-
COV-MobNets: a mobile networks ensemble model for diagnosis of COVID-19 based on chest X-ray images.BMC Medical Imaging Jun 2023The medical profession is facing an excessive workload, which has led to the development of various Computer-Aided Diagnosis (CAD) systems as well as Mobile-Aid...
BACKGROUND
The medical profession is facing an excessive workload, which has led to the development of various Computer-Aided Diagnosis (CAD) systems as well as Mobile-Aid Diagnosis (MAD) systems. These technologies enhance the speed and accuracy of diagnoses, particularly in areas with limited resources or remote regions during the pandemic. The primary purpose of this research is to predict and diagnose COVID-19 infection from chest X-ray images by developing a mobile-friendly deep learning framework, which has the potential for deployment in portable devices such as mobile or tablet, especially in situations where the workload of radiology specialists may be high. Moreover, this could improve the accuracy and transparency of population screening to assist radiologists during the pandemic.
METHODS
In this study, the Mobile Networks ensemble model called COV-MobNets is proposed to classify positive COVID-19 X-ray images from negative ones and can have an assistant role in diagnosing COVID-19. The proposed model is an ensemble model, combining two lightweight and mobile-friendly models: MobileViT based on transformer structure and MobileNetV3 based on Convolutional Neural Network. Hence, COV-MobNets can extract the features of chest X-ray images in two different methods to achieve better and more accurate results. In addition, data augmentation techniques were applied to the dataset to avoid overfitting during the training process. The COVIDx-CXR-3 benchmark dataset was used for training and evaluation.
RESULTS
The classification accuracy of the improved MobileViT and MobileNetV3 models on the test set has reached 92.5% and 97%, respectively, while the accuracy of the proposed model (COV-MobNets) has reached 97.75%. The sensitivity and specificity of the proposed model have also reached 98.5% and 97%, respectively. Experimental comparison proves the result is more accurate and balanced than other methods.
CONCLUSION
The proposed method can distinguish between positive and negative COVID-19 cases more accurately and quickly. The proposed method proves that utilizing two automatic feature extractors with different structures as an overall framework of COVID-19 diagnosis can lead to improved performance, enhanced accuracy, and better generalization to new or unseen data. As a result, the proposed framework in this study can be used as an effective method for computer-aided diagnosis and mobile-aided diagnosis of COVID-19. The code is available publicly for open access at https://github.com/MAmirEshraghi/COV-MobNets .
Topics: Humans; COVID-19; COVID-19 Testing; X-Rays; SARS-CoV-2; Deep Learning
PubMed: 37322450
DOI: 10.1186/s12880-023-01039-w -
CytoJournal 2013Pancreatic neuroendocrine tumors (PNETs) are rare tumors of the pancreas, which are increasingly diagnosed by endoscopic ultrasound-guided fine needle aspiration...
BACKGROUND
Pancreatic neuroendocrine tumors (PNETs) are rare tumors of the pancreas, which are increasingly diagnosed by endoscopic ultrasound-guided fine needle aspiration (EUS-FNA). In this retrospective study, we assessed the performance of EUS-FNA in diagnosing PNETs.
MATERIALS AND METHODS
We identified 48 cases of surgically resected PNETs in which pre-operative EUS-FNA was performed. The clinical features, cytological diagnoses, and surgical follow-up were retrospectively reviewed. The diagnostic performance of EUS-FNA was analyzed as compared to the diagnosis in the follow-up. The cases with discrepancies between cytological diagnosis and surgical follow-up were analyzed and diagnostic pitfalls in discrepant cases were discussed.
RESULTS
The patients were 20 male and 28 female with ages ranging from 15 years to 81 years (mean 57 years). The tumors were solid and cystic in 41 and 7 cases, respectively, with sizes ranging from 0.5 cm to 11 cm (mean 2.7 cm). Based on cytomorphologic features and adjunct immunocytochemistry results, when performed, 38 patients (79%) were diagnosed with PNET, while a diagnosis of suspicious for PNET or a diagnosis of neoplasm with differential diagnosis including PNET was rendered in the 3 patients (6%). One case was diagnosed as mucinous cystic neoplasm (2%). The remaining 6 patients (13%) had non-diagnostic, negative or atypical diagnosis.
