Development and multicenter validation of deep convolutional neural network-based detection of colorectal cancer on abdominal CT.

OBJECTIVES: This study aims to develop computer-aided detection (CAD) for colorectal cancer (CRC) using abdominal CT based on a deep convolutional neural network. METHODS: This retrospective study included consecutive patients with colorectal adenocarcinoma who underwent abdominal CT before CRC resection surgery (training set = 379, test set = 103). We customized the 3D U-Net of nnU-Net (CUNET) for CRC detection, which was trained with fivefold cross-validation using annotated CT images. CUNET was validated using datasets covering various clinical situations and institutions: an internal test set (n = 103), internal patients with CRC first determined by CT (n = 54) and asymptomatic CRC (n = 51), and an external validation set from two institutions (n = 60). During each validation, data from the healthy population were added (internal = 60; external = 130). CUNET was compared with other deep CNNs: residual U-Net and EfficientDet. The CAD performances were evaluated using per-CRC sensitivity (true positive/all CRCs), free-response receiver operating characteristic (FROC), and jackknife alternative FROC (JAFROC) curves. RESULTS: CUNET showed a higher maximum per-CRC sensitivity than residual U-Net and EfficientDet (internal test set 91.3% vs. 61.2%, and 64.1%). The per-CRC sensitivity of CUNET at false-positive rates of 3.0 was as follows: internal CRC determined by CT, 89.3%; internal asymptomatic CRC, 87.3%; and external validation, 89.6%. CUNET detected 69.2% (9/13) of CRCs missed by radiologists and 89.7% (252/281) of CRCs from all validation sets. CONCLUSIONS: CUNET can detect CRC on abdominal CT in patients with various clinical situations and from external institutions. KEY POINTS: • Customized 3D U-Net of nnU-Net (CUNET) can be applied to the opportunistic detection of colorectal cancer (CRC) in abdominal CT, helping radiologists detect unexpected CRC. • CUNET showed the best performance at false-positive rates ≥ 3.0, and 30.1% of false-positives were in the colorectum. CUNET detected 69.2% (9/13) of CRCs missed by radiologists and 87.3% (48/55) of asymptomatic CRCs. • CUNET detected CRCs in multiple validation sets composed of varying clinical situations and from different institutions, and CUNET detected 89.7% (252/281) of CRCs from all validation sets.

In vitro blood sample assessment: investigating correlation of laboratory hemoglobin and spectral properties of dual-energy CT measurements (ρ/Z).

OBJECTIVES: Our study comprised a single-center retrospective in vitro correlation between spectral properties, namely ρ/Z values, derived from scanning blood samples using dual-energy computed tomography (DECT) with the corresponding laboratory hemoglobin/hematocrit (Hb/Hct) levels and assessed the potential in anemia-detection. METHODS: DECT of 813 patient blood samples from 465 women and 348 men was conducted using a standardized scan protocol. Electron density relative to water (ρ or rho), effective atomic number (Zeff), and CT attenuation (Hounsfield unit) were measured. RESULTS: Positive correlation with the Hb/Hct was shown for ρ (r-values 0.37-0.49) and attenuation (r-values 0.59-0.83) while no correlation was observed for Zeff (r-values -0.04 to 0.08). Significant differences in attenuation and ρ values were detected for blood samples with and without anemia in both genders (p value < 0.001) with area under the curve ranging from 0.7 to 0.95. Depending on the respective CT parameters, various cutoff values for CT-based anemia detection could be determined. CONCLUSION: In summary, our study investigated the correlation between DECT measurements and Hb/Hct levels, emphasizing novel aspects of ρ and Zeff values. Assuming that quantitative changes in the number of hemoglobin proteins might alter the mean Zeff values, the results of our study show that there is no measurable correlation on the atomic level using DECT. We established a positive in vitro correlation between Hb/Hct values and ρ. Nevertheless, attenuation emerged as the most strongly correlated parameter with identifiable cutoff values, highlighting its preference for CT-based anemia detection. CLINICAL RELEVANCE STATEMENT: By scanning multiple blood samples with dual-energy CT scans and comparing the measurements with standard laboratory blood tests, we were able to underscore the potential of CT-based anemia detection and its advantages in clinical practice. KEY POINTS: Prior in vivo studies have found a correlation between aortic blood pool and measured hemoglobin and hematocrit. Hemoglobin and hematocrit correlated with electron density relative to water and attenuation but not Zeff. Dual-energy CT has the potential for additional clinical benefits, such as CT-based anemia detection.

