ABSTRACT
This observational study investigated the potential of radiomics as a non-invasive adjunct to CT in distinguishing COVID-19 lung nodules from other benign and malignant lung nodules. Lesion segmentation, feature extraction, and machine learning algorithms, including decision tree, support vector machine, random forest, feed-forward neural network, and discriminant analysis, were employed in the radiomics workflow. Key features such as Idmn, skewness, and long-run low grey level emphasis were identified as crucial in differentiation. The model demonstrated an accuracy of 83% in distinguishing COVID-19 from other benign nodules and 88% from malignant nodules. This study concludes that radiomics, through machine learning, serves as a valuable tool for non-invasive discrimination between COVID-19 and other benign and malignant lung nodules. The findings suggest the potential complementary role of radiomics in patients with COVID-19 pneumonia exhibiting lung nodules and suspicion of concurrent lung pathologies. The clinical relevance lies in the utilization of radiomics analysis for feature extraction and classification, contributing to the enhanced differentiation of lung nodules, particularly in the context of COVID-19.
Subject(s)
COVID-19 , Lung Neoplasms , Multiple Pulmonary Nodules , Humans , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/pathology , Radiomics , COVID-19/diagnostic imaging , Tomography, X-Ray Computed , Retrospective StudiesABSTRACT
In an observational study conducted from 2016 to 2021, we assessed the utility of radiomics in differentiating between benign and malignant lung nodules detected on computed tomography (CT) scans. Patients in whom a final diagnosis regarding the lung nodules was available according to histopathology and/or 2017 Fleischner Society guidelines were included. The radiomics workflow included lesion segmentation, region of interest (ROI) definition, pre-processing, and feature extraction. Employing random forest feature selection, we identified ten important radiomic features for distinguishing between benign and malignant nodules. Among the classifiers tested, the Decision Tree model demonstrated superior performance, achieving 79% accuracy, 75% sensitivity, 85% specificity, 82% precision, and 90% F1 score. The implementation of the XGBoost algorithm further enhanced these results, yielding 89% accuracy, 89% sensitivity, 89% precision, and an F1 score of 89%, alongside a specificity of 85%. Our findings highlight tumor texture as the primary predictor of malignancy, emphasizing the importance of texture-based features in computational oncology. Thus, our study establishes radiomics as a powerful, non-invasive adjunct to CT scans in the differentiation of lung nodules, with significant implications for clinical decision-making, especially for indeterminate nodules, and the enhancement of diagnostic and predictive accuracy in this clinical context.
Subject(s)
Adenocarcinoma of Lung , Lung Neoplasms , Humans , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/pathology , Retrospective Studies , Tomography, X-Ray Computed/methods , Adenocarcinoma of Lung/pathology , Lung/diagnostic imaging , Lung/pathologyABSTRACT
Objectives: Our aim is to describe the utility of magnetic resonance imaging (MRI) in the evaluation of pathologies affecting large intracranial arteries. Materials and Methods: We performed a prospective and observational study from 2018 to 2020 using 1.5 T MRI. Our study included 75 patients who were referred for MRI brain with clinical features of stroke or having tumors/infection involving large intracranial arteries (vertebral, basilar, and internal carotid arteries) on initial MRI. Correlation of MRI diagnosis was done with final diagnosis. Results: Atherothrombosis was the most common pathology involving all the intracranial large arteries and was most commonly seen in elderly male patients. The second most common pathology involving the internal carotid, vertebral, and basilar arteries was tumors, dissection, and aneurysms, respectively. The most common artery involved by atherothrombosis, tumor, and infection/inflammation was internal carotid artery, whereas it was basilar artery and vertebral artery in cases of aneurysm and dissection, respectively. Conclusion: MRI is an extremely useful modality to study large intracranial arteries. It is useful to demonstrate the site of abnormality, vessel lumen and caliber, vessel wall changes, and perivascular areas. This can help in arriving at correct diagnosis and thereby guide appropriate timely management.
