Your browser doesn't support javascript.
loading
Mostrar: 20 | 50 | 100
Resultados 1 - 20 de 177
Filtrar
1.
J Neurooncol ; 2021 Oct 14.
Artigo em Inglês | MEDLINE | ID: mdl-34648115

RESUMO

PURPOSE: In glioma, molecular alterations are closely associated with disease prognosis. This study aimed to develop a radiomics-based multiple gene prediction model incorporating mutual information of each genetic alteration in glioblastoma and grade 4 astrocytoma, IDH-mutant. METHODS: From December 2014 through January 2020, we enrolled 418 patients with pathologically confirmed glioblastoma (based on the 2016 WHO classification). All selected patients had preoperative MRI and isocitrate dehydrogenase (IDH) mutation, O-6-methylguanine-DNA methyltransferase (MGMT) promoter methylation, epidermal growth factor receptor amplification, and alpha-thalassemia/mental retardation syndrome X-linked (ATRX) loss status. Patients were randomly split into training and test sets (7:3 ratio). Enhancing tumor and peritumoral T2-hyperintensity were auto-segmented, and 660 radiomics features were extracted. We built binary relevance (BR) and ensemble classifier chain (ECC) models for multi-label classification and compared their performance. In the classifier chain, we calculated the mean absolute Shapley value of input features. RESULTS: The micro-averaged area under the curves (AUCs) for the test set were 0.804 and 0.842 in BR and ECC models, respectively. IDH mutation status was predicted with the highest AUCs of 0.964 (BR) and 0.967 (ECC). The ECC model showed higher AUCs than the BR model for ATRX (0.822 vs. 0.775) and MGMT promoter methylation (0.761 vs. 0.653) predictions. The mean absolute Shapley values suggested that predicted outcomes from the prior classifiers were important for better subsequent predictions along the classifier chains. CONCLUSION: We built a radiomics-based multiple gene prediction chained model that incorporates mutual information of each genetic alteration in glioblastoma and grade 4 astrocytoma, IDH-mutant and performs better than a simple bundle of binary classifiers using prior classifiers' prediction probability.

2.
J Cardiovasc Magn Reson ; 23(1): 100, 2021 09 03.
Artigo em Inglês | MEDLINE | ID: mdl-34479603

RESUMO

BACKGROUND: The prevalence of abnormal cardiovascular magnetic resonance (CMR) findings in recovered coronavirus disease 2019 (COVID-19) patients is unclear. This study aimed to investigate the prevalence of abnormal CMR findings in recovered COVID-19 patients. METHODS: A systematic literature search was performed to identify studies that report the prevalence of abnormal CMR findings in recovered COVID-19 patients. The number of patients with abnormal CMR findings and diagnosis of myocarditis on CMR (based on the Lake Louise criteria) and each abnormal CMR parameter were extracted. Subgroup analyses were performed according to patient characteristics (athletes vs. non-athletes and normal vs. undetermined cardiac enzyme levels). The pooled prevalence and 95% confidence interval (CI) of each CMR finding were calculated. Study heterogeneity was assessed, and meta-regression analysis was performed to investigate factors associated with heterogeneity. RESULTS: In total, 890 patients from 16 studies were included in the analysis. The pooled prevalence of one or more abnormal CMR findings in recovered COVID-19 patients was 46.4% (95% CI 43.2%-49.7%). The pooled prevalence of myocarditis and late gadolinium enhancement (LGE) was 14.0% (95% CI 11.6%-16.8%) and 20.5% (95% CI 17.7%-23.6%), respectively. Further, heterogeneity was observed (I2 > 50%, p < 0.1). In the subgroup analysis, the pooled prevalence of abnormal CMR findings and myocarditis was higher in non-athletes than in athletes (62.5% vs. 17.1% and 23.9% vs. 2.5%, respectively). Similarly, the pooled prevalence of abnormal CMR findings and LGE was higher in the undetermined than in the normal cardiac enzyme level subgroup (59.4% vs. 35.9% and 45.5% vs. 8.3%, respectively). Being an athlete was a significant independent factor related to heterogeneity in multivariate meta-regression analysis (p < 0.05). CONCLUSIONS: Nearly half of recovered COVID-19 patients exhibited one or more abnormal CMR findings. Athletes and patients with normal cardiac enzyme levels showed a lower prevalence of abnormal CMR findings than non-athletes and patients with undetermined cardiac enzyme levels. Trial registration The study protocol was registered in the PROSPERO database (registration number: CRD42020225234).