The application of lasers in vital pulp therapy: clinical and radiographic outcomes.

The main purpose of vital pulp therapy (VPT) is to preserve the integrity and function of the pulp. A wide variety of materials and techniques have been proposed to improve treatment outcomes, and among them, the utilization of lasers has gained significant attention. The application of lasers in different stages of VPT has witnessed remarkable growth in recent years, surpassing previous approaches.This study aimed to review the applications of lasers in different steps of VPT and evaluate associated clinical and radiographic outcomes. An electronic search using Scopus, MEDLINE, Web of Science and Google Scholar databases from 2000 to 2023 was carried out by two independent researchers. The focus was on human studies that examined the clinical and/or radiographic effects of different laser types in VPT. A total of 4243 studies were included in this narrative review article. Based on the compiled data, it can be concluded that although current literature suggests laser may be proposed as an adjunct modality for some procedural steps in VPT, more research with standardized methodologies and criteria is needed to obtain more reliable and conclusive results.

Quality of radiomics for predicting microvascular invasion in hepatocellular carcinoma: a systematic review.

OBJECTIVES: To comprehensively evaluate the reporting quality, risk of bias, and radiomics methodology quality of radiomics models for predicting microvascular invasion in hepatocellular carcinoma. METHODS: A systematic search of available literature was performed in PubMed, Embase, Web of Science, Scopus, and the Cochrane Library up to January 21, 2022. Studies that developed and/or validated machine learning models based on radiomics data to predict microvascular invasion in hepatocellular carcinoma were included. These studies were reviewed by two investigators and the consensus data were used for analyzing. The reporting quality, risk of bias, and radiomics methodological quality were evaluated by Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD), Prediction model Risk of Bias Assessment Tool, and Radiomics Quality Score (RQS), respectively. RESULTS: A total of 30 studies met eligibility criteria with 24 model developing studies and 6 model developing and external validation studies. The median overall TRIPOD adherence was 75.4% (range 56.7-94.3%). All studies were at high risk of bias with at least 2 of 20 sources of bias. Furthermore, 28 studies showed unclear risks of bias in up to 5 signaling questions because of the lack of specified reports. The median RQS score was 37.5% (range 25-61.1%). CONCLUSION: Current radiomic models for MVI-status prediction have moderate to good reporting quality, moderate radiomics methodology quality, and high risk of bias in model development and validation. KEY POINTS: • Current microvascular invasion prediction radiomics studies have moderate to good reporting quality, moderate radiomics methodology quality, and high risk of bias in model development and validation. • Data representativeness, feature robustness, events-per-variable ratio, evaluation metrics, and appropriate validation are five main aspects futures studies should focus more on to improve the quality of radiomics. • Both Radiomics Quality Score and Prediction model Risk of Bias Assessment Tool are needed to comprehensively evaluate a radiomics study.

Performance Test of a Well-Trained Model for Meningioma Segmentation in Health Care Centers: Secondary Analysis Based on Four Retrospective Multicenter Data Sets.

BACKGROUND: Convolutional neural networks (CNNs) have produced state-of-the-art results in meningioma segmentation on magnetic resonance imaging (MRI). However, images obtained from different institutions, protocols, or scanners may show significant domain shift, leading to performance degradation and challenging model deployment in real clinical scenarios. OBJECTIVE: This research aims to investigate the realistic performance of a well-trained meningioma segmentation model when deployed across different health care centers and verify the methods to enhance its generalization. METHODS: This study was performed in four centers. A total of 606 patients with 606 MRIs were enrolled between January 2015 and December 2021. Manual segmentations, determined through consensus readings by neuroradiologists, were used as the ground truth mask. The model was previously trained using a standard supervised CNN called Deeplab V3+ and was deployed and tested separately in four health care centers. To determine the appropriate approach to mitigating the observed performance degradation, two methods were used: unsupervised domain adaptation and supervised retraining. RESULTS: The trained model showed a state-of-the-art performance in tumor segmentation in two health care institutions, with a Dice ratio of 0.887 (SD 0.108, 95% CI 0.903-0.925) in center A and a Dice ratio of 0.874 (SD 0.800, 95% CI 0.854-0.894) in center B. Whereas in the other health care institutions, the performance declined, with Dice ratios of 0.631 (SD 0.157, 95% CI 0.556-0.707) in center C and 0.649 (SD 0.187, 95% CI 0.566-0.732) in center D, as they obtained the MRI using different scanning protocols. The unsupervised domain adaptation showed a significant improvement in performance scores, with Dice ratios of 0.842 (SD 0.073, 95% CI 0.820-0.864) in center C and 0.855 (SD 0.097, 95% CI 0.826-0.886) in center D. Nonetheless, it did not overperform the supervised retraining, which achieved Dice ratios of 0.899 (SD 0.026, 95% CI 0.889-0.906) in center C and 0.886 (SD 0.046, 95% CI 0.870-0.903) in center D. CONCLUSIONS: Deploying the trained CNN model in different health care institutions may show significant performance degradation due to the domain shift of MRIs. Under this circumstance, the use of unsupervised domain adaptation or supervised retraining should be considered, taking into account the balance between clinical requirements, model performance, and the size of the available data.

