Dataset Information

Robustness of Deep Learning Algorithm to Varying Imaging Conditions in Detecting Low Contrast Objects In Computed Tomography Phantom Images: in Comparison to 12 Radiologists.

ABSTRACT: The performance of deep learning algorithm (DLA) to that of radiologists was compared in detecting low contrast objects in CT phantom images under various imaging conditions. For training, 10,000 images were created using American College of Radiology CT phantom as the background. In half of the images, objects of 3-20 mm size and 5-30 HU contrast difference were generated in random locations. Binary responses were used as the ground truth. For testing, 640 images of Catphan® phantom were used, half of which had objects of either 5 or 9 mm size with 10 HU contrast difference. Twelve radiologists evaluated the presence of objects on a five-point scale. The performances of the DLA and radiologists were compared across different imaging conditions in terms of area under receiver operating characteristics curve (AUC). Multi-reader multi-case AUC and Hanley and McNeil tests were used. We performed post-hoc analysis using bootstrapping and verified that the DLA is less affected by the changing imaging conditions. The AUC of DLA was consistently higher than those of the radiologists across different imaging conditions (p < 0.0001), and it was less affected by varying imaging conditions. The DLA outperformed the radiologists and showed more robust performance under varying imaging conditions.

SUBMITTER: Kim HY

PROVIDER: S-EPMC7997324 | biostudies-literature | 2021 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Robustness of Deep Learning Algorithm to Varying Imaging Conditions in Detecting Low Contrast Objects In Computed Tomography Phantom Images: in Comparison to 12 Radiologists.

Kim Hae Young HY Lee Kyeorye K Chang Won W Kim Youngjune Y Lee Sungsoo S Oh Dong Yul DY Sunwoo Leonard L Lee Yoon Jin YJ Kim Young Hoon YH

Diagnostics (Basel, Switzerland) 20210228 3

The performance of deep learning algorithm (DLA) to that of radiologists was compared in detecting low contrast objects in CT phantom images under various imaging conditions. For training, 10,000 images were created using American College of Radiology CT phantom as the background. In half of the images, objects of 3-20 mm size and 5-30 HU contrast difference were generated in random locations. Binary responses were used as the ground truth. For testing, 640 images of Catphan<sup>®</sup> phantom ...[more]

PMID: 33670866

Similar Datasets

Project description:Study designRetrospective diagnostic study.ObjectiveTo automatically detect osteolytic bone metastasis lesions in the thoracolumbar region using conventional computed tomography (CT) scans, we developed a new deep learning (DL)-based computer-aided detection model.Summary of background dataRadiographic detection of bone metastasis is often difficult, even for orthopedic surgeons and diagnostic radiologists, with a consequent risk for pathologic fracture or spinal cord injury. If we can improve detection rates, we will be able to prevent the deterioration of patients' quality of life at the end stage of cancer.Materials and methodsThis study included CT scans acquired at Tokyo Medical and Dental University (TMDU) Hospital between 2016 and 2022. A total of 263 positive CT scans that included at least one osteolytic bone metastasis lesion in the thoracolumbar spine and 172 negative CT scans without bone metastasis were collected for the datasets to train and validate the DL algorithm. As a test data set, 20 positive and 20 negative CT scans were separately collected from the training and validation datasets. To evaluate the performance of the established artificial intelligence (AI) model, sensitivity, precision, F1-score, and specificity were calculated. The clinical utility of our AI model was also evaluated through observer studies involving six orthopaedic surgeons and six radiologists.ResultsOur AI model showed a sensitivity, precision, and F1-score of 0.78, 0.68, and 0.72 (per slice) and 0.75, 0.36, and 0.48 (per lesion), respectively. The observer studies revealed that our AI model had comparable sensitivity to orthopaedic or radiology experts and improved the sensitivity and F1-score of residents.ConclusionWe developed a novel DL-based AI model for detecting osteolytic bone metastases in the thoracolumbar spine. Although further improvement in accuracy is needed, the current AI model may be applied to current clinical practice.Level of evidenceLevel III.

Project description:IntroductionIn veterinary medicine, abdominal computer tomographic (CT) examinations regularly require a minimum of two scans, with a native scan (true unenhanced, TUE) as a reference for the subsequent contrast-enhanced CT scan (CECT). Spectral detector CT (SDCT) offers the possibility to calculate virtual non-contrast (VNC) images from the post-contrast scan, but this has not yet been investigated in veterinary medicine. The purpose of this study was to assess the reliability of VNC images for abdominal organs in 44 dogs without abdominal pathologies by evaluating their quantitative and qualitative parameters compared to TUE images. We hypothesized that the subtraction of iodine is sufficient in the VNC series compared to the TUE series and that the image quality of the SDCT series is superior to conventional CT images.MethodsCorresponding attenuation values in the VNC and TUE series regarding the regions of interest (ROI) in different parenchymal organs and major vessels of the abdominal cavity were assessed by means of a two one-sided t-test (TOST) and Bland-Altman plots. Additionally, the signal-to-noise ratio (SNR) was calculated for each ROI in the different series. In the second step, two board-certified veterinary radiologists made a qualitative assessment of VNC images vs. TUE images in consensus by rating the iodine subtraction, image noise, and image quality of VNC images based on a specific 5-point Likert scale.ResultsThe difference in corresponding Hounsfield units (HUs) between TUE and VNC images was less than 10 HU in 78.67% of all ROIs. Regarding the limit of less than 10 HU, in the performed TOST, significant p-values of < 0.05 were reached for the liver, spleen, pancreas, and musculature, implying equivalence of both modalities. The quality of spectral base image (SBI) data was rated equivalent to calculated conventional images in the subjective assessment by reaching an average Likert scale score of 3.2 points.DiscussionVNC images calculated from SDCT data prove a valid alternative to conventional TUE images in the abdominal organs of canine patients without abdominal pathology. VNC offers the possibility to reduce time under general anesthesia and minimize radiation exposure. Future studies are needed to prove the application of this method in clinically diseased patients.

