Liver and Spleen Volumetry with Quantitative MR Imaging and Dual-Space Clustering Segmentation

Children Are Not Small Adults: Addressing Limited Generalizability of an Adult Deep Learning CT Organ Segmentation Model to the Pediatric Population

Deep learning (DL) tools developed on adult data sets may not generalize well to pediatric patients, posing potential safety risks. We evaluated the performance of TotalSegmentator, a state-of-the-art adult-trained CT organ segmentation model, on a subset of organs in a pediatric CT dataset and explored optimization strategies to improve pediatric segmentation performance. TotalSegmentator was retrospectively evaluated on abdominal CT scans from an external adult dataset (n = 300) and an external pediatric data set (n = 359). Generalizability was quantified by comparing Dice scores between adult and pediatric external data sets using Mann-Whitney U tests. Two DL optimization approaches were then evaluated: (1) 3D nnU-Net model trained on only pediatric data, and (2) an adult nnU-Net model fine-tuned on the pediatric cases. Our results show TotalSegmentator had significantly lower overall mean Dice scores on pediatric vs. adult CT scans (0.73 vs. 0.81, P < .001) demonstrating limited generalizability to pediatric CT scans. Stratified by organ, there was lower mean pediatric Dice score for four organs (P < .001, all): right and left adrenal glands (right adrenal, 0.41 [0.39-0.43] vs. 0.69 [0.66-0.71]; left adrenal, 0.35 [0.32-0.37] vs. 0.68 [0.65-0.71]); duodenum (0.47 [0.45-0.49] vs. 0.67 [0.64-0.69]); and pancreas (0.73 [0.72-0.74] vs. 0.79 [0.77-0.81]). Performance on pediatric CT scans improved by developing pediatric-specific models and fine-tuning an adult-trained model on pediatric images where both methods significantly improved segmentation accuracy over TotalSegmentator for all organs, especially for smaller anatomical structures (e.g., > 0.2 higher mean Dice for adrenal glands; P < .001).

Reproducibility and efficiency of liver volumetry using manual method and liver analysis software

BACKGROUND

For liver volumetry, manual tracing on computed tomography (CT) images is time-consuming and operator dependent. To overcome these disadvantages, several three-dimensional simulation software programs have been developed; however, their efficacy has not fully been evaluated.

METHODS

Three physicians performed liver volumetry on preoperative CT images on 30 patients who underwent formal right hepatectomy, using manual tracing volumetry and two simulation software programs, SYNAPSE and syngo.via. The future liver remnant (FLR) was calculated using each method of volumetry. The primary endpoint was reproducibility and secondary outcomes were calculation time and learning curve.

RESULTS

The mean FLR was significantly lower for manual volumetry than for SYNAPSE or syngo.via; there was no significant difference in mean FLR between the two software-based methods. Reproducibility was lower for the manual method than for the software-based methods. Mean calculation time was shortest for SYNAPSE. For the two physicians unfamiliar with the software, no obvious learning curve was observed for using SYNAPSE, whereas learning curves were observed for using syngo.via.

CONCLUSIONS

Liver volumetry was more reproducible and faster with three-dimensional simulation software, especially SYNAPSE software, than with the conventional manual tracing method. Software can help even inexperienced physicians learn quickly how to perform liver volumetry.

Artificial Intelligence for Improved Hepatosplenomegaly Diagnosis

Hepatosplenomegaly is commonly diagnosed by radiologists based on single dimension measurements and heuristic cut-offs. Volumetric measurements may be more accurate for diagnosing organ enlargement. Artificial intelligence techniques may be able to automatically calculate liver and spleen volume and facilitate more accurate diagnosis. After IRB approval, 2 convolutional neural networks (CNN) were developed to automatically segment the liver and spleen on a training dataset comprised of 500 single-phase, contrast-enhanced CT abdomen and pelvis examinations. A separate dataset of ten thousand sequential examinations at a single institution was segmented with these CNNs. Performance was evaluated on a 1% subset and compared with manual segmentations using Sorensen-Dice coefficients and Pearson correlation coefficients. Radiologist reports were reviewed for diagnosis of hepatomegaly and splenomegaly and compared with calculated volumes. Abnormal enlargement was defined as greater than 2 standard deviations above the mean. Median Dice coefficients for liver and spleen segmentation were 0.988 and 0.981, respectively. Pearson correlation coefficients of CNN-derived estimates of organ volume against the gold-standard manual annotation were 0.999 for the liver and spleen (P < 0.001). Average liver volume was 1556.8 ± 498.7 cc and average spleen volume was 194.6 ± 123.0 cc. There were significant differences in average liver and spleen volumes between male and female patients. Thus, the volume thresholds for ground-truth determination of hepatomegaly and splenomegaly were determined separately for each sex. Radiologist classification of hepatomegaly was 65% sensitive, 91% specific, with a positive predictive value (PPV) of 23% and an negative predictive value (NPV) of 98%. Radiologist classification of splenomegaly was 68% sensitive, 97% specific, with a positive predictive value (PPV) of 50% and a negative predictive value (NPV) of 99%. Convolutional neural networks can accurately segment the liver and spleen and may be helpful to improve radiologist accuracy in the diagnosis of hepatomegaly and splenomegaly.

Imaging-calculated splenic volume is associated with response in melanoma patients treated with immunotherapy

Background: Predicting the roughly 50% of melanoma patients that will respond to immunotherapy is challenging. We tested if splenic volume could be a predictive biomarker. Methods: Splenic volume was measured by a semiautomated commercial software tool in pre- and post-treatment PET/CT, CT or MRI in 50 melanoma patients treated with immune checkpoint inhibitors. Results: Subjects with smaller spleens had better progression-free survival (median not achieved after 30.6 months of follow-up vs median 11.2 months; p = 0.0213) than their counterparts. A cut-off of <244 cm³ yielded a sensitivity of 83% and specificity of 54% to identify responders. Conclusion: Measuring splenic volume on imaging scans is feasible. Smaller pretreatment spleen volume is associated with better responses to immune checkpoint inhibitors.