Concept map-based learning in an oral radiographic interpretation course: Dental students' perceptions of its role as a learning tool.

OBJECTIVES: Concept mapping is used to promote critical thinking and meaningful learning in health professions education. This study aimed to evaluate dental undergraduates' perceptions of concept mapping as a learning tool for oral radiographic interpretation competency. METHODS: A total of 39 third-year undergraduates at a dental college participated in the oral radiological interpretation session. The students received 1 hour of instruction before creating their concept map (CM). Participants were provided an assignment of four relevant clinical cases to draw an individualised CM. The students' draft and final CM were submitted. The instructor evaluated the CMs to identify whether students could analyse, connect and organise the information in a meaningful way. Among them, 37 completed the questionnaire about their experiences and responses to concept mapping. RESULTS: Students perceived concept mapping positively. The instructor's feedback helped them recognise their misconceptions and fostered greater motivation to learn. Students attempted to integrate basic biomedical knowledge and clinical features through cross-linking. However, they also expressed negative attitudes toward concept mapping regarding time consumption and heavy burden. CONCLUSIONS: Concept mapping motivated students' directed learning and gave them the opportunity to recognise their misconceptions. This study suggests the potential use of CMs as an adjunctive learning strategy in the oral radiographic interpretation class.

Assessment of hepatic arterial hemodynamics with 4D flow MRI: in vitro analysis of motion and spatial resolution related error and in vivo feasibility study in 20 volunteers.

OBJECTIVES: To assess the ability of four-dimensional (4D) flow MRI to measure hepatic arterial hemodynamics by determining the effects of spatial resolution and respiratory motion suppression in vitro and its applicability in vivo with comparison to two-dimensional (2D) phase-contrast MRI. METHODS: A dynamic hepatic artery phantom and 20 consecutive volunteers were scanned. The accuracies of Cartesian 4D flow sequences with k-space reordering and navigator gating at four spatial resolutions (0.5- to 1-mm isotropic) and navigator acceptance windows (± 8 to ± 2 mm) and one 2D phase-contrast sequence (0.5-mm in -plane) were assessed in vitro at 3 T. Two sequences centered on gastroduodenal and hepatic artery branches were assessed in vivo for intra - and interobserver agreement and compared to 2D phase-contrast. RESULTS: In vitro, higher spatial resolution led to a greater decrease in error than narrower navigator window (30.5 to -4.67% vs -6.64 to -4.67% for flow). In vivo, hepatic and gastroduodenal arteries were more often visualized with the higher resolution sequence (90 vs 71%). Despite similar interobserver agreement (κ = 0.660 and 0.704), the higher resolution sequence had lower variability for area (CV = 20.04 vs 30.67%), flow (CV = 34.92 vs 51.99%), and average velocity (CV = 26.47 vs 44.76%). 4D flow had lower differences between inflow and outflow at the hepatic artery bifurcation (11.03 ± 5.05% and 15.69 ± 6.14%) than 2D phase-contrast (28.77 ± 21.01%). CONCLUSION: High-resolution 4D flow can assess hepatic artery anatomy and hemodynamics with improved accuracy, greater vessel visibility, better interobserver reliability, and internal consistency. KEY POINTS: • Motion-suppressed Cartesian four-dimensional (4D) flow MRI with higher spatial resolution provides more accurate measurements even when accepted respiratory motion exceeds voxel size. • 4D flow MRI with higher spatial resolution provides substantial interobserver agreement for visualization of hepatic artery branches. • Lower peak and average velocities and a trend toward better internal consistency were observed with 4D flow MRI as compared to 2D phase-contrast.

