MRI-based radiomic features of the urinary bladder wall identify patients with moderate-to-severe international prostate symptom score.

BACKGROUND: The International Prostate Symptom Score (IPSS) is a patient-reported measurement to assess the lower urinary tract symptoms of bladder outlet obstruction. Bladder outlet obstruction induces molecular and morphological alterations in the urothelium, suburothelium, detrusor smooth muscle cells, detrusor extracellular matrix, and nerves. We sought to analyze MRI-based radiomics features of the urinary bladder wall and their association with IPSS. METHOD: In this retrospective study, 87 patients who had pelvic MRI scans were identified. A biomarker discovery approach based on the optimal biomarker (OBM) method was used to extract features of the bladder wall from MR images, including morphological, intensity-based, and texture-based features, along with clinical variables. Mathematical models were created using subsets of features and evaluated based on their ability to discriminate between low and moderate-to-severe IPSS (less than 8 vs. equal to or greater than 8). RESULTS: Of the 7,666 features per patient, four highest-ranking optimal features were derived (all texture-based features), which provided a classification accuracy of 0.80 with a sensitivity, specificity, and area under the receiver operating characteristic curve of 0.81, 0.81, and 0.87, respectively. CONCLUSION: A highly independent set of urinary bladder wall features derived from MRI scans were able to discriminate between patients with low vs. moderate-to-severe IPSS with accuracy of 80%. Such differences in MRI-based properties of the bladder wall in patients with varying IPSS's might reflect differences in underlying molecular and morphological alterations that occur in the setting of chronic bladder outlet obstruction.

Preoperative Computed Tomography Morphological Features Indicative of Incisional Hernia Formation After Abdominal Surgery.

OBJECTIVE: To investigate key morphometric features identifiable on routine preoperative computed tomography (CT) imaging indicative of incisional hernia (IH) formation following abdominal surgery. BACKGROUND: IH is a pervasive surgical disease that impacts all surgical disciplines operating in the abdominopelvic region and affecting 13% of patients undergoing abdominal surgery. Despite the significant costs and disability associated with IH, there is an incomplete understanding of the pathophysiology of hernia. METHODS: A cohort of patients (n=21,501) that underwent colorectal surgery was identified, and clinical data and demographics were extracted, with a primary outcome of IH. Two datasets of case-control matched pairs were created for feature measurement, classification, and testing. Morphometric linear and volumetric measurements were extracted as features from anonymized preoperative abdominopelvic CT scans. Multivariate Pearson testing was performed to assess correlations among features. Each feature's ability to discriminate between classes was evaluated using 2-sided paired t testing. A support vector machine was implemented to determine the predictive accuracy of the features individually and in combination. RESULTS: Two hundred and twelve patients were analyzed (106 matched pairs). Of 117 features measured, 21 features were capable of discriminating between IH and non-IH patients. These features are categorized into three key pathophysiologic domains: 1) structural widening of the rectus complex, 2) increased visceral volume, 3) atrophy of abdominopelvic skeletal muscle. Individual prediction accuracy ranged from 0.69 to 0.78 for the top 3 features among 117. CONCLUSIONS: Three morphometric domains identifiable on routine preoperative CT imaging were associated with hernia: widening of the rectus complex, increased visceral volume, and body wall skeletal muscle atrophy. This work highlights an innovative pathophysiologic mechanism for IH formation hallmarked by increased intra-abdominal pressure and compromise of the rectus complex and abdominopelvic skeletal musculature.

Skeletal muscle adiposity and outcomes in candidates for lung transplantation: a lung transplant body composition cohort study.

CT measurement of body composition may improve lung transplant candidate selection. We assessed whether skeletal muscle adipose deposition on abdominal and thigh CT scans was associated with 6 min walk distance (6MWD) and wait-list survival in lung transplant candidates. Each ½-SD decrease in abdominal muscle attenuation (indicating greater lipid content) was associated with 14 m decrease in 6MWD (95% CI -20 to -8) and 20% increased risk of death or delisting (95% CI 10% to 40%). Each ½-standard deviation decrease in thigh muscle attenuation was associated with 15 m decrease in 6MWD (95% CI -21 to -10). CT imaging may improve candidate risk stratification.

Understanding Respiratory Restrictions as a Function of the Scoliotic Spinal Curve in Thoracic Insufficiency Syndrome: A 4D Dynamic MR Imaging Study.

