Search

Purpose Using common datasets, to estimate and compare the diagnostic performance of image-based denoising techniques or iterative reconstruction algorithms for the task of detecting hepatic metastases. Methods Datasets from contrast-enhanced CT scans of the liver were provided to participants in an NIH-, AAPM- and Mayo Clinic-sponsored Low Dose CT Grand Challenge. Training data included full-dose and quarter-dose scans of the ACR CT accreditation phantom and 10 patient examinations; both images and projections were provided in the training data. Projection data were supplied in a vendor-neutral standardized format (DICOM-CT-PD). Twenty quarter-dose patient datasets were provided to each participant for testing the performance of their technique. Images were provided to sites intending to perform denoising in the image domain. Fully preprocessed projection data and statistical noise maps were provided to sites intending to perform iterative reconstruction. Upon return of the denoised or iteratively reconstructed quarter-dose images, randomized, blinded evaluation of the cases was performed using a Latin Square study design by 11 senior radiology residents or fellows, who marked the locations of identified hepatic metastases. Markings were scored against reference locations of clinically or pathologically demonstrated metastases to determine a per-lesion normalized score and a per-case normalized score (a faculty abdominal radiologist established the reference location using clinical and pathological information). Scores increased for correct detections; scores decreased for missed or incorrect detections. The winner for the competition was the entry that produced the highest total score (mean of the per-lesion and per-case normalized score). Reader confidence was used to compute a Jackknife alternative free-response receiver operating characteristic (JAFROC) figure of merit, which was used for breaking ties. Results 103 participants from 90 sites and 26 countries registered to participate. Training data were shared with 77 sites that completed the data sharing agreements. Subsequently, 41 sites downloaded the 20 test cases, which included only the 25% dose data (CTDIvol = 3.0 ± 1.8 mGy, SSDE = 3.5 ± 1.3 mGy). 22 sites submitted results for evaluation. One site provided binary images and one site provided images with severe artifacts; cases from these sites were excluded from review and the participants removed from the challenge. The mean (range) per-lesion and per-case normalized scores were −24.2% (−75.8%, 3%) and 47% (10%, 70%), respectively. Compared to reader results for commercially reconstructed quarter-dose images with no noise reduction, 11 of the 20 sites showed a numeric improvement in the mean JAFROC figure of merit. Notably two sites performed comparably to the reader results for full-dose commercial images. The study was not designed for these comparisons, so wide confidence intervals surrounded these figures of merit and the results should be used only to motivate future testing. Conclusion Infrastructure and methodology were developed to rapidly estimate observer performance for liver metastasis detection in low-dose CT examinations of the liver after either image-based denoising or iterative reconstruction. The results demonstrated large differences in detection and classification performance between noise reduction methods, although the majority of methods provided some improvement in performance relative to the commercial quarter-dose images with no noise reduction applied.

Purpose Low contrast (LC) detectability is a common test criterion for diagnostic radiologic quality control (QC) programs. Automation of this test is desirable in order to reduce human variability and to speed up analysis. However, automation is challenging due to the complexity of the human visual perception system and the ability to create algorithms that mimic this response. This paper describes the development and testing of an automated LC detection algorithm for use in the analysis of magnetic resonance (MR) images of the American College of Radiology (ACR) QC phantom. Methods The detection algorithm includes fuzzy logic decision processes and various edge detection methods to quantify LC detectability. Algorithm performance was first evaluated using a single LC phantom MR image with the addition of incremental zero mean Gaussian noise resulting in a total of 200 images. A c-statistic was calculated to determine the role of CNR to indicate when the algorithm would detect ten spokes. To evaluate inter-rater agreement between experienced observers and the algorithm, a blinded observer study was performed on 196 LC phantom images acquired from nine clinical MR scanners. The nine scanners included two MR manufacturers and two field strengths (1.5 T, 3.0 T). Inter-rater and algorithm-rater agreement was quantified using Krippendorff's alpha. Results For the Gaussian noise added data, CNR ranged from 0.519 to 11.7 with CNR being considered an excellent discriminator of algorithm performance (c-statistic = 0.9777). Reviewer scoring of the clinical phantom data resulted in an inter-rater agreement of 0.673 with the agreement between observers and algorithm equal to 0.652, both of which indicate significant agreement. Conclusions This study demonstrates that the detection of LC test patterns for MR imaging QC programs can be successfully developed and that their response can model the human visual detection system of expert MR QC readers.

Purpose: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. Methods: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. Results: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.78–0.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. Conclusions: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

Purpose: Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes. Methods: A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status. Results: The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.78–0.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813. Conclusions: Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