Deep learning-based algorithm improved radiologists' performance in bone metastases detection on CT.

OBJECTIVES: To develop and evaluate a deep learning-based algorithm (DLA) for automatic detection of bone metastases on CT. METHODS: This retrospective study included CT scans acquired at a single institution between 2009 and 2019. Positive scans with bone metastases and negative scans without bone metastasis were collected to train the DLA. Another 50 positive and 50 negative scans were collected separately from the training dataset and were divided into validation and test datasets at a 2:3 ratio. The clinical efficacy of the DLA was evaluated in an observer study with board-certified radiologists. Jackknife alternative free-response receiver operating characteristic analysis was used to evaluate observer performance. RESULTS: A total of 269 positive scans including 1375 bone metastases and 463 negative scans were collected for the training dataset. The number of lesions identified in the validation and test datasets was 49 and 75, respectively. The DLA achieved a sensitivity of 89.8% (44 of 49) with 0.775 false positives per case for the validation dataset and 82.7% (62 of 75) with 0.617 false positives per case for the test dataset. With the DLA, the overall performance of nine radiologists with reference to the weighted alternative free-response receiver operating characteristic figure of merit improved from 0.746 to 0.899 (p < .001). Furthermore, the mean interpretation time per case decreased from 168 to 85 s (p = .004). CONCLUSION: With the aid of the algorithm, the overall performance of radiologists in bone metastases detection improved, and the interpretation time decreased at the same time. KEY POINTS: • A deep learning-based algorithm for automatic detection of bone metastases on CT was developed. • In the observer study, overall performance of radiologists in bone metastases detection improved significantly with the aid of the algorithm. • Radiologists' interpretation time decreased at the same time.

Use of the deep learning approach to measure alveolar bone level.

AIM: The goal was to use a deep convolutional neural network to measure the radiographic alveolar bone level to aid periodontal diagnosis. MATERIALS AND METHODS: A deep learning (DL) model was developed by integrating three segmentation networks (bone area, tooth, cemento-enamel junction) and image analysis to measure the radiographic bone level and assign radiographic bone loss (RBL) stages. The percentage of RBL was calculated to determine the stage of RBL for each tooth. A provisional periodontal diagnosis was assigned using the 2018 periodontitis classification. RBL percentage, staging, and presumptive diagnosis were compared with the measurements and diagnoses made by the independent examiners. RESULTS: The average Dice Similarity Coefficient (DSC) for segmentation was over 0.91. There was no significant difference in the RBL percentage measurements determined by DL and examiners ( p=.65 ). The area under the receiver operating characteristics curve of RBL stage assignment for stages I, II, and III was 0.89, 0.90, and 0.90, respectively. The accuracy of the case diagnosis was 0.85. CONCLUSIONS: The proposed DL model provides reliable RBL measurements and image-based periodontal diagnosis using periapical radiographic images. However, this model has to be further optimized and validated by a larger number of images to facilitate its application.

The impact of cardiopulmonary hemodynamic factors in volumetry for pulmonary nodule management.

BACKGROUND: The acceptance of coronary CT angiogram (CCTA) scans in the management of stable angina has led to an exponential increase in studies performed and reported incidental findings, including pulmonary nodules (PN). Using low-dose CT scans, volumetry tools are used in growth assessment and risk stratification of PN between 5 and 8 mm in diameter. Volumetry of PN could also benefit from the increased temporal resolution of CCTA scans, potentially expediting clinical decisions when an incidental PN is first detected on a CCTA scan, and allow for better resource management and planning in a Radiology department. This study aims to investigate how cardiopulmonary hemodynamic factors impact the volumetry of PN using CCTA scans. These factors include the cardiac phase, vascular distance from the main pulmonary artery (MPA) to the nodule, difference of the MPA diameter between systole and diastole, nodule location, and cardiomegaly presence. MATERIALS AND METHODS: Two readers reviewed all CCTA scans performed from 2016 to 2019 in a tertiary hospital and detected PN measuring between 5 and 8 mm in diameter. Each observer measured each nodule using two different software packages and in systole and diastole. A multiple linear regression model was applied, and inter-observer and inter-software agreement were assessed using intraclass correlation. RESULTS: A total of 195 nodules from 107 patients were included in this retrospective, cross-sectional and observational study. The regression model identified the vascular distance (p < 0.001), the difference of the MPA diameter between systole and diastole (p < 0.001), and the location within the lower or posterior thirds of the field of view (p < 0.001 each) as affecting the volume measurement. The cardiac phase was not significant in the model. There was a very high inter-observer agreement but no reasonable inter-software agreement between measurements. CONCLUSIONS: PN volumetry using CCTA scans seems to be sensitive to cardiopulmonary hemodynamic changes independently of the cardiac phase. These might also be relevant to non-gated scans, such as during PN follow-up. The cardiopulmonary hemodynamic changes are a new limiting factor to PN volumetry. In addition, when a patient experiences an acute or deteriorating cardiopulmonary disease during PN follow-up, these hemodynamic changes could affect the PN growth estimation.