BACKGROUND: Over the past 100 years, many procedures have been developed for correcting restrictive thoracic deformities which cause thoracic insufficiency syndrome. However, none of them have been assessed by a robust metric incorporating thoracic dynamics. In this paper, we investigate the relationship between radiographic spinal curve and lung volumes derived from thoracic dynamic magnetic resonance imaging (dMRI). Our central hypothesis is that different anteroposterior major spinal curve types induce different restrictions on the left and right lungs and their dynamics. METHODS: Retrospectively, we included 25 consecutive patients with thoracic insufficiency syndrome (14 neuromuscular, 7 congenital, 4 other) who underwent vertical expandable prosthetic titanium rib surgery and received preimplantation and postimplantation thoracic dMRI for clinical care. We measured thoracic and lumbar major curves by the Cobb measurement method from anteroposterior radiographs and classified the curves as per Scoliosis Research Society (SRS)-defined curve types. From 4D dMRI images, we derived static volumes and tidal volumes of left and right lung, along with left and right chest wall and left and right diaphragm tidal volumes (excursions), and analyzed their association with curve type and major curve angles. RESULTS: Thoracic and lumbar major curve angles ranged from 0 to 136 and 0 to 116 degrees, respectively. A dramatic postoperative increase in chest wall and diaphragmatic excursion was seen qualitatively. All components of volume increased postoperatively by up to 533%, with a mean of 70%. As the major curve, main thoracic curve (MTC) was associated with higher tidal volumes (effect size range: 0.7 to 1.0) than thoracolumbar curve (TLC) in preoperative and postoperative situation. Neither MTC nor TLC showed any meaningful correlation between volumes and major curve angles preoperatively or postoperatively. Moderate correlations (0.65) were observed for specific conditions like volumes at end-inspiration or end-expiration. CONCLUSIONS: The relationships between component tidal volumes and the spinal curve type are complex and are beyond intuitive reasoning and guessing. TLC has a much greater influence on restricting chest wall and diaphragm tidal volumes than MTC. Major curve angles are not indicative of passive resting volumes or tidal volumes. LEVEL OF EVIDENCE: Level II-diagnostic.

Quantitative Dynamic Thoracic MRI: Application to Thoracic Insufficiency Syndrome in Pediatric Patients.

Background Available methods to quantify regional dynamic thoracic function in thoracic insufficiency syndrome (TIS) are limited. Purpose To evaluate the use of quantitative dynamic MRI to depict changes in regional dynamic thoracic function before and after surgical correction of TIS. Materials and Methods Images from free-breathing dynamic MRI in pediatric patients with TIS (July 2009-August 2015) were retrospectively evaluated before and after surgical correction by using vertical expandable prosthetic titanium rib (VEPTR). Eleven volumetric parameters were derived from lung, chest wall, and diaphragm segmentations, and parameter changes before versus after operation were correlated with changes in clinical parameters. Paired analysis from Student t test on MRI parameters and clinical parameters was performed to detect if changes (from preoperative to postoperative condition) were statistically significant. Results Left and right lung volumes at end inspiration and end expiration increased substantially after operation in pediatric patients with thoracic insufficiency syndrome, especially right lung volume with 22.9% and 26.3% volume increase at end expiration (P = .001) and end inspiration (P = .002), respectively. The average lung tidal volumes increased after operation for TIS; there was a 43.8% and 55.3% increase for left lung tidal volume and right lung tidal volume (P < .001 for both), respectively. However, clinical parameters did not show significant changes from pre- to posttreatment states. Thoracic and lumbar Cobb angle were poor predictors of MRI tidal volumes (chest wall, diaphragm, and left and right separately), but assisted ventilation rating and forced vital capacity showed moderate correlations with tidal volumes (chest wall, diaphragm, and left and right separately). Conclusion Vertical expandable prosthetic titanium rib operation was associated with postoperative increases in all components of tidal volume (left and right chest wall and diaphragm, and left and right lung tidal volumes) measured at MRI. Clinical parameters did not demonstrate improvements in postoperative tidal volumes. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Paltiel in this issue.

Predicting the effort required to manually mend auto-segmentations.

Auto-segmentation is one of the critical and foundational steps for medical image analysis. The quality of auto-segmentation techniques influences the efficiency of precision radiology and radiation oncology since high-quality auto-segmentations usually require limited manual correction. Segmentation metrics are necessary and important to evaluate auto-segmentation results and guide the development of auto-segmentation techniques. Currently widely applied segmentation metrics usually compare the auto-segmentation with the ground truth in terms of the overlapping area (e.g., Dice Coefficient (DC)) or the distance between boundaries (e.g., Hausdorff Distance (HD)). However, these metrics may not well indicate the manual mending effort required when observing the auto-segmentation results in clinical practice. In this article, we study different segmentation metrics to explore the appropriate way of evaluating auto-segmentations with clinical demands. The mending time for correcting auto-segmentations by experts is recorded to indicate the required mending effort. Five well-defined metrics, the overlapping area-based metric DC, the segmentation boundary distance-based metric HD, the segmentation boundary length-based metrics surface DC (surDC) and added path length (APL), and a newly proposed hybrid metric Mendability Index (MI) are discussed in the correlation analysis experiment and regression experiment. In addition to these explicitly defined metrics, we also preliminarily explore the feasibility of using deep learning models to predict the mending effort, which takes segmentation masks and the original images as the input. Experiments are conducted using datasets of 7 objects from three different institutions, which contain the original computed tomography (CT) images, the ground truth segmentations, the auto-segmentations, the corrected segmentations, and the recorded mending time. According to the correlation analysis and regression experiments for the five well-defined metrics, the variety of MI shows the best performance to indicate the mending effort for sparse objects, while the variety of HD works best when assessing the mending effort for non-sparse objects. Moreover, the deep learning models could well predict efforts required to mend auto-segmentations, even without the need of ground truth segmentations, demonstrating the potential of a novel and easy way to evaluate and boost auto-segmentation techniques.

