Retrospective Diagnostic Study Research Articles

Oropharyngeal Cancer Staging Health Record Extraction Using Artificial Intelligence

Accurate, timely, and cost-effective methods for staging oropharyngeal cancers are crucial for patient prognosis and treatment decisions, but staging documentation is often inaccurate or incomplete. With the emergence of artificial intelligence in medicine, data abstraction may be associated with reduced costs but increased efficiency and accuracy of cancer staging. To evaluate an algorithm using an artificial intelligence engine capable of extracting essential information from medical records of patients with oropharyngeal cancer and assigning tumor, nodal, and metastatic stages according to American Joint Committee on Cancer eighth edition guidelines. This retrospective diagnostic study was conducted among a convenience sample of 806 patients with oropharyngeal squamous cell carcinoma. Medical records of patients with staged oropharyngeal squamous cell carcinomas who presented to a single tertiary care center between January 1, 2010, and August 1, 2020, were reviewed. A ground truth cancer stage dataset and comprehensive staging rule book consisting of 135 rules encompassing p16 status, tumor, and nodal and metastatic stage were developed. Subsequently, 4 distinct models were trained: model T (entity relationship extraction) for anatomical location and invasion state, model S (numerical extraction) for lesion size, model M (sequential classification) for metastasis detection, and a p16 model for p16 status. For validation, results were compared against ground truth established by expert reviewers, and accuracy was reported. Data were analyzed from March to November 2023. The accuracy of algorithm cancer stages was compared with ground truth. Among 806 patients with oropharyngeal cancer (mean [SD] age, 63.6 [10.6] years; 651 males [80.8%]), 421 patients (52.2%) were positive for human papillomavirus. The artificial intelligence engine achieved accuracies of 55.9% (95% CI, 52.5%-59.3%) for tumor, 56.0% (95% CI, 52.5%-59.4%) for nodal, and 87.6% (95% CI, 85.1%-89.7%) for metastatic stages and 92.1% (95% CI, 88.5%-94.6%) for p16 status. Differentiation between localized (stages 1-2) and advanced (stages 3-4) cancers achieved 80.7% (95% CI, 77.8%-83.2%) accuracy. This study found that tumor and nodal staging accuracies were fair to good and excellent for metastatic stage and p16 status, with clinical relevance in assigning optimal treatment and reducing toxic effect exposures. Further model refinement and external validation with electronic health records at different institutions are necessary to improve algorithm accuracy and clinical applicability.

Diagnostic Accuracy of Transthoracic Echocardiography for Acute Type A Aortic Syndrome: A Systematic Review and Meta-Analysis

Transthoracic echocardiography (TTE) is currently recognized as the potential first-line imaging test for patients with suspected acute type A aortic syndrome (AAAS). Direct TTE sign for detecting AAAS is positive if there is an intimal flap separating two aortic lumens or aortic wall thickening seen in the ascending aorta. Indirect TTE sign indicates high-risk features of AAAS, such as aortic root dilatation, pericardial effusion, and aortic regurgitation. Our aim is to summarize the existing clinical evidence regarding the diagnostic accuracy of TTE and to evaluate its potential role in the management of patients with suspected AAAS. We included prospective or retrospective diagnostic cohort studies, written in any language, that specifically focused on using TTE to diagnose AAAS from databases such as PubMed, EMBASE, MEDLINE, and the Cochrane Library. The pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio [1], and hierarchical summary receiver-operating characteristic (HSROC) curve were calculated for TTE in diagnosing AAAS. We applied Quality Assessment of Diagnostic Accuracy (QUADAS-2) tool and Grading of Recommendations, Assessment, Development and Evaluation (GRADE) quality assessment criteria. Ten studies (2886 patients) were included in the meta-analysis. The pooled sensitivity and specificity of direct TTE signs were 58% (95% CI, 38-76%) and 94% (95% CI, 89-97%). For any TTE signs, the pooled sensitivity and specificity were 91% (95% CI, 85-94%) and 74% (95% CI, 61-84%). The diagnostic accuracy of direct TTE signs was significantly higher than that of any TTE signs, as measured by the area under the HSROC curve [0.95 (95% CI, 0.92-0.96) vs. 0.87 (95% CI, 0.84-0.90)] in four studies. Our study suggests that TTE could serve as the initial imaging test for patients with suspected AAAS. Given its high specificity, the presence of direct TTE signs may indicate AAAS, whereas the absence of any TTE signs, combined with low clinical suspicion, could suggest a lower likelihood of AAAS.

Evaluation of postoperative pancreatic fistula prediction scales following pancreatoduodenectomies based on magnetic resonance imaging: A diagnostic test study.

