Nonalcoholic fatty liver disease (NAFLD) detection and deep learning in a Chinese community-based population.
We aimed to develop and validate a deep learning system (DLS) by using an auxiliary section that extracts and outputs specific ultrasound diagnostic features to improve the explainable, clinical relevant utility of using DLS for detecting NAFLD.In a community-based study of 4144 participants with abdominal ultrasound scan in Hangzhou, China, we sampled 928 (617 [66.5%] females, mean age: 56 years ± 13 [standard deviation]) participants (2 images per participant) to develop and validate DLS, a two-section neural network (2S-NNet). Radiologists' consensus diagnosis classified hepatic steatosis as none steatosis, mild, moderate, and severe. We also explored the NAFLD detection performance of six one-section neural network models and five fatty liver indices on our data set. We further evaluated the influence of participants' characteristics on the correctness of 2S-NNet by logistic regression.Area under the curve (AUROC) of 2S-NNet for hepatic steatosis was 0.90 for ≥ mild, 0.85 for ≥ moderate, and 0.93 for severe steatosis, and was 0.90 for NAFLD presence, 0.84 for moderate to severe NAFLD, and 0.93 for severe NAFLD. The AUROC of NAFLD severity was 0.88 for 2S-NNet, and 0.79-0.86 for one-section models. The AUROC of NAFLD presence was 0.90 for 2S-NNet, and 0.54-0.82 for fatty liver indices. Age, sex, body mass index, diabetes, fibrosis-4 index, android fat ratio, and skeletal muscle via dual-energy X-ray absorptiometry had no significant impact on the correctness of 2S-NNet (p > 0.05).By using two-section design, 2S-NNet had improved the performance for detecting NAFLD with more explainable, clinical relevant utility than using one-section design.• Based on the consensus review derived from radiologists, our DLS (2S-NNet) had an AUROC of 0.88 by using two-section design and yielded better performance for detecting NAFLD than using one-section design with more explainable, clinical relevant utility. • The 2S-NNet outperformed five fatty liver indices with the highest AUROCs (0.84-0.93 vs. 0.54-0.82) for different NAFLD severity screening, indicating screening utility of deep learning-based radiology may perform better than blood biomarker panels in epidemiology. • The correctness of 2S-NNet was not significantly influenced by individual's characteristics, including age, sex, body mass index, diabetes, fibrosis-4 index, android fat ratio, and skeletal muscle via dual-energy X-ray absorptiometry.
View details for DOI 10.1007/s00330-023-09515-1
View details for PubMedID 36892645
View details for PubMedCentralID 6770992
- Pitfalls and challenging Cases of Elbow Imaging: How to Triumph and Make the Diagnosis. InPractice ARRS 2023
Diagnostic Performance of the Ovarian-Adnexal Reporting and Data System (O-RADS) Ultrasound Risk Score in Women in the United States.
JAMA network open
2022; 5 (6): e2216370
The American College of Radiology (ACR) Ovarian-Adnexal Reporting and Data System (O-RADS) ultrasound (US) risk scoring system has been studied in a selected population of women referred for suspected or known adnexal lesions. This population has a higher frequency of malignant neoplasms than women presenting to radiology departments for pelvic ultrasonography for a variety of indications, potentially impacting the diagnostic performance of the risk scoring system.To evaluate the risk of malignant neoplasm and diagnostic performance of O-RADS US risk scoring system in a multi-institutional, nonselected cohort.This multi-institutional cohort study included a population of nonselected women in the United States who presented to radiology departments for routine pelvic ultrasonography between 2011 and 2014, with pathology confirmation imaging follow up or 2 years of clinical follow up.Analysis of 1014 adnexal lesions using the O-RADS US risk stratification system.Frequency of ovarian cancer and diagnostic performance of the O-RADS US risk stratification system.This study included 913 women with 1014 adnexal lesions. The mean (SD) age of the patients was 42.4 (13.9 years), and 674 of 913 (73.8%) were premenopausal. The overall frequency of malignant neoplasm was 8.4% (85 of 1014 adnexal lesions). The frequency of malignant neoplasm for O-RADS US 2 was 0.5% (3 of 657 lesions; <1% expected); O-RADS US 3, 4.5% (5 of 112 lesions; <10% expected); O-RADS US 4, 11.6% (18 of 155; 10%-50% expected); and O-RADS 5, 65.6% (59 of 90 lesions; >50% expected). O-RADS US 4 was the optimum cutoff for diagnosing cancer with sensitivity of 90.6% (95% CI, 82.3%-95.9%), specificity of 81.9% (95% CI, 79.3%-84.3%), positive predictive value of 31.4% (95% CI, 25.7%-37.7%) and negative predictive value of 99.0% (95% CI, 98.0%-99.6%).In this cohort study of a nonselected patient population, the O-RADS US risk stratification system performed within the expected range as published by the ACR O-RADS US committee. The frequency of malignant neoplasm was at the lower end of the published range, partially because of the lower prevalence of cancer in a nonselected population. However, a high negative predictive value was maintained, and when a lesion can be classified as an O-RADS US 2, the risk of cancer is low, which is reassuring for both clinician and patient.