Cross-sectional imaging in patients with primary sclerosing cholangitis: Single time-point liver or spleen volume is associated with survival

AIM

To evaluate the association between single time-point quantitative liver and spleen volumes in patients with PSC and transplant-free survival, independent of Mayo risk score.

MATERIALS AND METHODS

This HIPAA-compliant retrospective study included 165 PSC patients in a hospital. Total (T), and lobar (right [R], left [L], and caudate [C]) liver volumes and spleen volume (S) were measured. Adverse outcome was identified as being on liver transplantation list, transplantation or death (outcome 1), and transplantation or death (outcome 2). Cox-regression was performed to assess the predictive value of volumetric parameters to predict transplant-free survival with and without Mayo risk score. Stratified analysis by Mayo risk score categories was performed to assess the discriminative value of volumes in the model. Prediction models were developed dependent of Mayo score, based on patients demographics, lab values and volumetric measures for both defined outcomes. Kaplan-Meier curves were depicted for different liver and spleen volumes. P value <0.05 was considered statistically significant.

RESULTS

In this cohort (age 43 ± 17 years; 59 % men) 51 % of patients had adverse outcome. Cox-regression analysis demonstrated statistically significant association between values of T, L, R, C, S, L/T, and C/T and outcome 1; and also statistically significant association between values C, S, and C/T and outcome 2. Prediction models included age, INR, total bilirubin, AST, variceal bleeding, S, and C for outcome 1 and age, INR, total bilirubin, AST, variceal bleeding, and S for outcome 2.

CONCLUSIONS

Based on our observational study, quantitative liver and spleen volumes may be associated with transplant-free survival in patients with PSC and may have the potential for predicting the outcome but this should be validated by randomized clinical trial studies.

A protocol of hepatic volume measurement using magnetic resonance imaging in individuals from the Eastern Brazilian Amazon population

Determination of hepatic volume is an important preoperative procedure and is done through imaging exams or standard liver volume (SLV) formulas developed based on the biotype of each population. In the absence of a specific SLV formula for the Brazilian Eastern Amazon population, the measurement of liver volume is made with reference values from other populations. The aim of study was to compare the hepatic volume in healthy residents from the Brazilian Eastern Amazon population obtained with magnetic resonance imaging (MRI) and recommended SLV formulas validated to other populations. This was a Observational, cross-sectional study. Anthropometric data of 42 healthy individuals aged 18–60 years of both sexes was collected to measure the liver volume through SLV formulas calculations and MRI. MRI shows similarity with the Western European SLV liver volume values and significant differences with the Japan SLV formula, mainly for women, with a moderate-to-weak correlation with the MRI measurements. There was a strong correlation between weight and body surface area in male patients analysed with measurements of the liver volume by the MRI and SLV formulas. The SLV formula based on the Western European population could be used in the absence of a specific formula for individuals living in the Amazon region. The results suggest that liver measurement formulas should take into consideration the sex of individuals, as well as the development of a specific SLV formula for the Eastern Amazon population and the conduction of similar studies in other Brazilian regions.

Splenomegaly Segmentation on Multi-Modal MRI Using Deep Convolutional Networks

The findings of splenomegaly, abnormal enlargement of the spleen, is a non-invasive clinical biomarker for liver and spleen diseases. Automated segmentation methods are essential to efficiently quantify splenomegaly from clinically acquired abdominal magnetic resonance imaging (MRI) scans. However, the task is challenging due to: 1) large anatomical and spatial variations of splenomegaly; 2) large inter- and intra-scan intensity variations on multi-modal MRI; and 3) limited numbers of labeled splenomegaly scans. In this paper, we propose the Splenomegaly Segmentation Network (SS-Net) to introduce the deep convolutional neural network (DCNN) approaches in multi-modal MRI splenomegaly segmentation. Large convolutional kernel layers were used to address the spatial and anatomical variations, while the conditional generative adversarial networks were employed to leverage the segmentation performance of SS-Net in an end-to-end manner. A clinically acquired cohort containing both T1-weighted (T1w) and T2-weighted (T2w) MRI splenomegaly scans was used to train and evaluate the performance of multi-atlas segmentation (MAS), 2D DCNN networks, and a 3-D DCNN network. From the experimental results, the DCNN methods achieved superior performance to the state-of-the-art MAS method. The proposed SS-Net method has achieved the highest median and mean Dice scores among the investigated baseline DCNN methods.

A whole-body FDG PET/MR atlas for multiparametric voxel-based analysis

Quantitative multiparametric imaging is a potential key application for Positron Emission Tomography/Magnetic Resonance (PET/MR) hybrid imaging. To enable objective and automatic voxel-based multiparametric analysis in whole-body applications, the purpose of this study was to develop a multimodality whole-body atlas of functional 18F-fluorodeoxyglucose (FDG) PET and anatomical fat-water MR data of adults. Image registration was used to transform PET/MR images of healthy control subjects into male and female reference spaces, producing a fat-water MR, local tissue volume and FDG PET whole-body normal atlas consisting of 12 male (66.6 ± 6.3 years) and 15 female (69.5 ± 3.6 years) subjects. Manual segmentations of tissues and organs in the male and female reference spaces confirmed that the atlas contained adequate physiological and anatomical values. The atlas was applied in two anomaly detection tasks as proof of concept. The first task automatically detected anomalies in two subjects with suspected malignant disease using FDG data. The second task successfully detected abnormal liver fat infiltration in one subject using fat fraction data.

Automated CT and MRI Liver Segmentation and Biometry Using a Generalized Convolutional Neural Network

Purpose

To assess feasibility of training a convolutional neural network (CNN) to automate liver segmentation across different imaging modalities and techniques used in clinical practice and apply this to enable automation of liver biometry.