BackgroundPostoperative pancreatic fistula (POPF) is one of the most feared and common complications following pancreatoduodenectomies. This study aims to evaluate the performance of different scales in predicting POPF using magnetic resonance imaging (MRI), including estimation of the pancreatic duct diameter, pancreatic texture, main duct index, relation to the portal vein, and intra-abdominal fat thickness. Materials and MethodsA retrospective diagnostic test study was designed. Between January 2017 and December 2021, 133 pancreatoduodenectomies were performed at our institution. The performance for predicting overall POPF and clinically relevant POPF (CR-POPF) was evaluated using a receiver operating characteristic (ROC) curve. ResultsA total of 96 patients were included in the study, of whom 26 patients experienced overall POPF, and 8 patients had CR-POPF. When analyzing the predictive value of each of the different scores applied, the Birmingham score showed the highest performance for predicting overall POPF and CR-POPF with an AUC (area under the curve) of 0.815 (95% CI 0.725 – 0.906) and 0.813 (0.679 – 0.947), respectively. ConclusionThe Birmingham scale demonstrated the highest predictive performance for POPF. It is a simple scale with only two variables that can be obtained preoperatively using MRI. Based on these results, we recommend its use in patients undergoing pancreatoduodenectomy.

CPAK classification cannot be used to determine segmental coronal extra-articular knee deformity.

While the coronal plane alignment of the knee (CPAK) classification serves as a useful guide in personalising total knee arthroplasty (TKA), the extent of its correlation with segmental coronal extra-articular knee deformities remains uncertain. This study aims to investigate the potential correlation between CPAK matrix groups and segmental coronal extra-articular deformities in prearthritic knees, shedding light on the relationship between these two factors that seems to be both essential to perform personalised TKA. A radiological assessment of 1240 nonarthritic knees was performed by evaluating lower limb measurements following the protocol established by Paley et al. Subsequently, all knees were classified into their respective CPAK matrix groups. In our quest to discern any correlation between the CPAK matrix groups and the presence of segmental coronal extra-articular knee deformities, nine potential coronal extra-articular deformity phenotype(CEDP)groupswere identified based on medial proximal tibial angle (MPTA) and mechanical lateral distal femoral angle (mLDFA). Neutral values for MPTA and mLDFAwere set at 90.0° ± 3.0°and then at 87.0° ± 2.0°. Each CPAK matrix group underwent detailed coronal morphology analysisand then, segmental coronal extra-articular deformities were assessed by comparing them with the CEDP groups. The study revealed a mean hip-knee-ankle angle (HKA) of 178.6° ± 4.4°,mLDFA of 86.9° ± 2.5°, MPTAof 85.4° ± 2.4°, arithmetic HKAof -1.4° ± 3.2°and joint line obliquityof 172.5° ± 3.7°. The varus CPAK groups (I/IV/VII) included 435 patients, the neutral groups (II/V/VIII) comprised 630 patientsand the valgus groups (III/VI/IX) had 175 patients. Notably, CPAK matrix groups were not distinctly associated with specific coronal extra-articular deformity phenotype (CEDP) groups. Particularly among the most common CPAK matrix groups (I/II/III/IV/V), there was a significant variation in segmental coronal extra-articular deformity patterns. Moreover, when neutral MPTA/mLDFA values were set at 87.0° ± 2.0°, the CPAK matrix groups exhibited even greater variability in coronal extra-articular deformities. The CPAK matrix groups do not exhibit a direct correlation with a specific extra-articular deformity pattern (CEDP), thus rendering them unsuitable for determining segmental coronal extra-articular knee deformities. Level III, retrospective diagnostic study.

Dual-energy computed tomography in a multiparametric regression model for diagnosing lymph node metastases in pancreatic ductal adenocarcinoma

ObjectiveTo investigate the diagnostic value of dual-energy computed tomography (DECT) quantitative parameters in the identification of regional lymph node metastasis in pancreatic ductal adenocarcinoma (PDAC).MethodsThis retrospective diagnostic study assessed 145 patients with pathologically confirmed pancreatic ductal adenocarcinoma from August 2016–October 2020. Quantitative parameters for targeted lymph nodes were measured using DECT, and all parameters were compared between benign and metastatic lymph nodes to determine their diagnostic value. A logistic regression model was constructed; the receiver operator characteristics curve was plotted; the area under the curve (AUC) was calculated to evaluate the diagnostic efficacy of each energy DECT parameter; and the DeLong test was used to compare AUC differences. Model evaluation was used for correlation analysis of each DECT parameter.ResultsStatistical differences in benign and metastatic lymph nodes were found for several parameters. Venous phase iodine density had the highest diagnostic efficacy as a single parameter, with AUC 0.949 [95% confidence interval (CI):0.915–0.972, threshold: 3.95], sensitivity 79.80%, specificity 96.00%, and accuracy 87.44%. Regression models with multiple parameters had the highest diagnostic efficacy, with AUC 0.992 (95% CI: 0.967–0.999), sensitivity 95.96%, specificity 96%, and accuracy 94.97%, which was higher than that for a single DECT parameter, and the difference was statistically significant.ConclusionAmong all DECT parameters for regional lymph node metastasis in PDAC, venous phase iodine density has the highest diagnostic efficacy as a single parameter, which is convenient for use in clinical settings, whereas a multiparametric regression model has higher diagnostic value compared with the single-parameter model.