Project description:PurposeThe deformable nature of the liver can make focal treatment challenging and is not adequately addressed with simple rigid registration techniques. More advanced registration techniques can take deformations into account (eg, biomechanical modeling) but require segmentations of the whole liver for each scan, which is a time-intensive process. We hypothesize that fully convolutional networks can be used to rapidly and accurately autosegment the liver, removing the temporal bottleneck for biomechanical modeling.Methods and materialsManual liver segmentations on computed tomography scans from 183 patients treated at our institution and 30 scans from the Medical Image Computing & Computer Assisted Intervention challenges were collected for this study. Three architectures were investigated for rapid automated segmentation of the liver (VGG-16, DeepLabv3 +, and a 3-dimensional UNet). Fifty-six cases were set aside as a final test set for quantitative model evaluation. Accuracy of the autosegmentations was assessed using Dice similarity coefficient and mean surface distance. Qualitative evaluation was also performed by 3 radiation oncologists on 50 independent cases with previously clinically treated liver contours.ResultsThe mean (minimum-maximum) mean surface distance for the test groups with the final model, DeepLabv3 +, were as follows: μContrast(N = 17): 0.99 mm (0.47-2.2), μNon_Contrast(N = 19)l: 1.12 mm (0.41-2.87), and μMiccai(N = 30)t: 1.48 mm (0.82-3.96). The qualitative evaluation showed that 30 of 50 autosegmentations (60%) were preferred to manual contours (majority voting) in a blinded comparison, and 48 of 50 autosegmentations (96%) were deemed clinically acceptable by at least 1 reviewing physician.ConclusionsThe autosegmentations were preferred compared with manually defined contours in the majority of cases. The ability to rapidly segment the liver with high accuracy achieved in this investigation has the potential to enable the efficient integration of biomechanical model-based registration into a clinical workflow.

Project description:ObjectivesQuantitative computed tomography (CT) plays an increasingly important role in phenotyping airway diseases. Lung parenchyma and airway inflammation could be quantified by contrast enhancement at CT, but its investigation by multiphasic examinations is limited. We aimed to quantify lung parenchyma and airway wall attenuation in a single contrast-enhanced spectral detector CT acquisition.MethodsFor this cross-sectional retrospective study, 234 lung-healthy patients who underwent spectral CT in four different contrast phases (non-enhanced, pulmonary arterial, systemic arterial, and venous phase) were recruited. Virtual monoenergetic images were reconstructed from 40-160 keV, on which attenuations of segmented lung parenchyma and airway walls combined for 5th-10th subsegmental generations were assessed in Hounsfield Units (HU) by an in-house software. The spectral attenuation curve slope between 40 and 100 keV (λHU) was calculated.ResultsMean lung density was higher at 40 keV compared to that at 100 keV in all groups (p < 0.001). λHU of lung attenuation was significantly higher in the systemic (1.7 HU/keV) and pulmonary arterial phase (1.3 HU/keV) compared to that in the venous phase (0.5 HU/keV) and non-enhanced (0.2 HU/keV) spectral CT (p < 0.001). Wall thickness and wall attenuation were higher at 40 keV compared to those at 100 keV for the pulmonary and systemic arterial phase (p ≤ 0.001). λHU for wall attenuation was significantly higher in the pulmonary arterial (1.8 HU/keV) and systemic arterial (2.0 HU/keV) compared to that in the venous (0.7 HU/keV) and non-enhanced (0.3 HU/keV) phase (p ≤ 0.002).ConclusionsSpectral CT may quantify lung parenchyma and airway wall enhancement with a single contrast phase acquisition, and may separate arterial and venous enhancement. Further studies are warranted to analyze spectral CT for inflammatory airway diseases.Key points• Spectral CT may quantify lung parenchyma and airway wall enhancement with a single contrast phase acquisition. • Spectral CT may separate arterial and venous enhancement of lung parenchyma and airway wall. • The contrast enhancement can be quantified by calculating the spectral attenuation curve slope from virtual monoenergetic images.

Dataset Information

Robustness of Deep Learning Algorithm to Varying Imaging Conditions in Detecting Low Contrast Objects In Computed Tomography Phantom Images: in Comparison to 12 Radiologists.

Publications

Robustness of Deep Learning Algorithm to Varying Imaging Conditions in Detecting Low Contrast Objects In Computed Tomography Phantom Images: in Comparison to 12 Radiologists.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets