been well recognized in adults but rarely in children. Data on causes and outcome of intussusception among Saudi children are lacking in the literature. Our objectives were to characterize the pattern of IS among Saudi children and investigate the frequency, clinical presentation, and outcome of intussusception among children with CD. We searched the hospital's picture archiving and communications system for abdominal imaging studies (ultrasound, magnetic resonance imaging, computed tomography scan, and barium contrast studies), performed between 2008 and 2019, using "intussusception" as a search key word. The hospital medical records of the identified cases of intussusception (aged 0-14 years) were then retrospectively reviewed to collect demographic, clinical, laboratory and imaging findings, management, and outcome. During the study period, 57 cases were identified as confirmed IS (31 boys, median age 1.95 years, range 0.33-11 years). Abdominal ultrasound was the diagnostic imaging study in 93%. An underlying cause (secondary IS) could be identified in 19 (33.3%) cases: CD in 6, malignancy and Henoch-Schoenlein purpura, 5 each, and Meckel diverticulum in 3; the remaining 38 (66.6%) cases of IS were idiopathic (primary IS). The presence of hypoalbuminemia and abdominal distension were significantly associated with secondary IS as compared with primary IS ( P < 0.001, P = 0.006, respectively). All of the 6 cases of IS associated with CD resolved spontaneously, but 3 were recurrent. Secondary causes contributed to a large proportion of IS in our study cohort (33%) as compared with 5% to 10% in the literature. Celiac disease is an underrecognized cause of IS among children. A child with IS and hypoalbuminemia, anemia, or chronic diarrhea needs to be investigated for CD to avoid unnecessary surgery.

Fusion extracted features from deep learning for identification of multiple positioning errors in dental panoramic imaging.

Dental panoramic imaging plays a pivotal role in dentistry for diagnosis and treatment planning. However, correctly positioning patients can be challenging for technicians due to the complexity of the imaging equipment and variations in patient anatomy, leading to positioning errors. These errors can compromise image quality and potentially result in misdiagnoses. This research aims to develop and validate a deep learning model capable of accurately and efficiently identifying multiple positioning errors in dental panoramic imaging. This retrospective study used 552 panoramic images selected from a hospital Picture Archiving and Communication System (PACS). We defined six types of errors (E1-E6) namely, (1) slumped position, (2) chin tipped low, (3) open lip, (4) head turned to one side, (5) head tilted to one side, and (6) tongue against the palate. First, six Convolutional Neural Network (CNN) models were employed to extract image features, which were then fused using transfer learning. Next, a Support Vector Machine (SVM) was applied to create a classifier for multiple positioning errors, using the fused image features. Finally, the classifier performance was evaluated using 3 indices of precision, recall rate, and accuracy. Experimental results show that the fusion of image features with six binary SVM classifiers yielded high accuracy, recall rates, and precision. Specifically, the classifier achieved an accuracy of 0.832 for identifying multiple positioning errors. This study demonstrates that six SVM classifiers effectively identify multiple positioning errors in dental panoramic imaging. The fusion of extracted image features and the employment of SVM classifiers improve diagnostic precision, suggesting potential enhancements in dental imaging efficiency and diagnostic accuracy. Future research should consider larger datasets and explore real-time clinical application.

An evaluation of the practice of image-guided lumbar puncture.

The purpose of this study was to evaluate the technical success and complication rates of image-guided lumbar puncture (IGLP) and to evaluate for differences in approach employed to help establish the optimum technique. A retrospective search of the hospital picture archiving and communications system was performed to identify all IGLPs that had taken place over a 5-year period. Radiology reports and the electronic medical record were examined to identify technical parameters and complications associated with each procedure. The technical success rate was 96% (219/228). 69.4% (n = 161) had a previously failed bedside attempt. The rate of complications was 0.01% (n = 2). No major complications were observed. There was no difference in the rates of failure (2.4% vs 3.6%, p = 0.68) or complications (0.008% vs 0.012%, p = 1) between interlaminar and interspinous approaches. IGLP is a safe procedure with a high rate of technical success. Where a difficult bedside attempt is anticipated, it is reasonable to forego this and proceed directly to IGLP. This paper helps to confirm what is already assumed about a common radiological procedure. This is important as there has been a shift from bedside technique to most lumbar punctures being performed via image guidance.