Diagnostic performance of the WHO definition of probable dengue within the first 5 days of symptoms on Reunion Island.

The relevance of the World Health Organization (WHO) criteria for defining probable dengue had not yet been evaluated in the context of dengue endemicity on Reunion Island. The objective of this retrospective diagnostic study was to evaluate the diagnostic performance of the 2009 WHO definition of probable dengue and to propose an improvement thereof. From the medical database, we retrieved the data of subjects admitted to the emergency department of the University Hospital of Reunion Island in 2019 with suspected dengue fever (DF) within a maximum of 5 days post symptom onset, and whose diagnosis was confirmed by a Reverse Transcriptase Polymerase Chain Reaction (RT-PCR). The intrinsic characteristics of probable dengue definitions were reported in terms of sensitivity, specificity, positive and negative likelihood ratios (LR+ and LR-), using RT-PCR as the gold standard. Of the 1,181 subjects who exhibited a positive RT-PCR, 652 (55%) were classified as probable dengue. The WHO definition of probable dengue yielded a sensitivity of 64% (95%CI 60-67%), a specificity of 57% (95%CI 52-61%), a LR+ of 1.49 (95%CI 1.33-1.67), and a LR- of 0.63 (95%CI 0.56-0.72). The sensitivity and LR- for diagnosing and ruling out probable dengue could be improved by the addition of lymphopenia on admission (74% [95%CI: 71-78%] and 0.54 [95%CI: 0.46-0.63] respectively), at the cost of slight reductions of specificity and LR+ (48% [95%CI: 44-53%] and 1.42 [95%CI: 1.29-1.57], respectively). In the absence of, or when rapid diagnostic testing is unreliable, the use of the improved 2009 WHO definition of probable dengue could facilitate the identification of subjects who require further RT-PCR testing, which should encourage the development of patient management, while also optimizing the count and quarantine of cases, and guiding disease control.

Comparing three-dimensional and two-dimensional deep-learning, radiomics, and fusion models for predicting occult lymph node metastasis in laryngeal squamous cell carcinoma based on CT imaging: a multicentre, retrospective, diagnostic study

The occult lymph node metastasis (LNM) of laryngeal squamous cell carcinoma (LSCC) affects the treatment and prognosis of patients. This study aimed to comprehensively compare the performance of the three-dimensional and two-dimensional deep learning models, radiomics model, and the fusion models for predicting occult LNM in LSCC. In this retrospective diagnostic study, a total of 553 patients with clinical N0 stage LSCC, who underwent surgical treatment without distant metastasis and multiple primary cancers, were consecutively enrolled from four Chinese medical centres between January 01, 2016 and December 30, 2020. The participant data were manually retrieved from medical records, imaging databases, and pathology reports. The study cohort was divided into a training set (n=300), an internal test set (n=89), and two external test sets (n=120 and 44, respectively). The three-dimensional deep learning (3D DL), two-dimensional deep learning (2D DL), and radiomics model were developed using CT images of the primary tumor. The clinical model was constructed based on clinical and radiological features. Two fusion strategies were utilized to develop the fusion model: the feature-based DLRad_FB model and the decision-based DLRad_DB model. The discriminative ability and correlation of 3D DL, 2D DL and radiomics features were analysed comprehensively. The performances of the predictive models were evaluated based on the pathological diagnosis. The 3D DL features had superior discriminative ability and lower internal redundancy compared to 2D DL and radiomics features. The DLRad_DB model achieved the highest AUC (0.89-0.90) among all the study sets, significantly outperforming the clinical model (AUC=0.73-0.78, P=0.0001-0.042, Delong test). Compared to the DLRad_DB model, the AUC values for the DLRad_FB, 3D DL, 2D DL, and radiomics models were 0.82-0.84 (P=0.025-0.46), 0.86-0.89 (P=0.75-0.97), 0.83-0.86 (P=0.029-0.66), and 0.79-0.82 (P=0.0072-0.10), respectively in the study sets. Additionally, the DLRad_DB model exhibited the best sensitivity (82-88%) and specificity (79-85%) in the test sets. The decision-based fusion model DLRad_DB, which combines 3D DL, 2D DL, radiomics, and clinical data, can be utilized to predict occult LNM in LSCC. This has the potential to minimize unnecessary lymph node dissection and prophylactic radiotherapy in patients with cN0 disease. National Natural Science Foundation of China, Natural Science Foundation of Shandong Province.

Added value of non-contrast CT for the diagnosis of acute pyelonephritis in older patients with suspected infection with an unknown focus: a retrospective diagnostic study

ObjectivesIn older patients, the diagnosis of acute pyelonephritis (APN) is challenging. The aim was to evaluate the added value of CT to history, physical examination and urinalysis for the diagnosis...