CONCLUSIONS
Our data demonstrated that EUS-FNA has a relatively high sensitivity for diagnosing PNETs. Lack of additional materials for immunocytochemical studies could lead to a less definite diagnosis. Non-diagnostic or false negative FNA diagnosis can be seen in a limited number of cases, especially in those small sized tumors.
PubMed: 23858320
DOI: 10.4103/1742-6413.112648 -
American Family Physician Aug 2001Early and accurate detection of eye disorders in children can present a challenge for family physicians. Visual acuity screening, preferably performed before four years... (Review)
Review
Early and accurate detection of eye disorders in children can present a challenge for family physicians. Visual acuity screening, preferably performed before four years of age, is essential for diagnosing amblyopia. Cover testing may disclose small-angle or intermittent strabismus. Leukocoria, which is detected with an ophthalmoscope, may indicate retinoblastoma or cataract. Children with glaucoma may have light sensitivity and enlargement of the cornea, and conjunctivitis that does not respond quickly to treatment may reflect more serious ocular inflammation. Children with serious eye injuries often present to the primary care physician. Nystagmus and many systemic conditions are associated with specific eye findings.
Topics: Amblyopia; Child; Diagnosis, Differential; Diagnostic Errors; Eye Diseases; Eye Injuries; Glaucoma; Humans; Inflammation; Strabismus; Vision Disorders; Vision Screening
PubMed: 11529261
DOI: No ID Found -
BMC Musculoskeletal Disorders Aug 2013Previous reviews of the diagnostic performances of physical tests of the hip in orthopedics have drawn limited conclusions because of the low to moderate quality of... (Review)
Review
BACKGROUND
Previous reviews of the diagnostic performances of physical tests of the hip in orthopedics have drawn limited conclusions because of the low to moderate quality of primary studies published in the literature. This systematic review aims to build on these reviews by assessing a broad range of hip pathologies, and employing a more selective approach to the inclusion of studies in order to accurately gauge diagnostic performance for the purposes of making recommendations for clinical practice and future research. It specifically identifies tests which demonstrate strong and moderate diagnostic performance.
METHODS
A systematic search of Medline, Embase, Embase Classic and CINAHL was conducted to identify studies of hip tests. Our selection criteria included an analysis of internal and external validity. We reported diagnostic performance in terms of sensitivity, specificity, predictive values and likelihood ratios. Likelihood ratios were used to identify tests with strong and moderate diagnostic utility.
RESULTS
Only a small proportion of tests reported in the literature have been assessed in methodologically valid primary studies. 16 studies were included in our review, producing 56 independent test-pathology combinations. Two tests demonstrated strong clinical utility, the patellar-pubic percussion test for excluding radiologically occult hip fractures (negative LR 0.05, 95% Confidence Interval [CI] 0.03-0.08) and the hip abduction sign for diagnosing sarcoglycanopathies in patients with known muscular dystrophies (positive LR 34.29, 95% CI 10.97-122.30). Fifteen tests demonstrated moderate diagnostic utility for diagnosing and/or excluding hip fractures, symptomatic osteoarthritis and loosening of components post-total hip arthroplasty.
CONCLUSIONS
We have identified a number of tests demonstrating strong and moderate diagnostic performance. These findings must be viewed with caution as there are concerns over the methodological quality of the primary studies from which we have extracted our data. Future studies should recruit larger, representative populations and allow for the construction of complete 2×2 contingency tables.
Topics: Biomechanical Phenomena; Hip Joint; Humans; Joint Diseases; Orthopedics; Physical Examination; Predictive Value of Tests; Prognosis; Reproducibility of Results; Severity of Illness Index
PubMed: 23987589
DOI: 10.1186/1471-2474-14-257 -
Malaria Journal Jun 2012Accurate diagnosis and prompt treatment of pregnancy-associated malaria (PAM) are key aspects in averting adverse pregnancy outcomes. Microscopy is the gold standard in... (Comparative Study)
Comparative Study
BACKGROUND
Accurate diagnosis and prompt treatment of pregnancy-associated malaria (PAM) are key aspects in averting adverse pregnancy outcomes. Microscopy is the gold standard in malaria diagnosis, but it has limited detection and availability. When used appropriately, rapid diagnostic tests (RDTs) could be an ideal diagnostic complement to microscopy, due to their ease of use and adequate sensitivity in detecting even sub-microscopic infections. Polymerase chain reaction (PCR) is even more sensitive, but it is mainly used for research purposes. The accuracy and reliability of RDTs in diagnosing PAM was evaluated using microscopy and PCR.