Methods

We trained a 2D U-Net CNN for liver segmentation in two stages using 330 abdominal MRI and CT exams acquired at our institution. First, we trained the neural network with non-contrast multi-echo spoiled-gradient-echo (SGPR)images with 300 MRI exams to provide multiple signal-weightings. Then, we used transfer learning to generalize the CNN with additional images from 30 contrast-enhanced MRI and CT exams.We assessed the performance of the CNN using a distinct multi-institutional data set curated from multiple sources (n = 498 subjects). Segmentation accuracy was evaluated by computing Dice scores. Utilizing these segmentations, we computed liver volume from CT and T1-weighted (T1w) MRI exams, and estimated hepatic proton- density-fat-fraction (PDFF) from multi-echo T2*w MRI exams. We compared quantitative volumetry and PDFF estimates between automated and manual segmentation using Pearson correlation and Bland-Altman statistics.

Results

Dice scores were 0.94 ± 0.06 for CT (n = 230), 0.95 ± 0.03 (n = 100) for T1w MR, and 0.92 ± 0.05 for T2*w MR (n = 169). Liver volume measured by manual and automated segmentation agreed closely for CT (95% limit-of-agreement (LoA) = [-298 mL, 180 mL]) and T1w MR (LoA = [-358 mL, 180 mL]). Hepatic PDFF measured by the two segmentations also agreed closely (LoA = [-0.62%, 0.80%]).

Conclusions

Utilizing a transfer-learning strategy, we have demonstrated the feasibility of a CNN to be generalized to perform liver segmentations across different imaging techniques and modalities. With further refinement and validation, CNNs may have broad applicability for multimodal liver volumetry and hepatic tissue characterization.

Robust Multicontrast MRI Spleen Segmentation for Splenomegaly Using Multi-Atlas Segmentation

OBJECTIVE

Magnetic resonance imaging (MRI) is an essential imaging modality in noninvasive splenomegaly diagnosis. However, it is challenging to achieve spleen volume measurement from three-dimensional MRI given the diverse structural variations of human abdomens as well as the wide variety of clinical MRI acquisition schemes. Multi-atlas segmentation (MAS) approaches have been widely used and validated to handle heterogeneous anatomical scenarios. In this paper, we propose to use MAS for clinical MRI spleen segmentation for splenomegaly.

METHODS

First, an automated segmentation method using the selective and iterative method for performance level estimation (SIMPLE) atlas selection is used to address the concerns of inhomogeneity for clinical splenomegaly MRI. Then, to further control outliers, semiautomated craniocaudal spleen length-based SIMPLE atlas selection (L-SIMPLE) is proposed to integrate a spatial prior in a Bayesian fashion and guide iterative atlas selection. Last, a graph cuts refinement is employed to achieve the final segmentation from the probability maps from MAS.

RESULTS

A clinical cohort of 55 MRI volumes (28 T1 weighted and 27 T2 weighted) was used to evaluate both automated and semiautomated methods.

CONCLUSION

The results demonstrated that both methods achieved median Dice , and outliers were alleviated by the L-SIMPLE (≍1 min manual efforts per scan), which achieved 0.97 Pearson correlation of volume measurements with the manual segmentation.

SIGNIFICANCE

In this paper, spleen segmentation on MRI splenomegaly using MAS has been performed.

Correlation between quantitative liver and spleen volumes and disease severity in primary sclerosing cholangitis as determined by Mayo risk score

OBJECTIVES

To correlate total and lobar liver and spleen volume with disease severity in primary sclerosing cholangitis (PSC) as determined by Mayo risk score.

METHODS

This HIPAA-compliant single center retrospective study included 147 PSC patients with available imaging studies (MRCP/CT) and laboratory data between January 2003 and January 2018. Total and lobar (right, left and caudate) liver volume and spleen volume were measured. ANOVA test was performed to assess the differences in volumes between low, intermediate and high-risk groups (Mayo risk score <0, >0 and <2, >2, respectively). Correlations between volumes and Mayo risk score were calculated. ROC analysis was performed to assess the accuracy of the variable with the strongest correlation to PSC severity to predict Mayo risk score. P value <0.05 was considered statistically significant.

RESULTS

The mean age of this cohort was 45 ± 17 years; 58% were men. Absolute volumes of left lobe, caudate and spleen and volume ratios of left lobe and caudate to total liver volume of the high-risk group were significantly higher compared to those of low and intermediate risk groups (p < 0.05). Left lobe to total liver volume ratio had the highest correlation to Mayo risk score (Pearson correlation coefficient 0.61, p < 0.05) and on ROC analysis it had 84.4% accuracy in detecting high-risk PSC.

CONCLUSIONS

In this single institution large cohort study, the left lobe to total liver volume ratio was the best quantifiable volumetric biomarker to correlate with severity of PSC as identified by Mayo risk score.

Synthetic MRI of the Knee: Phantom Validation and Comparison with Conventional MRI

Purpose To test the hypothesis that synthetic MRI of the knee generates accurate and repeatable quantitative maps and produces morphologic MR images with similar quality and detection rates of structural abnormalities than does conventional MRI. Materials and Methods Data were collected prospectively between January 2017 and April 2018 and were retrospectively analyzed. An International Society for Magnetic Resonance in Medicine-National Institute of Standards and Technology phantom was used to determine the accuracy of T1, T2, and proton density (PD) quantification. Statistical models were applied for correction. Fifty-four participants (24 men, 30 women; mean age, 40 years; range, 18-62 years) underwent synthetic and conventional 3-T MRI twice on the same day. Fifteen of 54 participants (28%) repeated the protocol within 9 days. The intra- and interday agreements of quantitative cartilage measurements were assessed. Contrast-to-noise (CNR) ratios, image quality, and structural abnormalities were assessed on corresponding synthetic and conventional images. Statistical analyses included the Wilcoxon test, χ² test, and Cohen Kappa. P values less than or equal to .01 were considered to indicate a statistically significant difference. Results Synthetic MRI quantification of T1, T2, and PD values had an overall model-corrected error margin of 0.8%. The synthetic MRI interday repeatability of articular cartilage quantification had native and model-corrected error margins of 3.3% and 3.5%, respectively. The cartilage-to-fluid CNR and menisci-to-fluid CNR was higher on synthetic than conventional MR images (P ≤ .001, respectively). Synthetic MRI improved short-tau inversion recovery fat suppression (P ˂ .01). Intermethod agreements of structural abnormalities were good (kappa, 0.621-0.739). Conclusion Synthetic MRI of the knee is accurate for T1, T2, and proton density quantification, and simultaneously generated morphologic MR images have detection rates of structural abnormalities similar to those of conventional MR images, with similar acquisition time. © RSNA, 2018.

Muscles of the trunk and pelvis are responsive to testosterone administration: data from testosterone dose-response study in young healthy men

Testosterone dose-dependently increases appendicular muscle mass. However, the effects of testosterone administration on the core muscles of the trunk and the pelvis have not been evaluated. The present study evaluated the effects of testosterone administration on truncal and pelvic muscles in a dose-response trial. Participants were young healthy men aged 18-50 years participating in the 5α-Reductase (5aR) Trial. All participants received monthly injections of 7.5 mg leuprolide acetate to suppress endogenous testosterone production and weekly injections of 50, 125, 300, or 600 mg of testosterone enanthate and were randomized to receive either 2.5 mg dutasteride (5aR inhibitor) or placebo daily for 20 weeks. Muscles of the trunk and the pelvis were measured at baseline and the end of treatment using 1.5-Tesla magnetic resonance imaging. The dose effect of testosterone on changes in the psoas major muscle area was the primary outcome; secondary outcomes included changes in paraspinal, abdominal, pelvic floor, ischiocavernosus, and obturator internus muscles. The association between changes in testosterone levels and muscle area was also assessed. Testosterone dose-dependently increased areas of all truncal and pelvic muscles. The estimated change (95% confidence interval) of muscle area increase per 100 mg of testosterone enanthate dosage increase was 0.622 cm² (0.394, 0.850) for psoas; 1.789 cm² (1.317, 2.261) for paraspinal muscles, 2.530 cm² (1.627, 3.434) for total abdominal muscles, 0.455 cm² (0.233, 0.678) for obturator internus, and 0.082 cm² (0.003, 0.045) for ischiocavernosus; the increase in these volumes was significantly associated with the changes in on-treatment total and free serum testosterone concentrations. In conclusion, core muscles of the trunk and pelvis are responsive to testosterone administration. Future trials should evaluate the potential role of testosterone administration in frail men who are predisposed to falls and men with pelvic floor dysfunction.

Determination of splenomegaly by coronal oblique length on CT

PURPOSE

The aims of this study were (a) to determine whether a coronal oblique length (COL) > 12 cm, which is often used to detect splenomegaly (SM) on ultrasound, can be used as a marker of SM on computed tomography (CT), (b) to compare the diagnostic accuracy of COL with other unidimensional linear measurements (ULM) in identifying SM, (c) to determine which ULM most closely correlates with splenic volume (SVol) according to the splenic index on CT, (d) to assess the relationship between SVol and patient's gender, age and body parameters (height, weight), and (e) to determine whether there is a difference between non-contrast-enhanced and contrast-enhanced CT images in identifying SM.

MATERIALS AND METHODS

The body parameters and ULM (width, length, thickness, COL) were obtained from patients who underwent CT for various indications from April 2016 to April 2017. SVol and body mass indexes were calculated for each patient.

RESULTS

Of the 1078 patients [male/female = 526/552; 47.57 (mean) ± 19.21 (standard deviation) years], 392 patients had SM. The sensitivity, specificity, positive and negative predictive values of COL > 12 cm for diagnosing SM were 44.6, 95.6, 85.3 and 75.1%, respectively (p < 0.001). SVol correlated with all ULM (p < 0.001). In the non-SM group (n = 686), the mean SVol was 331.7 ± 92.2 cm³ and females had smaller spleens than males (p < 0.001). SVol showed correlation with gender, age and height (p < 0.001).

CONCLUSION

COL > 12 cm is not superior to other ULM for the detection of SM, but it is very successful in determining normal spleens. The unidimensional measurements and volume of the spleen should be calculated by taking gender and body parameters into account for different ethnic populations. Non-contrast-enhanced CT can also be used to detect SM.

Principles of Quantitative MR Imaging with Illustrated Review of Applicable Modular Pulse Diagrams

Continued improvements in diagnostic accuracy using magnetic resonance (MR) imaging will require development of methods for tissue analysis that complement traditional qualitative MR imaging studies. Quantitative MR imaging is based on measurement and interpretation of tissue-specific parameters independent of experimental design, compared with qualitative MR imaging, which relies on interpretation of tissue contrast that results from experimental pulse sequence parameters. Quantitative MR imaging represents a natural next step in the evolution of MR imaging practice, since quantitative MR imaging data can be acquired using currently available qualitative imaging pulse sequences without modifications to imaging equipment. The article presents a review of the basic physical concepts used in MR imaging and how quantitative MR imaging is distinct from qualitative MR imaging. Subsequently, the article reviews the hierarchical organization of major applicable pulse sequences used in this article, with the sequences organized into conventional, hybrid, and multispectral sequences capable of calculating the main tissue parameters of T1, T2, and proton density. While this new concept offers the potential for improved diagnostic accuracy and workflow, awareness of this extension to qualitative imaging is generally low. This article reviews the basic physical concepts in MR imaging, describes commonly measured tissue parameters in quantitative MR imaging, and presents the major available pulse sequences used for quantitative MR imaging, with a focus on the hierarchical organization of these sequences. ^©RSNA, 2017.

Automatized spleen segmentation in non-contrast-enhanced MR volume data using subject-specific shape priors

To develop the first fully automated 3D spleen segmentation framework derived from T1-weighted magnetic resonance (MR) imaging data and to verify its performance for spleen delineation and volumetry. This approach considers the issue of low contrast between spleen and adjacent tissue in non-contrast-enhanced MR images. Native T1-weighted MR volume data was performed on a 1.5 T MR system in an epidemiological study. We analyzed random subsamples of MR examinations without pathologies to develop and verify the spleen segmentation framework. The framework is modularized to include different kinds of prior knowledge into the segmentation pipeline. Classification by support vector machines differentiates between five different shape types in computed foreground probability maps and recognizes characteristic spleen regions in axial slices of MR volume data. A spleen-shape space generated by training produces subject-specific prior shape knowledge that is then incorporated into a final 3D level set segmentation method. Individually adapted shape-driven forces as well as image-driven forces resulting from refined foreground probability maps steer the level set successfully to the segment the spleen. The framework achieves promising segmentation results with mean Dice coefficients of nearly 0.91 and low volumetric mean errors of 6.3%. The presented spleen segmentation approach can delineate spleen tissue in native MR volume data. Several kinds of prior shape knowledge including subject-specific 3D prior shape knowledge can be used to guide segmentation processes achieving promising results.

Does coronal oblique length of spleen on CT reflect splenic index?

AIMS AND OBJECTIVES

Splenic dimension of >12 cm on coronal plane on ultrasound is considered to represent splenomegaly. We sought to estimate the accuracy of similar coronal oblique length of spleen on CT in identifying splenomegaly by comparing it with CT splenic index. We also sought to establish the splenic width, craniocaudal dimension (CC), thickness, and coronal oblique length in both normal and splenomegaly groups.

MATERIALS AND METHODS

319 consecutive patients undergoing CT abdomen were included in the study and measurements of width (W), CC, thickness (T), and coronal oblique length (L) made. Splenic index was calculated based on the formula CC × W × T. Diagnostic accuracy of coronal oblique length in identifying splenomegaly on the basis of splenic index was assessed. Patients with splenic trauma were excluded. Patients with perisplenic collection that precluded proper measurement of spleen were excluded.

RESULTS

Of the 319 patients, 41 patients had splenomegaly with splenic index >480. 278 patients showed normal splenic index less than or equal to 480. Sensitivity and specificity of coronal oblique length as ≥12 cm in identifying splenomegaly were 97.8% and 34.1%, respectively. Positive and negative predictive values for the same were 91% and 70%, respectively (p value-0.000). There is overlap of numerical values for thickness and coronal oblique length in the normal and splenomegaly groups for a 95% confidence interval, whereas there is no significant overlap between these groups with respect to the craniocaudal length and width. Hence, craniocaudal length of 9.5 cm and width of 10.6 cm may be used as upper cutoff limit for normal spleen.

CONCLUSION

Coronal oblique dimension of spleen >12 cm is highly sensitive and shows good positive predictive value in diagnosing splenomegaly but has poor specificity and negative predictive value. Ideally, splenic index calculated using the CC, width, and thickness is the most reliable measurement for diagnosing splenomegaly. But if single measurements are to be used for identifying splenomegaly, craniocaudal length >9.5 cm, width of >10.6 cm, and coronal oblique dimension >12 cm may be used.

Comparison of MRI- and CT-based semiautomated liver segmentation: a validation study

PURPOSE

To compare the repeatability, agreement, and efficiency of MRI- and CT-based semiautomated liver segmentation for the assessment of total and subsegmental liver volume.

METHODS

This retrospective study was conducted in 31 subjects who underwent contemporaneous liver MRI and CT. Total and subsegmental liver volumes were segmented from contrast-enhanced 3D gradient-recalled echo MRI sequences and CT images. Semiautomated segmentation was based on variational interpolation and Laplacian mesh optimization. All segmentations were repeated after 2 weeks. Manual segmentation of CT images using an active contour tool was used as the reference standard. Repeatability and agreement of the methods were evaluated with intra-class correlation coefficients (ICC) and Bland-Altman analysis. Total interaction time was recorded.

RESULTS

Intra-reader ICC were ≥0.987 for MRI and ≥0.995 for CT. Intra-reader repeatability was 30 ± 217 ml (bias ± 1.96 SD) (95% limits of agreement: -187 to 247 ml) for MRI and -10 ± 143 ml (-153 to 133 ml) for CT. Inter-method ICC between semiautomated and manual volumetry were ≥0.995 for MRI and ≥0.986 for CT. Inter-method segmental ICC varied between 0.584 and 0.865 for MRI and between 0.596 and 0.890 for CT. Inter-method agreement was -14 ± 136 ml (-150 to 122 ml) for MRI and 50 ± 226 ml (-176 to 276 ml) for CT. Inter-method segmental agreement ranged from 10 ± 47 ml (-37 to 57 ml) to 2 ± 214 ml (-212 to 216 ml) for MRI and 9 ± 45 ml (-36 to 54 ml) to -46 ± 183 ml (-229 to 137 ml) for CT. Interaction time (mean ± SD) was significantly shorter for MRI-based semiautomated segmentation (7.2 ± 0.1 min, p < 0.001) and for CT-based semiautomated segmentation (6.5 ± 0.2 min, p < 0.001) than for CT-based manual segmentation (14.5 ± 0.4 min).

CONCLUSION

MRI-based semiautomated segmentation provides similar repeatability and agreement to CT-based segmentation for total liver volume.

Gd-EOB enhanced MRI T1-weighted 3D-GRE with and without elevated flip angle modulation for threshold-based liver volume segmentation

BACKGROUND

Despite novel software solutions, liver volume segmentation is still a time-consuming procedure and often requires further manual optimization. With the high signal intensity of the liver parenchyma in Gd-EOB enhanced magnetic resonance imaging (MRI), liver volume segmentation may be improved.