Diagnosis of superficial endometriosis on transvaginal ultrasound by visualization of peritoneum of pouch of Douglas.

Around 80% of women with endometriosis have superficial endometriosis (SE) rather than ovarian or deep endometriosis (DE). However, to date, advances in non-invasive, imaging-based diagnosis have been limited to DE or ovarian disease. The objective of this study was to determine whether we can detect SE on transvaginal ultrasound scan (TVS) by assessing the peritoneum of the pouch of Douglas (POD). This was a retrospective diagnostic test study following a change in practice to include POD peritoneum assessment for SE during TVS at a tertiary London hospital. Eligible patients underwent TVS by a single clinician trained in endometriosis scanning and a subsequent surgical procedure (laparoscopy) between April 2018 and September 2021. Participants formed a consecutive series. The TVS findings were compared with those of laparoscopy as the gold standard. Comparison of TVS findings with intraoperative findings was performed by calculating the diagnostic test performance measures (sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and positive and negative likelihood ratios). The study included a total of 100 patients. We found that 43/100 (43.0%) patients had no endometriosis, 33/100 (33.0%) had SE and 24/100 (24.0%) had DE on laparoscopy. SE was correctly detected on TVS in 17/33 patients, with a sensitivity of 51.5% (95% CI, 33.5-69.2%), specificity of 94.0% (95% CI, 85.4-98.4%), PPV of 81.0% (95% CI, 60.8-92.1%) and NPV of 79.7% (95% CI, 73.4-84.9%). DE was correctly diagnosed in 20/24 cases, including all ovarian cases, with a sensitivity of 83.3% (95% CI, 62.3-95.3%), specificity of 97.4% (95% CI, 90.8-99.7%), PPV of 90.9% (95% CI, 71.6-97.5%) and NPV of 94.9% (95% CI, 88.3-97.8%). The detection of SE on TVS was most accurate in the POD (sensitivity, 50.0%; specificity, 96.4%; PPV, 76.9%; NPV, 88.9%). This study shows that the detection of SE in the POD is possible using routine TVS. While negative TVS does not reliably confirm the absence of disease or replace diagnostic laparoscopy, positive TVS facilitates non-invasive diagnosis for a much larger group of women than was previously possible. This should help to reduce the time from the onset of symptoms to diagnosis and enable initiation of medical treatment without the risk, cost and delay associated with a surgical diagnosis. © 2023 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

A New Deep Learning Algorithm for Detecting Spinal Metastases on Computed Tomography Images.

Retrospective diagnostic study. To automatically detect osteolytic bone metastasis lesions in the thoracolumbar region using conventional computed tomography (CT) scans, we developed a new deep learning (DL)-based computer-aided detection model. Radiographic detection of bone metastasis is often difficult, even for orthopedic surgeons and diagnostic radiologists, with a consequent risk for pathologic fracture or spinal cord injury. If we can improve detection rates, we will be able to prevent the deterioration of patients' quality of life at the end stage of cancer. This study included CT scans acquired at Tokyo Medical and Dental University (TMDU) Hospital between 2016 and 2022. A total of 263 positive CT scans that included at least one osteolytic bone metastasis lesion in the thoracolumbar spine and 172 negative CT scans without bone metastasis were collected for the datasets to train and validate the DL algorithm. As a test data set, 20 positive and 20 negative CT scans were separately collected from the training and validation datasets. To evaluate the performance of the established artificial intelligence (AI) model, sensitivity, precision, F1-score, and specificity were calculated. The clinical utility of our AI model was also evaluated through observer studies involving six orthopaedic surgeons and six radiologists. Our AI model showed a sensitivity, precision, and F1-score of 0.78, 0.68, and 0.72 (per slice) and 0.75, 0.36, and 0.48 (per lesion), respectively. The observer studies revealed that our AI model had comparable sensitivity to orthopaedic or radiology experts and improved the sensitivity and F1-score of residents. We developed a novel DL-based AI model for detecting osteolytic bone metastases in the thoracolumbar spine. Although further improvement in accuracy is needed, the current AI model may be applied to current clinical practice. Level III.

Machine learning detects symptomatic patients with carotid plaques based on 6-type calcium configuration classification on CT angiography.