KNEE ALIGNMENT COMPARISON IN SINGLE VS DOUBLE LEG WEIGHTBEARING X-RAY

Background: One of the leading causes of disability in elders is osteoarthritis of knee. The functional outcome of surgical procedures depends on proper alignment assessment. Our study aims to access the knee alignment parameters is x-rays taken in double standing and single standing weight bearing x-rays. One hundred patients who visited Chettinad hospital and research inst Methods: itute, Chennai, with knee pain and knee deformity were selected for the study. All patients were subjected to single and double-leg standing Anteroposterior (AP) full-limb radiographs to access the varus or valgus deformity. Radiology technicians followed a standard protocol to obtain these radiographs. All X-rays were uploaded to the hospital picture archiving and communicating system and Hip-Knee-Ankle Angle (HKAA), Medial-Proximal-Tibial Angle (MPTA), LateralDistal-Femoral Angle (LDFA) and Joint-Line-Convergence Angle (JLCA) measured. The data was analysed using, MedCalc v18.2.1 The Results: mean age was 56.2 years (range 45- 68 years) mean BMI was 27.1 kg/m2. We were able to nd a signicant difference in JLCA and HKAA between double and single-leg standing x-rays. HKAA and JLCA were affected by the patie Conclusion: nt's position. In single-leg weight-bearing, the force line shifts medial to the knee joint centre, therefore the compressive stress is increased in the medial aspect. Followed by this change the angles are higher while the patient is standing on a single leg and varus deformity is augmented in this position

Relationship Between Nasal Septal Deviation Angles and Turbinates: A Computed Tomography Study.

Background This study aimed to evaluate inferior turbinate hypertrophy caused by nasal septum deviation, nasal septal deviation (SD) angles, and age differences with the help of paranasal computed tomography (CT) and to investigate the relationship between these parameters. Methodology The paranasal sinus CT images of 100 patients (50 males and 50 females) were retrieved from the hospital's picture archiving and communication system. In this retrospective study, patients were examined in two groups. There were 50 patients aged >35 years in group 1 and 50 patients aged <35 years in group 2. The SD side was determined using a coronal image and was mentioned as the ipsilateral side. The contralateral side of the SD side was mentioned as the contralateral side. Additionally, the SD curve angle (SDCA), superior SD angle (SSDA), and diameters and mucosal thicknesses of the inferior turbinates were measured. Concomitant ipsilateral sinusitis and ipsilateral concha bullosa (in the middle concha) were also noted as present or absent. Results In our study, the SDCA values of the ≥35-year age group were significantly higher than those of the <35-year age group (p < 0.05). Furthermore, the SSDA values of the ≥35-year age group were significantly lower than those ofthe <35-year age group (p < 0.05). In each of the age groups, ipsilateral inferior turbinate mucosal thickness and ipsilateral inferior turbinate diameter values were significantly higher than those of the contralateral sides (p < 0.05). Ipsilateral concha bullosa was present in 30.0% of the <35-year age group and 18.0% of the ≥35-year age group.Ipsilateral sinusitis was present in 34.0% of the <35-year age group and 52.0% of the ≥35-year age group. Conclusions SD and inferior turbinate hypertrophies should be evaluated together and measured with paranasal CT to provide more efficient nasal aeration. Studies with larger patient series are needed to elucidate the etiology.

Artificial Intelligence Assisted Computational Tomographic Detection of Lung Nodules for Prognostic Cancer Examination: A Large-Scale Clinical Trial.

Low-dose computed tomography (LDCT) has emerged as a standard method for detecting early-stage lung cancer. However, the tedious computer tomography (CT) slide reading, patient-by-patient check, and lack of standard criteria to determine the vague but possible nodule leads to variable outcomes of CT slide interpretation. To determine the artificial intelligence (AI)-assisted CT examination, AI algorithm-assisted CT screening was embedded in the hospital picture archiving and communication system, and a 200 person-scaled clinical trial was conducted at two medical centers. With AI algorithm-assisted CT screening, the sensitivity of detecting nodules sized 4-5 mm, 6~10 mm, 11~20 mm, and &gt;20 mm increased by 41%, 11.2%, 10.3%, and 18.7%, respectively. Remarkably, the overall sensitivity of detecting varied nodules increased by 20.7% from 67.7% to 88.4%. Furthermore, the sensitivity increased by 18.5% from 72.5% to 91% for detecting ground glass nodules (GGN), which is challenging for radiologists and physicians. The free-response operating characteristic (FROC) AI score was ≥0.4, and the AI algorithm standalone CT screening sensitivity reached &gt;95% with an area under the localization receiver operating characteristic curve (LROC-AUC) of &gt;0.88. Our study demonstrates that AI algorithm-embedded CT screening significantly ameliorates tedious LDCT practices for doctors.