METHODS
A cohort of pregnant women in north-eastern Tanzania was followed throughout pregnancy for detection of plasmodial infection using venous and placental blood samples evaluated by histidine rich protein 2 (HRP-2) and parasite lactate dehydrogenase (pLDH) based RDTs (Parascreen™) or HRP-2 only (Paracheck Pf® and ParaHIT®f), microscopy and nested Plasmodium species diagnostic PCR.
RESULTS
From a cohort of 924 pregnant women who completed the follow up, complete RDT and microscopy data was available for 5,555 blood samples and of these 442 samples were analysed by PCR. Of the 5,555 blood samples, 49 ((proportion and 95% confidence interval) 0.9% [0.7 -1.1]) samples were positive by microscopy and 91 (1.6% [1.3-2.0]) by RDT. Forty-six (50.5% [40.5 - 60.6]) and 45 (49.5% [39.4 - 59.5]) of the RDT positive samples were positive and negative by microscopy, respectively, whereas nineteen (42.2% [29.0 - 56.7]) of the microscopy negative, but RDT positive, samples were positive by PCR. Three (0.05% [0.02 - 0.2]) samples were positive by microscopy but negative by RDT. 351 of the 5,461 samples negative by both RDT and microscopy were tested by PCR and found negative. There was no statistically significant difference between the performances of the different RDTs.
CONCLUSIONS
Microscopy underestimated the real burden of malaria during pregnancy and RDTs performed better than microscopy in diagnosing PAM. In areas where intermittent preventive treatment during pregnancy may be abandoned due to low and decreasing malaria risk and instead replaced with active case management, screening with RDT is likely to identify most infections in pregnant women and out-performs microscopy as a diagnostic tool.
Topics: Adolescent; Adult; Antigens, Protozoan; Blood; Clinical Laboratory Techniques; Cohort Studies; Diagnostic Tests, Routine; Female; Humans; Malaria; Parasitology; Pregnancy; Pregnancy Complications, Infectious; Prospective Studies; Sensitivity and Specificity; Tanzania; Young Adult
PubMed: 22720788
DOI: 10.1186/1475-2875-11-211 -
Medicine Nov 2020To compare distant visual acuity (DVA) and near visual acuity (NVA) in amblyopia and evaluate if NVA can be used to diagnose amblyopia.A retrospective study was...
To compare distant visual acuity (DVA) and near visual acuity (NVA) in amblyopia and evaluate if NVA can be used to diagnose amblyopia.A retrospective study was performed on 73 patients diagnosed with amblyopia based on DVA, by measuring their NVA and comparing the DVA and NVA. The NVA was measured by Snellen chart at 30 cm and the DVA was measured by Dr Hahn vision test chart at 5m. The patients' age, type of amblyopia, spherical equivalent, the difference between spherical equivalent and the fellow eye spherical equivalent spherical, and prism diopter (PD) were evaluated and their relationship with the difference between the DVA and NVA was analyzed.The NVA was significantly better than the DVA in amblyopia (P = .000). The difference between the DVA and NVA was not significantly related to the type of amblyopia (P = .600) or the patients' age(P = .351). Also, the difference between the DVA and NVA was not significantly affected by the spherical equivalent (P = .425) or the difference between spherical equivalent and the fellow eye spherical equivalent (P = .212) in anisometropia amblyopia, and also not by the PD (P = .882) in strabismus amblyopia.In amblyopes, the NVA was better than the DVA before amblyopia treatment. The difference between the DVA and NVA was not affected by age, type of amblyopia, spherical equivalent, the difference between spherical equivalent and the fellow eye spherical equivalent spherical, or PD. Therefore, it should be taken into consideration that NVA could underestimate the severity of amblyopia and affect the accuracy at diagnosing amblyopia.