While the link between carotid plaque composition and cerebrovascular vascular (CVE) events is recognized, the role of calcium configuration remains unclear. This study aimed to develop and validate a CT angiography (CTA)-based machine learning (ML) model that uses carotid plaques 6-type calcium grading, and clinical parameters to identify CVE patients with bilateral plaques. We conducted a multicenter, retrospective diagnostic study (March 2013-May 2020) approved by the institutional review board. We included adults (18 +) with bilateral carotid artery plaques, symptomatic patients having recently experienced a carotid territory ischemic event, and asymptomatic patients either after 3months from symptom onset or with no such event. Four ML models (clinical factors, calcium configurations, and both with and without plaque grading [ML-All-G and ML-All-NG]) and logistic regression on all variables identified symptomatic patients. Internal validation assessed discrimination and calibration. External validation was also performed, and identified important variables and causes of misclassifications. We included 790 patients (median age 72, IQR [61-80], 42% male, 64% symptomatic) for training and internal validation, and 159 patients (age 68 [63-76], 36% male, 39% symptomatic) for external testing. The ML-All-G model achieved an area-under-ROC curve of 0.71 (95% CI 0.58-0.78; p < .001) and sensitivity 80% (79-81). Performance was comparable on external testing. Calcified plaque, especially the positive rim sign on the right artery in older and hyperlipidemic patients, had a major impact on identifying symptomatic patients. The developed model can identify symptomatic patients using plaques calcium configuration data and clinical information with reasonable diagnostic accuracy. The analysis of the type of calcium configuration in carotid plaques into 6 classes, combined with clinical variables, allows for an effective identification of symptomatic patients. • While the association between carotid plaques composition and cerebrovascular events is recognized, the role of calcium configuration remains unclear. • Machine learning of 6-type plaque grading can identify symptomatic patients. Calcified plaques on the right artery, advanced age, and hyperlipidemia were the most important predictors. • Fast acquisition of CTA enables rapid grading of plaques upon the patient's arrival at the hospital, which streamlines the diagnosis of symptoms using ML.

Prediction of Parkinson's disease by transcranial sonography-based deep learning.

Transcranial sonography has been used as a valid neuroimaging tool to diagnose Parkinson's disease (PD). This study aimed to develop a modified transcranial sonography (TCS) technique based on a deep convolutional neural network (DCNN) model to predict Parkinson's disease. This retrospective diagnostic study was conducted using 1529 transcranial sonography images collected from 854 patients with PD and 775 normal controls admitted to the Second Affiliated Hospital of Soochow University (Suzhou, Jiangsu, China) between September 2019 and May 2022. The data set was divided into training cohorts (570 PD patients and 541 normal controls), and the validation set (184 PD patients and 234 normal controls). Using these datasets, we developed four different DCNN models (ResNet18, ResNet50, ResNet152, and DenseNet121). We then assessed their diagnostic performance, including the area under the receiver operating characteristic (AUROC) curve, specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV), and F1 score and compared with traditional diagnostic criteria. Among the 1529 TCS images, 570 PD patients and 541 normal controls from 4 of 6 sonographers of the TCS team were selected as the training cohort, and 184 PD patients and 234 normal controls from the other 2 sonographers were chosen as the validation cohort. There were no sex and age differences between PD patients and normal control subjects in the training and validation cohorts (P values > 0.05). All DCNN models achieved good performance in distinguishing PD patients from normal control subjects on the validation datasets, with diagnostic AUROCs and accuracy of 0.949 (95% CI 0.925, 0.965) and 86.60 for the RestNet18 model, 0.949 (95% CI 0.929, 0.971) and 87.56 for ResNet50, 0.945 (95% CI 0.931, 0.969) and 88.04 for ResNet152, 0.953 (95% CI 0.935, 0.971) and 87.80 for DenseNet121, respectively. On the other hand, the diagnostic accuracy of the traditional diagnostic method was 82.30. The accuracy of all DCNN models was higher than that of traditional diagnostic method. Moreover, the 5k-fold cross-validation results in train datasets showed that these DCNN models are robust. The developed transcranial sonography-based DCNN models performed better than traditional diagnostic criteria, thus improving the sonographer's accuracy in diagnosing PD.

Limited clinical validity of univariate resting-state EEG markers for classifying seizure disorders.

Differentiating between epilepsy and psychogenic non-epileptic seizures presents a considerable challenge in clinical practice, resulting in frequent misdiagnosis, unnecessary treatment and long diagnostic delays. Quantitative markers extracted from resting-state EEG may reveal subtle neurophysiological differences that are diagnostically relevant. Two observational, retrospective diagnostic accuracy studies were performed to test the clinical validity of univariate resting-state EEG markers for the differential diagnosis of epilepsy and psychogenic non-epileptic seizures. Clinical EEG data were collected for 179 quasi-consecutive patients (age > 18) with a suspected diagnosis of epilepsy or psychogenic non-epileptic seizures who were medication-naïve at the time of EEG; 148 age- and gender-matched patients subsequently received a diagnosis from specialist clinicians and were included in the analyses. Study 1 is a hypothesis-driven study testing the ability of theta power and peak alpha frequency to classify people with epilepsy and people with psychogenic non-epileptic seizures, with an advanced machine learning pipeline. The next study (Study 2) is data-driven; a high number of quantitative EEG features are extracted and a similar machine learning approach as Study 1 assesses whether previously unexplored univariate EEG measures show promise as diagnostic markers. The results of Study 1 suggest that EEG markers that were previously identified as promising diagnostic indicators (i.e. theta power and peak alpha frequency) have limited clinical validity for the classification of epilepsy and psychogenic non-epileptic seizures (mean accuracy: 48%). The results of Study 2 indicate that identifying univariate markers that show good correlation with a categorical diagnostic label is challenging (mean accuracy: 45-60%). This is due to a considerable overlap in neurophysiological features between the diagnostic classes considered in this study, and to the presence of more dominant EEG dynamics such as alterations due to temporal proximity to epileptiform discharges. Markers that were identified in the context of previous epilepsy research using visually normal resting-state EEG were found to have limited clinical validity for the classification task of distinguishing between people with epilepsy and people with psychogenic non-epileptic seizures. A search for alternative diagnostic markers uncovered the challenges involved and generated recommendations for further research.