Assuntos
COVID-19/epidemiologia , Doenças Cardiovasculares/diagnóstico , Imagem Cinética por Ressonância Magnética/métodos , Miocárdio/patologia , COVID-19/diagnóstico , Doenças Cardiovasculares/epidemiologia , Comorbidade , Saúde Global , Humanos , Pandemias , Valor Preditivo dos Testes , Prevalência , SARS-CoV-2
3.
Acta Radiol ; : 2841851211038802, 2021 Sep 25.
Artigo em Inglês | MEDLINE | ID: mdl-34565214

RESUMO

BACKGROUND: Signal intensity (SI) of predominant fibroid (F1) on T2-weighted (T2W) images is useful for predicting the volume reduction response after gonadotropin-releasing hormone (GnRH)-agonist treatment. Few studies have been published regarding when and how to use GnRH agonist before UAE. PURPOSE: To investigate magnetic resonance imaging (MRI) prediction of volume reduction rate (VRR) of large fibroids after GnRH-agonist treatment before uterine artery embolization (UAE) as well as the efficacy of UAE based on MRI. MATERIAL AND METHODS: Data from 30 patients with a large fibroid and MRI results both before and after GnRH-agonist treatment were retrospectively analyzed. Indications for GnRH-agonist treatment are fibroids with a maximum diameter ≥10 cm or pedunculated submucosal fibroids ≥8 cm. GnRH agonist (3.75 mg leuprolide acetate) was administered subcutaneously once per month 2-6 times. SI of F1 on T2W imaging was measured: the SI was referenced to the SI of the rectus abdominis muscle (F/R). RESULTS: Mean maximum fibroid diameter was 11.1 ± 1.9 cm (range = 8.0-15.5 cm). Mean number of GnRH-agonist injections before UAE was 2.8 (range = 2-6). For predicting VRR ≥50% and <30%, the optimal cut-off values of F/R were 2.58 (sensitivity 80%, specificity 80%) and 1.69 (sensitivity 100%, specificity 70%), respectively. Of the 30 patients, fibroid infarction was complete in 29 (96.7%). CONCLUSION: SI of F1 on T2W imaging is useful for predicting the volume reduction response after GnRH-agonist treatment. After GnRH-agonist treatment for large fibroids, UAE is effective to achieve complete infarction of fibroids.

4.
Sci Rep ; 11(1): 17450, 2021 08 31.
Artigo em Inglês | MEDLINE | ID: mdl-34465816

RESUMO

We aimed to determine the proper modified thresholds for detecting and weighting CAC scores at 100 kV through histogram matching in comparison with 120 kV as a standard reference. From the training set (680 participants), modified thresholds at 100 kV were obtained through histogram matching of calcium pixels to 120 kV. From the validation set (213 participants), a standard CAC score at 120 kV, and modified CAC score at 100 kV using modified thresholds were compare through the paired t test and the Bland-Altman plot. Agreement for risk categories (no, minimal, mild, moderate, and severe) was evaluated using kappa statistics. Radiation doses were also compared. For the validation set, there was no significant difference between standard (median, 18.7; IQR, 0.0-207.0) and modified (median, 17.3; IQR, 0.0-220.9) CAC scores (P = 0.689). A small bias was achieved (0.74) with 95% limits of agreement from - 52.35 to 53.83. Agreements for risk categories were excellent (κ = 0.994). The mean dose-length-product of 100-kV scanning (30.1 ± 0.8 mGy * cm) was significantly decreased compared to 120-kV scanning (42.9 ± 0.6 mGy * cm) (P < 0.001). Histogram-derived modified thresholds at 100 kV can enable accurate CAC scoring while reducing radiation exposure.

5.
PLoS One ; 16(8): e0256152, 2021.
Artigo em Inglês | MEDLINE | ID: mdl-34383858

RESUMO

This study aims to determine how randomly splitting a dataset into training and test sets affects the estimated performance of a machine learning model and its gap from the test performance under different conditions, using real-world brain tumor radiomics data. We conducted two classification tasks of different difficulty levels with magnetic resonance imaging (MRI) radiomics features: (1) "Simple" task, glioblastomas [n = 109] vs. brain metastasis [n = 58] and (2) "difficult" task, low- [n = 163] vs. high-grade [n = 95] meningiomas. Additionally, two undersampled datasets were created by randomly sampling 50% from these datasets. We performed random training-test set splitting for each dataset repeatedly to create 1,000 different training-test set pairs. For each dataset pair, the least absolute shrinkage and selection operator model was trained and evaluated using various validation methods in the training set, and tested in the test set, using the area under the curve (AUC) as an evaluation metric. The AUCs in training and testing varied among different training-test set pairs, especially with the undersampled datasets and the difficult task. The mean (±standard deviation) AUC difference between training and testing was 0.039 (±0.032) for the simple task without undersampling and 0.092 (±0.071) for the difficult task with undersampling. In a training-test set pair with the difficult task without undersampling, for example, the AUC was high in training but much lower in testing (0.882 and 0.667, respectively); in another dataset pair with the same task, however, the AUC was low in training but much higher in testing (0.709 and 0.911, respectively). When the AUC discrepancy between training and test, or generalization gap, was large, none of the validation methods helped sufficiently reduce the generalization gap. Our results suggest that machine learning after a single random training-test set split may lead to unreliable results in radiomics studies especially with small sample sizes.

6.
AJR Am J Roentgenol ; 2021 Aug 11.
Artigo em Inglês | MEDLINE | ID: mdl-34378399

RESUMO

Background: Postoperative mammograms present interpretive challenges due to postoperative distortion and hematomas. The application of digital breast tomosynthesis (DBT) and artificial intelligence-based computer-aided detection (AI-CAD) after breast concerving therapy (BCT) has not been widely investigated. Objective: To assess the impact of additional DBT or AI-CAD on recall rate and diagnostic performance in women undergoing mammographic surveillance after BCT. Methods: This retrospective study included 314 women (mean age 53.2±10.6 years; 4 with bilateral breast cancer) who underwent BCT followed by DBT (mean interval from surgery to DBT of 15.2±15.4 months). Three breast radiologists independently reviewed images in three sessions: digital mammography (DM), DM with DBT (DM+DBT), and DM with AI-CAD (DM+AI-CAD). Recall rates and diagnostic performance were compared between DM, DM+DBT, and DM+AI-CAD, using readers' mean results. Results: Of the 314 women, 6 breast recurrences (3 ipsilateral, 3 contralateral) developed at the time of surveillance mammography. Ipsilateral breast recall rate was lower for DM+AI-CAD (1.9%) than for DM (11.2%) or DM+DBT (4.1%) (p<.001). Contralateral breast recall rate was lower for DM+AI-CAD (1.5%, p<.001) than for DM (6.6%) but not DM+DBT (2.7%, p=.08). In ipsilateral breast, accuracy was higher for DM+AI-CAD (97.0%) than for DM (88.5%) or DM+DBT (94.8%) (p<.05); specificity was higher for DM+AICAD (98.3%) than for DM (89.3%) or DM+DBT (96.1%) (p<.05); sensitivity was lower for DM+AI-CAD (22.2%) than for DM (66.7%, p=.03) but not DM+DBT (22.2%, p>.99). In contralateral breast, accuracy was higher for DM+AI-CAD (97.1%) than for DM (92.5%, p<.001) but not DM+DBT (96.1%, p=.25); specificity was higher for DM+AI-CAD (98.6%) than for DM (93.7%, p<.001) but not DM+DBT (97.5%) (p=.09); sensitivity was not different between DM (33.3%), DM+DBT (22.2%), and DM+AI-CAD (11.1%) (p>.05). Conclusion: After BCT, adjunct DBT or AI-CAD reduced recall rates and improved accuracy in the ipsilateral and contralateral breasts compared with DM. In the ipsilateral breast, addition of AI-CAD resulted in lower recall rate and higher accuracy than addition of DBT. Clinical Impact: AI-CAD may help address the challenges of post-BCT surveillance mammograms.

8.
J Comput Assist Tomogr ; 45(3): 395-402, 2021.
Artigo em Inglês | MEDLINE | ID: mdl-34297510

RESUMO

OBJECTIVE: This study aimed to compare the prognostic performance of Coronary Artery Disease (CAD)-Reporting and Data System (CAD-RADS) score with those of clinical risk factors and the extent of CAD classification for predicting major adverse cardiac events in emergency department patients. METHODS: A total of 779 patients with acute chest pain at low to intermediate risk for CAD underwent cardiac computed tomography angiography. The primary end point was early and late major adverse cardiac events. We developed the following models: model 1, clinical risk factors; model 2, clinical risk factors and CAD-RADS scores; model 3, clinical risk factors and extent of CAD. RESULTS: The C-statistics revealed that both CAD-RADS score and CAD extent improved risk stratification over the clinical risk factors (C-index for early events: C-index: 0.901 vs 0.814 and 0.911 vs 0.814; C-index for late events: 0.897 vs 0.808 and 0.905 vs 0.808; all P < 0.05). CONCLUSIONS: The CAD-RADS score had additional risk prediction benefits over clinical risk factors for emergency department patients.


Assuntos
Dor no Peito/etiologia , Doença da Artéria Coronariana/diagnóstico por imagem , Sistemas de Apoio a Decisões Clínicas , Sistemas de Informação em Radiologia , Adulto , Idoso , Doença da Artéria Coronariana/mortalidade , Serviço Hospitalar de Emergência , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Prognóstico , Medição de Risco , Fatores de Risco , Tomografia Computadorizada por Raios X
9.
Korean J Radiol ; 2021 Jul 14.
Artigo em Inglês | MEDLINE | ID: mdl-34269534