Patterns of Associated Knee Ligament and Chondral Injuries in First-Time Traumatic Patellar Dislocation: A Retrospective Magnetic Resonance Imaging (MRI) -Based Study.

Introduction First-time acute traumatic patellar dislocation, when managed without a knee magnetic resonance imaging (MRI) scan, may lead to missed diagnoses of important associated knee injuries. The aim of this study was to ascertain the incidence of associated ligamentous and cartilaginous injuries in first-time traumatic patella dislocation. Methods This was a five-year retrospective study on patients aged 16-45 who had knee MRI scans showing the characteristic bone bruise patterns seen in traumatic lateral patellar dislocation. Anonymized data from the hospital picture archiving and communication system (PACS) was obtained with each scan reviewed by a consultant radiologist, a fellowship-trained orthopaedic knee specialist, and an orthopaedic registrar or resident. Results A total of 200 knee MRI scans were screened. 61 eligible knee MRI scans were included in the study. The patients' ages ranged from 16 to 42 years old, with a mean of 25 years. 73.8% were male. A medial patellofemoral ligament (MPFL) tear or rupture occurred in 58 of 61 knees (95%) with MPFL attenuation in three (5%) injured knees. Meniscal injuries were identified in 5 of 61 knees (8.2%), medial collateral ligament (MCL) injuries in 11 of 61 knees (18%), osteochondral injuries and loose bodies in 17 of 61 knees (27.9%), and anterior cruciate ligament (ACL) injury in one knee (1.6%). Conclusions This single-centre MRI-based study has provided information on the incidence of associated chondral and ligamentous injuries in patients with first-time acute traumatic patellar dislocation. This information will be useful for clinicians when counselling patients and will add to the available literature on this injury. An MRI scan should be obtained in cases of suspected first-time traumatic patellar dislocations, especially in active young patients, due to the incidence of other associated traumatic knee lesions that might need surgical treatment and lead to persisting knee symptoms if neglected.

Deployment and validation of an AI system for detecting abnormal chest radiographs in clinical settings.

BackgroundThe purpose of this paper is to demonstrate a mechanism for deploying and validating an AI-based system for detecting abnormalities on chest X-ray scans at the Phu Tho General Hospital, Vietnam. We aim to investigate the performance of the system in real-world clinical settings and compare its effectiveness to the in-lab performance.MethodThe AI system was directly integrated into the Hospital's Picture Archiving and Communication System (PACS) after being trained on a fixed annotated dataset from other sources. The system's performance was prospectively measured by matching and comparing the AI results with the radiology reports of 6,285 chest X-ray examinations extracted from the Hospital Information System (HIS) over the last 2 months of 2020. The normal/abnormal status of a radiology report was determined by a set of rules and served as the ground truth.ResultsOur system achieves an F1 score—the harmonic average of the recall and the precision—of 0.653 (95% CI 0.635, 0.671) for detecting any abnormalities on chest X-rays. This corresponds to an accuracy of 79.6%, a sensitivity of 68.6%, and a specificity of 83.9%.ConclusionsComputer-Aided Diagnosis (CAD) systems for chest radiographs using artificial intelligence (AI) have recently shown great potential as a second opinion for radiologists. However, the performances of such systems were mostly evaluated on a fixed dataset in a retrospective manner and, thus, far from the real performances in clinical practice. Despite a significant drop from the in-lab performance, our result establishes a reasonable level of confidence in applying such a system in real-life situations.

DICOM Image ANalysis and Archive (DIANA): an Open-Source System for Clinical AI Applications.