Augmented Vision for Dental Students' Education in Detecting Proximal Carious Lesions on Bitewing Radiographs: A Randomized Controlled Trial.

This two-arm, parallel, randomized controlled trial aimed to assess the effect of augmented vision (AV, using interactive color overlays) on the education of dental students in detecting proximal carious lesions on bitewing radiographs compared to black-and-white textbook-like illustrations. Forty-eight preclinical third-year dental students were randomized using a random number generator into two learning groups: test (AV, allowing interaction with color-highlighted carious lesions, n = 24) and control (showing the native radiograph and a black-and-white illustration displaying the carious lesion, n = 24). First, students had 2 weeks to assess 50 bitewings (lesion prevalence on the tooth level: 54.5%) in the test or control. Due to the nature of the intervention, participants could not be blinded toward the intervention. After that, they were asked to detect lesions on 10 independent bitewings and to assess lesion extent (outer/inner enamel; outer/middle/inner dentin). The reference test was constituted by two experienced dentists. No significant differences in accuracy (test 0.84 [95% CI: 0.79, 0.88]; control 0.83 [0.78, 0.87]), AUC (test 0.82 [0.81, 0.84]; control 0.81 [0.80, 0.83]) and F1 score (test 0.79 [0.75, 0.82]; control 0.77 [0.72, 0.81]) were observed between groups. Students of both groups showed difficulties in differentiating enamel from dentin carious lesions. While AV was reported to be motivating by students, it did not increase their accuracy.

Application of deep machine learning for the radiographic diagnosis of periodontitis.

OBJECTIVE: Successful application of deep machine learning could reduce time-consuming and labor-intensive clinical work of calculating the amount of radiographic bone loss (RBL) in diagnosing and treatment planning for periodontitis. This study aimed to test the accuracy of RBL classification by machine learning. MATERIALS AND METHODS: A total of 236 patients with standardized full mouth radiographs were included. Each tooth from the periapical films was evaluated by three calibrated periodontists for categorization of RBL and radiographic defect morphology. Each image was pre-processed and augmented to ensure proper data balancing without data pollution, then a novel multitasking InceptionV3 model was applied. RESULTS: The model demonstrated an average accuracy of 0.87 ± 0.01 in the categorization of mild (< 15%) or severe (≥ 15%) bone loss with fivefold cross-validation. Sensitivity, specificity, positive predictive, and negative predictive values of the model were 0.86 ± 0.03, 0.88 ± 0.03, 0.88 ± 0.03, and 0.86 ± 0.02, respectively. CONCLUSIONS: Application of deep machine learning for the detection of alveolar bone loss yielded promising results in this study. Additional data would be beneficial to enhance model construction and enable better machine learning performance for clinical implementation. CLINICAL RELEVANCE: Higher accuracy of radiographic bone loss classification by machine learning can be achieved with more clinical data and proper model construction for valuable clinical application.

AI-based improvement in lung cancer detection on chest radiographs: results of a multi-reader study in NLST dataset.