RESUMO

OBJECTIVE: To assess the feasibility of quantitatively assessing pancreatic steatosis using magnetic resonance imaging (MRI) and its correlation with obesity and metabolic risk factors in pediatric patients. MATERIALS AND METHODS: Pediatric patients (≤ 18 years) who underwent liver fat quantification MRI between January 2016 and June 2019 were retrospectively included and divided into the obesity and control groups. Pancreatic proton density fat fraction (P-PDFF) was measured as the average value for three circular regions of interest (ROIs) drawn in the pancreatic head, body, and tail. Age, weight, laboratory results, and mean liver MRI values including liver PDFF (L-PDFF), stiffness on MR elastography, and T2* values were assessed for their correlation with P-PDFF using linear regression analysis. The associations between P-PDFF and metabolic risk factors, including obesity, hypertension, diabetes mellitus (DM), and dyslipidemia, were assessed using logistic regression analysis. RESULTS: A total of 172 patients (male:female = 125:47; mean ± standard deviation [SD], 13.2 ± 3.1 years) were included. The mean P-PDFF was significantly higher in the obesity group than in the control group (mean ± SD, 4.2 ± 2.5% vs. 3.4 ± 2.4%; p = 0.037). L-PDFF and liver stiffness values showed no significant correlation with P-PDFF (p = 0.235 and p = 0.567, respectively). P-PDFF was significantly associated with obesity (odds ratio 1.146, 95% confidence interval 1.006-1.307, p = 0.041), but there was no significant association with hypertension, DM, and dyslipidemia. CONCLUSION: MRI can be used to quantitatively measure pancreatic steatosis in children. P-PDFF is significantly associated with obesity in pediatric patients.

10.
Abdom Radiol (NY) ; 46(10): 4729-4735, 2021 10.
Artigo em Inglês | MEDLINE | ID: mdl-34216244

RESUMO

PURPOSE: To assess how different driver power amplitudes affect the measurement of liver stiffness in pediatric liver magnetic resonance elastography (MRE). METHODS: From January 2018 to May 2018, pediatric patients (≤ 18 years) who underwent liver MRE with 20% and 56% driver power amplitudes were included in this retrospective study. Region-of-interests (ROIs) were drawn on four stiffness maps to include the largest area of the liver parenchyma. Intraclass correlation coefficients (ICCs) were used to assess agreements for the area, mean, maximum, minimum and standard deviation of liver stiffness between the driver power amplitudes. RESULTS: 128 MRE stiffness maps from 16 patients (M:F = 10:6, median 12.5 years old) were included. On MRE, median ROI areas of liver were 83.1 cm2 (range, 46.9-144.1 cm2) and 63.0 cm2 (range, 5.4-123.4 cm2) for the driver power amplitudes of 20% and 56%, respectively. Median liver stiffness values were 2.3 kPa (range, 1.7-8.0 kPa) and 2.8 kPa (range, 1.7-8.5 kPa). Maximum and minimum liver stiffness values were 5.3 kPa and 1.0 kPa for 20% and 7.8 kPa and 1.1 kPa for 56%. Standard deviation was 0.6 kPa for 20% and 1.0 kPa for 56%. ICC values between the two power amplitudes were 0.33-0.51 for the ROI area and the maximum, minimum and standard deviation values of liver stiffness. The ICC value for liver stiffness was 0.857 (95% confidence interval, 0.760-0.915). CONCLUSION: Liver stiffness with two driver power amplitudes on MRE showed good reliability in pediatric patients. Driver power amplitudes need to be optimized according to the pediatric liver size.


Assuntos
Técnicas de Imagem por Elasticidade , Criança , Imagem Ecoplanar , Humanos , Fígado/diagnóstico por imagem , Imageamento por Ressonância Magnética , Reprodutibilidade dos Testes , Estudos Retrospectivos
11.
Radiology ; 301(1): 81-90, 2021 10.
Artigo em Inglês | MEDLINE | ID: mdl-34282972

RESUMO

Background The Coronary Artery Disease Reporting and Data System (CAD-RADS) was established in 2016 to standardize the reporting of coronary artery disease at coronary CT angiography (CCTA). Purpose To assess the prognostic value of CAD-RADS at CCTA for major adverse cardiovascular events (MACEs) in patients presenting to the emergency department with chest pain. Materials and Methods This multicenter retrospective observational cohort study was conducted at four qualifying university teaching hospitals. Patients presenting to the emergency department with acute chest pain underwent CCTA between January 2010 and December 2017. Multivariable Cox regression analysis was used to evaluate risk factors for MACEs, including clinical factors, coronary artery calcium score (CACS), and CAD-RADS categories. The prognostic value compared with clinical risk factors and CACS was also assessed. Results A total of 1492 patients were evaluated (mean age, 58 years ± 14 years [standard deviation]; 759 men). During a median follow-up period of 31.5 months, 103 of the 1492 patients (7%) experienced MACEs. Multivariable Cox regression analysis showed that a moderate to severe CACS was associated with MACEs after adjusting for clinical risk factors (hazard ratio [HR] range, 2.3-4.4; P value range, <.001 to <.01). CAD-RADS categories from 3 to 4 or 5 (HR range, 3.2-8.5; P < .001) and high-risk plaques (HR = 3.6, P < .001) were also associated with MACEs. The C statistics revealed that the CAD-RADS score improved risk stratification more than that using clinical risk factors alone or combined with CACS (C-index, 0.85 vs 0.63 [P < .001] and 0.76 [P < .01], respectively). Conclusion The Coronary Artery Disease Reporting and Data System classification had an incremental prognostic value compared with the coronary artery calcium score in the prediction of major adverse cardiovascular events in patients presenting to the emergency department with acute chest pain. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Vliegenthart in this issue.