ABSTRACT
Objectives The aim of the study is to assess the role of ultrafast (UF) magnetic resonance (MR) sequences in stroke imaging. Material and Methods We prospectively studied 85 patients having clinical suspicion of stroke referred for MR imaging (MRI) during August 2016 to July 2018. These patients were subjected to both conventional and UF MRI sequences. The patients were divided into six categories based on the pathologies encountered. Further subclassification was done based on the size of the lesions as ≤10 mm and >10 mm as seen separately in both UF and conventional MR sequences. The number and visibility of these lesions on conventional and UF MRI were compared. The image quality of all the subjects was also compared based on a scale categorized into excellent, satisfactory, and poor. The findings on conventional and UF imaging sequences were correlated with the final clinical diagnosis arrived at the time of discharge. Results In our study comprising 85 patients, 57 showed pathologies. The patients showing pathologies were assigned into the six categories as acute infarct (34 cases), acute hemorrhagic infarct (six cases), chronic infarct (17 cases), chronic hemorrhagic infarct (four cases), subacute infarct (three cases), and chronic hemorrhage (one case). The number of lesions seen on conventional and UF sequences were the same although there was a slight decrease in the size of the lesions on UF sequences as compared with conventional counterparts. The image quality using UF sequences was better in motion prone patients while conventional imaging showed better image quality in cooperative patients. Conclusion In motion prone patients, UF sequences are a suitable alternative for conventional sequences as they help in arriving at the diagnosis in lesser time, with reasonably good image quality, and without motion artifacts. In cooperative stroke patients, it is better to use conventional MR sequences as the image quality is better.
ABSTRACT
Purpose: Preterm infants are at increased risk of adverse neurodevelopmental outcome (NDO). Cranial ultrasound has limited predictability. The purpose of the study is to evaluate whether magnetic resonance imaging (MRI) done at term equivalent age (TEA) predicts NDO at 18-22 months of corrected gestational age (CGA).Materials and methods: This cohort study of preterm infants born at ≤32 weeks of gestation and/or birth weight <1500 grams between April 2011 and August 2012 was conducted in a tertiary care institute in India. MRI done at TEA was reported using objective scoring. NDO at 18-22-month CGA was assessed using Bayley Scale of Infant Development (BSID) version III. Composite score (CS) < 85 in motor, language, or cognition domain was taken as adverse NDO. Association between individual MRI subscores and NDO was evaluated using multiple linear regressions by backward elimination method. Validity of MRI abnormality in predicting adverse NDO was assessed.Results: Out of 94 infants who had MRI at TEA, 56 (60%) underwent BSID III. Mean gestational age was 29.8 ± 2.1 weeks. Median CS of all domains was lower with higher total MRI score. Predictive accuracy for various subscores ranged from 55 to 73%. By multiple regression analysis, signal abnormality was associated with motor delay (ß -8.4; p .02) and cystic white matter (WM) changes with motor delay (ß -7.3; p .003) and cognitive delay (ß -6.1; p .005).Conclusions: Although specificity and negative predictive value were moderate to high across all subscores in MRI to predict the NDO, the accuracy has been only low to moderate, which limits its use as sole predictor.
Subject(s)
Child Development , Infant, Premature, Diseases/diagnostic imaging , Magnetic Resonance Imaging , Neurodevelopmental Disorders/diagnostic imaging , Female , Follow-Up Studies , Gestational Age , Humans , Infant , Infant, Newborn , Infant, Premature , Male , Neuropsychological Tests , Predictive Value of Tests , Prognosis , Prospective Studies , Sensitivity and SpecificityABSTRACT
OBJECTIVE: The aim of this study is to compare the images obtained from standard ultrafast magnetic resonance (MR) imaging sequences with gradient (GRE) sequence images in identifying fetal intracranial hemorrhage (ICH). MATERIALS AND METHODS: MR images of fetal brains with ICH done between October 2012 and September 2015 were reviewed. The images obtained from four ultrafast MR sequences- Turbo Fast Low Angle Shot (Turbo FLASH) T1-weighted images, Half Fourier Acquisition single-shot turbo spin echo (HASTE) T2-weighted images, b0 images of diffusion-weighted imaging (DWI) and b800 images of DWI were compared with images obtained from GRE sequence in depicting fetal ICH. RESULTS: Out of the 212 fetuses during the study period, 15 fetuses had ICH. In the 15 fetuses with ICH as detected on GRE, Grade1 germinal matrix hemorrhage was seen in 5 fetuses, Grade 2 in 4 fetuses, Grade 3 in 3 fetuses, and Grade 4 in two fetuses. Subdural hemorrhage was seen in 1 fetus. In comparison to GRE sequence, b0 of DWI sequence was almost equal in the depiction of ICH. T2 HASTE sequence also delineated hemorrhage, although not as effectively as GRE and b0 images of images DWI. T1 Turbo FLASH and b800 images of DWI were less reliable in the depiction of fetal ICH but were useful in predicting the stage of hemorrhage. CONCLUSION: As compared to GRE sequence, b0 images of DWI followed by HASTE are the two preferred ultrafast sequences in the diagnosis of fetal ICH.