Topics: Amblyopia; Anisometropia; Child; Diagnostic Techniques, Ophthalmological; Dimensional Measurement Accuracy; Early Diagnosis; Female; Humans; Male; Severity of Illness Index; Strabismus; Visual Acuity
PubMed: 33181679
DOI: 10.1097/MD.0000000000023111 -
Respirology (Carlton, Vic.) Apr 2017Tissue diagnosis of peripheral pulmonary lesions (PPLs) can be challenging. In the past, flexible bronchoscopy was commonly performed for this purpose but its diagnostic... (Meta-Analysis)
Meta-Analysis Review
Tissue diagnosis of peripheral pulmonary lesions (PPLs) can be challenging. In the past, flexible bronchoscopy was commonly performed for this purpose but its diagnostic yield is suboptimal. This has led to the development of new bronchoscopic modalities such as radial endobronchial ultrasound (R-EBUS), electromagnetic navigation bronchoscopy (ENB) and virtual bronchoscopy (VB). We performed this meta-analysis using data from previously published R-EBUS studies, to determine its diagnostic yield and other performance characteristics. Ovid MEDLINE and PubMed databases were searched for R-EBUS studies in September 2016. Diagnostic yield was calculated by dividing the number of successful diagnoses by the total number of lesions. Meta-analysis was performed using MedCalc (Version 16.8). Inverse variance weighting was used to aggregate diagnostic yield proportions across studies. Publication bias was assessed using funnel plot and Duval and Tweedie's test. 57 studies with a total of 7872 lesions were included in the meta-analysis. These were published between October 2002 and August 2016. Overall weighted diagnostic yield for R-EBUS was 70.6% (95% CI: 68-73.1%). The diagnostic yield was significantly higher for lesions >2 cm in size, malignant in nature and those associated with a bronchus sign on computerized tomography (CT) scan. Diagnostic yield was also higher when R-EBUS probe was within the lesion as opposed to being adjacent to it. Overall complication rate was 2.8%. This is the largest meta-analysis performed to date, assessing the performance of R-EBUS for diagnosing PPLs. R-EBUS has a high diagnostic yield (70.6%) with a very low complication rate.
Topics: Bronchi; Bronchoscopy; Endosonography; Humans; Lung Neoplasms; Tomography, X-Ray Computed; Tumor Burden
PubMed: 28177181
DOI: 10.1111/resp.12980 -
Journal of Clinical Sleep Medicine :... Feb 2022To compare type 2 polysomnography (T2PSG) to the gold standard type 1 in-laboratory polysomnography (T1PSG) for diagnosing obstructive sleep apnea (OSA) in children;...
Comparison of home ambulatory type 2 polysomnography with a portable monitoring device and in-laboratory type 1 polysomnography for the diagnosis of obstructive sleep apnea in children.
STUDY OBJECTIVES
To compare type 2 polysomnography (T2PSG) to the gold standard type 1 in-laboratory polysomnography (T1PSG) for diagnosing obstructive sleep apnea (OSA) in children; validate home T2PSG in children with suspected OSA.
METHODS
Eighty-one participants (ages 6-18) with suspected OSA had simultaneous T1PSG and T2PSG in the sleep laboratory, 47 participants (ages 5-16) had T1PSG in the sleep laboratory and T2PSG performed at home. Sleep scientists staged and scored polysomnography data, and pediatric sleep physicians assigned a diagnosis of normal or OSA. Participant demographics, polysomnography variables, and diagnoses were compared using chi-square and Fisher's exact tests for nominal variables, test for continuous variables and Cohen's kappa to assess concordance.
RESULTS
Acceptable recordings were obtained for every home T2PSG. When T1PSG and T2PSG were simultaneous, correlation between the number of arousals, respiratory disturbance index, and sleep stages was excellent. T2PSG at home demonstrated less stage 2 sleep, more rapid eye movement sleep, and higher sleep efficiency. Comparison of home T2PSG to T1PSG for diagnosing OSA showed a false-positive rate of 6.6% and false-negative rate of 3% for those performed at home.
CONCLUSIONS
T2PSG in the home is feasible with excellent concordance with T1PSG for the purposes of diagnosing OSA in children aged 5-18 years. Home T2PSG may be more representative of a "normal" night for children and could benefit those suspected of having OSA by reducing waiting times for laboratory PSG, improving access to PSG and possibly reducing costs of investigating and treating OSA.
CITATION
Withers A, Maul J, Rosenheim E, O'Donnell A, Wilson A, Stick S. Comparison of home ambulatory type 2 polysomnography with a portable monitoring device and in-laboratory type 1 polysomnography for the diagnosis of obstructive sleep apnea in children. . 2022;18(2):393-402.
