Radiomics and PD-L1 expression predict immunotherapy benefits in patients with head and neck squamous cell carcinoma.

Aim: To evaluate the performance of MRI-derived radiomic risk score (RRS) and PD-L1 expression to predict overall survival (OS) and progression-free survival (PFS) of patients with recurrent head and neck squamous cell carcinoma receiving nivolumab therapy. Materials & methods: Three hundred forty radiomic features from pretreatment MRI were used to construct the RRS. The integrated area under the receiver operating characteristic curve (iAUC) was calculated to evaluate the performance of the RRS and PD-L1. Results: The RRS showed iAUCs of 0.69 and 0.57 for OS and PFS, respectively. PD-L1 expression showed iAUCs of 0.61 and 0.62 for OS and PFS, respectively. Conclusion: RRS and PD-L1 potentially predict the OS and PFS of patients with recurrent head and neck squamous cell carcinoma receiving nivolumab therapy.

[Box: see text].

Latent diffusion model-based MRI superresolution enhances mild cognitive impairment prognostication and Alzheimer's disease classification.

INTRODUCTION: Timely diagnosis and prognostication of Alzheimer's disease (AD) and mild cognitive impairment (MCI) are pivotal for effective intervention. Artificial intelligence (AI) in neuroradiology may aid in such appropriate diagnosis and prognostication. This study aimed to evaluate the potential of novel diffusion model-based AI for enhancing AD and MCI diagnosis through superresolution (SR) of brain magnetic resonance (MR) images. METHODS: 1.5T brain MR scans of patients with AD or MCI and healthy controls (NC) from Alzheimer's Disease Neuroimaging Initiative 1 (ADNI1) were superresolved to 3T using a novel diffusion model-based generative AI (d3T*) and a convolutional neural network-based model (c3T*). Comparisons of image quality to actual 1.5T and 3T MRI were conducted based on signal-to-noise ratio (SNR), naturalness image quality evaluator (NIQE), and Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE). Voxel-based volumetric analysis was then conducted to study whether 3T* images offered more accurate volumetry than 1.5T images. Binary and multiclass classifications of AD, MCI, and NC were conducted to evaluate whether 3T* images offered superior AD classification performance compared to actual 1.5T MRI. Moreover, CNN-based classifiers were used to predict conversion of MCI to AD, to evaluate the prognostication performance of 3T* images. The classification performances were evaluated using accuracy, sensitivity, specificity, F1 score, Matthews correlation coefficient (MCC), and area under the receiver-operating curves (AUROC). RESULTS: Analysis of variance (ANOVA) detected significant differences in image quality among the 1.5T, c3T*, d3T*, and 3T groups across all metrics. Both c3T* and d3T* showed superior image quality compared to 1.5T MRI in NIQE and BRISQUE with statistical significance. While the hippocampal volumes measured in 3T* and 3T images were not significantly different, the hippocampal volume measured in 1.5T images showed significant difference. 3T*-based AD classifications showed superior performance across all performance metrics compared to 1.5T-based AD classification. Classification performance between d3T* and actual 3T was not significantly different. 3T* images offered superior accuracy in predicting the conversion of MCI to AD than 1.5T images did. CONCLUSIONS: The diffusion model-based MRI SR enhances the resolution of brain MR images, significantly improving diagnostic and prognostic accuracy for AD and MCI. Superresolved 3T* images closely matched actual 3T MRIs in quality and volumetric accuracy, and notably improved the prediction performance of conversion from MCI to AD.

Revisiting prognostic factors of gliomatosis cerebri in adult-type diffuse gliomas.