View details for DOI 10.1001/jamanetworkopen.2022.16370
View details for PubMedID 35679042
- Imaging Approach to Pulmonary Infections in the Immunocompromised Patient INDIAN JOURNAL OF RADIOLOGY AND IMAGING 2022
A robust 3D fast spin-echo technique for fast examination of the brachial plexus.
To introduce a 3D fast spin-echo (FSE) sequence technique that may replace conventional clinical 2D FSE sequences for examining the brachial plexus.A 3D FSE sequence with motion-sensitized driven equilibrium magnitude preparation, triple-echo Dixon, and outer-volume suppression techniques, dubbed as MSDE-CUBE-fTED, was compared with clinical 2D T2-weighted and T1-weighted FSE sequences on the conventional brachial plexus exam of 14 volunteers. The resulting images were evaluated by two radiologists for fat suppression, blood flow suppression, nerve visualization, scalene muscle shape, surrounding fat planes, and diagnostic confidence. The inter-rater agreement of the reviewers was also measured. In addition, the signal magnitude ratios and contrast-to-noise ratios between nerve-to-vessel, nerve-to-muscle, and fat-to-muscle were compared.The MSDE-CUBE-fTED sequence scored significantly higher than the T2-weighed FSE sequence in all visualization categories (P < 0.05). Its score was not significantly different from that of the T1-weighted FSE in muscle and fat visualization (P ≥ 0.5). The inter-rater agreements were substantial (Gwet's agreement coefficient ≥ 0.7). The signal magnitude and contrast ratios were significantly higher in the MSDE-CUBE-fTED sequence (P < 0.05).Our results suggest that the MSDE-CUBE-fTED sequence can make a potential alternative to standard T2- and T1-weighted FSE sequences for examining the brachial plexus.
View details for DOI 10.1007/s00256-022-04021-9
View details for PubMedID 35347408
Color Doppler Imaging of Vascular Abnormalities of the Uterus
2022; 38 (1): 72-82
View details for DOI 10.1097/RUQ.0000000000000578
Ovarian Cancer Detection in Average-Risk Women: Classic- versus Nonclassic-appearing Adnexal Lesions at US.
Background Several US risk stratification schemas for assessing adnexal lesions exist. These multiple-subcategory systems may be more multifaceted than necessary for isolated adnexal lesions in average-risk women. Purpose To explore whether a US-based classification scheme of classic versus nonclassic appearance can be used to help appropriately triage women at average risk of ovarian cancer without compromising diagnostic performance. Materials and Methods This retrospective multicenter study included isolated ovarian lesions identified at pelvic US performed between January 2011 and June 2014, reviewed between September 2019 and September 2020. Lesions were considered isolated in the absence of ascites or peritoneal implants. Lesions were classified as classic or nonclassic based on sonographic appearance. Classic lesions included simple cysts, hemorrhagic cysts, endometriomas, and dermoids. Otherwise, lesions were considered nonclassic. Outcomes based on histopathologic results or clinical or imaging follow-up were recorded. Diagnostic performance and frequency of malignancy were calculated. Frequency of malignancy between age groups was compared using the χ2 test, and Poisson regression was used to explore relationships between imaging features and malignancy. Results A total of 970 isolated lesions in 878 women (mean age, 42 years ± 14 [SD]) were included. The malignancy rate for classic lesions was less than 1%. Of 970 lesions, 53 (6%) were malignant. The malignancy rate for nonclassic lesions was 32% (33 of 103) when blood flow was present and 8% (16 of 194) without blood flow (P < .001). For women older than 60 years, the malignancy rate was 50% (10 of 20 lesions) when blood flow was present and 13% (five of 38) without blood flow (P = .004). The sensitivity, specificity, positive predictive value, and negative predictive value of the classic-versus-nonclassic schema was 93% (49 of 53 lesions), 73% (669 of 917 lesions), 17% (49 of 297 lesions), and 99% (669 of 673 lesions), respectively, for detection of malignancy. Conclusion Using a US classification schema of classic- or nonclassic-appearing adnexal lesions resulted in high sensitivity and specificity in the diagnosis of malignancy in ovarian cancer. The highest risk of cancer was in isolated nonclassic lesions with blood flow in women older than 60 years. © RSNA, 2022 See also the editorial by Baumgarten in this issue.