Predicting human effort needed to correct auto-segmentations.

Medical image auto-segmentation techniques are basic and critical for numerous image-based analysis applications that play an important role in developing advanced and personalized medicine. Compared with manual segmentations, auto-segmentations are expected to contribute to a more efficient clinical routine and workflow by requiring fewer human interventions or revisions to auto-segmentations. However, current auto-segmentation methods are usually developed with the help of some popular segmentation metrics that do not directly consider human correction behavior. Dice Coefficient (DC) focuses on the truly-segmented areas, while Hausdorff Distance (HD) only measures the maximal distance between the auto-segmentation boundary with the ground truth boundary. Boundary length-based metrics such as surface DC (surDC) and Added Path Length (APL) try to distinguish truly-predicted boundary pixels and wrong ones. It is uncertain if these metrics can reliably indicate the required manual mending effort for application in segmentation research. Therefore, in this paper, the potential use of the above four metrics, as well as a novel metric called Mendability Index (MI), to predict the human correction effort is studied with linear and support vector regression models. 265 3D computed tomography (CT) samples for 3 objects of interest from 3 institutions with corresponding auto-segmentations and ground truth segmentations are utilized to train and test the prediction models. The five-fold cross-validation experiments demonstrate that meaningful human effort prediction can be achieved using segmentation metrics with varying prediction errors for different objects. The improved variant of MI, called MIhd, generally shows the best prediction performance, suggesting its potential to indicate reliably the clinical value of auto-segmentations.

Optimal strategies for modeling anatomy in a hybrid intelligence framework for auto-segmentation of organs.

Organ segmentation is a fundamental requirement in medical image analysis. Many methods have been proposed over the past 6 decades for segmentation. A unique feature of medical images is the anatomical information hidden within the image itself. To bring natural intelligence (NI) in the form of anatomical information accumulated over centuries into deep learning (DL) AI methods effectively, we have recently introduced the idea of hybrid intelligence (HI) that combines NI and AI and a system based on HI to perform medical image segmentation. This HI system has shown remarkable robustness to image artifacts, pathology, deformations, etc. in segmenting organs in the Thorax body region in a multicenter clinical study. The HI system utilizes an anatomy modeling strategy to encode NI and to identify a rough container region in the shape of each object via a non-DL-based approach so that DL training and execution are applied only to the fuzzy container region. In this paper, we introduce several advances related to modeling of the NI component so that it becomes substantially more efficient computationally, and at the same time, is well integrated with the DL portion (AI component) of the system. We demonstrate a 9-40 fold computational improvement in the auto-segmentation task for radiation therapy (RT) planning via clinical studies obtained from 4 different RT centers, while retaining state-of-the-art accuracy of the previous system in segmenting 11 objects in the Thorax body region.

GA-Net: A geographical attention neural network for the segmentation of body torso tissue composition.

PURPOSE: Body composition analysis (BCA) of the body torso plays a vital role in the study of physical health and pathology and provides biomarkers that facilitate the diagnosis and treatment of many diseases, such as type 2 diabetes mellitus, cardiovascular disease, obstructive sleep apnea, and osteoarthritis. In this work, we propose a body composition tissue segmentation method that can automatically delineate those key tissues, including subcutaneous adipose tissue, skeleton, skeletal muscle tissue, and visceral adipose tissue, on positron emission tomography/computed tomography scans of the body torso. METHODS: To provide appropriate and precise semantic and spatial information that is strongly related to body composition tissues for the deep neural network, first we introduce a new concept of the body area and integrate it into our proposed segmentation network called Geographical Attention Network (GA-Net). The body areas are defined following anatomical principles such that the whole body torso region is partitioned into three non-overlapping body areas. Each body composition tissue of interest is fully contained in exactly one specific minimal body area. Secondly, the proposed GA-Net has a novel dual-decoder schema that is composed of a tissue decoder and an area decoder. The tissue decoder segments the body composition tissues, while the area decoder segments the body areas as an auxiliary task. The features of body areas and body composition tissues are fused through a soft attention mechanism to gain geographical attention relevant to the body tissues. Thirdly, we propose a body composition tissue annotation approach that takes the body area labels as the region of interest, which significantly improves the reproducibility, precision, and efficiency of delineating body composition tissues. RESULTS: Our evaluations on 50 low-dose unenhanced CT images indicate that GA-Net outperforms other architectures statistically significantly based on the Dice metric. GA-Net also shows improvements for the 95% Hausdorff Distance metric in most comparisons. Notably, GA-Net exhibits more sensitivity to subtle boundary information and produces more reliable and robust predictions for such structures, which are the most challenging parts to manually mend in practice, with potentially significant time-savings in the post hoc correction of these subtle boundary placement errors. Due to the prior knowledge provided from body areas, GA-Net achieves competitive performance with less training data. Our extension of the dual-decoder schema to TransUNet and 3D U-Net demonstrates that the new schema significantly improves the performance of these classical neural networks as well. Heatmaps obtained from attention gate layers further illustrate the geographical guidance function of body areas for identifying body tissues. CONCLUSIONS: (i) Prior anatomic knowledge supplied in the form of appropriately designed anatomic container objects significantly improves the segmentation of bodily tissues. (ii) Of particular note are the improvements achieved in the delineation of subtle boundary features which otherwise would take much effort for manual correction. (iii) The method can be easily extended to existing networks to improve their accuracy for this application.

Interactive Segmentation of Lung Tissue and Lung Excursion in Thoracic Dynamic MRI Based on Shape-guided Convolutional Neural Networks.

Purpose: Lung tissue and lung excursion segmentation in thoracic dynamic magnetic resonance imaging (dMRI) is a critical step for quantitative analysis of thoracic structure and function in patients with respiratory disorders such as Thoracic Insufficiency Syndrome (TIS). However, the complex variability of intensity and shape of anatomical structures and the low contrast between the lung and surrounding tissue in MR images seriously hamper the accuracy and robustness of automatic segmentation methods. In this paper, we develop an interactive deep-learning based segmentation system to solve this problem. Material & Methods: Considering the significant difference in lung morphological characteristics between normal subjects and TIS subjects, we utilized two independent data sets of normal subjects and TIS subjects to train and test our model. 202 dMRI scans from 101 normal pediatric subjects and 92 dMRI scans from 46 TIS pediatric subjects were acquired for this study and were randomly divided into training, validation, and test sets by an approximate ratio of 5:1:4. First, we designed an interactive region of interest (ROI) strategy to detect the lung ROI in dMRI for accelerating the training speed and reducing the negative influence of tissue located far away from the lung on lung segmentation. Second, we utilized a modified 2D U-Net to segment the lung tissue in lung ROIs, in which the adjacent slices are utilized as the input data to take advantage of the spatial information of the lungs. Third, we extracted the lung shell from the lung segmentation results as the shape feature and inputted the lung ROIs with shape feature into another modified 2D U-Net to segment the lung excursion in dMRI. To evaluate the performance of our approach, we computed the Dice coefficient (DC) and max-mean Hausdorff distance (MM-HD) between manual and automatic segmentations. In addition, we utilized Coefficient of Variation (CV) to assess the variability of our method on repeated dMRI scans and the differences of lung tidal volumes computed from the manual and automatic segmentation results. Results: The proposed system yielded mean Dice coefficients of 0.96±0.02 and 0.89±0.05 for lung segmentation in dMRI of normal subjects and TIS subjects, respectively, demonstrating excellent agreement with manual delineation results. The Coefficient of Variation and p-values show that the estimated lung tidal volumes of our approach are statistically indistinguishable from those derived by manual segmentations. Conclusions: The proposed approach can be applied to lung tissue and lung excursion segmentation from dynamic MR images with high accuracy and efficiency. The proposed approach has the potential to be utilized in the assessment of patients with TIS via dMRI routinely.

An Anthropomorphic Diagnosis System of Pulmonary Nodules using Weak Annotation-Based Deep Learning.

