Journal of Medical Imaging最新文献

Purpose: Medical imaging-based machine learning (ML) for computer-aided diagnosis of in vivo lesions consists of two basic components or modules of (i) feature extraction from non-invasively acquired medical images and (ii) feature classification for prediction of malignancy of lesions detected or localized in the medical images. This study investigates their individual performances for diagnosis of low-dose computed tomography (CT) screening-detected lesions of pulmonary nodules and colorectal polyps.

Approach: Three feature extraction methods were investigated. One uses the mathematical descriptor of gray-level co-occurrence image texture measure to extract the Haralick image texture features (HFs). One uses the convolutional neural network (CNN) architecture to extract deep learning (DL) image abstractive features (DFs). The third one uses the interactions between lesion tissues and X-ray energy of CT to extract tissue-energy specific characteristic features (TFs). All the above three categories of extracted features were classified by the random forest (RF) classifier with comparison to the DL-CNN method, which reads the images, extracts the DFs, and classifies the DFs in an end-to-end manner. The ML diagnosis of lesions or prediction of lesion malignancy was measured by the area under the receiver operating characteristic curve (AUC). Three lesion image datasets were used. The lesions' tissue pathological reports were used as the learning labels.

Results: Experiments on the three datasets produced AUC values of 0.724 to 0.878 for the HFs, 0.652 to 0.965 for the DFs, and 0.985 to 0.996 for the TFs, compared to the DL-CNN of 0.694 to 0.964. These experimental outcomes indicate that the RF classifier performed comparably to the DL-CNN classification module and the extraction of tissue-energy specific characteristic features dramatically improved AUC value.

Conclusions: The feature extraction module is more important than the feature classification module. Extraction of tissue-energy specific characteristic features is more important than extraction of image abstractive and characteristic features.

Purpose: Tetralogy of Fallot (TOF) is a congenital heart disease, and patients undergo surgical repair early in their lives. The evaluation of TOF patients is continuous through their adulthood. The use of cardiac magnetic resonance imaging (CMR) is vital for the evaluation of TOF patients. We aim to correlate advanced MRI sequences [parametric longitudinal relaxation time (T1), extracellular volume (ECV) mapping] with cardiac functionality to provide biomarkers for the evaluation of these patients.

Methods: A complete CMR examination with the same imaging protocol was conducted in a total of 11 TOF patients and a control group of 25 healthy individuals. A Modified Look-Locker Inversion recovery (MOLLI) sequence was included to acquire the global T1 myocardial relaxation times of the left ventricular (LV) pre and post-contrast administration. Appropriate software (Circle cmr42) was used for the CMR analysis and the calculation of native, post-contrast T1, and ECV maps. A regression analysis was conducted for the correlation between global LV T1 values and right ventricular (RV) functional indices.

Results: Statistically significant results were obtained for RV cardiac index [RV_CI= -32.765 + 0.029 × T1 native; $p=0.003$ ], RV end diastolic volume [RV_EDV/BSA = -1023.872 + 0.902 × T1 native; $p=0.001$ ], and RV end systolic volume [RV_ESV/BSA = -536.704 + 0.472 × T1 native; $p=0.011$ ].

Conclusions: We further support the diagnostic importance of T1 mapping as a structural imaging tool in CMR. In addition to the well-known affected RV function in TOF patients, the LV structure is also impaired as there is a strong correlation between LV T1 mapping and RV function, evoking that the heart operates as an entity.

Purpose: Recent research explores using neural networks to reconstruct undersampled magnetic resonance imaging. Because of the complexity of the artifacts in the reconstructed images, there is a need to develop task-based approaches to image quality. We compared conventional global quantitative metrics to evaluate image quality in undersampled images generated by a neural network with human observer performance in a detection task. The purpose is to study which acceleration (2×, 3×, 4×, 5×) would be chosen with the conventional metrics and compare it to the acceleration chosen by human observer performance.

Approach: We used common global metrics for evaluating image quality: the normalized root mean squared error (NRMSE) and structural similarity (SSIM). These metrics are compared with a measure of image quality that incorporates a subtle signal for a specific task to allow for image quality assessment that locally evaluates the effect of undersampling on a signal. We used a U-Net to reconstruct under-sampled images with 2×, 3×, 4×, and 5× one-dimensional undersampling rates. Cross-validation was performed for a 500- and a 4000-image training set with both SSIM and MSE losses. A two-alternative forced choice (2-AFC) observer study was carried out for detecting a subtle signal (small blurred disk) from images with the 4000-image training set.

Results: We found that for both loss functions, the human observer performance on the 2-AFC studies led to a choice of a 2× undersampling, but the SSIM and NRMSE led to a choice of a 3× undersampling.

Conclusions: For this detection task using a subtle small signal at the edge of detectability, SSIM and NRMSE led to an overestimate of the achievable undersampling using a U-Net before a steep loss of image quality between 2×, 3×, 4×, 5× undersampling rates when compared to the performance of human observers in the detection task.