View details for DOI 10.1148/radiol.212338
View details for PubMedID 35315722
Elbow Imaging: Variants and Asymptomatic Findings.
Seminars in musculoskeletal radiology
2021; 25 (4): 546-557
One of the key principles in the interpretation of radiology images is the ability to differentiate between normal and abnormal findings. This article provides a comprehensive overview of normal structures and anatomical variants occurring around the elbow including potential diagnostic pitfalls. We discuss frequently observed anatomical variants found in routine clinical practice associated with osseous, ligamentous, musculotendinous, and neurovascular structures at the elbow that may simulate pathology or predispose to symptoms under specific circumstances.
View details for DOI 10.1055/s-0041-1729960
View details for PubMedID 34706384
Niche role of MRI in the evaluation of female infertility
INDIAN JOURNAL OF RADIOLOGY AND IMAGING
2020; 30 (1): 32–45
Infertility is a major social and clinical problem affecting 13-15% of couples worldwide. The pelvic causes of female infertility are categorized as ovarian disorders, tubal, peritubal disorders, and uterine disorders. Appropriate selection of an imaging modality is essential to accurately diagnose the aetiology of infertlity, since the imaging diagnosis directs the appropriate treatment to be instituted. Imaging evaluation begins with hystero- salpingography (HSG), to evaluate fallopian tube patency. Uterine filling defects and contour abnormalities may be discovered at HSG but usually require further characterization with pelvic ultrasound (US), sono-hysterography (syn: hystero-sonography/saline infusion sonography) or pelvic magnetic resonance imaging (MRI), when US remains inconclusive. The major limitation of hysterographic US, is its inability to visualize extraluminal pathologies, which are better evaluated by pelvic US and MRI. Although pelvic US is a valuable modality in diagnosing entities comprising the garden variety, however, extensive pelvic inflammatory disease, complex tubo-ovarian pathologies, deep-seated endometriosis deposits with its related complications, Mulllerian duct anomalies, uterine synechiae and adenomyosis, often remain unresolved by both transabdominal and transvaginal US. Thus, MRI comes to the rescue and has a niche role in resolving complex adnexal masses, endometriosis, and Mullerian duct anomalies with greater ease. This is a review, based on the authors' experience at tertiary care teaching hospitals and aims to provide an imaging approach towards the abnormalities which are not definitively diagnosed by ultrasound alone.
View details for DOI 10.4103/ijri.IJRI_377_19
View details for Web of Science ID 000524755700007
View details for PubMedID 32476748
View details for PubMedCentralID PMC7240899
Ovarian Cancer - Current Status of Blood Biomarker and Imaging Screening Strategies
Ovarian Cancer - Current Status of Blood Biomarker and Imaging Screening Strategies
View details for DOI 10.18103/mra.v8i6.2116
Do not forget the brachial plexus-prevalence of distal brachial plexus pathology on routine shoulder MRI.