Serum Anti-Mullerian Hormone Levels Might Indicate Premenopausal Endometrial Lesions.

Endometrial proliferative lesions (EPL) usually refer to endometrial hyperplasia (EH) and endometrial cancer (EC). Among patients with premenopausal EPL who wish to preserve their fertility, only those with EH and early-stage EC have the possibility to undergo fertility preservation therapy. However, there is currently a lack of specific and reliable screening criteria and models for identifying these patients. This study utilized a retrospective diagnostic study design. The training set included medical record information that met the criteria between August 2017 and October 2022, while the validation set consisted of medical record information that met the criteria from November 2022 to May 2023. The endometrial pathological test served as the gold standard. The serum anti-Mullerian hormone (AMH) level before endometrial sampling and a regression model were employed to predict EPL. The study included a total of 1209 patients with PCOS (1119 in the control group and 90 in the endometrial proliferative lesion group) and 5366 women without PCOS (5249 in the control group and 117 in the proliferative lesion group). In the case of PCOS patients aged 20-39 years, the most effective screening threshold for AMH was found to be a serum AMH level of ≤5.39 ng/mL. The model used for this group was logit(p) = -2.562 - 0.430 × AMH + 0.127 × BMI + 1.512 × hypertension + 0.956 × diabetes -1.145 × regular menstruation. On the other hand, for non-PCOS women aged 20-39 years, the optimal screening threshold for AMH was determined to be a serum AMH value of ≤2.18 ng/mL. The model used for this group was logit(p) = -3.778 - 0.823 × AMH + 0.176 × BMI + 2.660 × diabetes -1.527 × regular menstruation -1.117 × dysmenorrhea. It is important to note that all of these findings have successfully passed internal verification. For PCOS and non-PCOS women aged 20-39 years, the serum AMH test and related multiple regression models were obtained for the warning of EPL.

A deep learning framework for intracranial aneurysms automatic segmentation and detection on magnetic resonance T1 images.

To design a deep learning-based framework for automatic segmentation and detection of intracranial aneurysms (IAs) on magnetic resonance T1 images and test the robustness and performance of framework. A retrospective diagnostic study was conducted based on 159 IAs from 136 patients who underwent the T1 images. Among them, 127 cases were randomly selected for training and validation, and 32 cases were used to assess the accuracy and consistency of our algorithm. We developed and assembled three convolutional neural networks for the segmentation and detection of IAs. The segmentation and detection performance of the model were compared with the ground truth, and various metrics were calculated at the voxel level, IAs level, and patient level to show the performance of our framework. Our assembled model achieved overall Dice, voxel-level sensitivity, specificity, balanced accuracy, and F1 score of 0.802, 0.874, 0.9998, 0.937, and 0.802, respectively. A coincidence greater than 0.7 between the aneurysms predicted by the model and the ground truth was considered as a true positive. For IAs detection, the sensitivity reached 90.63% with 0.58 false positives per case. The volume of IAs segmented by our model showed a high agreement and consistency with the volume of IAs labeled by experts. The deep learning framework is achievable and robust for IAs segmentation and detection. Our model offers more clinical application opportunities compared to digital subtraction angiography (DSA)-based, CTA-based, and MRA-based methods. Our deep learning framework effectively detects and segments intracranial aneurysms using clinical routine T1 sequences, showing remarkable effectiveness and offering great potential for improving the detection of latent intracranial aneurysms and enabling early identification. •There is no segmentation method based on clinical routine T1 images. Our study shows that the proper deep learning framework can effectively localize the intracranial aneurysms. •The T1-based segmentation and detection method is more universal than other angiography-based detection methods, which can potentially reduce missed diagnoses caused by the absence of angiography images. •The deep learning framework is robust and has the potential to be applied in a clinical setting.