PURPOSE: There is lack of comprehensive analysis evaluating the impact of clinical, molecular, imaging, and surgical data on survival of patients with gliomatosis cerebri (GC). This study aimed to investigate prognostic factors of GC in adult-type diffuse glioma patients. METHODS: Retrospective chart and imaging review was performed in 99 GC patients from adult-type diffuse glioma (among 1,211 patients; 6 oligodendroglioma, 16 IDH-mutant astrocytoma, and 77 IDH-wildtype glioblastoma) from a single institution between 2005 and 2021. Predictors of overall survival (OS) of entire patients and IDH-wildtype glioblastoma patients were determined. RESULTS: The median OS was 16.7 months (95% confidence interval [CI] 14.2-22.2) in entire patients and 14.3 months (95% CI 12.2-61.9) in IDH-wildtype glioblastoma patients. In entire patients, KPS (hazard ratio [HR] = 0.98, P = 0.004), no 1p/19q codeletion (HR = 10.75, P = 0.019), MGMTp methylation (HR = 0.54, P = 0.028), and hemorrhage (HR = 3.45, P = 0.001) were independent prognostic factors on multivariable analysis. In IDH-wildtype glioblastoma patients, KPS (HR = 2.24, P = 0.075) was the only independent prognostic factor on multivariable analysis. In subgroup of IDH-wildtype glioblastoma with CE tumors, total resection of CE tumor did not remain as a significant prognostic factor (HR = 1.13, P = 0.685). CONCLUSIONS: The prognosis of GC patients is determined by its underlying molecular type and patient performance status. Compared with diffuse glioma without GC, aggressive surgery of CE tumor in GC patients does not improve survival.

Radiomics using non-contrast CT to predict hemorrhagic transformation risk in stroke patients undergoing revascularization.

OBJECTIVES: This study explores whether textural features from initial non-contrast CT scans of infarcted brain tissue are linked to hemorrhagic transformation susceptibility. MATERIALS AND METHODS: Stroke patients undergoing thrombolysis or thrombectomy from Jan 2012 to Jan 2022 were analyzed retrospectively. Hemorrhagic transformation was defined using follow-up magnetic resonance imaging. A total of 94 radiomic features were extracted from the infarcted tissue on initial NCCT scans. Patients were divided into training and test sets (7:3 ratio). Two models were developed with fivefold cross-validation: one incorporating first-order and textural radiomic features, and another using only textural radiomic features. A clinical model was also constructed using logistic regression with clinical variables, and test set validation was performed. RESULTS: Among 362 patients, 218 had hemorrhagic transformations. The LightGBM model with all radiomics features had the best performance, with an area under the receiver operating characteristic curve (AUROC) of 0.986 (95% confidence interval [CI], 0.971-1.000) on the test dataset. The ExtraTrees model performed best when textural features were employed, with an AUROC of 0.845 (95% CI, 0.774-0.916). Minimum, maximum, and ten percentile values were significant predictors of hemorrhagic transformation. The clinical model showed an AUROC of 0.544 (95% CI, 0.431-0.658). The performance of the radiomics models was significantly better than that of the clinical model on the test dataset (p < 0.001). CONCLUSIONS: The radiomics model can predict hemorrhagic transformation using NCCT in stroke patients. Low Hounsfield unit was a strong predictor of hemorrhagic transformation, while textural features alone can predict hemorrhagic transformation. CLINICAL RELEVANCE STATEMENT: Using radiomic features extracted from initial non-contrast computed tomography, early prediction of hemorrhagic transformation has the potential to improve patient care and outcomes by aiding in personalized treatment decision-making and early identification of at-risk patients. KEY POINTS: • Predicting hemorrhagic transformation following thrombolysis in stroke is challenging since multiple factors are associated. • Radiomics features of infarcted tissue on initial non-contrast CT are associated with hemorrhagic transformation. • Textural features on non-contrast CT are associated with the frailty of the infarcted tissue.

Clinical, qualitative imaging biomarkers, and tumor oxygenation imaging biomarkers for differentiation of midline-located IDH wild-type glioblastomas and H3 K27-altered diffuse midline gliomas in adults.

PURPOSE: To compare the clinical, qualitative and quantitative imaging phenotypes, including tumor oxygenation characteristics of midline-located IDH-wildtype glioblastomas (GBMs) and H3 K27-altered diffuse midline gliomas (DMGs) in adults. METHODS: Preoperative MRI data of 55 adult patients with midline-located IDH-wildtype GBM or H3 K27-altered DMG (32 IDH-wildtype GBM and 23 H3 K27-altered DMG patients) were included. Qualitative imaging assessment was performed. Quantitative imaging assessment including the tumor volume, normalized cerebral blood volume, capillary transit time heterogeneity (CTH), oxygen extraction fraction (OEF), relative cerebral metabolic rate of oxygen values, and mean ADC value were performed from the tumor mask via automatic segmentation. Univariable and multivariable logistic analyses were performed. RESULTS: On multivariable analysis, age (odds ratio [OR] = 0.92, P = 0.015), thalamus or medulla location (OR = 10.48, P = 0.013), presence of necrosis (OR = 0.15, P = 0.038), and OEF (OR = 0.01, P = 0.042) were independent predictors to differentiate H3 K27-altered DMG from midline-located IDH-wildtype GBM. The area under the curve, accuracy, sensitivity, and specificity of the multivariable model were 0.88 (95 % confidence interval: 0.77-0.95), 81.8 %, 82.6 %, and 81.3 %, respectively. CONCLUSIONS: Along with younger age, tumor location, less frequent necrosis, and lower OEF may be useful imaging biomarkers to differentiate H3 K27-altered DMG from midline-located IDH-wildtype GBM. Tumor oxygenation imaging biomarkers may reflect the less hypoxic nature of H3 K27-altered DMG than IDH-wildtype GBM and may contribute to differentiation.