Purpose: To develop an anthropomorphic diagnosis system of pulmonary nodules (PN) based on Deep learning (DL) that is trained by weak annotation data and has comparable performance to full-annotation based diagnosis systems. Methods: The proposed system uses deep learning (DL) models to classify PNs (benign vs. malignant) with weak annotations, which eliminates the need for time-consuming and labor-intensive manual annotations of PNs. Moreover, the PN classification networks, augmented with handcrafted shape features acquired through the ball-scale transform technique, demonstrate capability to differentiate PNs with diverse labels, including pure ground-glass opacities, part-solid nodules, and solid nodules. Results: The experiments were conducted on two lung CT datasets: (1) public LIDC-IDRI dataset with 1,018 subjects, (2) In-house dataset with 2740 subjects. Through 5-fold cross-validation on two datasets, the system achieved the following results: (1) an Area Under Curve (AUC) of 0.938 for PN localization and an AUC of 0.912 for PN differential diagnosis on the LIDC-IDRI dataset of 814 testing cases, (2) an AUC of 0.943 for PN localization and an AUC of 0.815 for PN differential diagnosis on the in-house dataset of 822 testing cases. These results demonstrate comparable performance to full annotation-based diagnosis systems. Conclusions: Our system can efficiently localize and differentially diagnose PNs even in resource-limited environments with good robustness across different grade and morphology sub-groups in the presence of deviations due to the size, shape, and texture of the nodule, indicating its potential for future clinical translation. Summary: An anthropomorphic diagnosis system of pulmonary nodules (PN) based on deep learning and weak annotation was found to achieve comparable performance to full-annotation dataset-based diagnosis systems, significantly reducing the time and the cost associated with the annotation. Key Points: A fully automatic system for the diagnosis of PN in CT scans using a suitable deep learning model and weak annotations was developed to achieve comparable performance (AUC = 0.938 for PN localization, AUC = 0.912 for PN differential diagnosis) with the full-annotation based deep learning models, reducing around 30%∼80% of annotation time for the experts.The integration of the hand-crafted feature acquired from human experts (natural intelligence) into the deep learning networks and the fusion of the classification results of multi-scale networks can efficiently improve the PN classification performance across different diameters and sub-groups of the nodule.

Quantifying Normal Diaphragmatic Motion and Shape and their Developmental Changes via Dynamic MRI.

Background: The diaphragm is a critical structure in respiratory function, yet in-vivo quantitative description of its motion available in the literature is limited. Research Question: How to quantitatively describe regional hemi-diaphragmatic motion and curvature via free-breathing dynamic magnetic resonance imaging (dMRI)? Study Design and Methods: In this prospective cohort study we gathered dMRI images of 177 normal children and segmented hemi-diaphragm domes in end-inspiration and end-expiration phases of the constructed 4D image. We selected 25 points uniformly located on each 3D hemi-diaphragm surface. Based on the motion and local shape of hemi-diaphragm at these points, we computed the velocities and sagittal and coronal curvatures in 13 regions on each hemi-diaphragm surface and analyzed the change in these properties with age and gender. Results: Our cohort consisted of 94 Females, 6-20 years (12.09 + 3.73), and 83 Males, 6-20 years (11.88 + 3.57). We observed velocity range: ∼2mm/s to ∼13mm/s; Curvature range -Sagittal: ∼3m -1 to ∼27m -1 ; Coronal: ∼6m -1 to ∼20m -1 . There was no significant difference in velocity between genders, although the pattern of change in velocity with age was different for the two groups. Strong correlations in velocity were observed between homologous regions of right and left hemi-diaphragms. There was no significant difference in curvatures between genders or change in curvatures with age. Interpretation: Regional motion/curvature of the 3D diaphragmatic surface can be estimated using free-breathing dynamic MRI. Our analysis sheds light on here-to-fore unknown matters such as how the pediatric 3D hemi-diaphragm motion/shape varies regionally, between right and left hemi-diaphragms, between genders, and with age.

Do Rib-Based Anchors Impair Chest Wall Motion in Early Onset Scoliosis (EOS)?

Purpose: There is a concern in pediatric surgery practice that rib-based fixation may limit chest wall motion in early onset scoliosis (EOS). The purpose of this study is to address the above concern by assessing the contribution of chest wall excursion to respiration before and after surgery. Methods: Quantitative dynamic magnetic resonance imaging (QdMRI) is performed on EOS patients (before and after surgery) and normal children in this retrospective study. QdMRI is purely an image-based approach and allows free breathing image acquisition. Tidal volume parameters for chest walls (CWtv) and hemi-diaphragms (Dtv) were analyzed on concave and convex sides of the spinal curve. EOS patients (1-14 years) and normal children (5-18 years) were enrolled, with an average interval of two years for dMRI acquisition before and after surgery. Results: CWtv significantly increased after surgery in the global comparison including all EOS patients (p < 0.05). For main thoracic curve (MTC) EOS patients, CWtv significantly improved by 50.24% (concave side) and 35.17% (convex side) after age correction (p < 0.05) after surgery. The average ratio of Dtv to CWtv on the convex side in MTC EOS patients was not significantly different from that in normal children (p=0.78), although the concave side showed the difference to be significant. Conclusion: Chest wall component tidal volumes in EOS patients measured via QdMRI did not decrease after rib-based surgery, suggesting that rib-based fixation does not impair chest wall motion in pediatric patients with EOS.