OBJECTIVE: Assess if deep learning-based artificial intelligence (AI) algorithm improves reader performance for lung cancer detection on chest X-rays (CXRs). METHODS: This reader study included 173 images from cancer-positive patients (n = 98) and 346 images from cancer-negative patients (n = 196) selected from National Lung Screening Trial (NLST). Eight readers, including three radiology residents, and five board-certified radiologists, participated in the observer performance test. AI algorithm provided image-level probability of pulmonary nodule or mass on CXRs and a heatmap of detected lesions. Reader performance was compared with AUC, sensitivity, specificity, false-positives per image (FPPI), and rates of chest CT recommendations. RESULTS: With AI, the average sensitivity of readers for the detection of visible lung cancer increased for residents, but was similar for radiologists compared to that without AI (0.61 [95% CI, 0.55-0.67] vs. 0.72 [95% CI, 0.66-0.77], p = 0.016 for residents, and 0.76 [95% CI, 0.72-0.81] vs. 0.76 [95% CI, 0.72-0.81, p = 1.00 for radiologists), while false-positive findings per image (FPPI) was similar for residents, but decreased for radiologists (0.15 [95% CI, 0.11-0.18] vs. 0.12 [95% CI, 0.09-0.16], p = 0.13 for residents, and 0.24 [95% CI, 0.20-0.29] vs. 0.17 [95% CI, 0.13-0.20], p < 0.001 for radiologists). With AI, the average rate of chest CT recommendation in patients positive for visible cancer increased for residents, but was similar for radiologists (54.7% [95% CI, 48.2-61.2%] vs. 70.2% [95% CI, 64.2-76.2%], p < 0.001 for residents and 72.5% [95% CI, 68.0-77.1%] vs. 73.9% [95% CI, 69.4-78.3%], p = 0.68 for radiologists), while that in cancer-negative patients was similar for residents, but decreased for radiologists (11.2% [95% CI, 9.6-13.1%] vs. 9.8% [95% CI, 8.0-11.6%], p = 0.32 for residents and 16.4% [95% CI, 14.7-18.2%] vs. 11.7% [95% CI, 10.2-13.3%], p < 0.001 for radiologists). CONCLUSIONS: AI algorithm can enhance the performance of readers for the detection of lung cancers on chest radiographs when used as second reader. KEY POINTS: • Reader study in the NLST dataset shows that AI algorithm had sensitivity benefit for residents and specificity benefit for radiologists for the detection of visible lung cancer. • With AI, radiology residents were able to recommend more chest CT examinations (54.7% vs 70.2%, p < 0.001) for patients with visible lung cancer. • With AI, radiologists recommended significantly less proportion of unnecessary chest CT examinations (16.4% vs. 11.7%, p < 0.001) in cancer-negative patients.

Radiomics approach for survival prediction in chronic obstructive pulmonary disease.

OBJECTIVES: To apply radiomics analysis for overall survival prediction in chronic obstructive pulmonary disease (COPD), and evaluate the performance of the radiomics signature (RS). METHODS: This study included 344 patients from the Korean Obstructive Lung Disease (KOLD) cohort. External validation was performed on a cohort of 112 patients. In total, 525 chest CT-based radiomics features were semi-automatically extracted. The five most useful features for survival prediction were selected by least absolute shrinkage and selection operation (LASSO) Cox regression analysis and used to generate a RS. The ability of the RS for classifying COPD patients into high or low mortality risk groups was evaluated with the Kaplan-Meier survival analysis and Cox proportional hazards regression analysis. RESULTS: The five features remaining after the LASSO analysis were %LAA-950, AWT_Pi10_6th, AWT_Pi10_heterogeneity, %WA_heterogeneity, and VA18mm. The RS demonstrated a C-index of 0.774 in the discovery group and 0.805 in the validation group. Patients with a RS greater than 1.053 were classified into the high-risk group and demonstrated worse overall survival than those in the low-risk group in both the discovery (log-rank test, < 0.001; hazard ratio [HR], 5.265) and validation groups (log-rank test, < 0.001; HR, 5.223). For both groups, RS was significantly associated with overall survival after adjustments for patient age and body mass index. CONCLUSIONS: A radiomics approach for survival prediction and risk stratification in COPD patients is feasible, and the constructed radiomics model demonstrated acceptable performance. The RS derived from chest CT data of COPD patients was able to effectively identify those at increased risk of mortality. KEY POINTS: • A total of 525 chest CT-based radiomics features were extracted and the five radiomics features of %LAA-950, AWT_Pi10_6th, AWT_Pi10_heterogeneity, %WA_heterogeneity, and VA18mm were selected to generate a radiomics model. • A radiomics model for predicting survival of COPD patients demonstrated reliable performance with a C-index of 0.774 in the discovery group and 0.805 in the validation group. • Radiomics approach was able to effectively identify COPD patients with an increased risk of mortality, and patients assigned to the high-risk group demonstrated worse overall survival in both the discovery and validation groups.

Implications of US radiomics signature for predicting malignancy in thyroid nodules with indeterminate cytology.