PURPOSE

To evaluate the practicability of threshold-based segmentation of the liver volume using Gd-EOB-enhanced MRI including a customized three-dimensional (3D) sequence.

MATERIAL AND METHODS

A total of 20 patients examined with Gd-EOB MRI (hepatobiliary phase T1-weighted (T1W) 3D sequence [VIBE]; flip angle [FA], 10° and 30°) were enrolled in this retrospective study. The datasets were independently processed by two blinded observers (O1 and O2) in two ways: manual (man) and threshold-based (thresh; study method) segmentation of the liver each followed by an optimization step (man+opt and thresh+opt; man+opt [FA10°] served as reference method). Resulting liver volumes and segmentation times were compared. A liver conversion factor was calculated in percent, describing the non-hepatocellular fraction of the total liver volume, i.e. bile ducts and vessels.

RESULTS

Thresh+opt (FA10°) was significantly faster compared to the reference method leading to a median volume overestimation of 4%/8% (P < 0.001). Using thresh+opt (FA30°), segmentation was even faster (P < 0.001) and even reduced median volume deviation of 0%/2% (O1/O2; both P > 0.2). The liver conversion factor was found to be 10%.

CONCLUSION

Threshold-based liver segmentation employing Gd-EOB-enhanced hepatobiliary phase standard T1W 3D sequence is accurate and time-saving. The performance of this approach can be further improved by increasing the FA.

Volumetric gain of the human pancreas after left partial pancreatic resection: A CT-scan based retrospective study

BACKGROUND/OBJECTIVES

Regeneration of the pancreas has been well characterized in animal models. However, there are conflicting data on the regenerative capacity of the human pancreas. The aim of the present study was to assess the regenerative capacity of the human pancreas.

METHODS

In a retrospective study, data from patients undergoing left partial pancreatic resection at a single center were eligible for inclusion (n = 185). Volumetry was performed based on 5 mm CT-scans acquired through a 256-slice CT-scanner using a semi-automated software.

RESULTS

Data from 24 patients (15 males/9 females) were included. Mean ± SD age was 68 ± 11 years (range, 40-85 years). Median time between surgery and the 1st postoperative CT was 9 days (range, 0-27 days; IQR, 7-13), 55 days (range, 21-141 days; IQR, 34-105) until the 2nd CT, and 191 days (range, 62-1902; IQR, 156-347) until the 3rd CT. The pancreatic volumes differed significantly between the first and the second postoperative CT scans (median volume 25.6 mL and 30.6 mL, respectively; p = 0.008) and had significantly increased further by the 3rd CT scan (median volume 37.9 mL; p = 0.001 for comparison with 1st CT scan and p = 0.003 for comparison with 2nd CT scan).

CONCLUSIONS

The human pancreas shows a measurable and considerable potential of volumetric gain after partial resection. Multidetector-CT based semi-automated volume analysis is a feasible method for follow-up of the volume of the remaining pancreatic parenchyma after partial pancreatectomy. Effects on exocrine and endocrine pancreatic function have to be evaluated in a prospective manner.

Computerized segmentation of liver in hepatic CT and MRI by means of level-set geodesic active contouring

Computerized liver volumetry has been studied, because the current "gold-standard" manual volumetry is subjective and very time-consuming. Liver volumetry is done in either CT or MRI. A number of researchers have developed computerized liver segmentation in CT, but there are fewer studies on ones for MRI. Our purpose in this study was to develop a general framework for liver segmentation in both CT and MRI. Our scheme consisted of 1) an anisotropic diffusion filter to reduce noise while preserving liver structures, 2) a scale-specific gradient magnitude filter to enhance liver boundaries, 3) a fast-marching algorithm to roughly determine liver boundaries, and 4) a geodesic-active-contour model coupled with a level-set algorithm to refine the initial boundaries. Our CT database contained hepatic CT scans of 18 liver donors obtained under a liver transplant protocol. Our MRI database contains 23 patients with 1.5T MRI scanners. To establish "gold-standard" liver volumes, radiologists manually traced the contour of the liver on each CT or MR slice. We compared our computer volumetry with "gold-standard" manual volumetry. Computer volumetry in CT and MRI reached excellent agreement with manual volumetry (intra-class correlation coefficient = 0.94 and 0.98, respectively). Average user time for computer volumetry in CT and MRI was 0.57 ± 0.06 and 1.0 ± 0.13 min. per case, respectively, whereas those for manual volumetry were 39.4 ± 5.5 and 24.0 ± 4.4 min. per case, respectively, with statistically significant difference (p < .05). Our computerized liver segmentation framework provides an efficient and accurate way of measuring liver volumes in both CT and MRI.

Testosterone dose-response relationships with cardiovascular risk markers in androgen-deficient women: a randomized, placebo-controlled trial

OBJECTIVE

To determine dose-dependent effects of T administration on cardiovascular risk markers in women with low T levels.

METHODS

Seventy-one hysterectomized women with or without oophorectomy with total T < 31 ng/dL and/or free T < 3.5 pg/mL received a standardized transdermal estradiol regimen during the 12-week run-in period and were then randomized to receive weekly im injections of placebo or 3-, 6.25-, 12.5-, or 25-mg T enanthate for 24 weeks. Total and free T levels were measured by liquid chromatography-tandem mass spectrometry and equilibrium dialysis, respectively. Insulin resistance and inflammatory markers were measured at baseline and 24 weeks. In a subset of women, magnetic resonance imaging of the abdomen was performed to quantify abdominal fat volume.