Virtual Growing Child (VGC): A general normative comparative system via quantitative dynamic MRI for quantifying pediatric regional respiratory anomalies with application in thoracic insufficiency syndrome (TIS).

Background: A normative database of regional respiratory structure and function in healthy children does not exist. Methods: VGC provides a database with four categories of regional respiratory measurement parameters including morphological, architectural, dynamic, and developmental. The database has 3,820 3D segmentations (around 100,000 2D slices with segmentations). Age and gender group analysis and comparisons for healthy children were performed using those parameters via two-sided t-testing to compare mean measurements, for left and right sides at end-inspiration (EI) and end-expiration (EE), for different age and gender specific groups. We also apply VGC measurements for comparison with TIS patients via an extrapolation approach to estimate the association between measurement and age via a linear model and to predict measurements for TIS patients. Furthermore, we check the Mahalanobis distance between TIS patients and healthy children of corresponding age. Findings: The difference between male and female groups (10-12 years) behave differently from that in other age groups which is consistent with physiology/natural growth behavior related to adolescence with higher right lung and right diaphragm tidal volumes for females(p<0.05). The comparison of TIS patients before and after surgery show that the right and left components are not symmetrical, and the left side diaphragm height and tidal volume has been significantly improved after surgery (p <0.05). The left lung volume at EE, and left diaphragm height at EI of TIS patients after surgery are closer to the normal children with a significant smaller Mahalanobis distance (MD) after surgery (p<0.05). Interpretation: The VGC system can serve as a reference standard to quantify regional respiratory abnormalities on dMRI in young patients with various respiratory conditions and facilitate treatment planning and response assessment. Funding: The grant R01HL150147 from the National Institutes of Health (PI Udupa).

Assessment of 3D hemi-diaphragmatic motion via free-breathing dynamic MRI in pediatric thoracic insufficiency syndrome.

Purpose: Thoracic insufficiency syndrome (TIS) affects ventilatory function due to spinal and thoracic deformities limiting lung space and diaphragmatic motion. Corrective orthopedic surgery can be used to help normalize skeletal anatomy, restoring lung space and diaphragmatic motion. This study employs free-breathing dynamic MRI (dMRI) and quantifies the 3D motion of each hemi-diaphragm surface in normal and TIS patients, and evaluates effects of surgical intervention. Materials and Methods: In a retrospective study of 149 pediatric patients with TIS and 190 healthy children, we constructed 4D images from free-breathing dMRI and manually delineated the diaphragm at end-expiration (EE) and end-inspiration (EI) time points. We automatically selected 25 points uniformly on each hemi-diaphragm surface, calculated their relative velocities between EE and EI, and derived mean velocities in 13 homologous regions for each hemi-diaphragm to provide measures of regional 3D hemi-diaphragm motion. T-testing was used to compare velocity changes before and after surgery, and to velocities in healthy controls. Results: The posterior-central region of the right hemi-diaphragm exhibited the highest average velocity post-operatively. Posterior regions showed greater velocity changes after surgery in both right and left hemi-diaphragms. Surgical reduction of thoracic Cobb angle displayed a stronger correlation with changes in diaphragm velocity than reduction in lumbar Cobb angle. Following surgery, the anterior regions of the left hemi-diaphragm tended to approach a more normal state. Conclusion: Quantification of regional motion of the 3D diaphragm surface in normal subjects and TIS patients via free-breathing dMRI is feasible. Derived measurements can be assessed in comparison to normal subjects to study TIS and the effects of surgery.

Auto-segmentation of thoraco-abdominal organs in pediatric dynamic MRI.