Mesial temporal atrophy in preoperative MRI rather than steep Trendelenburg position is associated with postoperative delirium in patients undergoing a major urologic surgery.

PURPOSE: To investigate whether steep Trendelenburg in a major urologic surgery is associated with postoperative delirium, and to examine other potential clinical and radiologic factors predictive of postoperative delirium. METHODS: 182 patients who received a major urologic surgery and underwent a 3.0-T brain MRI scan within 1 year prior to the date of surgery were retrospectively enrolled. Preoperative brain MRIs were used to analyze features related to small vessel disease burden and mesial temporal atrophy. Presence of a significant mesial temporal atrophy was defined as Scheltens' scale ≥ 2. Patients' clinico-demographic data and MRI features were used to identify significant predictors of postoperative delirium using the logistic regression analysis. Independent predictors found significant in the univariate analysis were further evaluated in the multivariate analysis. RESULTS: Incidence of postoperative delirium was 6.0%. Patients with postoperative delirium had lower body mass index (21.3 vs. 25.0 kg/m2, P = 0.003), prolonged duration of anesthesia (362.7 vs. 224.7 min, P < 0.001) and surgery (302.2 vs. 174.5 min, P < 0.001), and had more significant mesial temporal atrophy (64% vs. 30%, P = 0.046). In the univariate analysis, female sex, type of surgery (radical prostatectomy over cystectomy), prolonged duration of anesthesia (≥ 6 h), and presence of a significant mesial temporal atrophy were significant predictors (all P-values < 0.050), but only the presence of significant mesial temporal atrophy was significant in the multivariate analysis [odds ratio (OR), 3.69; 95% CI 0.99-13.75; P = 0.046]. CONCLUSION: Steep Trendelenburg was not associated with postoperative delirium. Significant mesial temporal atrophy (Scheltens' scale ≥ 2) in preoperative brain MRI was predictive of postoperative delirium. TRIAL REGISTRATION: Not applicable.

MRI-Based Radiomics Approach for Differentiating Juvenile Myoclonic Epilepsy from Epilepsy with Generalized Tonic-Clonic Seizures Alone.

BACKGROUND: The clinical presentation of juvenile myoclonic epilepsy (JME) and epilepsy with generalized tonic-clonic seizures alone (GTCA) is similar, and MRI scans are often perceptually normal in both conditions making them challenging to differentiate. PURPOSE: To develop and validate an MRI-based radiomics model to accurately diagnose JME and GTCA, as well as to classify prognostic groups. STUDY TYPE: Retrospective. POPULATION: 164 patients (127 with JME and 37 with GTCA) patients (age 24.0 ± 9.6; 50% male), divided into training (n = 114) and test (n = 50) sets in a 7:3 ratio with the same proportion of JME and GTCA patients kept in both sets. FIELD STRENGTH/SEQUENCE: 3T; 3D T1-weighted spoiled gradient-echo. ASSESSMENT: A total of 17 region-of-interest in the brain were identified as having clinical evidence of association with JME and GTCA, from where 1581 radiomics features were extracted for each subject. Forty-eight machine-learning combinations of oversampling, feature selection, and classification algorithms were explored to develop an optimal radiomics model. The performance of the best radiomics models for diagnosis and for classification of the favorable outcome group were evaluated in the test set. STATISTICAL TESTS: Model performance measured using area under the curve (AUC) of receiver operating characteristic (ROC) curve. Shapley additive explanations (SHAP) analysis to estimate the contribution of each radiomics feature. RESULTS: The AUC (95% confidence interval) of the best radiomics models for diagnosis and for classification of favorable outcome group were 0.767 (0.591-0.943) and 0.717 (0.563-0.871), respectively. SHAP analysis revealed that the first-order and textural features of the caudate, cerebral white matter, thalamus proper, and putamen had the highest importance in the best radiomics model. CONCLUSION: The proposed MRI-based radiomics model demonstrated the potential to diagnose JME and GTCA, as well as to classify prognostic groups. MRI regions associated with JME, such as the basal ganglia, thalamus, and cerebral white matter, appeared to be important for constructing radiomics models. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 3.