RESULTS

Fifty-nine women who completed the 24-week intervention were included in the final analysis. The five groups were similar at baseline. Mean on-treatment nadir total T concentrations were 14, 79, 105, 130, and 232 ng/dL in the placebo group and the 3-, 6.25-, 12.5-, and 25-mg groups, respectively. No significant changes in fasting glucose, fasting insulin, homeostatic model assessment of insulin resistance, high sensitivity C-reactive protein, adiponectin, blood pressure, and heart rate were observed at any T dose when compared to placebo. Similarly, no dose- or concentration-dependent changes were observed in abdominal fat on magnetic resonance imaging.

CONCLUSION

Short-term T administration over a wide range of doses for 24 weeks in women with low T levels was not associated with worsening of cardiovascular risk markers.

Computerized liver volumetry on MRI by using 3D geodesic active contour segmentation

OBJECTIVE

Our purpose was to develop an accurate automated 3D liver segmentation scheme for measuring liver volumes on MRI.

SUBJECTS AND METHODS

Our scheme for MRI liver volumetry consisted of three main stages. First, the preprocessing stage was applied to T1-weighted MRI of the liver in the portal venous phase to reduce noise and produce the boundary-enhanced image. This boundary-enhanced image was used as a speed function for a 3D fast-marching algorithm to generate an initial surface that roughly approximated the shape of the liver. A 3D geodesic-active-contour segmentation algorithm refined the initial surface to precisely determine the liver boundaries. The liver volumes determined by our scheme were compared with those manually traced by a radiologist, used as the reference standard.

RESULTS

The two volumetric methods reached excellent agreement (intraclass correlation coefficient, 0.98) without statistical significance (p = 0.42). The average (± SD) accuracy was 99.4% ± 0.14%, and the average Dice overlap coefficient was 93.6% ± 1.7%. The mean processing time for our automated scheme was 1.03 ± 0.13 minutes, whereas that for manual volumetry was 24.0 ± 4.4 minutes (p < 0.001).

CONCLUSION

The MRI liver volumetry based on our automated scheme agreed excellently with reference-standard volumetry, and it required substantially less completion time.

Effect of Breath Holding on Spleen Volume Measured by Magnetic Resonance Imaging

Objective

Ultrasonographic studies have demonstrated transient reduction in spleen volume in relation to apnea diving. We measured spleen volume under various respiratory conditions by MR imaging to accurately determine the influence of ordinary breath holding on spleen volumetry.

Materials and Methods

Twelve healthy adult volunteers were examined. Contiguous MR images of the spleen were acquired during free breathing and during respiratory manipulations, including breath holding at the end of normal expiration, breath holding at deep inspiration, and the valsalva maneuver, and spleen volume was measured from each image set based on the sum-of-areas method. Acquisition during free breathing was performed with respiratory triggering. The duration of each respiratory manipulation was 30 s, and five sets of MR images were acquired serially during each manipulation.

Results

Baseline spleen volume before respiratory manipulation was 173.0 ± 79.7 mL, and the coefficient of variance for two baseline measures was 1.4% ± 1.6%, suggesting excellent repeatability. Spleen volume decreased significantly just after the commencement of respiratory manipulation, remained constant during the manipulation, and returned to the control value 2 min after the cessation of the manipulation, irrespective of manipulation type. The percentages of volume reduction were 10.2% ± 2.9%, 10.2% ± 3.5%, and 13.3% ± 5.7% during expiration breath holding, deep-inspiration breath holding, and the valsalva maneuver, respectively, and these values did not differ significantly.

Conclusions

Spleen volume is reduced during short breath-hold apnea in healthy adults. Physiological responses of the spleen to respiratory manipulations should be considered in the measurement and interpretation of spleen volume.

Assessing splenomegaly: automated volumetric analysis of the spleen

RATIONALE AND OBJECTIVES

To define systematic volumetric thresholds to identify and grade splenomegaly and retrospectively evaluate the performance of radiologists to assess splenomegaly in computed tomography (CT) image data.

MATERIALS AND METHODS

A clinical tool was developed to segment spleens from 172 contrast-enhanced clinical CT studies. There were 45 normal and 127 splenomegaly cases confirmed by radiological reports. Spleen volumes were compared to manual measurements using overlap/error. Volumetric thresholds for mild/massive splenomegaly were defined at 1/2.5 standard deviations above the average splenic volume of the healthy population. The thresholds were validated against consensus reports. The performance of radiologists in assessing splenomegaly was retrospectively evaluated.

RESULTS

The automated segmentation of spleens was robust with volume overlap/error of 95.2/3.3%. There were no significant differences (P > .2) between manual and automated segmentations for either normal/splenomegaly subgroups. Comparable correlations between interobserver and manual-automated measurements were found (r = 0.99 for all). The average volume of normal spleens was 236.89 ± 77.58 mL. For splenomegaly, average volume was 1004.75 ± 644.27 mL. Volumetric thresholds of 314.47/430.84 mL were used to define mild/massive splenomegaly (±18.86 mL, 95% CI). Radiologists disagreed in 23.25% (n = 40) of the diagnosed cases. The area under the receiver operating characteristic curve of the volumetric criterion for splenomegaly detection was 0.96. Using the volumetric thresholds as the reference standard, the sensitivity of radiologists in detecting all/mild/massive splenomegaly was 95.0/66.6/99.0% at 78.0% specificity, respectively.

CONCLUSION

Thresholds for the identification and grading of splenomegaly from automatic volumetric spleen assessment were introduced. The volumetric thresholds match well with clinical interpretations for splenomegaly and may improve splenomegaly detection compared with splenic cephalocaudal height measurements or visual inspection commonly used in current clinical practice.