ABSTRACT
Viral encephalitis by definition is the result of human virus affecting the brain and sparing the meninges. The other nervous system manifestations are meningitis, meningoencephalitis, encephalomyelitis, and encephalomyeloradiculitis. Encephalitis can involve any age group from children to old people. The severity of the disease depends on the viral agent and the host immune system. The patient can present with fever, headache, seizure, neurological deficit, or altered sensorium. Laboratory investigations, imaging, and cerebrospinal fluid analysis are crucial in the diagnosis of encephalitis. Magnetic resonance imaging (MRI) findings may be nonspecific or specific and plays a major role in the diagnosis of encephalitis and predicting the possible cause. This pictorial essay reviews the MRI findings of common types of viral encephalitis.
ABSTRACT
BACKGROUND & OBJECTIVES: Renal artery stenosis (RAS) is an important cause of severe hypertension in patients with chronic kidney disease (CKD). It is important to detect RAS early as it can reverse hypertension and stop rapid deterioration of renal function. The potential drawbacks of various imaging modalities used to detect RAS including contrast-related adverse effects associated with diagnostic angiography have led to increasing interest in unenhanced magnetic resonance (MR) renal angiography. The aim of this study was to detect and grade RAS in patients with CKD and suspected renovascular hypertension using unenhanced MR angiography (UMRA) and to identify patients with significant RAS (>70%) who would subsequently require further investigation and revascularization. METHODS: Thirty five CKD patients with suspected RAS were subjected to UMRA using non-contrast MR angiography of ArTery and VEins 3D True fast imaging with steady state precession technique over a three year period. Patients with RAS >70 per cent on UMRA were subjected to digital subtraction angiography (DSA) with intervention if indicated. RESULTS: In all, 76 renal arteries were evaluated using UMRA in 35 patients, of which 18 arteries showed stenosis and 11 were haemodynamically significant (eight patients). Seven patients (10 renal arteries) underwent DSA. INTERPRETATION & CONCLUSIONS: An association between UMRA and DSA findings was obtained in six patients (nine renal arteries), and these patients were stented. Post-procedure follow up showed good improvement in blood pressure and renal function. UMRA was found to be a useful non-invasive imaging modality to detect RAS in CKD patients. It can identify patients who require further invasive angiography and revascularization.
Subject(s)
Hypertension, Renovascular/diagnostic imaging , Magnetic Resonance Angiography/methods , Renal Artery/diagnostic imaging , Renal Insufficiency, Chronic/diagnostic imaging , Adult , Aged , Cross-Sectional Studies , Female , Humans , Hypertension, Renovascular/physiopathology , Male , Middle Aged , Renal Artery Obstruction/diagnostic imaging , Renal Insufficiency, Chronic/physiopathologyABSTRACT
This review paper focuses on the neonatal brain segmentation algorithms in the literature. It provides an overview of clinical magnetic resonance imaging (MRI) of the newborn brain and the challenges in automated tissue classification of neonatal brain MRI. It presents a complete survey of the existing segmentation methods and their salient features. The different approaches are categorized into intracranial and brain tissue segmentation algorithms based on their level of tissue classification. Further, the brain tissue segmentation techniques are grouped based on their atlas usage into atlas-based, augmented atlas-based and atlas-free methods. In addition, the research gaps and lacunae in literature are also identified.
Subject(s)
Brain/anatomy & histology , Image Processing, Computer-Assisted/methods , Magnetic Resonance Imaging/methods , Algorithms , Humans , Infant, NewbornABSTRACT
Unilateral agenesis of internal carotid artery (ICA) with intercavernous anastomosis is a rare congenital anomaly. We present a case of a 25-year-old female with 2-month history of holocranial headache. Neurological examination was unremarkable. Magnetic resonance imaging (MRI) with magnetic resonance angiography (MRA) showed absence of left ICA with an abnormal intercavernous vessel in the sella. Computed tomography (CT) showed absence of the left carotid canal. Doppler ultrasonography (USG) showed high resistance flow in the left common carotid artery (CCA). Since no hemorrhage or aneurysm was seen, patient was managed conservatively and is on regular follow-up. Based on our knowledge, this is the first case to demonstrate the features of unilateral agenesis of ICA with intercavernous anastomosis in X-ray, Doppler USG, CT, and MRI scans of the brain.
ABSTRACT
Combination of middle and inner ear anomaly is extremely rare. We describe a case where there was inner and middle ear anomaly on the right side and inner ear anomaly on the left side. This patient also had anomalous interposition of air cells between the middle and inner ear on the right side which precluded the use of cochlear implant. No such findings have been reported till date to the best of our knowledge.
