-
The Journal of International Medical... Jan 2021To construct deep learning (DL) models to improve the accuracy and efficiency of thyroid disease diagnosis by thyroid scintigraphy.
OBJECTIVE
To construct deep learning (DL) models to improve the accuracy and efficiency of thyroid disease diagnosis by thyroid scintigraphy.
METHODS
We constructed DL models with AlexNet, VGGNet, and ResNet. The models were trained separately with transfer learning. We measured each model's performance with six indicators: recall, precision, negative predictive value (NPV), specificity, accuracy, and F1-score. We also compared the diagnostic performances of first- and third-year nuclear medicine (NM) residents with assistance from the best-performing DL-based model. The Kappa coefficient and average classification time of each model were compared with those of two NM residents.
RESULTS
The recall, precision, NPV, specificity, accuracy, and F1-score of the three models ranged from 73.33% to 97.00%. The Kappa coefficient of all three models was >0.710. All models performed better than the first-year NM resident but not as well as the third-year NM resident in terms of diagnostic ability. However, the ResNet model provided "diagnostic assistance" to the NM residents. The models provided results at speeds 400 to 600 times faster than the NM residents.
CONCLUSION
DL-based models perform well in diagnostic assessment by thyroid scintigraphy. These models may serve as tools for NM residents in the diagnosis of Graves' disease and subacute thyroiditis.
Topics: Deep Learning; Graves Disease; Humans; Radionuclide Imaging; Thyroid Diseases
PubMed: 33445994
DOI: 10.1177/0300060520982842 -
International Journal of Pediatric... Feb 2022This study aims to investigate the inter-rater reliability and agreement of the diagnosis of otitis media with effusion, acute otitis media, and no effusion cases based...
OBJECTIVES
This study aims to investigate the inter-rater reliability and agreement of the diagnosis of otitis media with effusion, acute otitis media, and no effusion cases based on an otoscopy image and in some cases an additional wideband tympanometry measurement of the patient.
METHODS
1409 cases were examined and diagnosed by an otolaryngologist in the clinic, and otoscopy examination and wideband tympanometry (WBT) measurement were conducted. Afterwards, four otolaryngologists (Ear, Nose, and Throat doctors, ENTs), who did not perform the acute examination of the patients, evaluated the otoscopy images and WBT measurements results for diagnosis (acute otitis media, otitis media with effusion, or no effusion). They also specified their diagnostic certainty for each case, and reported whether they used the image, wideband tympanometry, or both, for diagnosis.
RESULTS
All four ENTs agreed on the diagnosis in 57% of the cases, with a pairwise agreement of 74%, and a Light's Kappa of 0.58. There are, however, large differences in agreement and certainty between the three diagnoses. Acute otitis media yields the highest agreement (77% between all four ENTs) and certainty (0.90), while no effusion shows much lower agreement and certainty (34% and 0.58, respectively). There is a positive correlation between certainty and agreement between the ENTs across all cases, and both certainty and agreement increase for cases where a WBT measurement is shown in addition to the otoscopy image.
CONCLUSIONS
The inter-rater reliability between four ENTs was high when diagnosing acute otitis media and lower when diagnosing otitis media with effusion. However, WBT can add valuable information to get closer to the ground-truth diagnosis without myringotomy. Furthermore, the diagnostic certainty increases when the WBT is examined together with the otoscopy image.
Topics: Acoustic Impedance Tests; Humans; Infant; Otitis Media; Otitis Media with Effusion; Otoscopes; Otoscopy; Reproducibility of Results
PubMed: 35033784
DOI: 10.1016/j.ijporl.2021.111034 -
Cancer Cytopathology Jun 2022The American Thyroid Association recommends either repeat fine-needle aspiration biopsy (FNAB) or molecular testing (eg, ThyroSeq) of Bethesda category III (atypia of...
BACKGROUND
The American Thyroid Association recommends either repeat fine-needle aspiration biopsy (FNAB) or molecular testing (eg, ThyroSeq) of Bethesda category III (atypia of undetermined significance/follicular lesion of undetermined significance [AUS/FLUS]) nodules to provide further risk stratification. How a testing algorithm that uses ancillary molecular tests performs as a reflex test for repeat sampling of indeterminant nodules remains unclear.
METHODS
Thyroid FNABs performed over a 24-month period that received a diagnosis of AUS/FLUS and underwent subsequent FNAB were analyzed.