In the era of data-driven medicine, rapid access and accurate interpretation of medical images are becoming increasingly important. The DICOM Image ANalysis and Archive (DIANA) system is an open-source, lightweight, and scalable Python interface that enables users to interact with hospital Picture Archiving and Communications Systems (PACS) to access such data. In this work, DIANA functionality was detailed and evaluated in the context of retrospective PACS data retrieval and two prospective clinical artificial intelligence (AI) pipelines: bone age (BA) estimation and intra-cranial hemorrhage (ICH) detection. DIANA orchestrates activity beginning with post-acquisition study discovery and ending with online notifications of findings. For AI applications, system latency (exam completion to system report time) was quantified and compared to that of clinicians (exam completion to initial report creation time). Mean DIANA latency was 9.04 ± 3.83 and 20.17 ± 10.16min compared to clinician latency of 51.52 ± 58.9 and 65.62 ± 110.39min for BA and ICH, respectively, with DIANA latencies being significantly lower (p < 0.001). DIANA's capabilities were also explored and found effective in retrieving and anonymizing protected health information for "big-data" medical imaging research and analysis. Mean per-image retrieval times were 1.12 ± 0.50 and 0.08 ± 0.01s across x-ray and computed tomography studies, respectively. The data herein demonstrate that DIANA can flexibly integrate into existing hospital infrastructure and improve the process by which researchers/clinicians access imaging repository data. This results in a simplified workflow for large data retrieval and clinical integration of AI models.

Important landmarks and distances for posterior fossa surgery measured by temporal MDCT.

In this retrospective study, we aimed to present important anatomical structures and distances for posterior fossa surgery by temporal multidetector computed tomography (MDCT). The temporal MDCT images of 317 adult patients (158 males and 159 females) were retrieved from the hospital's picture archiving and communication system (PACS). In the coronal temporal MDCT views, the cochlea-carotid canal and jugular bulb-mastoid bone outer surface were measured. In the axial MDCT views, the carotid canal-jugular bulb and carotid canal-posterior fossa distances were measured; the carotid canal and jugular bulb anterior-posterior (AP) and transverse dimensions were also measured. The bilateral cochlea-carotid canal, jugular bulb-mastoid bone outer surface, and right carotid canal-jugular bulb distances were significantly greater in the males than those in the females (p < 0.05). The carotid canal-posterior fossa distance was not different in both genders (p > 0.05). The carotid canal-jugular bulb and the carotid canal-posterior fossa distances were greater on the left side than those on the right side in both genders (p < 0.05). In males, the outer surface distance was greater on the left jugular bulb-mastoid bone than that on the right side of that bone (p < 0.05). The difference between the carotid canal AP dimensions was not significant between males and females (p > 0.05). However, the carotid canal transverse dimension, jugular bulb AP, and transverse dimensions were significantly greater in the males than those in the females, bilaterally (p < 0.05). In each gender separately, the carotid canal AP and transverse dimensions were greater on the left side and the jugular bulb AP and transverse dimensions were greater on the right side than those on the left side (p < 0.05). Positive correlations were found between the cochlea-carotid canal, the jugular bulb-mastoid bone outer surface, and the carotid canal-jugular bulb distances as well as between the jugular bulb-mastoid bone outer surface and the carotid canal-posterior fossa distances (p < 0.05). In older patients, the carotid canal-posterior fossa distances were shorter on the left side (p < 0.05). Vascular and neural localizations should be well understood in the operative area before applying the surgical approach in the posterior fossa. Computed tomography (CT) has a greater role in the evaluation of bone structures and vascular canals in this area.

Morphometric examination of the styloid process by 3D-CT in patients with Eagle syndrome

This study aimed to obtain the length, medial-anterior angulation and basis-apex coordinates of the styloid process in patients with Eagle syndrome by three-dimensional computed tomography. This study was performed on three-dimensional computed tomography images obtained from the hospital picture archiving and communication system (by obtaining 3D images by a RadiAnt DICOM viewer 4.6.9 version). In this study, the anterior and lateral lengths of the styloid process, its anterior and medial angulation, and the coordinate values on the x, y, and z-axes of 24 patients (14 males, 10 females) diagnosed with Eagle syndrome were examined. The mean anterior lengths were found to be 41.45mm on the right and 36.07mm on the left sides, while the mean lateral lengths were found to be 42.15mm on the right and 37.59mm on the left sides. The mean medial angulation was measured to be 62.91° on the right and 63.42° on the left, while mean anterior angulation was measured to be 28.01° on the right and 30.43° on the left. The styloid process basis coordinates were determined as (right: x = - 41.30, y = 0, z = 0, left: x = 40.79, y = 0, z = 0), and apex coordinates were determined as (right: x = - 22.61, y = - 36.86, z = - 19.52, left: x = 24.90, y = - 32.14, z = - 18.65). Knowing the styloid process basis and apex coordinates in addition to knowing the its length and angulation will be useful in diagnosing Eagle syndrome. We think that these results in relation to the coordinates of the styloid process will bring a new perspective to clinicians who investigate the length and angulation of the styloid process.