THU588 Sparing Confirmatory Testing In Primary Aldosteronism: The Combination Of Renin, Aldosterone And Potassium Levels

Abstract Disclosure: A.C. Ong Lopez: None. D.M. Dimayuga: None. L.E. Tiu: None. O.C. Dampil: None. E. Mendoza: None. M.L. Villa: None. A.M. Macabuag-Oliva: None. Background and Objective: The diagnosis of primary aldosteronism (PA) is comprehensive which includes case-detecting testing, case confirmation followed by subtype classification. In certain instances, one may not proceed with further confirmatory testing in the setting of spontaneous hypokalemia, suppressed renin, and PAC &amp;gt; 15 ng/dL. However, this quality of evidence is very low. This study sought to determine the optimal laboratory combination patterns that can spare confirmatory testing for primary aldosteronism by evaluating the diagnostic performances of the various combination settings and correlating to the recent evidence and clinical practice guidelines. Methodology: This is a multi-center, retrospective diagnostic accuracy cohort-selected cross-sectional study with a sampling total of 116 participants aged 18 years and above who underwent saline infusion test between January 2010 to March 2023. The main outcome measures comprise the diagnostic performances of the different index test combinations (baseline PAC, baseline PRA and presence of spontaneous hypokalemia) which include the sensitivity, specificity, negative predictive value, positive predictive value, positive likelihood ratio, negative likelihood ratio, and diagnostic accuracy. Data analysis was performed using SPSS 29.0.0.0 &amp; MedCalc 20.218. Results: Of the 116 patients who underwent saline infusion test, 76 (65.52%) were diagnosed with PA. A PAC of &amp;gt; 25 ng/dL plus PRA &amp;lt; 1.0 ng/dL/hr with spontaneous hypokalemia showed the highest specificity at 100% (95% CI 89.11%, 100.00%) and positive predictive value at 100% (95% CI 82.35, 100.00%). The optimal combination criteria with the best balance of the diagnostic performance indices were determined to be a PAC of &amp;gt; 20 ng/dL plus PRA &amp;lt; 0.6 ng/dL/hr, and presence of spontaneous hypokalemia. It has a sensitivity of 34.72% (95% CI 23.88% , 46.86%), specificity of 93.75% (95% CI 79.19%, 99.23%), positive predictive value (PPV) of 92.59% (95% CI 75.71%, 99.09%), negative predictive value (NPV) of 38.96% (95% CI 28.05%, 50.75%), positive likelihood ratio (LR+) of 5.56 (95% CI 1.40, 22.06), negative likelihood ratio (LR-) of 0.70 (95% CI 0.58, 0.84), and a diagnostic accuracy of 52.88% (95% CI 42.85%, 62.75%). Conclusion: Hypertensive patients with spontaneous hypokalemia who had initial screening test results of PAC &amp;gt; 20 ng/dL and PRA &amp;lt; 0.6 ng/ml/hr can be diagnosed as overt primary aldosteronism and may not proceed with dynamic confirmatory testing. This is in line with the 2016 Endocrine Society Clinical Practice Guideline recommendation. Presentation: Thursday, June 15, 2023

Generative Artificial Intelligence for Chest Radiograph Interpretation in the Emergency Department

Multimodal generative artificial intelligence (AI) methodologies have the potential to optimize emergency department care by producing draft radiology reports from input images. To evaluate the accuracy and quality of AI-generated chest radiograph interpretations in the emergency department setting. This was a retrospective diagnostic study of 500 randomly sampled emergency department encounters at a tertiary care institution including chest radiographs interpreted by both a teleradiology service and on-site attending radiologist from January 2022 to January 2023. An AI interpretation was generated for each radiograph. The 3 radiograph interpretations were each rated in duplicate by 6 emergency department physicians using a 5-point Likert scale. The primary outcome was any difference in Likert scores between radiologist, AI, and teleradiology reports, using a cumulative link mixed model. Secondary analyses compared the probability of each report type containing no clinically significant discrepancy with further stratification by finding presence, using a logistic mixed-effects model. Physician comments on discrepancies were recorded. A total of 500 ED studies were included from 500 unique patients with a mean (SD) age of 53.3 (21.6) years; 282 patients (56.4%) were female. There was a significant association of report type with ratings, with post hoc tests revealing significantly greater scores for AI (mean [SE] score, 3.22 [0.34]; P < .001) and radiologist (mean [SE] score, 3.34 [0.34]; P < .001) reports compared with teleradiology (mean [SE] score, 2.74 [0.34]) reports. AI and radiologist reports were not significantly different. On secondary analysis, there was no difference in the probability of no clinically significant discrepancy between the 3 report types. Further stratification of reports by presence of cardiomegaly, pulmonary edema, pleural effusion, infiltrate, pneumothorax, and support devices also yielded no difference in the probability of containing no clinically significant discrepancy between the report types. In a representative sample of emergency department chest radiographs, results suggest that the generative AI model produced reports of similar clinical accuracy and textual quality to radiologist reports while providing higher textual quality than teleradiologist reports. Implementation of the model in the clinical workflow could enable timely alerts to life-threatening pathology while aiding imaging interpretation and documentation.

Is symmetric dimethylarginine superior to creatinine for assessing glomerular filtration rate for cats with kidney disease?