OBJECTIVES: The purpose of this study was to evaluate the role of the radiomics score using US images to predict malignancy in AUS/FLUS and FN/SFN nodules. METHODS: One hundred fifty-five indeterminate thyroid nodules in 154 patients who received initial US-guided FNA for diagnostic purposes were included in this retrospective study. A representative US image of each tumor was acquired, and square ROIs covering the whole nodule were drawn using the Paint program of Windows 7. Texture features were extracted by in-house texture analysis algorithms implemented in MATLAB 2019b. The LASSO logistic regression model was used to choose the most useful predictive features, and ten-fold cross-validation was performed. Two prediction models were constructed using multivariable logistic regression analysis: one based on clinical variables, and the other based on clinical variables with the radiomics score. Predictability of the two models was assessed with the AUC of the ROC curves. RESULTS: Clinical characteristics did not significantly differ between malignant and benign nodules, except for mean nodule size. Among 730 candidate texture features generated from a single US image, 15 features were selected. Radiomics signatures were constructed with a radiomics score, using selected features. In multivariable logistic regression analysis, higher radiomics score was associated with malignancy (OR = 10.923; p < 0.001). The AUC of the malignancy prediction model composed of clinical variables with the radiomics score was significantly higher than the model composed of clinical variables alone (0.839 vs 0.583). CONCLUSIONS: Quantitative US radiomics features can help predict malignancy in thyroid nodules with indeterminate cytology.

Deep learning-based algorithm to detect primary hepatic malignancy in multiphase CT of patients at high risk for HCC.

OBJECTIVES: To develop and evaluate a deep learning-based model capable of detecting primary hepatic malignancies in multiphase CT images of patients at high risk for hepatocellular carcinoma (HCC). METHODS: A total of 1350 multiphase CT scans of 1280 hepatic malignancies (1202 HCCs and 78 non-HCCs) in 1320 patients at high risk for HCC were retrospectively analyzed. Following the delineation of the focal hepatic lesions according to reference standards, the CT scans were categorized randomly into the training (568 scans), tuning (193 scans), and test (589 scans) sets. Multiphase CT information was subjected to multichannel integration, and livers were automatically segmented before model development. A deep learning-based model capable of detecting malignancies was developed using a mask region-based convolutional neural network. The thresholds of the prediction score and the intersection over union were determined on the tuning set corresponding to the highest sensitivity with < 5 false-positive cases per CT scan. The sensitivity and the number of false-positives of the proposed model on the test set were calculated. Potential causes of false-negatives and false-positives on the test set were analyzed. RESULTS: This model exhibited a sensitivity of 84.8% with 4.80 false-positives per CT scan on the test set. The most frequent potential causes of false-negatives and false-positives were determined to be atypical enhancement patterns for HCC (71.7%) and registration/segmentation errors (42.7%), respectively. CONCLUSIONS: The proposed deep learning-based model developed to automatically detect primary hepatic malignancies exhibited an 84.8% of sensitivity with 4.80 false-positives per CT scan in the test set. KEY POINTS: • Image processing, including multichannel integration of multiphase CT and automatic liver segmentation, enabled the application of a deep learning-based model to detect primary hepatic malignancy. • Our model exhibited a sensitivity of 84.8% with a false-positive rate of 4.80 per CT scan.

Test-retest reproducibility of a deep learning-based automatic detection algorithm for the chest radiograph.

OBJECTIVES: To perform test-retest reproducibility analyses for deep learning-based automatic detection algorithm (DLAD) using two stationary chest radiographs (CRs) with short-term intervals, to analyze influential factors on test-retest variations, and to investigate the robustness of DLAD to simulated post-processing and positional changes. METHODS: This retrospective study included patients with pulmonary nodules resected in 2017. Preoperative CRs without interval changes were used. Test-retest reproducibility was analyzed in terms of median differences of abnormality scores, intraclass correlation coefficients (ICC), and 95% limits of agreement (LoA). Factors associated with test-retest variation were investigated using univariable and multivariable analyses. Shifts in classification between the two CRs were analyzed using pre-determined cutoffs. Radiograph post-processing (blurring and sharpening) and positional changes (translations in x- and y-axes, rotation, and shearing) were simulated and agreement of abnormality scores between the original and simulated CRs was investigated. RESULTS: Our study analyzed 169 patients (median age, 65 years; 91 men). The median difference of abnormality scores was 1-2% and ICC ranged from 0.83 to 0.90. The 95% LoA was approximately ± 30%. Test-retest variation was negatively associated with solid portion size (ß, - 0.50; p = 0.008) and good nodule conspicuity (ß, - 0.94; p < 0.001). A small fraction (15/169) showed discordant classifications when the high-specificity cutoff (46%) was applied to the model outputs (p = 0.04). DLAD was robust to the simulated positional change (ICC, 0.984, 0.996), but relatively less robust to post-processing (ICC, 0.872, 0.968). CONCLUSIONS: DLAD was robust to the test-retest variation. However, inconspicuous nodules may cause fluctuations of the model output and subsequent misclassifications. KEY POINTS: • The deep learning-based automatic detection algorithm was robust to the test-retest variation of the chest radiographs in general. • The test-retest variation was negatively associated with solid portion size and good nodule conspicuity. • High-specificity cutoff (46%) resulted in discordant classifications of 8.9% (15/169; p = 0.04) between the test-retest radiographs.