Purpose: Analysis of the abnormal motion of thoraco-abdominal organs in respiratory disorders such as the Thoracic Insufficiency Syndrome (TIS) and scoliosis such as adolescent idiopathic scoliosis (AIS) or early onset scoliosis (EOS) can lead to better surgical plans. We can use healthy subjects to find out the normal architecture and motion of a rib cage and associated organs and attempt to modify the patient's deformed anatomy to match to it. Dynamic magnetic resonance imaging (dMRI) is a practical and preferred imaging modality for capturing dynamic images of healthy pediatric subjects. In this paper, we propose an auto-segmentation set-up for the lungs, kidneys, liver, spleen, and thoraco-abdominal skin in these dMRI images which have their own challenges such as poor contrast, image non-standardness, and similarity in texture amongst gas, bone, and connective tissue at several inter-object interfaces. Methods: The segmentation set-up has been implemented in two steps: recognition and delineation using two deep neural network (DL) architectures (say DL-R and DL-D) for the recognition step and delineation step, respectively. The encoder-decoder framework in DL-D utilizes features at four different resolution levels to counter the challenges involved in the segmentation. We have evaluated on dMRI sagittal acquisitions of 189 (near-)normal subjects. The spatial resolution in all dMRI acquisitions is 1.46 mm in a sagittal slice and 6.00 mm between sagittal slices. We utilized images of 89 (10) subjects at end inspiration for training (validation). For testing we experimented with three scenarios: utilizing (1) the images of 90 (=189-89-10) different (remaining) subjects at end inspiration for testing, (2) the images of the aforementioned 90 subjects at end expiration for testing, and (3) the images of the aforesaid 99 (=89+10) subjects but at end expiration for testing. In some situations, we can take advantage of already available ground truth (GT) of a subject at a particular respiratory phase to automatically segment the object in the image of the same subject at a different respiratory phase and then refining the segmentation to create the final GT. We anticipate that this process of creating GT would require minimal post hoc correction. In this spirit, we conducted separate experiments where we assume to have the ground truth of the test subjects at end expiration for scenario (1), end inspiration for (2), and end inspiration for (3). Results: Amongst these three scenarios of testing, for the DL-R, we achieve a best average location error (LE) of about 1 voxel for the lungs, kidneys, and spleen and 1.5 voxels for the liver and the thoraco- abdominal skin. The standard deviation (SD) of LE is about 1 or 2 voxels. For the delineation approach, we achieve an average Dice coefficient (DC) of about 0.92 to 0.94 for the lungs, 0.82 for the kidneys, 0.90 for the liver, 0.81 for the spleen, and 0.93 for the thoraco-abdominal skin. The SD of DC is lower for the lungs, liver, and the thoraco-abdominal skin, and slightly higher for the spleen and kidneys. Conclusions: Motivated by applications in surgical planning for disorders such as TIS, AIS, and EOS, we have shown an auto-segmentation system for thoraco-abdominal organs in dMRI acquisitions. This proposed setup copes with the challenges posed by low resolution, motion blur, inadequate contrast, and image intensity non-standardness quite well. We are in the process of testing its effectiveness on TIS patient dMRI data.

Post-acquisition Standardization of Positron Emission Tomography Images.

Purpose: Tissue radiotracer activity measured from positron emission tomography (PET) images is an important biomarker that is clinically utilized for diagnosis, staging, prognostication, and treatment response assessment in patients with cancer and other clinical disorders. Using PET image values to define a normal range of metabolic activity for quantification purposes is challenging due to variations in patient-related factors and technical factors. Although the formulation of standardized uptake value (SUV) has compensated for some of these variabilities, significant non-standardness still persists. We propose an image processing method to substantially mitigate these variabilities. Methods: The standardization method is similar for activity concentration (AC) PET and SUV PET images with some differences and consists of two steps. The calibration step is performed only once for each of AC PET or SUV PET, employs a set of images of normal subjects, and requires a reference object, while the transformation step is executed for each patient image to be standardized. In the calibration step, a standardized scale is determined along with 3 key image intensity landmarks defined on it including the minimum percentile intensity smin, median intensity sm, and high percentile intensity smax. smin and sm are estimated based on image intensities within the body region in the normal calibration image set. The optimal value of the maximum percentile ß corresponding to the intensity smax is estimated via an optimization process by using the reference object to optimally separate the highly variable high uptake values from the normal uptake intensities. In the transformation step, the first two landmarks - the minimum percentile intensity pα(I), and the median intensity pm(I) - are found for the given image I for the body region, and the high percentile intensity pß(I) is determined corresponding to the optimally estimated high percentile value ß. Subsequently, intensities of I are mapped to the standard scale piecewise linearly for different segments.We employ three strategies for evaluation and comparison with other standardization methods: (i) Comparing coefficient of variation (CVO) of mean intensity within test objects O across different normal test subjects before and after standardization; (ii) Comparing mean absolute difference (MDO) of mean intensity within test objects O across different subjects in repeat scans before and after standardization; (iii) Comparing CVO of mean intensity across different normal subjects before and after standardization where the scans came from different brands of scanners. Results: Our data set consisted of 84 FDG-PET/CT scans of the body torso including 38 normal subjects and two repeat-scans of 23 patients. We utilized one of two objects - liver and spleen - as a reference object and the other for testing. The proposed standardization method reduced CVO and MDO by a factor of 3-8 in comparison to other standardization methods and no standardization. Upon standardization by our method, the image intensities (both for AC and SUV) from two different brands of scanners become statistically indistinguishable, while without standardization, they differ significantly and by a factor of 3-9. Conclusions: The proposed method is automatic, outperforms current standardization methods, and effectively overcomes the residual variation left over in SUV and inter-scanner variations.