Most of the shoulder magnetic resonance imaging (MRI) examination focuses on internal joint structures but disregarding other structures like the distal brachial plexus, which may miss important findings. Hereby, we attempt to evaluate the prevalence of distal brachial plexus abnormalities and/or muscular denervation changes seen on routine shoulder MRI examinations and discuss common pathologies affecting the distal brachial plexus.A total of 701 routine shoulder MRI studies were evaluated. The evaluation of each exam was focused on the visualized brachial plexus elements and musculature abnormalities in each case. If any abnormalities of plexus and/or musculature were found, potential underlying etiologies such as paralabral or spinoglenoid notch cysts, infiltrative/primary masses on imaging, history of prior viral illness, and radiation therapy were searched. It was then confirmed whether the abnormal findings were mentioned in the exam reports or not.Thirty-four cases (4.85%) demonstrated abnormal findings of the visualized brachial plexus cords or branches and/or musculature. It was observed that in 35.3% of exam reports these findings were not mentioned, mainly missing subtle nerve abnormalities, but correctly reporting and interpreting the encountered muscle abnormalities.The distal brachial plexus and its branches should be included in the search pattern for shoulder MRI examinations.• Normal T2 signal of the brachial plexus is iso- to slightly hyperintense to muscle but less signal intense than fluid. • Diffuse, geographic muscle edema is an indirect sign of brachial plexus pathology. • Increased T2-weighted nerve signal with or without caliber or course change should be reported and followed up to find the underlying etiology.
View details for DOI 10.1007/s00330-020-07476-3
View details for PubMedID 33236205
Approach to pediatric renal tumors: an imaging review.
Abdominal radiology (New York)
2019; 44 (2): 619-641
Renal tumors comprise 7% of all childhood cancers. A wide variety of renal tumors can affect the pediatric kidneys, which can be broadly classified as primary benign tumors, primary malignant tumors, and metastatic lesions. This article aims to enumerate usual benign and malignant renal tumors that can occur in childhood and emphasizes the characteristic imaging appearances which aid in their differential diagnosis. Additionally, the leading role of the Radiologist in primary diagnosis of renal infiltration by hematological malignancies and contiguous invasion by neuroblastoma is also introduced and unraveled. Imaging protocol comprises initial Ultrasound evaluation with subsequent computed tomography (CT) and/or Magnetic resonance imaging (MRI), all of which are invaluable in confirming the diagnosis, documenting the organ of origin, describing extent of local and distant spread. The complimentary role of nuclear medicine studies in delineating differential renal function, post-operative complications, and metastasis is also highlighted.
View details for DOI 10.1007/s00261-018-1773-z
View details for PubMedID 30311048
Magnetic resonance imaging spectrum of intracranial tubercular lesions: one disease, many faces
POLISH JOURNAL OF RADIOLOGY
2018; 83: E524–E535
Tuberculosis is a devastating disease and has shown resurgence in recent years with the advent of acquired immunodeficiency syndrome. Central nervous system involvement is the most devastating form of the disease, comprising 10% of all tuberculosis cases. The causative organism, Mycobacterium tuberculosis, incites a granulomatous inflammatory response in the brain, the effects of which can be appreciated on magnetic resonance imaging (MRI), which can thus be used for diagnosis of the same. Neurotuberculosis can present in various patterns, which can be identified on MRI. The meningeal forms include leptomeningitis and pachymeningitis. Parenchymal forms of neurotuberculosis include tuberculoma in its various stages, tubercular cerebritis and abscess, tubercular rhombencephalitis, and tubercular encephalopathy. Each pattern has characteristic MRI appearances and differential diagnoses on imaging. Complications of neurotuberculosis, usually of tubercular meningitis, include hydrocephalus, vasculitis, and infarcts as well as cranial nerve palsies. Various MRI sequences besides the conventional ones can provide additional insight into the disease, help in quantifying the disease load, and help in differentiation of neurotuberculosis from conditions with similar imaging appearances and presentations. These can enable accurate and timely diagnosis by the radiologist and early institution of treatment in order to reduce the likelihood of permanent neurological sequelae.
View details for DOI 10.5114/pjr.2018.81408
View details for Web of Science ID 000459542600001
View details for PubMedID 30800191
View details for PubMedCentralID PMC6384409
A Case of Mistaken Identity: Glutaric Aciduria Type I Masquerading as Postmeningitic Hydrocephalus.
Journal of clinical imaging science
2018; 8: 50
We report the characteristic neuroimaging features of a rare metabolic leukodystrophy in an 8-year-old boy, born of consanguineous parenthood. The child presented with macrocrania, regression of milestones, and dystonia. The patient was referred for magnetic resonance imaging with a clinical diagnosis of postmeningitic hydrocephalus. Imaging revealed ventriculomegaly, diffuse brain atrophy, bilaterally symmetric widened sylvian fissure with temporal lobe hypoplasia, periventricular white-matter hyperintensities, and atrophy with hyperintensity in bilateral basal ganglia was also seen. These imaging features were signatory to arrive at a diagnosis of glutaric aciduria type 1. This disorder may mimic other neurological diseases such as postmeningitic hydrocephalus, which delays the diagnosis. Since early diagnosis and treatment can arrest progression, increased awareness about this condition among radiologists will certainly prevent erroneous diagnosis as had occurred in our patient.
