Deep learning CT reconstruction improves liver metastases detection.

OBJECTIVES: Detection of liver metastases is crucial for guiding oncological management. Computed tomography through iterative reconstructions is widely used in this indication but has certain limitations. Deep learning image reconstructions (DLIR) use deep neural networks to achieve a significant noise reduction compared to iterative reconstructions. While reports have demonstrated improvements in image quality, their impact on liver metastases detection remains unclear. Our main objective was to determine whether DLIR affects the number of detected liver metastasis. Our secondary objective was to compare metastases conspicuity between the two reconstruction methods. METHODS: CT images of 121 patients with liver metastases were reconstructed using a 50% adaptive statistical iterative reconstruction (50%-ASiR-V), and three levels of DLIR (DLIR-low, DLIR-medium, and DLIR-high). For each reconstruction, two double-blinded radiologists counted up to a maximum of ten metastases. Visibility and contour definitions were also assessed. Comparisons between methods for continuous parameters were performed using mixed models. RESULTS: A higher number of metastases was detected by one reader with DLIR-high: 7 (2-10) (median (Q1-Q3); total 733) versus 5 (2-10), respectively for DLIR-medium, DLIR-low, and ASiR-V (p < 0.001). Ten patents were detected with more metastases with DLIR-high simultaneously by both readers and a third reader for confirmation. Metastases visibility and contour definition were better with DLIR than ASiR-V. CONCLUSION: DLIR-high enhanced the detection and visibility of liver metastases compared to ASiR-V, and also increased the number of liver metastases detected. CRITICAL RELEVANCE STATEMENT: Deep learning-based reconstruction at high strength allowed an increase in liver metastases detection compared to hybrid iterative reconstruction and can be used in clinical oncology imaging to help overcome the limitations of CT. KEY POINTS: Detection of liver metastases is crucial but limited with standard CT reconstructions. More liver metastases were detected with deep-learning CT reconstruction compared to iterative reconstruction. Deep learning reconstructions are suitable for hepatic metastases staging and follow-up.

Applying Self-Supervised Learning to Image Quality Assessment in Chest CT Imaging.

Many new reconstruction techniques have been deployed to allow low-dose CT examinations. Such reconstruction techniques exhibit nonlinear properties, which strengthen the need for a task-based measure of image quality. The Hotelling observer (HO) is the optimal linear observer and provides a lower bound of the Bayesian ideal observer detection performance. However, its computational complexity impedes its widespread practical usage. To address this issue, we proposed a self-supervised learning (SSL)-based model observer to provide accurate estimates of HO performance in very low-dose chest CT images. Our approach involved a two-stage model combining a convolutional denoising auto-encoder (CDAE) for feature extraction and dimensionality reduction and a support vector machine for classification. To evaluate this approach, we conducted signal detection tasks employing chest CT images with different noise structures generated by computer-based simulations. We compared this approach with two supervised learning-based methods: a single-layer neural network (SLNN) and a convolutional neural network (CNN). The results showed that the CDAE-based model was able to achieve similar detection performance to the HO. In addition, it outperformed both SLNN and CNN when a reduced number of training images was considered. The proposed approach holds promise for optimizing low-dose CT protocols across scanner platforms.

Comparison of supervised-learning approaches for designing a channelized observer for image quality assessment in CT.

BACKGROUND: The current paradigm for evaluating computed tomography (CT) system performance relies on a task-based approach. As the Hotelling observer (HO) provides an upper bound of observer performances in specific signal detection tasks, the literature advocates HO use for optimization purposes. However, computing the HO requires calculating the inverse of the image covariance matrix, which is often intractable in medical applications. As an alternative, dimensionality reduction has been extensively investigated to extract the task-relevant features from the raw images. This can be achieved by using channels, which yields the channelized-HO (CHO). The channels are only considered efficient when the channelized observer (CO) can approximate its unconstrained counterpart. Previous work has demonstrated that supervised learning-based methods can usually benefit CO design, either for generating efficient channels using partial least squares (PLS) or for replacing the Hotelling detector with machine-learning (ML) methods. PURPOSE: Here we investigated the efficiency of a supervised ML-algorithm used to design a CO for predicting the performance of unconstrained HO. The ML-algorithm was applied either (1) in the estimator for dimensionality reduction, or (2) in the detector function. METHODS: A channelized support vector machine (CSVM) was employed and compared against the CHO in terms of ability to predict HO performances. Both the CSVM and the CHO were estimated with channels derived from the singular value decomposition (SVD) of the system operator, principal component analysis (PCA), and PLS. The huge variety of regularization strategies proposed by CT system vendors for statistical image reconstruction (SIR) make the generalization capability of an observer a key point to consider upfront of implementation in clinical practice. To evaluate the generalization properties of the observers, we adopted a 2-step testing process: (1) achieved with the same regularization strategy (as in the training phase) and (2) performed using different reconstruction properties. We generated simulated- signal-known-exactly/background-known-exactly (SKE/BKE) tasks in which different noise structures were generated using Markov random field (MRF) regularizations using either a Green or a quadratic, function. RESULTS: The CSVM outperformed the CHO for all types of channels and regularization strategies. Furthermore, even though both COs generalized well to images reconstructed with the same regularization strategy as the images considered in the training phase, the CHO failed to generalize to images reconstructed differently whereas the CSVM managed to successfully generalize. Lastly, the proposed CSVM observer used with PCA channels outperformed the CHO with PLS channels while using a smaller training data set. CONCLUSION: These results argue for introducing the supervised-learning paradigm in the detector function rather than in the operator of the channels when designing a CO to provide an accurate estimate of HO performance. The CSVM with PCA channels proposed here could be used as a surrogate for HO in image quality assessment.