Fixation of ankle fractures - a major trauma centre's experience in improving quality.

Ankle fracture malreduction has been shown to result in poor long-term functional outcomes. Varying methods can be used to change practice and thereby outcomes. We present over four years' worth of results with the effects of different techniques for change. Two audit cycles were performed incorporating three audit data collections; an initial standard setting audit in 2013, with re-audits in 2015 and 2017. Between the first and second audit was a period of education and reflection. Between the second and third audit there was a change in process in ankle fracture management supported by education. Image intensifier films were reviewed on the hospital picture archiving and communication system, by at least two blinded observers in each cycle. These were scored based on the criteria published by Pettrone et al. In the initial audit in 2013, there were 94 patients, with a malreduction rate of 33%. In the second audit in 2015, there were 68 patients, with an increase in malreduction rate to 43.8%. In the third audit in 2017, there were 205 patients, with a significant decrease in malreduction rate to 2.4%. The final major complication rate was 0.98%. The rate of deep infection was 0.5%. By recognising and addressing the need to improve the quality of ankle fracture fixation, we have made significant improvements in radiological outcomes. Education alone, without system change, was not successful in our department in achieving improved outcomes.

Deep Learning in the Prediction of Ischaemic Stroke Thrombolysis Functional Outcomes: A Pilot Study

Intravenous thrombolysis decision-making and obtaining of consent would be assisted by an individualized risk-benefit ratio. Deep learning (DL) models may be able to assist with this patient selection. Clinical data regarding consecutive patients who received intravenous thrombolysis across two tertiary hospitals over a 7-year period were extracted from existing databases. The noncontrast computed tomography brain scans for these patients were then retrieved with hospital picture archiving and communication systems. Using a combination of convolutional neural networks (CNN) and artificial neural networks (ANN) several models were developed to predict either improvement in the National Institutes of Health Stroke Scale of ≥4 points at 24 hours ("NIHSS24"), or modified Rankin Scale 0-1 at 90 days ("mRS90"). The developed CNN and ANN were then applied to a test set. The THRIVE, HIAT, and SPAN-100 scores were also calculated for the patients in the test set and used to predict NIHSS24 and mRS90. Data from 204 individuals were included in the project. The best performing DL model for prediction of mRS90 was a combination CNN + ANN based on clinical data and computed tomography brain (accuracy = 0.74, F1 score = 0.69). The best performing model for NIHSS24 prediction was also the combination CNN + ANN (accuracy = 0.71, F1 score = 0.74). DL models may aid in the prediction of functional thrombolysis outcomes. Further investigation with larger datasets and additional imaging sequences is indicated.

Imaging Anatomic Research of Oblique Lumbar Interbody Fusion in a Chinese Population Based on Magnetic Resonance

To provide anatomic evidence of preoperative assessment of oblique lumbar interbody fusion (OLIF) for Chinese patients. From the hospital picture archiving and communication system, 400 lumbar magnetic resonance imaging studies of adults performed between November 2016 and January 2017 were selected. L2-3, L3-4, L4-5, and L5-S1 transverse and sagittal images were studied, and anatomic parameters associated with OLIF surgery, including bare window and psoas window, were measured and recorded. SPSS software was used for data summarization, sorting, and analysis to explore the significance of various anatomic parameters. OLIF surgical corridors to the L2-S1 discs were found in most magnetic resonance imaging scans studied. The size of the psoas affects the difficulty of psoas muscle traction. It is relatively easy to perform OLIF surgery in older women. Most of the human iliac arteries were bifurcated and aggregated in front of the L4-5 intervertebral disc. The lower the aggregate level of the common iliac vein, the less likely it was to have the OLIF surgical corridor in the L5-S1 segment. The most frequently used lengths for a lumbar interbody cage for OLIF for Chinese patients are 50 mm and 55 mm. OLIF can be a good choice for lumbar intervertebral fusion, including L5-S1 segment, in most Chinese patients. Older women are likely to have more scope of OLIF surgery. As a routine preoperative examination, lumbar magnetic resonance imaging is of great importance to OLIF surgery preoperative assessment.