Multiparametric MRI-based radiomics model for predicting human papillomavirus status in oropharyngeal squamous cell carcinoma: optimization using oversampling and machine learning techniques.

OBJECTIVES: To develop and validate a multiparametric MRI-based radiomics model with optimal oversampling and machine learning techniques for predicting human papillomavirus (HPV) status in oropharyngeal squamous cell carcinoma (OPSCC). METHODS: This retrospective, multicenter study included consecutive patients with newly diagnosed and pathologically confirmed OPSCC between January 2017 and December 2020 (110 patients in the training set, 44 patients in the external validation set). A total of 293 radiomics features were extracted from three sequences (T2-weighted images [T2WI], contrast-enhanced T1-weighted images [CE-T1WI], and ADC). Combinations of three feature selection, five oversampling, and 12 machine learning techniques were evaluated to optimize its diagnostic performance. The area under the receiver operating characteristic curve (AUC) of the top five models was validated in the external validation set. RESULTS: A total of 154 patients (59.2 ± 9.1 years; 132 men [85.7%]) were included, and oversampling was employed to account for data imbalance between HPV-positive and HPV-negative OPSCC (86.4% [133/154] vs. 13.6% [21/154]). For the ADC radiomics model, the combination of random oversampling and ridge showed the highest diagnostic performance in the external validation set (AUC, 0.791; 95% CI, 0.775-0.808). The ADC radiomics model showed a higher trend in diagnostic performance compared to the radiomics model using CE-T1WI (AUC, 0.604; 95% CI, 0.590-0.618), T2WI (AUC, 0.695; 95% CI, 0.673-0.717), and a combination of both (AUC, 0.642; 95% CI, 0.626-0.657). CONCLUSIONS: The ADC radiomics model using random oversampling and ridge showed the highest diagnostic performance in predicting the HPV status of OPSCC in the external validation set. CLINICAL RELEVANCE STATEMENT: Among multiple sequences, the ADC radiomics model has a potential for generalizability and applicability in clinical practice. Exploring multiple oversampling and machine learning techniques was a valuable strategy for optimizing radiomics model performance. KEY POINTS: • Previous radiomics studies using multiparametric MRI were conducted at single centers without external validation and had unresolved data imbalances. • Among the ADC, CE-T1WI, and T2WI radiomics models and the ADC histogram models, the ADC radiomics model was the best-performing model for predicting human papillomavirus status in oropharyngeal squamous cell carcinoma. • The ADC radiomics model with the combination of random oversampling and ridge showed the highest diagnostic performance.

Evaluating diagnostic content of AI-generated chest radiography: A multi-center visual Turing test.

BACKGROUND: Accurate interpretation of chest radiographs requires years of medical training, and many countries face a shortage of medical professionals to meet such requirements. Recent advancements in artificial intelligence (AI) have aided diagnoses; however, their performance is often limited due to data imbalance. The aim of this study was to augment imbalanced medical data using generative adversarial networks (GANs) and evaluate the clinical quality of the generated images via a multi-center visual Turing test. METHODS: Using six chest radiograph datasets, (MIMIC, CheXPert, CXR8, JSRT, VBD, and OpenI), starGAN v2 generated chest radiographs with specific pathologies. Five board-certified radiologists from three university hospitals, each with at least five years of clinical experience, evaluated the image quality through a visual Turing test. Further evaluations were performed to investigate whether GAN augmentation enhanced the convolutional neural network (CNN) classifier performances. RESULTS: In terms of identifying GAN images as artificial, there was no significant difference in the sensitivity between radiologists and random guessing (result of radiologists: 147/275 (53.5%) vs result of random guessing: 137.5/275, (50%); p = .284). GAN augmentation enhanced CNN classifier performance by 11.7%. CONCLUSION: Radiologists effectively classified chest pathologies with synthesized radiographs, suggesting that the images contained adequate clinical information. Furthermore, GAN augmentation enhanced CNN performance, providing a bypass to overcome data imbalance in medical AI training. CNN based methods rely on the amount and quality of training data; the present study showed that GAN augmentation could effectively augment training data for medical AI.