CT-based deep learning model to differentiate invasive pulmonary adenocarcinomas appearing as subsolid nodules among surgical candidates: comparison of the diagnostic performance with a size-based logistic model and radiologists.

OBJECTIVES: To evaluate the deep learning models for differentiating invasive pulmonary adenocarcinomas (IACs) among subsolid nodules (SSNs) considered for resection in a retrospective diagnostic cohort in comparison with a size-based logistic model and expert radiologists. METHODS: This study included 525 patients (309 women; median, 62 years) to develop models, and an independent cohort of 101 patients (57 women; median, 66 years) was used for validation. A size-based logistic model and deep learning models using 2.5-dimension (2.5D) and three-dimension (3D) CT images were developed to discriminate IAC from less invasive pathologies. Overall performance, discrimination, and calibration were assessed. Diagnostic performances of the three thoracic radiologists were compared with those of the deep learning model. RESULTS: The overall performances of the deep learning models (Brier score, 0.122 for the 2.5D DenseNet and 0.121 for the 3D DenseNet) were superior to those of the size-based logistic model (Brier score, 0.198). The area under the receiver operating characteristic curve (AUC) of the 2.5D DenseNet (0.921) was significantly higher than that of the 3D DenseNet (0.835; p = 0.037) and the size-based logistic model (0.836; p = 0.009). At equally high sensitivities of 90%, the 2.5D DenseNet showed significantly higher specificity (88.2%; all p < 0.05) and positive predictive value (97.4%; all p < 0.05) than other models. Model calibration was poor for all models (all p < 0.05). The 2.5D DenseNet had a comparable performance with the radiologists (AUC, 0.848-0.910). CONCLUSION: The 2.5D DenseNet model could be used as a highly sensitive and specific diagnostic tool to differentiate IACs among SSNs for surgical candidates. KEY POINTS: • The deep learning model developed using 2.5D DenseNet showed higher overall performance and discrimination than the size-based logistic model for the differentiation of invasive adenocarcinomas among subsolid nodules for surgical candidates. • The 2.5D DenseNet demonstrated a thoracic radiologist-level diagnostic performance and had higher specificity (88.2%) at equal sensitivities (90%) than the size-based logistic model (specificity, 52.9%). • The 2.5D DenseNet could be used to reduce potential overtreatment for the indolent subsolid nodules or to select candidates for sublobar resection instead of the standard lobectomy.

Utility of Iodine Density Perfusion Maps From Dual-Energy Spectral Detector CT in Evaluating Cardiothoracic Conditions: A Primer for the Radiologist.

OBJECTIVE. The purpose of this article is to outline the utility of iodine density maps for evaluating cardiothoracic disease and abnormalities. Multiple studies have shown that the variety of images generated from dual-energy spectral detector CT (SDCT) improve identification of cardiothoracic conditions. CONCLUSION. Understanding the technique of SDCT and being familiar with the features of different cardiothoracic conditions on iodine density map images help the radiologist make a better diagnosis.

[Radiomics-AI-based image analysis]. / Radiomics ­ KI-basierte Bildanalyse.

Radiomics deals with the statistical analysis of radiologic image data. In this article, radiomics is introduced and some of its applications are presented. In particular, an example is used to demonstrate that pathology and radiology can work together for better diagnoses. There is no denying that artificial intelligence will find its place in radiology (and pathology). Deep learning in particular will increasingly find applications. However, the impact on clinical routine is more long term and probably gradual, so AI will initially only be used in the form of specialized tools to support everyday clinical practice until methods and programs improve to the extent that AI can also take on more general diagnoses. However, this will not replace pathologists and radiologists in the long term, but rather turn them into "information specialists" who interpret the results obtained and integrate them into clinical contours.