Topics: Adolescent; Child; Child, Preschool; Humans; Laboratories; Monitoring, Ambulatory; Polysomnography; Sleep; Sleep Apnea, Obstructive
PubMed: 34323688
DOI: 10.5664/jcsm.9576 -
Cancer Cytopathology Feb 2022Despite widespread clinical use, lymph node fine-needle aspiration cytology (LN-FNAC) lacks universal acceptance for definitively diagnosing lymphomas. This is likely...
BACKGROUND
Despite widespread clinical use, lymph node fine-needle aspiration cytology (LN-FNAC) lacks universal acceptance for definitively diagnosing lymphomas. This is likely due to reports of lower diagnostic performance, inconsistent terminology use in cytopathology diagnostic reports, and only limited data on the clinical implications of LN-FNAC diagnoses. Recently, a uniform LN-FNAC cytopathological diagnostic reporting system was proposed (the Sydney System). This study evaluated LN-FNAC diagnostic performance and risks of malignancy associated with the proposed diagnostic categories.
METHODS
LN-FNAC specimens obtained in 2018-2019, with and without concurrent core biopsy, to evaluate for suspected lymphoma were analyzed (n = 349). LN-FNAC diagnoses were compared with final diagnoses obtained via subsequent tissue biopsy and/or clinical assessment.
RESULTS
The mean patient age was 57.6 years, and 41% were female. LN-FNAC was the initial diagnostic test in 223 (63.9%), and it was used to evaluate for recurrence in 126 (36.1%). LN-FNAC diagnosed 202 hematological malignancies (57.9%), 23 nonhematological malignancies (6.6%), and 124 reactive processes (35.5%). Subsequent tissue biopsy was performed in 42 (12%). The risks of malignancy per diagnostic category were as follows: inadequate, 58.3%; benign, 6.4%; atypical, 69.2%; suspicious, 96.7%; and malignant, 99.3%. LN-FNAC demonstrated up to 96.3% sensitivity, 91.91% specificity, and 87.35% accuracy. Optimal specimen quality and the use of intradepartmental consultation reduced diagnostic error rates in FNA cases without concurrent core biopsy (P = .029 and P = .0002 respectively).
CONCLUSIONS
LN-FNAC is accurate and reliable for the diagnosis of lymphoma. Inadequate LN-FNAC samples should be resampled due to a significant associated risk of lymphoma. The diagnostic performance of LN-FNAC may be improved with good specimen quality and reviews by multiple pathologists. Understanding the risks of malignancy associated with LN-FNAC diagnostic categories will help to guide optimal patient management.
Topics: Biopsy, Fine-Needle; Biopsy, Large-Core Needle; Cytological Techniques; Female; Humans; Lymph Nodes; Male; Middle Aged; Neoplasms; Sensitivity and Specificity
PubMed: 34661975
DOI: 10.1002/cncy.22523 -
PloS One 2022Cancer is among the leading causes of death in the developed world, and lung cancer is the most lethal type. Early detection is crucial for better prognosis, but can be...
PURPOSE
Cancer is among the leading causes of death in the developed world, and lung cancer is the most lethal type. Early detection is crucial for better prognosis, but can be resource intensive to achieve. Automating tasks such as lung tumor localization and segmentation in radiological images can free valuable time for radiologists and other clinical personnel. Convolutional neural networks may be suited for such tasks, but require substantial amounts of labeled data to train. Obtaining labeled data is a challenge, especially in the medical domain.
METHODS
This paper investigates the use of a teacher-student design to utilize datasets with different types of supervision to train an automatic model performing pulmonary tumor segmentation on computed tomography images. The framework consists of two models: the student that performs end-to-end automatic tumor segmentation and the teacher that supplies the student additional pseudo-annotated data during training.
RESULTS
Using only a small proportion of semantically labeled data and a large number of bounding box annotated data, we achieved competitive performance using a teacher-student design. Models trained on larger amounts of semantic annotations did not perform better than those trained on teacher-annotated data. Our model trained on a small number of semantically labeled data achieved a mean dice similarity coefficient of 71.0 on the MSD Lung dataset.
CONCLUSIONS
Our results demonstrate the potential of utilizing teacher-student designs to reduce the annotation load, as less supervised annotation schemes may be performed, without any real degradation in segmentation accuracy.
Topics: Humans; Image Processing, Computer-Assisted; Lung Neoplasms; Neural Networks, Computer; Students; Tomography, X-Ray Computed
PubMed: 35381046
DOI: 10.1371/journal.pone.0266147