PICO Question Is symmetric dimethylarginine (SDMA) superior to creatinine for assessing glomerular filtration rate (GFR) for cats with chronic kidney disease? Clinical bottom line Category Diagnosis. Number and type of studies reviewed: The literature searches returned four relevant papers that answered this PICO. Three of the studies were retrospective diagnostic test studies, and one was a randomised, controlled, unblinded study. Strength of evidence Moderate. Outcomes reported Three out of the four studies analysed found that there was no significant difference between the correlation between symmetric dimethylarginine (SDMA) and glomerular filtration rate (GFR) and creatinine (sCr) and GFR, whilst the other identified a stronger correlation between SDMA and GFR. Two papers also evaluated the sensitivity and specificity of SDMA compared to creatinine. One paper found a similar sensitivity for both biomarkers and a significantly higher specificity for creatinine using the upper reference interval of 18 µg/dL for SDMA and 155.6 µmol/L (1.76 mg/dL) sCr, whilst the other found SDMA to have a superior sensitivity and both biomarkers to have a similar specificity using the upper reference interval of 14 µg/dL for SDMA and 185.64 µmol/L (2.1 mg/dL) for sCr. Conclusions In view of the strength of evidence we conclude that the majority of the studies analysed do not demonstrate that SDMA is superior to creatinine for the assessment of GFR in cats with chronic kidney disease. More research is needed with larger sample sizes to investigate this further. Standardisation of the upper reference intervals across studies for creatinine and SDMA would be advantageous for comparison of sensitivity and specificity in future studies. How to apply this evidence in practice The application of evidence into practice should take into account multiple factors, not limited to: individual clinical expertise, patient’s circumstances and owners’ values, country, location or clinic where you work, the individual case in front of you, the availability of therapies and resources. Knowledge Summaries are a resource to help reinforce or inform decision making. They do not override the responsibility or judgement of the practitioner to do what is best for the animal in their care.

Pre-Operative Score Development: Predicting Difficulty in Elective Laparoscopic Cholecystectomy

Objective: To develop and internally validate a pre-operative scoring system to predict difficulty in elective laparoscopic cholecystectomy (LC).Material and Methods: A retrospective diagnostic prediction study was conducted. Patients undergoing elective LC; from September 2016 to January 2023, at Hatyai Hospital in Southern Thailand were included. Patients were categorized by difficultly of LC according to the Nassar scale (grades 1-2 as non-difficult LC and grades 3-4 as difficult LC). Pre-operative data were compared between both groups; utilizing multivariable logistic regression. Internal validation was performed via the bootstrapping procedure.Results: In total, three hundred and eighteen patients were categorized into either; difficult LC 121 patients or non-difficult LC 197 patients. From this, 7 variables obtained from the multivariable logistic reduced model (male, cirrhosis, history of ERCP, ASA III, gallbladder wall ≥4 mm, dilated gallbladder, contracted gallbladder) were developed as a pre-operative score. The scoring (range: 0.0-16.6) was classified into 3 groups for clinical practicability. The positive predictive values (PPV) were 18.1 for low-risk, 38.0 for moderate-risk, and 76.0 for high-risk. Internal validation, via bootstrap technique, showed a C-statistic value of 0.76, and bootstrap shrinkage was 0.995. The prediction ability (AuROC) of the pre-operative score was 0.76.Conclusion: The developed of a pre-operative score had a good predictive performance, with fewer predictors for prediction difficulty of elective LC that can assist surgeons in surgical management selection.

Preoperative differentiation of mediastinum and retroperitoneum ganglioneuroma from schwannoma with clinical data and enhanced CT: developing a multivariable prediction model

To develop a multivariable prediction model for preoperative differentiation of ganglioneuroma (GN) from schwannoma in mediastinum and retroperitoneum based on clinical data and enhanced computed tomography (CT). This was a retrospective diagnostic study. Patients diagnosed with mediastinum or retroperitoneal GN or schwannoma at Zhongshan Hospital between July 2006 and March 2022 were divided into a training cohort and a validation cohort at a ratio of 7:3. Clinical information and CT features were collected. Histopathology was the reference standard for diagnosis. The model was developed using binary logistic regression. The predictive performance of the model was assessed using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). A total of 105 patients (47 men and 58 women; mean age of 41±15 years) were enrolled. There were significant differences in symptoms (p=0.006), location (p=0.008), ratio of the craniocaudal diameter (CC) to the major axis on axial images (CC/M; p=0.025), ratio of the CC to the diameter on axial images (CC/D; p<0.001), density homogeneity (p=0.001), enhancement homogeneity (p<0.001), enhancement degree (p<0.001), venous phase CT attenuation value (V; p=0.011), and blood vessels changes (p=0.045) between GN and schwannoma. The area under the ROC curve (AUC) and accuracy in the validation cohort were 0.841 (95% confidence interval [CI] 0.672, 1.000) and 0.839 (95% CI: 0.674, 0.929), respectively. Calibration curves and DCA showed that the model was beneficial for patients. The multivariable prediction model exhibited good predictive performance and may facilitate preoperative planning.