Purpose: Synthetic datasets hold the potential to offer cost-effective alternatives to clinical data, ensuring privacy protections and potentially addressing biases in clinical data. We present a method leveraging such datasets to train a machine learning algorithm applied as part of a computer-aided detection (CADe) system.

Approach: Our proposed approach utilizes clinically acquired computed tomography (CT) scans of a physical anthropomorphic phantom into which manufactured lesions were inserted to train a machine learning algorithm. We treated the training database obtained from the anthropomorphic phantom as a simplified representation of clinical data and increased the variability in this dataset using a set of randomized and parameterized augmentations. Furthermore, to mitigate the inherent differences between phantom and clinical datasets, we investigated adding unlabeled clinical data into the training pipeline.

Results: We apply our proposed method to the false positive reduction stage of a lung nodule CADe system in CT scans, in which regions of interest containing potential lesions are classified as nodule or non-nodule regions. Experimental results demonstrate the effectiveness of the proposed method; the system trained on labeled data from physical phantom scans and unlabeled clinical data achieves a sensitivity of 90% at eight false positives per scan. Furthermore, the experimental results demonstrate the benefit of the physical phantom in which the performance in terms of competitive performance metric increased by 6% when a training set consisting of 50 clinical CT scans was enlarged by the scans obtained from the physical phantom.

Conclusions: The scalability of synthetic datasets can lead to improved CADe performance, particularly in scenarios in which the size of the labeled clinical data is limited or subject to inherent bias. Our proposed approach demonstrates an effective utilization of synthetic datasets for training machine learning algorithms.

Purpose: Spatio-temporal variability in clinical fluoroscopy and cine angiography images combined with nonlinear image processing prevents the application of traditional image quality measurements in the cardiac catheterization laboratory. We aimed to develop and validate methods to measure human observer impressions of the image quality.

Approach: Multi-frame images of the thorax of a euthanized pig were acquired to provide an anatomical background. The detector dose was varied from 6 to 200 nGy (increments 2×), and 0.6 and 1.0 mm focal spots were used. Two coronary stents with/without 0.5 mm separation and a synthetic right coronary artery (RCA) with hemispherical defects were embedded into the background images as test objects. The quantitative observer ( $n=17$ ) performance was measured using a two-alternating forced-choice test of whether stents were separated and by a count of visible right coronary artery defects. Qualitative impressions of noise, spatial resolution, and overall image quality were measured using a visual analog scale (VAS). A paired $t$ -test and multinomial logistic regression model were used to identify statistically significant factors affecting the observer's impression image quality.

Results: The proportion of correct detection of stent separation and the number of reported right coronary artery defects changed significantly with detector dose increment in the 6 to 100 nGy ( $p<0.05$ ). Although a trend favored the 0.6 versus 1.0 mm focal spot for these quantitative assessments, this was insignificant. Visual analog scale measurements changed significantly with detector dose increments in the range of 24 to 100 nGy and focal spot size ( $p<0.05$ ). The application of multinomial logistic regression analysis to observer VAS scores demonstrated sensitivity matching of the paired $t$ -test applied to quantitative observer performance measurements.

Conclusions: Both quantitative and qualitative measurements of observer impression of the image quality were sensitive to image quality changes associated with changing the detector dose and focal spot size. These findings encourage future work that uses qualitative image quality measurements to assess clinical fluoroscopy and angiography image quality.

Purpose: We aim to compare the imaging performance of a cone-beam CT (CBCT) imaging system with noncircular scan protocols (sine-on-sphere) to a conventional circular orbit.

Approach: A biplane C-arm system (ARTIS Icono; Siemens Healthineers) capable of circular and noncircular CBCT acquisition was used, with the latter orbit (sine-on-sphere, "Sine Spin") executing a sinusoidal motion with $\pm 10\mathrm{deg}$ tilt amplitude over the half-scan orbit. A test phantom was used for the characterization of image uniformity, noise, noise-power spectrum (NPS), spatial resolution [modulation transfer function (MTF) in axial and oblique directions], and cone-beam artifacts. Findings were interpreted using an anthropomorphic head phantom with respect to pertinent tasks in skull base neurosurgery.

Results: The noncircular scan protocol exhibited several advantages associated with improved 3D sampling-evident in the NPS as filling of the null cone about the $f_z$ spatial frequency axis and reduction of cone-beam artifacts. The region of support at the longitudinal extrema was reduced from 16 to $\sim 12\mathrm{cm}$ at a radial distance of 6.5 cm. Circular and noncircular orbits exhibited nearly identical image uniformity and quantum noise, demonstrating cupping of $-16.7\%$ and overall noise of $\sim 27\mathrm{HU}$ . Although both the radially averaged axial MTF ( $f_{x,y}$ ) and 45 deg oblique MTF ( $f_{x,y,z}$ ) were $\sim 20\%$ lower for the noncircular orbit compared with the circular orbit at the default full reconstruction field of view (FOV), there was no difference in spatial resolution for the medium reconstruction FOV (smaller voxel size). Differences in the perceptual image quality for the anthropomorphic phantom reinforced the objective, quantitative findings, including reduced beam-hardening and cone-beam artifacts about structures of interest in the skull base.