Quick and simple; psoas density measurement is an independent predictor of anastomotic leak and other complications after colorectal resection

BackgroundRadiologically defined sarcopenia has been shown to predict negative outcomes after cancer surgery, however radiological assessment of sarcopenia often requires additional software and standardisation against anthropomorphic data. Measuring psoas density using hospital Picture Archiving and Communication Systems (PACS), universally available in the UK, may have advantages over methods requiring the use of additional specialist and often costly software. The aim of this study was to assess the association between radiologically defined sarcopenia measured by psoas density and postoperative outcome in patients having a colorectal cancer resection.MethodsAll patients having a resection for colorectal cancer, discussed by the colorectal multi-disciplinary team in one institution between 1/1/15 and 31/12/15, were retrospectively identified. Mean psoas density at the level of the L3 vertebra was analysed from preoperative computed tomography (CT) scans to define sarcopenia using the Picture Archiving and Communication Systems (PACS). Postoperative complications and mortality were recorded.ResultsOne hundred and sixty-nine patients had a colorectal resection for cancer and 140 of these had a primary anastomosis. Ninety-day mortality and 1-year mortality were 1.1% and 7.1%, respectively. Eighteen (10.7%) patients suffered a Clavien–Dindo grade 3 or 4 complication of which 6 (33%) were anastomotic leaks. In the whole cohort, sarcopenia was associated with an increased risk of Clavien–Dindo grade 3 or 4 complications [adjusted OR 6.33 (1.65–24.23) p = 0.007]. In those who had an anastomosis, sarcopenia was associated with an increased risk of anastomotic leak [adjusted OR 14.37 (1.37–150.04) p = 0.026].ConclusionsA quick and easy radiological assessment of sarcopenia by measuring psoas density on preoperative CT scan using software universally available in the UK is highly predictive of postoperative morbidity in colorectal cancer patients.

Convolutional Neural Networks for Automated Fracture Detection and Localization on Wrist Radiographs.

To demonstrate the feasibility and performance of an object detection convolutional neural network (CNN) for fracture detection and localization on wrist radiographs. Institutional review board approval was obtained with waiver of consent for this retrospective study. A total of 7356 wrist radiographic studies were extracted from a hospital picture archiving and communication system. Radiologists annotated all radius and ulna fractures with bounding boxes. The dataset was split into training (90%) and validation (10%) sets and used to train fracture localization models for frontal and lateral images. Inception-ResNet Faster R-CNN architecture was implemented as a deep learning model. The models were tested on an unseen test set of 524 consecutive emergency department wrist radiographic studies with two radiologists in consensus as the reference standard. Per-fracture, per-image (ie, per-view), and per-study sensitivity and specificity were determined. Area under the receiver operating characteristic curve (AUC) analysis was performed. The model detected and correctly localized 310 (91.2%) of 340 and 236 (96.3%) of 245 of all radius and ulna fractures on the frontal and lateral views, respectively. The per-image sensitivity, specificity, and AUC were 95.7% (95% confidence interval [CI]: 92.4%, 97.8%), 82.5% (95% CI: 77.4%, 86.8%), and 0.918 (95% CI: 0.894, 0.941), respectively, for the frontal view and 96.7% (95% CI: 93.6%, 98.6%), 86.4% (95% CI: 81.9%, 90.2%), and 0.933 (95% CI: 0.912, 0.954), respectively, for the lateral view. The per-study sensitivity, specificity, and AUC were 98.1% (95% CI: 95.6%, 99.4%), 72.9% (95% CI: 67.1%, 78.2%), and 0.895 (95% CI: 0.870, 0.920), respectively. The ability of an object detection CNN to detect and localize radius and ulna fractures on wrist radiographs with high sensitivity and specificity was demonstrated.© RSNA, 2019.