Estimating normal metabolic activity for disease quantification via PET/CT images.

In this paper, we propose a novel pipeline for conducting disease quantification in positron emission tomography/computed tomography (PET/CT) images on anatomically pre-defined objects. The pipeline is composed of standardized uptake value (SUV) standardization, object segmentation, and disease quantification (DQ). DQ is conducted on non-linearly standardized PET images and masks of target objects derived from CT images. Total lesion burden (TLB) is quantified by estimating normal metabolic activity (TMAn) in the object and subtracting this entity from total metabolic activity (TMA) of the object, thereby measuring the overall disease quantity of the region of interest without the necessity of explicitly segmenting individual lesions. TMAn is calculated with object-specific SUV distribution models. In the modeling stage, SUV models are constructed from a set of PET/CT images obtained from normal subjects with manually delineated masks of target objects. Two ways of SUV modeling are explored, where the mean of mean values of the modeling samples is utilized as a consistent normality value in the hard strategy, and the likelihood representing normal tissue is determined from the SUV distribution (histogram) for each SUV value in the fuzzy strategy. The evaluation experiments are conducted on a separate clinical dataset of normal subjects and a phantom dataset with lesions. The ratio of absolute TLB to TMA is taken as the metric, alleviating the individual difference of volume sizes and uptake levels. The results show that the ratios in normal objects are close to 0 and the ratios for lesions approach 1, demonstrating that contributions on TLB are minimal from the normal tissue and mainly from the lesion tissue.

Mendability Index: A new metric for estimating the effort required for manually editing auto-segmentations of objects of interest.

Auto-segmentation of medical images is critical to boost precision radiology and radiation oncology efficiency, thereby improving medical quality for both health care practitioners and patients. An appropriate metric to evaluate auto-segmentation results is one of the significant tools necessary for building an effective, robust, and practical auto-segmentation technique. However, by comparing the predicted segmentation with the ground truth, currently widely-used metrics usually focus on the overlapping area (Dice Coefficient) or the most severe shifting of the boundary (Hausdorff Distance), which seem inconsistent with human reader behaviors. Human readers usually verify and correct auto-segmentation contours and then apply the modified segmentation masks to guide clinical application in diagnosis or treatment. A metric called Mendability Index (MI) is proposed to better estimate the effort required for manually editing the auto-segmentations of objects of interest in medical images so that the segmentations become acceptable for the application at hand. Considering different human behaviors for different errors, MI classifies auto-segmented errors into three types with different quantitative behaviors. The fluctuation of human subjective delineation is also considered in MI. 505 3D computed tomography (CT) auto-segmentations consisting of 6 objects from 3 institutions with the corresponding ground truth and the recorded manual mending time needed by experts are used to validate the performance of the proposed MI. The correlation between the time for editing with the segmentation metrics demonstrates that MI is generally more suitable for indicating mending efforts than Dice Coefficient or Hausdorff Distance, suggesting that MI may be an effective metric to quantify the clinical value of auto-segmentations.

Prediction of lymphoma response to CAR T cells by deep learning-based image analysis.

Clinical prognostic scoring systems have limited utility for predicting treatment outcomes in lymphomas. We therefore tested the feasibility of a deep-learning (DL)-based image analysis methodology on pre-treatment diagnostic computed tomography (dCT), low-dose CT (lCT), and 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) images and rule-based reasoning to predict treatment response to chimeric antigen receptor (CAR) T-cell therapy in B-cell lymphomas. Pre-treatment images of 770 lymph node lesions from 39 adult patients with B-cell lymphomas treated with CD19-directed CAR T-cells were analyzed. Transfer learning using a pre-trained neural network model, then retrained for a specific task, was used to predict lesion-level treatment responses from separate dCT, lCT, and FDG-PET images. Patient-level response analysis was performed by applying rule-based reasoning to lesion-level prediction results. Patient-level response prediction was also compared to prediction based on the international prognostic index (IPI) for diffuse large B-cell lymphoma. The average accuracy of lesion-level response prediction based on single whole dCT slice-based input was 0.82+0.05 with sensitivity 0.87+0.07, specificity 0.77+0.12, and AUC 0.91+0.03. Patient-level response prediction from dCT, using the "Majority 60%" rule, had accuracy 0.81, sensitivity 0.75, and specificity 0.88 using 12-month post-treatment patient response as the reference standard and outperformed response prediction based on IPI risk factors (accuracy 0.54, sensitivity 0.38, and specificity 0.61 (p = 0.046)). Prediction of treatment outcome in B-cell lymphomas from pre-treatment medical images using DL-based image analysis and rule-based reasoning is feasible. This approach can potentially provide clinically useful prognostic information for decision-making in advance of initiating CAR T-cell therapy.