ABSTRACT
OBJECTIVES: To evaluate the possible role of Magnetic Resonance (MR) signal intensity measurements in diagnosing Central Nervous System (CNS) anomalies antenatally. METHODS: MR images of 110 fetal brains between 18 and 38 weeks were studied. Nine were excluded due to destroyed brain. 50 had CNS anomalies. 51 had normal CNS and were used as controls. Regions of interest (ROI) cursors were placed in Vitreous, cerebellar vermis, thalamus, frontal white matter, corona radiata, periventricular region and grey matter. The lateral ventricle diameters were also obtained. Signal intensity ratio (SIR) was calculated by the signal intensity of each of the above regions to that of the vitreous. SIR in controls were compared with fetuses having: (1) Hydrocephalus. (2) Arnold Chiari type-2 Malformation (ACM-2) (3) Non-progressive ventriculomegaly (4) Miscellaneous CNS anomalies. The correlation of the normalcy or abnormalcy of the brain was based on Clinical/Physical examination in 51, Ultrasound in 20, MRI in 2 and autopsy in 28. RESULTS: In hydrocephalus and ACM-2, the SIR of vermis and periventricular region were higher than controls whereas in non-progressive ventriculomegaly and miscellaneous CNS anomalies there was no significant difference. CONCLUSION: Signal intensity measurements are useful to differentiate physiological and non-progressive ventriculomegaly from hydrocephalus and ACM-2.
Subject(s)
Central Nervous System/abnormalities , Central Nervous System/diagnostic imaging , Image Interpretation, Computer-Assisted , Magnetic Resonance Imaging/methods , Nervous System Malformations/diagnostic imaging , Prenatal Diagnosis/methods , Cerebral Ventriculography/methods , Female , Gestational Age , Humans , Image Interpretation, Computer-Assisted/methods , Magnetic Resonance Imaging/statistics & numerical data , Pregnancy , Pregnancy Trimester, Second , Pregnancy Trimester, Third , Signal Processing, Computer-AssistedABSTRACT
Decubital ischemic fasciitis, also called atypical decubital fibroplasia, is a distinctive fibroplasia occurring in nonambulatory older patients who are confined to bed or wheelchair-bound. These lesions are seen in the subcutaneous plane at pressure points or bony prominences. We present a case of ischemic fasciitis in a 68-year-old male patient who presented with a hard swelling in the middle third of the posterolateral aspect of the right leg, which is an unusual site for ischemic fasciitis.
ABSTRACT
Esophageal duplication cysts are classified as a subgroup of foregut duplication cysts. They are very rare and are predominantly detected in children. Antenatal detection is very rare. We report a case of an esophageal duplication cyst that was accurately identified antenatally by USG and MRI.
ABSTRACT
OBJECTIVE: To compare antenatal sonography and magnetic resonance (MR) imaging (MRI) in the diagnosis of fetal head and trunk anomalies. METHODS: Forty pregnant women with fetal anomalies on ultrasound (US) examination underwent MRI. The MR examination was done by a radiologist who was provided with the US data. The MR images were then read by one of the two radiologists who were blinded to the US data. They were however told that the region of interest was (1) head-neck or (2) trunk or (3) both. Antenatal US and MRI findings were compared with postnatal diagnosis. Postnatal evaluation included plain radiograph, US, computed tomography, MRI, surgery, physical evaluation and autopsy. RESULTS: Sixty anomalies were detected in the 40 women studied. This included 36 central nervous system (CNS), 7 thoracic, 7 gastrointestinal, 8 genitourinary and 2 face-neck anomalies. In the evaluation of CNS and thoracic anomalies, more number of confident diagnoses could be obtained by MRI when compared with that by US. In the detection of gastrointestinal and genitourinary anomalies, there was no significant difference between the two modalities. CONCLUSION: More number of confident diagnoses could be obtained by MRI when compared with that by US, in the evaluation of fetal CNS and thoracic anomalies. MRI can be used in complex fetal anomalies as a supplementary tool following US.
Subject(s)
Abnormalities, Multiple/diagnosis , Fetus/abnormalities , Magnetic Resonance Imaging , Ultrasonography, Prenatal , Female , Humans , Pregnancy , Single-Blind MethodABSTRACT
Congenital infiltrating lipomatosis of the face is a rare condition characterized by diffuse fatty infiltration of the facial soft tissues. There may be muscle involvement along with associated bony hyperplasia. It is a type of lipomatous tumor that is congenital in origin; it is rare and seen usually in childhood. We recently saw an 11-year-old girl with this condition. She presented with a swelling of the right side of the face that had been present since birth; there were typical findings on plain radiographs, CT, and MRI. The patient underwent cosmetic surgery. Histopathological examination showed mature adipocytes without any capsule.