A deep learning algorithm to quantify AVF stenosis and predict 6-month primary patency: a pilot study.

Background: A deep convolutional neural network (DCNN) model that predicts the degree of arteriovenous fistula (AVF) stenosis and 6-month primary patency (PP) based on AVF shunt sounds was developed, and was compared with various machine learning (ML) models trained on patients' clinical data. Methods: Forty dysfunctional AVF patients were recruited prospectively, and AVF shunt sounds were recorded before and after percutaneous transluminal angioplasty using a wireless stethoscope. The audio files were converted to melspectrograms to predict the degree of AVF stenosis and 6-month PP. The diagnostic performance of the melspectrogram-based DCNN model (ResNet50) was compared with that of other ML models [i.e. logistic regression (LR), decision tree (DT) and support vector machine (SVM)], as well as the DCNN model (ResNet50) trained on patients' clinical data. Results: Melspectrograms qualitatively reflected the degree of AVF stenosis by exhibiting a greater amplitude at mid-to-high frequency in the systolic phase with a more severe degree of stenosis, corresponding to a high-pitched bruit. The proposed melspectrogram-based DCNN model successfully predicted the degree of AVF stenosis. In predicting the 6-month PP, the area under the receiver operating characteristic curve of the melspectrogram-based DCNN model (ResNet50) (≥0.870) outperformed that of various ML models based on clinical data (LR, 0.783; DT, 0.766; SVM, 0.733) and that of the spiral-matrix DCNN model (0.828). Conclusion: The proposed melspectrogram-based DCNN model successfully predicted the degree of AVF stenosis and outperformed ML-based clinical models in predicting 6-month PP.

Feasibility of Deep Learning-Based Analysis of Auscultation for Screening Significant Stenosis of Native Arteriovenous Fistula for Hemodialysis Requiring Angioplasty.

OBJECTIVE: To investigate the feasibility of using a deep learning-based analysis of auscultation data to predict significant stenosis of arteriovenous fistulas (AVF) in patients undergoing hemodialysis requiring percutaneous transluminal angioplasty (PTA). MATERIALS AND METHODS: Forty patients (24 male and 16 female; median age, 62.5 years) with dysfunctional native AVF were prospectively recruited. Digital sounds from the AVF shunt were recorded using a wireless electronic stethoscope before (pre-PTA) and after PTA (post-PTA), and the audio files were subsequently converted to mel spectrograms, which were used to construct various deep convolutional neural network (DCNN) models (DenseNet201, EfficientNetB5, and ResNet50). The performance of these models for diagnosing ≥ 50% AVF stenosis was assessed and compared. The ground truth for the presence of ≥ 50% AVF stenosis was obtained using digital subtraction angiography. Gradient-weighted class activation mapping (Grad-CAM) was used to produce visual explanations for DCNN model decisions. RESULTS: Eighty audio files were obtained from the 40 recruited patients and pooled for the study. Mel spectrograms of "pre-PTA" shunt sounds showed patterns corresponding to abnormal high-pitched bruits with systolic accentuation observed in patients with stenotic AVF. The ResNet50 and EfficientNetB5 models yielded an area under the receiver operating characteristic curve of 0.99 and 0.98, respectively, at optimized epochs for predicting ≥ 50% AVF stenosis. However, Grad-CAM heatmaps revealed that only ResNet50 highlighted areas relevant to AVF stenosis in the mel spectrogram. CONCLUSION: Mel spectrogram-based DCNN models, particularly ResNet50, successfully predicted the presence of significant AVF stenosis requiring PTA in this feasibility study and may potentially be used in AVF surveillance.

Predictive performance of ultrasonography-based radiomics for axillary lymph node metastasis in the preoperative evaluation of breast cancer.