Conclusions: Image quality differences between circular and noncircular CBCT orbits were quantitatively evaluated on a clinical system in the context of neurosurgery. The primary performance advantage for the noncircular orbit was the improved sampling and elimination of cone-beam artifacts.

Purpose: Our study investigates the potential benefits of incorporating prior anatomical knowledge into a deep learning (DL) method designed for the automated segmentation of lung lobes in chest CT scans.

Approach: We introduce an automated DL-based approach that leverages anatomical information from the lung's vascular system to guide and enhance the segmentation process. This involves utilizing a lung vessel connectivity (LVC) map, which encodes relevant lung vessel anatomical data. Our study explores the performance of three different neural network architectures within the nnU-Net framework: a standalone U-Net, a multitasking U-Net, and a cascade U-Net.

Results: Experimental findings suggest that the inclusion of LVC information in the DL model can lead to improved segmentation accuracy, particularly, in the challenging boundary regions of expiration chest CT volumes. Furthermore, our study demonstrates the potential for LVC to enhance the model's generalization capabilities. Finally, the method's robustness is evaluated through the segmentation of lung lobes in 10 cases of COVID-19, demonstrating its applicability in the presence of pulmonary diseases.

Conclusions: Incorporating prior anatomical information, such as LVC, into the DL model shows promise for enhancing segmentation performance, particularly in the boundary regions. However, the extent of this improvement has limitations, prompting further exploration of its practical applicability.

Purpose: Segmentation of ovarian/adnexal masses from surrounding tissue on ultrasound images is a challenging task. The separation of masses into different components may also be important for radiomic feature extraction. Our study aimed to develop an artificial intelligence-based automatic segmentation method for transvaginal ultrasound images that (1) outlines the exterior boundary of adnexal masses and (2) separates internal components.

Approach: A retrospective ultrasound imaging database of adnexal masses was reviewed for exclusion criteria at the patient, mass, and image levels, with one image per mass. The resulting 54 adnexal masses (36 benign/18 malignant) from 53 patients were separated by patient into training (26 benign/12 malignant) and independent test (10 benign/6 malignant) sets. U-net segmentation performance on test images compared to expert detailed outlines was measured using the Dice similarity coefficient (DSC) and the ratio of the Hausdorff distance to the effective diameter of the outline ( $R_{\mathrm{HD}\text{-}D}$ ) for each mass. Subsequently, in discovery mode, a two-level fuzzy c-means (FCM) unsupervised clustering approach was used to separate the pixels within masses belonging to hypoechoic or hyperechoic components.

Results: The DSC (median [95% confidence interval]) was 0.91 [0.78, 0.96], and $R_{\mathrm{HD}\text{-}D}$ was 0.04 [0.01, 0.12], indicating strong agreement with expert outlines. Clinical review of the internal separation of masses into echogenic components demonstrated a strong association with mass characteristics.

Conclusion: A combined U-net and FCM algorithm for automatic segmentation of adnexal masses and their internal components achieved excellent results compared with expert outlines and review, supporting future radiomic feature-based classification of the masses by components.

Purpose: Endometrial cancer (EC) is one of the most common types of cancer affecting women. While the hematoxylin-and-eosin (H&E) staining remains the standard for histological analysis, the immunohistochemistry (IHC) method provides molecular-level visualizations. Our study proposes a digital staining method to generate the hematoxylin-3,3'-diaminobenzidine (H-DAB) IHC stain of Ki-67 for the whole slide image of the EC tumor from its H&E stain counterpart.

Approach: We employed a color unmixing technique to yield stain density maps from the optical density (OD) of the stains and utilized the U-Net for end-to-end inference. The effectiveness of the proposed method was evaluated using the Pearson correlation between the digital and physical stain's labeling index (LI), a key metric indicating tumor proliferation. Two different cross-validation schemes were designed in our study: intraslide validation and cross-case validation (CCV). In the widely used intraslide scheme, the training and validation sets might include different regions from the same slide. The rigorous CCV validation scheme strictly prohibited any validation slide from contributing to training.

Results: The proposed method yielded a high-resolution digital stain with preserved histological features, indicating a reliable correlation with the physical stain in terms of the Ki-67 LI. In the intraslide scheme, using intraslide patches resulted in a biased accuracy (e.g., $R=0.98$ ) significantly higher than that of CCV. The CCV scheme retained a fair correlation (e.g., $R=0.66$ ) between the LIs calculated from the digital stain and its physical IHC counterpart. Inferring the OD of the IHC stain from that of the H&E stain enhanced the correlation metric, outperforming that of the baseline model using the RGB space.

Conclusions: Our study revealed that molecule-level insights could be obtained from H&E images using deep learning. Furthermore, the improvement brought via OD inference indicated a possible method for creating more generalizable models for digital staining via per-stain analysis.