RESULTS
In total, 187 patients were identified who received an FNAB diagnosis of AUS/FLUS and had repeat sampling. Of these patients, 64% received a subsequent indeterminant diagnosis on repeat biopsy: 7 (3.7%) repeat biopsies were diagnosed as nondiagnostic/unsatisfactory, 104 (55.6%) were diagnosed as AUS/FLUS, and 8 (4.3%) were diagnosed as follicular neoplasm/suspicious for follicular neoplasm. Of the repeat biopsied nodules, 63% underwent subsequent testing with ThyroSeq version 3. The diagnostic performance was calculated using only surgically confirmed nodules (sensitivity, 100%; specificity, 30%; positive predictive value, 41%; negative predictive value, 100%) and by assigning nonresected nodules with negative ThyroSeq or benign cytology results as benign (sensitivity, 100%; specificity, 88%; positive predictive value, 41%; negative predictive value, 100%).
CONCLUSIONS
In the majority of patients, repeat FNAB for AUS/FLUS did not preclude subsequent molecular ancillary testing because of the high rate of indeterminant results on repeat biopsy. The diagnostic performance of the testing algorithm reported here was very similar to other reports using either repeat biopsy or molecular testing alone. Ultimately, the algorithm of performing molecular testing on repeat indeterminant nodules increased the number of biopsies performed and lengthened the time to definitive risk stratification without a disproportionate decrease in the use of molecular testing or an appreciable improvement in diagnostic performance.
Topics: Adenocarcinoma, Follicular; Biopsy, Fine-Needle; Genomics; Humans; Reflex; Retrospective Studies; Thyroid Neoplasms; Thyroid Nodule
PubMed: 35045202
DOI: 10.1002/cncy.22552 -
Revista Brasileira de Ginecologia E... Jul 2018To estimate the cytological and colposcopic performances for the diagnosis of cervical neoplasias.
OBJECTIVE
To estimate the cytological and colposcopic performances for the diagnosis of cervical neoplasias.
METHODS
Cross-sectional retrospective study with data from patients' charts. The participants underwent colposcopy, guided biopsies, and excision when needed. The cytological and colposcopic categorization followed the Bethesda System and the international colposcopic terminologies. The cytology and colposcopy performances were evaluated by sensitivity (SE), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) analyses with 95% confidence interval (95% CI).
RESULTS
From 1,571 participants, a total of 1,154 (73.4%) were diagnosed with cervical squamous intraepithelial neoplasia grade 2 or worse (CIN 2), 114 (7.2%) with adenocarcinoma in situ or worse (AIS), 615 (39.2%) presented atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion or worse (ASC-H) cytology, and 934 (59.4%) presented major or suspicious for invasion colposcopic abnormalities. The SE, SP, PPV, and NPV of ASC-H for diagnoses of CIN 2 and AIS were, respectively: 44% (95% CI: 41-47) and 72% (95% CI: 67-76), 79% (95% CI: 77-81) and 79% (95% CI: 75-83), 88% (95% CI: 87-90) and 55% (95% CI: 50-60), and 28% (95% CI: 26-31) and 88% (95% CI: 85-91). The SE, SP, PPV, and NPV of major or suspicious for invasion colposcopic abnormalities for diagnoses of CIN 2 and AIS were, respectively: 62% (95% CI: 60-65) and 86% (95% CI: 83-89), 59% (95% CI: 57-62) and 59% (95% CI: 55-64), 85% (95% CI: 83-87) and 44% (95% CI: 40-49), and 29% (95% CI: 27-32) and 92% (95% CI: 89-94).
CONCLUSION
The SE analyses results of ASC-H and major or suspicious for invasion colposcopic abnormalities were higher for diagnoses of glandular neoplasias. These results confirm the role of cytology in identifying women at risk who will have their final diagnoses settled by colposcopy and histology.
Topics: Adolescent; Adult; Aged; Aged, 80 and over; Biopsy; Carcinoma, Squamous Cell; Cervix Uteri; Colposcopy; Cross-Sectional Studies; Female; Humans; Middle Aged; Neoplasms, Glandular and Epithelial; Predictive Value of Tests; Retrospective Studies; Young Adult
PubMed: 29996168
DOI: 10.1055/s-0038-1666995 -
Advances in Respiratory Medicine 2021Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) of mediastinal lymph nodes is currently considered to be the most effective minimally...
Comparison of the diagnostic effectiveness between an expert and a student in the diagnosis of sarcoidosis by transbronchial biopsy of mediastinal lymph nodes under the control of endosonography.
INTRODUCTION
Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) of mediastinal lymph nodes is currently considered to be the most effective minimally invasive diagnostic method in patients with suspected stage I and II sarcoidosis. However, diagnostic effectiveness depends on the experience and skills of the doctor which is dependent on the number of correctly performed procedures. The aim of the study is to compare the diagnostic effectiveness of the EBUS-TBNA test obtained by an expert in this field vs that of his student.