PURPOSE: The purpose of this study was to evaluate the predictive performance of ultrasonography (US)-based radiomics for axillary lymph node metastasis and to compare it with that of a clinicopathologic model. METHODS: A total of 496 patients (mean age, 52.5±10.9 years) who underwent breast cancer surgery between January 2014 and December 2014 were included in this study. Among them, 306 patients who underwent surgery between January 2014 and August 2014 were enrolled as a training cohort, and 190 patients who underwent surgery between September 2014 and December 2014 were enrolled as a validation cohort. To predict axillary lymph node metastasis in breast cancer, we developed a preoperative clinicopathologic model using multivariable logistic regression and constructed a radiomics model using 23 radiomic features selected via least absolute shrinkage and selection operator regression. RESULTS: In the training cohort, the areas under the curve (AUC) were 0.760, 0.812, and 0.858 for the clinicopathologic, radiomics, and combined models, respectively. In the validation cohort, the AUCs were 0.708, 0.831, and 0.810, respectively. The combined model showed significantly better diagnostic performance than the clinicopathologic model. CONCLUSION: A radiomics model based on the US features of primary breast cancers showed additional value when combined with a clinicopathologic model to predict axillary lymph node metastasis.

A Radiomics Approach for the Classification of Fibroepithelial Lesions on Breast Ultrasonography.

A radiomics-based classifier to distinguish phyllodes tumor and fibroadenoma on gray-scale breast ultrasonography was developed and validated. A total of 93 radiomics features were extracted from representative transverse plane ultrasound images of 182 fibroepithelial lesions initially diagnosed by core needle biopsy. High-throughput radiomics features were selected using the intra-class correlation coefficient between two radiologist readers and the Least Absolute Shrinkage and Selection Operator regression through 10-fold cross-validation. When applied to the validation set, the radiomics classifier for the differentiation of phyllodes tumors and benign/fibroadenomas achieved an area under the receiver operating characteristic curve of 0.765 (95% confidence interval [CI]: 0.597-0.888) with an accuracy of 0.703 (sensitivity: 0.857; specificity: 0.5). Our radiomics signature-based classifier may help predict phyllodes tumors among fibroepithelial lesions on breast ultrasonography.

Immunoglobulin G4-Related Disease of the Ovary Mimicking Bilateral Ovarian Malignancies.

Immunoglobulin G4-related disease (IgG4-RD) is a fibro-inflammatory condition characterized by several pathological features that can theoretically involve all organs. Ovarian involvement in IgG4-RD has been reported by two studies only. Herein, we report a pathologically confirmed case of ovarian involvement of IgG4-RD, which mimicked bilateral ovarian malignancies on computed tomography and magnetic resonance imaging.

Deep Convolutional Neural Network-based Software Improves Radiologist Detection of Malignant Lung Nodules on Chest Radiographs.

Background Multicenter studies are required to validate the added benefit of using deep convolutional neural network (DCNN) software for detecting malignant pulmonary nodules on chest radiographs. Purpose To compare the performance of radiologists in detecting malignant pulmonary nodules on chest radiographs when assisted by deep learning-based DCNN software with that of radiologists or DCNN software alone in a multicenter setting. Materials and Methods Investigators at four medical centers retrospectively identified 600 lung cancer-containing chest radiographs and 200 normal chest radiographs. Each radiograph with a lung cancer had at least one malignant nodule confirmed by CT and pathologic examination. Twelve radiologists from the four centers independently analyzed the chest radiographs and marked regions of interest. Commercially available deep learning-based computer-aided detection software separately trained, tested, and validated with 19 330 radiographs was used to find suspicious nodules. The radiologists then reviewed the images with the assistance of DCNN software. The sensitivity and number of false-positive findings per image of DCNN software, radiologists alone, and radiologists with the use of DCNN software were analyzed by using logistic regression and Poisson regression. Results The average sensitivity of radiologists improved (from 65.1% [1375 of 2112; 95% confidence interval {CI}: 62.0%, 68.1%] to 70.3% [1484 of 2112; 95% CI: 67.2%, 73.1%], P < .001) and the number of false-positive findings per radiograph declined (from 0.2 [488 of 2400; 95% CI: 0.18, 0.22] to 0.18 [422 of 2400; 95% CI: 0.16, 0.2], P < .001) when the radiologists re-reviewed radiographs with the DCNN software. For the 12 radiologists in this study, 104 of 2400 radiographs were positively changed (from false-negative to true-positive or from false-positive to true-negative) using the DCNN, while 56 of 2400 radiographs were changed negatively. Conclusion Radiologists had better performance with deep convolutional network software for the detection of malignant pulmonary nodules on chest radiographs than without. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Jacobson in this issue.