Assuntos
Dor no Peito/complicações , Angiografia por Tomografia Computadorizada/métodos , Sistemas de Informação em Radiologia , Calcificação Vascular/complicações , Calcificação Vascular/diagnóstico por imagem , Doença Aguda , Estudos de Coortes , Vasos Coronários/diagnóstico por imagem , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Estudos Retrospectivos , Medição de Risco
12.
J Cardiovasc Magn Reson ; 23(1): 76, 2021 06 17.
Artigo em Inglês | MEDLINE | ID: mdl-34134713

RESUMO

BACKGROUND: Chemotherapy-induced cardiotoxicity is a well-recognized adverse effect of chemotherapy. Quantitative T1-mapping cardiovascular magnetic resonance (CMR) is useful for detecting subclinical myocardial changes in anthracycline-induced cardiotoxicity. The aim of the present study was to histopathologically validate the T1 and T2 mapping parameters for the evaluation of diffuse myocardial changes in rat models of cardiotoxicity. METHODS: Rat models of cardiotoxicity were generated by injecting rats with doxorubicin (1 mg/kg, twice a week). CMR was performed with a 9.4 T ultrahigh-field scanner using cine, pre-T1, post-T1 and T2 mapping sequences to evaluate the left ventricular ejection fraction (LVEF), native T1, T2, and extracellular volume fraction (ECV). Histopathological examinations were performed and the association of histopathological changes with CMR parameters was assessed. RESULTS: Five control rats and 36 doxorubicin-treated rats were included and classified into treatment periods. In the doxorubicin-treated rats, the LVEF significantly decreased after 12 weeks of treatment (control vs. 12-week treated: 73 ± 4% vs. 59 ± 9%, P = 0.01).  Increased native T1 and ECV were observed after 6 weeks of treatment (control vs. 6-week treated: 1148 ± 58 ms, 14.3 ± 1% vs. 1320 ± 56 ms, 20.3 ± 3%; P = 0.005, < 0.05, respectively). T2 values also increased by six weeks of treatment (control vs. 6-week treated: 16.3 ± 2 ms vs. 10.3 ± 1 ms, P < 0.05). The main histopathological features were myocardial injury, interstitial fibrosis, inflammation, and edema. The mean vacuolar change (%), fibrosis (%), and inflammation score were significantly higher in 6-week treated rats than in the controls (P = 0.03, 0.03, 0.02, respectively). In the univariable analysis, vacuolar change showed the highest correlation with native T1 value (R = 0.60, P < 0.001), and fibrosis showed the highest correlation with ECV value (R = 0.78, P < 0.001). In the multiple linear regression analysis model, vacuolar change was a significant factor for change in native T1 (P = 0.01), and vacuolar change and fibrosis were significant factors for change in ECV (P = 0.006, P < 0.001, respectively) by adding other histopathological parameters (i.e., inflammation and edema scores) CONCLUSIONS: Quantitative T1 and T2 mapping CMR is a useful non-invasive tool reflecting subclinical histopathological changes in anthracycline-induced cardiotoxicity.

13.
J Neurooncol ; 154(1): 83-92, 2021 Aug.
Artigo em Inglês | MEDLINE | ID: mdl-34191225

RESUMO

PURPOSE: We predicted molecular profiles in newly diagnosed glioblastoma patients using magnetic resonance (MR) imaging features and explored the associations between imaging features and major molecular alterations. METHODS: This retrospective study included patients with newly diagnosed glioblastoma and available next-generation sequencing results. From preoperative MR imaging, Visually AcceSAble Rembrandt Images (VASARI) features, volumetric parameters, and apparent diffusion coefficient (ADC) values were obtained. First, univariate random forest was performed to identify gene abnormalities that could be predicted by imaging features with high accuracy and stability. Next, multivariate random forest was trained to predict the selected genes in the discovery cohort and was validated in the external cohort. Univariable logistic regression was performed to further explore the associations between imaging features and genes. RESULTS: Univariate random forest identified nine genes predicted by imaging features, with high accuracy and stability. The multivariate random forest model showed excellent performance in predicting IDH and PTPN11 mutations in the discovery cohort, which were validated in the external validation cohorts (areas under the receiver operator characteristic curve [AUCs] of 0.855 for IDH and 0.88 for PTPN11). ATRX loss and EGFR mutation were predicted with AUCs of 0.753 and 0.739, respectively, whereas PTEN could not be reliably predicted. Based on univariable logistic regression analyses, IDH, ATRX, and TP53 were clustered according to their shared imaging features, whereas EGFR and CDKN2A/B were clustered in the opposite direction. CONCLUSIONS: MR imaging features are related to specific molecular alterations and can be used to predict molecular profiles in patients with newly diagnosed glioblastoma.