Impact of iterative reconstruction algorithms on the applicability of Fourier-based detectability index for x-ray CT imaging.

PURPOSE: The increasing application of iterative reconstruction algorithms in clinical computed tomography to improve image quality and reduce radiation dose, elicits strong interest, and needs model observers to optimize CT scanning protocols objectively and efficiently. The current paradigm for evaluating imaging system performance relies on Fourier methods, which presuppose a linear, wide-sense stationary system. Long-range correlations introduced by iterative reconstruction algorithms may narrow the applicability of Fourier techniques. Differences in the implementation of reconstruction algorithms between manufacturers add further complexity. The present work set out to quantify the errors entailed by the use of Fourier methods, which can lead to design decisions that do not correlate with detectability. METHODS: To address this question, we evaluated the noise properties and the detectability index of the ideal linear observer using the spatial approach and the Fourier-based approach. For this purpose, a homogeneous phantom was imaged on two scanners: the Revolution CT (GE Healthcare) and the Somatom Definition AS+ (Siemens Healthcare) at different exposure levels. Images were reconstructed using different strength levels of IR algorithms available on the systems considered: Adaptative Statistical Iterative Reconstruction (ASIR-V) and Sinogram Affirmed Iterative Reconstruction (SAFIRE). RESULTS: Our findings highlight that the spatial domain estimate of the detectability index is higher than the Fourier domain estimate. This trend is found to be dependent on the specific regularization used by IR algorithms as well as the signal to be detected. The eigenanalysis of the noise covariance matrix and of its circulant approximation yields explanation about the evoked trends. In particular, this analysis suggests that the predictive power of the Fourier-based ideal linear observer depends on the ability of each basis analyzed to be relevant to the signal to be detected. CONCLUSION: The applicability of Fourier techniques is dependent on the specific regularization used by IR algorithms. These results argue for verifying the assumptions made when using Fourier methods since Fourier-task-based detectability index does not always correlate with signal detectability.

Optimized radiological alert thresholds based on device dosimetric information and peak skin dose in vascular fluoroscopically guided intervention.

OBJECTIVES: The National Council on Radiation Protection (NCRP) report no. 168 recommended that during fluoroscopically guided interventions (FGIs), each patient should be monitored when one of the following thresholds is reached: an air kerma > 5 Gy, a kerma area product (KAP) > 500 Gy.cm2, a fluoroscopy time > 60 min, or a peak skin dose (PSD) > 3 Gy. Whereas PSD is the most accurate metric regarding the prevention of radiological risks, it remains the most difficult parameter to assess. We aimed to evaluate the relevance of the other, more accessible metrics and propose new optimized threshold (OT) for improved patient follow-up. METHODS: Overall, 108 patients who underwent FGI in which at least one NCRP threshold was reached and PSD was measured were considered. The correlation between all metrics was assessed using principal component analysis (PCA). ROC curves and the sensitivity/specificity of both NCRP and OT to predict PSD > 3 Gy were evaluated. RESULTS: The PCA shows that FGI can be decomposed with two components based on time and dose variables. Only KAP and kerma were correlated with PSD. The overall sensitivity and specificity of the new OT regarding KAP (67.6/93.0), kerma (97.3/81.7), and time (62.2/62.0) were better compared with NCRP thresholds (97.3/16.9, 40.5/95.4, and 21.6/74.7). CONCLUSIONS: This study shows that fluoroscopy time is not a relevant metric when used to predict PSDs > 3 Gy. By adapting KAP and kerma thresholds to predict PSD over 3 Gy, patient follow-ups following vascular FGI can be improved. KEY POINTS: • In vascular fluoroscopically guided interventions, principal component analysis demonstrates that between fluoroscopy time, KAP, and kerma, only the two last were correlated to the peak skin dose. • Optimized thresholds replacing NRCP ones obtained with ROC curves analysis were 85,451 µGy.cm2, 2938 mGy, and 41 min for KAP, kerma, and fluoroscopy time respectively. • Improvements to trigger patient follow-up after vascular fluoroscopically guided interventions may be obtained by using the optimized thresholds.