View details for DOI 10.4103/jcis.JCIS_56_18
View details for PubMedID 30546934
View details for PubMedCentralID PMC6251246
Primary presentation of Jeune's syndrome as gastric motility disorder in an infant: A case report
INDIAN JOURNAL OF RADIOLOGY AND IMAGING
2018; 28 (1): 65–69
We report a case of a 4-week-old female neonate with Jeune's asphyxiating thoracic dystrophy (JATD) and coexistent situs anomaly, primarily presenting as gastric motility disorder. The child presented with abdominal distension and nonbilious vomiting since birth with failure to thrive. However, skeletal survey revealed JATD. Upper gastrointestinal contrast study showed situs inversus with delayed gastric emptying. Pyloric biopsy and intraoperative antro-duodenal manometry confirmed association of gastric motility disorder. Awareness of the unusual possibility of primary presentation of Jeune syndrome as gastric motility disorder will improve the management approach in such infants.
View details for DOI 10.4103/ijri.IJRI_303_17
View details for Web of Science ID 000428955300015
View details for PubMedID 29692530
View details for PubMedCentralID PMC5894323
Imaging Evaluation of Metastases to Breast: Rare Presentation of Carcinoma Gallbladder
Imaging Evaluation of Metastases to Breast: Rare Presentation of Carcinoma Gallbladder
View details for DOI 10.7860/JCDR/2017/31345.10893
Hepatic Venous Waveform, Splenoportal and Damping Index in Liver Cirrhosis: Correlation with Child Pugh's Score and Oesophageal Varices
JOURNAL OF CLINICAL AND DIAGNOSTIC RESEARCH
2016; 10 (2): TC1–TC5
Clinical assessment of chronic liver disease is done by Modified Child Pugh's and Model for end-stage liver disease scoring system. Measurement of hepatic venous pressure gradient (HVPG) and Upper GI Endoscopy are considered the gold standards for measurement of portal hypertension in cirrhotics. There is a need for non-invasive evaluation of portal hypertension. Ultrasonography with colour and spectral Doppler evaluation may be an effective, rapid and inexpensive alternative.To evaluate hepatic venous waveform, damping index, splenoportal index in patients of cirrhosis on Colour Doppler ultrasound, also predict severity of portal hypertension and presence of oesophageal varices.Thirty patients of chronic liver disease were included in the study. Ultrasound and colour Doppler was done to look hepatic venous waveform pattern, Damping Index (DI), and Splenoportal Index (SPI). Contrast-enhanced Computed Tomography scan (CT) was done if renal function tests were normal, else endoscopy when the renal function tests were deranged to look for oesophageal varices.Twenty two (73.3%) patients had monophasic waveform. Biphasic and triphasic waveforms were seen in 4 (13.3%) cases. Twenty two patients (73.3%) had monophasic waveforms and majority of them were in class C. This distribution of hepatic vein waveform was statistically significantly with the Child Pugh's class (p<0.05). Twenty patients (66.7%) had value of Damping index more than >0.6 where majority of patients (18) belonged to class C and 2 in class B. There was a positive correlation between Child Pugh's total score and Damping index (r=0.614; p<0.05). There was weak positive correlation between splenoportal index and Child Pugh's score (r=0.269; p=0.15).Change in triphasic to monophasic waveform and DI >0.6 suggests severe liver dysfunction and is associated with severe portal hypertension. Hepatic venous waveform pressure changes, DI and SPI have no value in predicting presence of oesophageal varices.
View details for DOI 10.7860/JCDR/2016/15706.7181
View details for Web of Science ID 000397847400119
View details for PubMedID 27042553
View details for PubMedCentralID PMC4800619
- Painful swelling on the side of the knee Pure or primary leprous neuritis (PLN) CLINICAL AND EXPERIMENTAL DERMATOLOGY 2015; 40 (5): 586–88