Efficient optimization of CT protocols demands a quantitative approach to predicting human observer performance on specific tasks at various scan and reconstruction settings. The goal of this work was to investigate how well a channelized Hotelling observer (CHO) can predict human observer performance on 2-alternative forced choice (2AFC) lesion-detection tasks at various dose levels and two different reconstruction algorithms: a filtered-backprojection (FBP) and an iterative reconstruction (IR) method.A 35 × 26 cm(2) torso-shaped phantom filled with water was used to simulate an average-sized patient. Three rods with different diameters (small: 3 mm; medium: 5 mm; large: 9 mm) were placed in the center region of the phantom to simulate small, medium, and large lesions. The contrast relative to background was -15 HU at 120 kV. The phantom was scanned 100 times using automatic exposure control each at 60, 120, 240, 360, and 480 quality reference mAs on a 128-slice scanner. After removing the three rods, the water phantom was again scanned 100 times to provide signal-absent background images at the exact same locations. By extracting regions of interest around the three rods and on the signal-absent images, the authors generated 21 2AFC studies. Each 2AFC study had 100 trials, with each trial consisting of a signal-present image and a signal-absent image side-by-side in randomized order. In total, 2100 trials were presented to both the model and human observers. Four medical physicists acted as human observers. For the model observer, the authors used a CHO with Gabor channels, which involves six channel passbands, five orientations, and two phases, leading to a total of 60 channels. The performance predicted by the CHO was compared with that obtained by four medical physicists at each 2AFC study.The human and model observers were highly correlated at each dose level for each lesion size for both FBP and IR. The Pearson's product-moment correlation coefficients were 0.986 [95% confidence interval (CI): 0.958-0.996] for FBP and 0.985 (95% CI: 0.863-0.998) for IR. Bland-Altman plots showed excellent agreement for all dose levels and lesions sizes with a mean absolute difference of 1.0% ± 1.1% for FBP and 2.1% ± 3.3% for IR.Human observer performance on a 2AFC lesion detection task in CT with a uniform background can be accurately predicted by a CHO model observer at different radiation dose levels and for both FBP and IR methods.

Purpose: The purpose of this study was to investigate the correlation between model observer and human observer performance in CT imaging for the task of lesion detection and localization when the lesion location is uncertain. Methods: Two cylindrical rods (3-mm and 5-mm diameters) were placed in a 35 × 26 cm torso-shaped water phantom to simulate lesions with −15 HU contrast at 120 kV. The phantom was scanned 100 times on a 128-slice CT scanner at each of four dose levels (CTDIvol = 5.7, 11.4, 17.1, and 22.8 mGy). Regions of interest (ROIs) around each lesion were extracted to generate images with signal-present, with each ROI containing 128 × 128 pixels. Corresponding ROIs of signal-absent images were generated from images without lesion mimicking rods. The location of the lesion (rod) in each ROI was randomly distributed by moving the ROIs around each lesion. Human observer studies were performed by having three trained observers identify the presence or absence of lesions, indicating the lesion location in each image and scoring confidence for the detection task on a 6-point scale. The same image data were analyzed using a channelized Hotelling model observer (CHO) with Gabor channels. Internal noise was added to the decision variables for the model observer study. Area under the curve (AUC) of ROC and localization ROC (LROC) curves were calculated using a nonparametric approach. The Spearman's rank order correlation between the average performance of the human observers and the model observer performance was calculated for the AUC of both ROC and LROC curves for both the 3- and 5-mm diameter lesions. Results: In both ROC and LROC analyses, AUC values for the model observer agreed well with the average values across the three human observers. The Spearman's rank order correlation values for both ROC and LROC analyses for both the 3- and 5-mm diameter lesions were all 1.0, indicating perfect rank ordering agreement of the figures of merit (AUC) between the average performance of the human observers and the model observer performance. Conclusions: In CT imaging of different sizes of low-contrast lesions (−15 HU), the performance of CHO with Gabor channels was highly correlated with human observer performance for the detection and localization tasks with uncertain lesion location in CT imaging at four clinically relevant dose levels. This suggests the ability of Gabor CHO model observers to meaningfully assess CT image quality for the purpose of optimizing scan protocols and radiation dose levels in detection and localization tasks for low-contrast lesions.

Purpose: Efficient optimization of CT protocols demands a quantitative approach to predicting human observer performance on specific tasks at various scan and reconstruction settings. The goal of this work was to investigate how well a channelized Hotelling observer (CHO) can predict human observer performance on 2-alternative forced choice (2AFC) lesion-detection tasks at various dose levels and two different reconstruction algorithms: a filtered-backprojection (FBP) and an iterative reconstruction (IR) method. Methods: A 35 × 26 cm2 torso-shaped phantom filled with water was used to simulate an average-sized patient. Three rods with different diameters (small: 3 mm; medium: 5 mm; large: 9 mm) were placed in the center region of the phantom to simulate small, medium, and large lesions. The contrast relative to background was −15 HU at 120 kV. The phantom was scanned 100 times using automatic exposure control each at 60, 120, 240, 360, and 480 quality reference mAs on a 128-slice scanner. After removing the three rods, the water phantom was again scanned 100 times to provide signal-absent background images at the exact same locations. By extracting regions of interest around the three rods and on the signal-absent images, the authors generated 21 2AFC studies. Each 2AFC study had 100 trials, with each trial consisting of a signal-present image and a signal-absent image side-by-side in randomized order. In total, 2100 trials were presented to both the model and human observers. Four medical physicists acted as human observers. For the model observer, the authors used a CHO with Gabor channels, which involves six channel passbands, five orientations, and two phases, leading to a total of 60 channels. The performance predicted by the CHO was compared with that obtained by four medical physicists at each 2AFC study. Results: The human and model observers were highly correlated at each dose level for each lesion size for both FBP and IR. The Pearson's product-moment correlation coefficients were 0.986 [95% confidence interval (CI): 0.958–0.996] for FBP and 0.985 (95% CI: 0.863–0.998) for IR. Bland-Altman plots showed excellent agreement for all dose levels and lesions sizes with a mean absolute difference of 1.0% ± 1.1% for FBP and 2.1% ± 3.3% for IR. Conclusions: Human observer performance on a 2AFC lesion detection task in CT with a uniform background can be accurately predicted by a CHO model observer at different radiation dose levels and for both FBP and IR methods.