MATERIAL AND METHODS
in patients with a clinical and radiological suspicion of sarcoidosis, EBUS-TBNA procedures were performed by an expert (over 1000 previously performed tests) and by his student who completed basic training (15 procedures performed). In the expert's opinion, the student was experienced enough to perform the EBUS-TBNA on his own. Previously, more than 100 conventional fibreoptic bronchoscopies had been performed by the student. During that time, he had been working in the department of pulmonary diseases and tuberculosis for two years. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), negative likelihood ratio (LR-), and accuracy of the EBUS-TBNA test in diagnosing sarcoidosis were calculated. Statistical evaluation was made using ROC curves for the expert and for the student.
RESULTS
The study included 215 patients between 22-68 years of age with suspected sarcoidosis who were diagnosed between 2013-2016. 124 EBUS-TBNA tests were performed by the expert, and 91 procedures were performed by the student. The presence of sarcoid granulomas was confirmed by a biopsy in 165 (76.7%) patients. In terms of the diagnosis of sarcoidosis, a higher sensitivity and accuracy of the EBUS test was found in the procedures done by the expert (76.7% and 95.3%, respectively) as compared to the results obtained by the student (66.1% and 93.1%, respectively). However, these differences were not statistically significant (p = 0.11). All tests were assessed in a hospital pathology unit, but not necessarily by one person, which may be a limitation of our research. In this study, only cytological smears were taken into consideration.
CONCLUSIONS
In the diagnosis of sarcoidosis, the student, after appropriate training by an expert, achieved a comparable level of diagnostic effectiveness with EBUS-TBNA after performing 90 tests independently.
Topics: Clinical Competence; Curriculum; Endoscopic Ultrasound-Guided Fine Needle Aspiration; Endosonography; Humans; Sarcoidosis, Pulmonary; Students, Medical
PubMed: 34196375
DOI: 10.5603/ARM.a2021.0050 -
Journal of the American Geriatrics... Jan 2018Brief cognitive screens lack the sensitivity to detect mild cognitive impairment (MCI) or support differential diagnoses. The objective of this study was to validate the...
BACKGROUND/OBJECTIVES
Brief cognitive screens lack the sensitivity to detect mild cognitive impairment (MCI) or support differential diagnoses. The objective of this study was to validate the 10-minute, tablet-based University of California, San Francisco (UCSF) Brain Health Assessment (BHA) to overcome these limitations.
DESIGN
Cross-sectional.
SETTING
UCSF Memory and Aging Center.
PARTICIPANTS
Older adults (N = 347) (neurologically healthy controls (n = 185), and individuals diagnosed with MCI (n = 99), dementia (n = 42), and as normal with concerns (n = 21)).
MEASUREMENTS
The BHA includes subtests of memory, executive function and speed, visuospatial skills, and language and an optional informant survey. Participants completed the Montreal Cognitive Assessment (MoCA) and criterion-standard neuropsychological tests. Standardized structural 3T brain magnetic resonance imaging was performed in 145 participants.
RESULTS
At a fixed 85% specificity rate, the BHA had 100% sensitivity to dementia and 84% to MCI; the MoCA had 75% sensitivity to dementia and 25% to MCI. The BHA had 83% sensitivity to MCI likely due to AD and 88% to MCI unlikely due to AD, and the MoCA had 58% sensitivity to MCI likely AD and 24% to MCI unlikely AD. The BHA subtests demonstrated moderate to high correlations with the criterion-standard tests from their respective cognitive domains. Memory test performance correlated with medial temporal lobe volumes; executive and speed with frontal, parietal, and basal ganglia volumes; and visuospatial with right parietal volumes.
CONCLUSION
The BHA had excellent combined sensitivity and specificity to detect dementia and MCI, including MCI due to diverse etiologies. The subtests provide efficient, valid measures of neurocognition that are critical in making a differential diagnosis.
Topics: Aged; Brain; Cross-Sectional Studies; Humans; Magnetic Resonance Imaging; Mass Screening; Neurocognitive Disorders; Neuropsychological Tests; San Francisco; Sensitivity and Specificity; Surveys and Questionnaires
PubMed: 29355911
DOI: 10.1111/jgs.15208 -
Clinical Characteristics of Patients Undergoing Right Heart Catheterizations in Community Hospitals.Journal of the American Heart... Sep 2022Background Recognition of precapillary pulmonary hypertension (PH) has significant implications for patient management. However, the low a priori chance to find this...
Background Recognition of precapillary pulmonary hypertension (PH) has significant implications for patient management. However, the low a priori chance to find this rare condition in community hospitals may create a barrier against performing a right heart catheterization (RHC). This could result in misclassification of PH and delayed diagnosis/treatment of precapillary PH. Therefore, we investigated patient characteristics and echocardiographic parameters associated with the decision whether to perform an RHC in patients with incident PH in 12 Dutch community hospitals. Methods and Results In total, 275 patients were included from the OPTICS (Optimizing PH Diagnostic Network in Community Hospitals) registry, a prospective cohort study with patients with incident PH; 157 patients were diagnosed with RHC (34 chronic thromboembolic PH, 38 pulmonary arterial hypertension, 81 postcapillary PH, 4 miscellaneous PH), while 118 patients were labeled as probable postcapillary PH without hemodynamic confirmation. Multivariable analysis showed that older age (>60 years), left ventricular diastolic dysfunction grade 2-3, left atrial dilatation were independently associated with the decision to not perform an RHC, while presence of prior venous thromboembolic events or pulmonary arterial hypertension-associated conditions, right atrial dilatation, and tricuspid regurgitation velocity ≥3.7 m/s favor an RHC performance. Conclusions Older age and echocardiographic parameters of left heart disease were independently associated with the decision to not perform an RHC, while presence of prior venous thromboembolic events or pulmonary arterial hypertension-associated conditions, right atrial dilation, and severe PH on echocardiography favored an RHC performance. As such, especially elderly patients may be at an increased risk of diagnostic delays and missed diagnoses of treatable precapillary PH, which could lead to a worse prognosis.
Topics: Aged; Cardiac Catheterization; Familial Primary Pulmonary Hypertension; Hospitals, Community; Humans; Hypertension, Pulmonary; Prospective Studies; Pulmonary Arterial Hypertension
PubMed: 36062610
DOI: 10.1161/JAHA.121.025143 -
Nigerian Journal of Clinical Practice Feb 2024Congenital diseases are still an important medical, social, and economic problem all over the world. In North Cyprus, in addition to other reasons, early prenatal...
BACKGROUND
Congenital diseases are still an important medical, social, and economic problem all over the world. In North Cyprus, in addition to other reasons, early prenatal diagnostic measures are undertaken to prevent births with thalassemia major, a locally widespread genetic disease.
AIM
This study aims to evaluate the results of prenatal invasive diagnostic tests performed in a private obstetrics clinic in Northern Cyprus and show the diagnosis process of thalassemia and chromosomal anomalies.
MATERIALS AND METHODS
This study is a retrospective, descriptive study. Chorionic villus sampling (CVS) results and the amniocentesis tests performed between 1990 and 2022 are evaluated. Thalassemia and chromosome analysis of samples obtained by CVS and amniocentesis tests were performed. To diagnose alpha or beta thalassemia and sickle cell, 239 CVS was performed. And to diagnose chromosomal anomalies, 396 CVS and amniocentesis were performed.
RESULTS
The mean age of the 480 pregnant women included in the study was 31.12 years (18-46) and 30% of them were older than 34 years. The most common indications for invasive prenatal diagnostic test (IPDT) were; mother/father thalassemia minor/major, advanced maternal age, high risk of ultrasonography erase findings, and the noninvasive screening test. The result of IPDT detected 7.3% chromosomal anomaly and 69.5% thalassemia and sickle cell anemia. Of the 239 CVS performed to diagnose alpha or beta thalasemia and sickle cell, 23.4% beta major, 42.3% beta minor, and 2.1% alpha minor were diagnosed. Of the 396 CVS and amniocentesis performed to diagnose chromosomal anormalies; 2.8% of Down syndrome and 4.54% of other chromosomal anomalies were diagnosed.
CONCLUSION
IPDT is important in correctly diagnosing fetal anomalies at the prenatal stage to help families decide at the right time.
Topics: Pregnancy; Female; Humans; Adolescent; Young Adult; Adult; Middle Aged; Retrospective Studies; Cyprus; Prenatal Diagnosis; Chorionic Villi Sampling; Chromosome Aberrations; Chromosome Disorders; Thalassemia; Anemia, Sickle Cell; Diagnostic Tests, Routine
PubMed: 38409146
DOI: 10.4103/njcp.njcp_540_23 -
PloS One 2022Deep learning techniques have achieved remarkable success in lesion segmentation and classification between benign and malignant tumors in breast ultrasound images....
Deep learning techniques have achieved remarkable success in lesion segmentation and classification between benign and malignant tumors in breast ultrasound images. However, existing studies are predominantly focused on devising efficient neural network-based learning structures to tackle specific tasks individually. By contrast, in clinical practice, sonographers perform segmentation and classification as a whole; they investigate the border contours of the tissue while detecting abnormal masses and performing diagnostic analysis. Performing multiple cognitive tasks simultaneously in this manner facilitates exploitation of the commonalities and differences between tasks. Inspired by this unified recognition process, this study proposes a novel learning scheme, called the cross-task guided network (CTG-Net), for efficient ultrasound breast image understanding. CTG-Net integrates the two most significant tasks in computerized breast lesion pattern investigation: lesion segmentation and tumor classification. Further, it enables the learning of efficient feature representations across tasks from ultrasound images and the task-specific discriminative features that can greatly facilitate lesion detection. This is achieved using task-specific attention models to share the prediction results between tasks. Then, following the guidance of task-specific attention soft masks, the joint feature responses are efficiently calibrated through iterative model training. Finally, a simple feature fusion scheme is used to aggregate the attention-guided features for efficient ultrasound pattern analysis. We performed extensive experimental comparisons on multiple ultrasound datasets. Compared to state-of-the-art multi-task learning approaches, the proposed approach can improve the Dice's coefficient, true-positive rate of segmentation, AUC, and sensitivity of classification by 11%, 17%, 2%, and 6%, respectively. The results demonstrate that the proposed cross-task guided feature learning framework can effectively fuse the complementary information of ultrasound image segmentation and classification tasks to achieve accurate tumor localization. Thus, it can aid sonographers to detect and diagnose breast cancer.
Topics: Breast; Breast Neoplasms; Female; Humans; Image Processing, Computer-Assisted; Neural Networks, Computer; Ultrasonography; Ultrasonography, Mammary
PubMed: 35951606
DOI: 10.1371/journal.pone.0271106 -
Diagnostic Value and Effect of Bedside Ultrasound in Acute Appendicitis in the Emergency Department.Academic Emergency Medicine : Official... May 2017Early and accurate diagnosis of acute appendicitis (AA) with ultrasound (US) can minimize the morbidity and mortality of the patients. In this regard, US can help... (Observational Study)
Observational Study
OBJECTIVE
Early and accurate diagnosis of acute appendicitis (AA) with ultrasound (US) can minimize the morbidity and mortality of the patients. In this regard, US can help emergency physicians (EPs) in the diagnosing process and clinical decision making for AA. Therefore, we primarily aimed to evaluate the effectiveness of point-of-care US (POCUS) in clinical decision making of EPs for the diagnostic evaluation for AA in the emergency department (ED).
METHODS
The study sample consisted of patients aged > 18 years who presented to the ED with abdominal pain and underwent diagnostic evaluation for AA. All patients were examined initially with POCUS by EPs and then with radiology-performed US (RADUS) by radiologists. Pre- and post-POCUS median diagnostic certainty values (MDCVs) for AA were determined with visual analog scale (VAS) scores (0 = not present, 100 = certainly present) by POCUS performers. Definitive diagnoses were determined by surgery, pathologic evaluation of appendectomy specimens, or clinical follow-up results. The sensitivity, specificity, positive likelihood ratio (PLR), and negative likelihood ratio (NLR) for POCUS and RADUS together with pre- and post-POCUS VAS scores for MDCVs were compared.
RESULTS
A total of 264 patients were included into the final analysis and 169 (64%) had a diagnosis of AA. The sensitivity, specificity, PLR, and NLR of US examinations were 92.3% (95% confidence interval [CI] = 87.2%-95.8%), 95.8% (89.5%-98.8%), 21.9 (8.4-57.2), and 0.08 (0.05-0.1) for POCUS and 76.9% (69.8-83%), 97.8% (84.9-99.7%), 36.4 (9.25-144.3), and 0.24 (0.18-0.31) for RADUS, respectively. Pre-POCUS and post-POCUS VAS scores for MDCVs were 60 (interquartile range [IQR] = 50-65) and 95 (IQR = 20-98), respectively (p = 0.000).
CONCLUSION
Point-of-care ultrasonography, when performed in ED for the diagnosis of AA, has high sensitivity and specificity and had a positive impact on the clinical decision making of EPs.
Topics: Abdominal Pain; Acute Disease; Adult; Appendicitis; Emergency Service, Hospital; Female; Humans; Male; Middle Aged; Point-of-Care Systems; Prospective Studies; Sensitivity and Specificity; Ultrasonography; Young Adult
PubMed: 28171688
DOI: 10.1111/acem.13169