14.
Clin Breast Cancer ; 2021 May 05.
Artigo em Inglês | MEDLINE | ID: mdl-34078566

RESUMO

BACKGROUND: Incidental breast cancers can be detected on chest computed tomography (CT) scans. With the use of deep learning, the sensitivity of incidental breast cancer detection on chest CT would improve. This study aimed to evaluate the performance of a deep learning algorithm to detect breast cancers on chest CT and to validate the results in the internal and external datasets. PATIENTS AND METHODS: This retrospective study collected 1170 preoperative chest CT scans after the diagnosis of breast cancer for algorithm development (n = 1070), internal test (n = 100), and external test (n = 100). A deep learning algorithm based on RetinaNet was developed and tested to detect breast cancer on chest CT. RESULTS: In the internal test set, the algorithm detected 96.5% of breast cancers with 13.5 false positives per case (FPs/case). In the external test set, the algorithm detected 96.1% of breast cancers with 15.6 FPs/case. When the candidate probability of 0.3 was used as the cutoff value, the sensitivities were 92.0% with 7.36 FPs/case for the internal test set and 93.0% with 8.85 FPs/case for the external test set. When the candidate probability of 0.4 was used as the cutoff value, the sensitivities were 88.5% with 5.24 FPs/case in the internal test set and 90.7% with 6.3 FPs/case in the external test set. CONCLUSION: The deep learning algorithm could sensitively detect breast cancer on chest CT in both the internal and external test sets.

15.
Eur Radiol ; 31(11): 8786-8796, 2021 Nov.
Artigo em Inglês | MEDLINE | ID: mdl-33970307

RESUMO

OBJECTIVE: To evaluate diagnostic performance of a radiomics model for classifying hepatic cyst, hemangioma, and metastasis in patients with colorectal cancer (CRC) from portal-phase abdominopelvic CT images. METHODS: This retrospective study included 502 CRC patients who underwent contrast-enhanced CT and contrast-enhanced liver MRI between January 2005 and December 2010. Portal-phase CT images of training (n = 386) and validation (n = 116) cohorts were used to develop a radiomics model for differentiating three classes of liver lesions. Among multiple handcrafted features, the feature selection was performed using ReliefF method, and random forest classifiers were used to train the selected features. Diagnostic performance of the developed model was compared with that of four radiologists. A subgroup analysis was conducted based on lesion size. RESULTS: The radiomics model demonstrated significantly lower overall and hemangioma- and metastasis-specific polytomous discrimination index (PDI) (overall, 0.8037; hemangioma-specific, 0.6653; metastasis-specific, 0.8027) than the radiologists (overall, 0.9622-0.9680; hemangioma-specific, 0.9452-0.9630; metastasis-specific, 0.9511-0.9869). For subgroup analysis, the PDI of the radiomics model was different according to the lesion size (< 10 mm, 0.6486; ≥ 10 mm, 0.8264) while that of the radiologists was relatively maintained. For classifying metastasis from benign lesions, the radiomics model showed excellent diagnostic performance, with an accuracy of 84.36% and an AUC of 0.9426. CONCLUSION: Albeit inferior to the radiologists, the radiomics model achieved substantial diagnostic performance when differentiating hepatic lesions from portal-phase CT images of CRC patients. This model was limited particularly to classifying hemangiomas and subcentimeter lesions. KEY POINTS: • Albeit inferior to the radiologists, the radiomics model could differentiate cyst, hemangioma, and metastasis with substantial diagnostic performance using portal-phase CT images of colorectal cancer patients. • The radiomics model demonstrated limitations especially in classifying hemangiomas and subcentimeter liver lesions.

16.
Radiology ; 300(2): 390-399, 2021 08.
Artigo em Inglês | MEDLINE | ID: mdl-34032515

RESUMO

Background Group comparison results associating cortical thinning and Parkinson disease (PD) dementia (PDD) are limited in their application to clinical settings. Purpose To investigate whether cortical thickness from MRI can help predict conversion from mild cognitive impairment (MCI) to dementia in PD at an individual level using a machine learning-based model. Materials and Methods In this retrospective study, patients with PD and MCI who underwent MRI from September 2008 to November 2016 were included. Features were selected from clinical and cortical thickness variables in 10 000 randomly generated training sets. Features selected 5000 times or more were used to train random forest and support vector machine models. Each model was trained and tested in 10 000 randomly resampled data sets, and a median of 10 000 areas under the receiver operating characteristic curve (AUCs) was calculated for each. Model performances were validated in an external test set. Results Forty-two patients progressed to PDD (converters) (mean age, 71 years ± 6 [standard deviation]; 22 women), and 75 patients did not progress to PDD (nonconverters) (mean age, 68 years ± 6; 40 women). Four PDD converters (mean age, 74 years ± 10; four men) and 20 nonconverters (mean age, 67 years ± 7; 11 women) were included in the external test set. Models trained with cortical thickness variables (AUC range, 0.75-0.83) showed fair to good performances similar to those trained with clinical variables (AUC range, 0.70-0.81). Model performances improved when models were trained with both variables (AUC range, 0.80-0.88). In pair-wise comparisons, models trained with both variables more frequently showed better performance than others in all model types. The models trained with both variables were successfully validated in the external test set (AUC range, 0.69-0.84). Conclusion Cortical thickness from MRI helped predict conversion from mild cognitive impairment to dementia in Parkinson disease at an individual level, with improved performance when integrated with clinical variables. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Port in this issue.


Assuntos
Disfunção Cognitiva/diagnóstico por imagem , Demência/diagnóstico por imagem , Aprendizado de Máquina , Imageamento por Ressonância Magnética/métodos , Doença de Parkinson/diagnóstico por imagem , Idoso , Disfunção Cognitiva/patologia , Demência/patologia , Progressão da Doença , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Doença de Parkinson/patologia
17.
Korean J Radiol ; 22(7): 1034-1043, 2021 07.
Artigo em Inglês | MEDLINE | ID: mdl-33856134

RESUMO

OBJECTIVE: The purpose of this meta-analysis was to investigate the pooled agreements of the coronary artery calcium (CAC) severities assessed by electrocardiogram (ECG)-gated and non-ECG-gated CT and evaluate the impact of the scan parameters. MATERIALS AND METHODS: PubMed, EMBASE, and the Cochrane library were systematically searched. A modified Quality Assessment of Diagnostic Accuracy Studies-2 tool was used to evaluate the quality of the studies. Meta-analytic methods were utilized to determine the pooled weighted bias, limits of agreement (LOA), and the correlation coefficient of the CAC scores or the weighted kappa for the categorization of the CAC severities detected by the two modalities. The heterogeneity among the studies was also assessed. Subgroup analyses were performed based on factors that could affect the measurement of the CAC score and severity: slice thickness, reconstruction kernel, and radiation dose for non-ECG-gated CT. RESULTS: A total of 4000 patients from 16 studies were included. The pooled bias was 62.60, 95% LOA were -36.19 to 161.40, and the pooled correlation coefficient was 0.94 (95% confidence interval [CI] = 0.89-0.97) for the CAC score. The pooled weighted kappa of the CAC severity was 0.85 (95% CI = 0.79-0.91). Heterogeneity was observed in the studies (I² > 50%, p < 0.1). In the subgroup analysis, the agreement between the CAC categorizations was better when the two CT examinations had reconstructions based on the same slice thickness and kernel. CONCLUSION: The pooled agreement of the CAC severities assessed by the ECG-gated and non-ECG-gated CT was excellent; however, it was significantly affected by scan parameters, such as slice thickness and the reconstruction kernel.

18.
Eur Radiol ; 31(9): 6929-6937, 2021 Sep.
Artigo em Inglês | MEDLINE | ID: mdl-33710372

RESUMO

OBJECTIVE: To compare the diagnostic agreement and performances of synthetic and conventional mammograms when artificial intelligence-based computer-assisted diagnosis (AI-CAD) is applied. MATERIAL AND METHOD: From January 2017 to April 2017, 192 patients (mean age 53.7 ± 11.7 years) diagnosed with 203 breast cancers were enrolled in this retrospective study. All patients underwent digital breast tomosynthesis (DBT) with digital mammograms (DM) simultaneously. Commercial AI-CAD was applied to the reconstructed synthetic mammograms (SM) from DBT and DM respectively and abnormality scores were calculated. We compared the median abnormality scores between DM and SM with the Wilcoxon signed-rank test and used the Bland-Altman analysis to evaluate agreements between the two mammograms and to investigate clinicopathological factors which might affect agreement. Diagnostic performances were compared using an area under the receiver operating characteristic curve (AUC). RESULT: The abnormality scores showed a mean difference (bias) of - 3.26 (95% limits of agreement: - 32.69, 26.18) between the two mammograms by the Bland-Altman analysis. The concordance correlation coefficient was 0.934 (95% CI: 0.92, 0.946), suggesting high reproducibility. SM showed higher abnormality scores in cancer with distortion and occult findings, T1 and N0 cancer, and luminal type cancer than DM (all p ≤ 0.001). Diagnostic performance did not differ between the mammograms (AUC 0.945 for conventional mammograms, 0.938 for synthetic mammograms, p = 0.499). CONCLUSION: AI-CAD can also work well on synthetic mammograms, showing good agreement and comparable diagnostic performance compared to its application to DM. KEY POINTS: • AI-CAD which was developed based on imaging findings of digital mammograms can also be applied to synthetic mammograms. • AI-CAD showed good agreement and similar diagnostic performance when applied to both synthetic and digital mammograms. • With AI-CAD, synthetic mammograms showed relatively higher abnormality scores in cancer with distortion and occult findings, T1 and N0 cancer, and luminal type cancer than digital mammograms.


Assuntos
Inteligência Artificial , Neoplasias da Mama , Adulto , Idoso , Mama/diagnóstico por imagem , Neoplasias da Mama/diagnóstico por imagem , Diagnóstico por Computador , Feminino , Humanos , Mamografia , Pessoa de Meia-Idade , Reprodutibilidade dos Testes , Estudos Retrospectivos
19.
Eur Radiol ; 31(9): 6686-6695, 2021 Sep.
Artigo em Inglês | MEDLINE | ID: mdl-33738598

RESUMO

OBJECTIVES: To evaluate whether a deep learning (DL) model using both three-dimensional (3D) black-blood (BB) imaging and 3D gradient echo (GRE) imaging may improve the detection and segmentation performance of brain metastases compared to that using only 3D GRE imaging. METHODS: A total of 188 patients with brain metastases (917 lesions) who underwent a brain metastasis MRI protocol including contrast-enhanced 3D BB and 3D GRE were included in the training set. DL models based on 3D U-net were constructed. The models were validated in the test set consisting of 45 patients with brain metastases (203 lesions) and 49 patients without brain metastases. RESULTS: The combined 3D BB and 3D GRE model yielded better performance than the 3D GRE model (sensitivities of 93.1% vs 76.8%, p < 0.001), and this effect was significantly stronger in subgroups with small metastases (p interaction < 0.001). For metastases < 3 mm, ≥ 3 mm and < 10 mm, and ≥ 10 mm, the sensitivities were 82.4%, 93.2%, and 100%, respectively. The combined 3D BB and 3D GRE model showed a false-positive per case of 0.59 in the test set. The combined 3D BB and 3D GRE model showed a Dice coefficient of 0.822, while 3D GRE model showed a lower Dice coefficient of 0.756. CONCLUSIONS: The combined 3D BB and 3D GRE DL model may improve the detection and segmentation performance of brain metastases, especially in detecting small metastases. KEY POINTS: • The combined 3D BB and 3D GRE model yielded better performance for the detection of brain metastases than the 3D GRE model (p < 0.001), with sensitivities of 93.1% and 76.8%, respectively. • The combined 3D BB and 3D GRE model showed a false-positive rate per case of 0.59 in the test set. • The combined 3D BB and 3D GRE model showed a Dice coefficient of 0.822, while the 3D GRE model showed a lower Dice coefficient of 0.756.


Assuntos
Neoplasias Encefálicas , Aprendizado Profundo , Afro-Americanos , Neoplasias Encefálicas/diagnóstico por imagem , Meios de Contraste , Humanos , Imageamento Tridimensional , Imageamento por Ressonância Magnética
20.
Korean J Radiol ; 22(6): 880-889, 2021 06.
Artigo em Inglês | MEDLINE | ID: mdl-33686816

RESUMO

OBJECTIVE: This study aimed to investigate the regional amyloid burden and myocardial deformation using T1 mapping and strain values in patients with cardiac amyloidosis (CA) according to late gadolinium enhancement (LGE) patterns. MATERIALS AND METHODS: Forty patients with CA were divided into 2 groups per LGE pattern, and 15 healthy subjects were enrolled. Global and regional native T1 and T2 mapping, extracellular volume (ECV), and cardiac magnetic resonance (CMR)-feature tracking strain values were compared in an intergroup and interregional manner. RESULTS: Of the patients with CA, 32 had diffuse global LGE (group 2), and 8 had focal patchy or no LGE (group 1). Global native T1, T2, and ECV were significantly higher in groups 1 and 2 than in the control group (native T1: 1384.4 ms vs. 1466.8 ms vs. 1230.5 ms; T2: 53.8 ms vs. 54.2 ms vs. 48.9 ms; and ECV: 36.9% vs. 51.4% vs. 26.0%, respectively; all, p < 0.001). Basal ECV (53.7%) was significantly higher than the mid and apical ECVs (50.1% and 50.0%, respectively; p < 0.001) in group 2. Basal and mid peak radial strains (PRSs) and peak circumferential strains (PCSs) were significantly lower than the apical PRS and PCS, respectively (PRS, 15.6% vs. 16.7% vs. 26.9%; and PCS, -9.7% vs. -10.9% vs. -15.0%; all, p < 0.001). Basal ECV and basal strain (2-dimensional PRS) in group 2 showed a significant negative correlation (r = -0.623, p < 0.001). Group 1 showed no regional ECV differences (basal, 37.0%; mid, 35.9%; and apical, 38.3%; p = 0.184). CONCLUSION: Quantitative T1 mapping parameters such as native T1 and ECV may help diagnose early CA. ECV, in particular, can reflect regional differences in the amyloid deposition in patients with advanced CA, and increased basal ECV is related to decreased basal strain. Therefore, quantitative CMR parameters may help diagnose CA and determine its severity in patients with or without LGE.


Assuntos
Amiloidose , Cardiomiopatias , Adulto , Idoso , Amiloidose/complicações , Amiloidose/diagnóstico por imagem , Cardiomiopatias/diagnóstico por imagem , Cardiomiopatias/etiologia , Meios de Contraste , Feminino , Gadolínio , Humanos , Imageamento por Ressonância Magnética , Imagem Cinética por Ressonância Magnética , Masculino , Pessoa de Meia-Idade , Miocárdio , Valor Preditivo dos Testes , Estudos Prospectivos , Função Ventricular Esquerda
SELEÇÃO DE REFERÊNCIAS
DETALHE DA PESQUISA
...