A simple method for accurate liver volume estimation by use of curve-fitting: a pilot study

In this paper, we describe the effectiveness of our curve-fitting method by comparing liver volumes estimated by our new technique to volumes obtained with the standard manual contour-tracing method. Hepatic parenchymal-phase images of 13 patients were obtained with multi-detector CT scanners after intravenous bolus administration of 120-150 mL of contrast material (300 mgI/mL). The liver contours of all sections were traced manually by an abdominal radiologist, and the liver volume was computed by summing of the volumes inside the contours. The section number between the first and last slice was then divided into 100 equal parts, and each volume was re-sampled by use of linear interpolation. We generated 13 model profile curves by averaging 12 cases, leaving out one case, and we estimated the profile curve for each patient by fitting the volume values at 4 points using a scale and translation transform. Finally, we determined the liver volume by integrating the sampling points of the profile curve. We used Bland-Altman analysis to evaluate the agreement between the volumes estimated with our curve-fitting method and the volumes measured by the manual contour-tracing method. The correlation between the volume measured by manual tracing and that estimated with our curve-fitting method was relatively high (r = 0.98; slope 0.97; p < 0.001). The mean difference between the manual tracing and our method was -22.9 cm(3) (SD of the difference, 46.2 cm(3)). Our volume-estimating technique that requires the tracing of only 4 images exhibited a relatively high linear correlation with the manual tracing technique.

Age-related apparent diffusion coefficient changes in the normal brain

PURPOSE

To measure the mean diffusional age-related changes of the brain over the full human life span by using diffusion-weighted spin-echo single-shot echo-planar magnetic resonance (MR) imaging and sequential whole-brain apparent diffusion coefficient (ADC) histogram analysis and, secondarily, to build mathematical models of these normal age-related changes throughout human life.

MATERIALS AND METHODS

After obtaining institutional review board approval, a HIPAA-compliant retrospective search was conducted for brain MR imaging studies performed in 2007 for various clinical indications. Informed consent was waived. The brain data of 414 healthy subjects (189 males and 225 females; mean age, 33.7 years; age range, 2 days to 89.3 years) were obtained with diffusion-weighted spin-echo single-shot echo-planar MR imaging. ADC histograms of the whole brain were generated. ADC peak values, histogram widths, and intracranial volumes were plotted against age, and model parameters were estimated by using nonlinear regression.

RESULTS

Four different stages were identified for aging changes in ADC peak values, as characterized by specific mathematical terms: There were age-associated exponential decays for the maturation period and the development period, a constant term for adulthood, and a linear increase for the senescence period. The age dependency of ADC peak value was simulated by using four-term six-coefficient function, including biexponential and linear terms. This model fit the data very closely (R(2) = 0.91).

CONCLUSION

Brain diffusivity as a whole demonstrated age-related changes through four distinct periods of life. These results could contribute to establishing an ADC baseline of the normal brain, covering the full human life span.

Assessing renal parenchymal volume on unenhanced CT as a marker for predicting renal function in patients with chronic kidney disease

OBJECTIVES

To estimate renal volume in chronic kidney disease (CKD) patients using a semiautomated software and compare them with split renal function estimates from radionuclide renogram (RR). We proposed that renal volume from unenhanced computed tomography (CT) scans may serve as surrogate marker for assessing renal function in CKD patients.

MATERIALS AND METHODS

Unenhanced multidetector CT scans of 26 patients with CKD (estimated glomerular filtration rate [eGFR] <60 mL/kg/body surface area [BSA]) and 10 controls (eGFR >60 mL/kg/BSA) were analyzed to calculate renal volumes using a semiautomated software (AMIRAV5.2.0). Volumes obtained were then correlated with corresponding eGFR and split renal function estimates from RR. Volumes were also compared with those obtained on enhanced scans in 10 cases (five disease group, five controls). Bland-Altman analysis was used to assess agreement between methods.

RESULTS

A moderately positive correlation was found between renal volume obtained on unenhanced CT and eGFR (r = 0.65, P < .0001), whereas a significantly high correlation with split function estimates from RR (r = 0.95, P < .001) was found. Bland-Altman analysis revealed a good agreement between renal volume from CT and renal function from RR (34/36 observations were within 95% CI and there were two outliers). Correlation between volumes obtained from unenhanced and enhanced CT scans was also significant (r = 0.96).

CONCLUSION

In patients with CKD, renal volume derived from unenhanced CT can possibly serve as a surrogate marker for assessing and monitoring renal function reserves to plan further management.

Assessing hepatomegaly: automated volumetric analysis of the liver

RATIONALE AND OBJECTIVES

The aims of this study were to define volumetric nomograms for identifying hepatomegaly and to retrospectively evaluate the performance of radiologists in assessing hepatomegaly.

MATERIALS AND METHODS

Livers were automatically segmented from 148 abdominal contrast-enhanced computed tomographic scans: 77 normal livers and 71 cases of hepatomegaly (diagnosed by visual inspection and/or linear liver height by radiologists). Quantified liver volumes were compared to manual measurements using volume overlap and error. Liver volumes were normalized to body surface area, from which hepatomegaly nomograms were defined (H scores) by analyzing the distribution of liver sizes in the healthy population. H scores were validated against consensus reports. The performance of radiologists in diagnosing hepatomegaly was retrospectively evaluated.

RESULTS

The automated segmentation of livers was robust, with volume overlap and error of 96.2% and 2.2%, respectively. There were no significant differences (P > .10) between manual and automated segmentation for either the normal or the hepatomegaly subgroup. The average volumes of normal and enlarged livers were 1.51 ± 0.25 and 2.32 ± 0.75 L, respectively. One-way analysis of variance found that body surface area (P = .004) and gender (P = .02), but not age, significantly affected normal liver volume. No significant effects were observed for two-way and three-way interactions among the three variables (P > .18). H-score cutoffs of 0.92 and 1.08 L/m2 were used to define mild and massive hepatomegaly (95% confidence interval, ± 0.02 L/m2). Using the H score as the reference standard, the sensitivity of radiologists in detecting all, mild, and massive hepatomegaly was 84.4%, 56.7%, and 100.0% at 90.1% specificity, respectively. Radiologists disagreed on 20.9% of the diagnosed cases (n = 31). The area under the receiver-operating characteristic curve of the H-score criterion for hepatomegaly detection was 0.98.

CONCLUSIONS

Nomograms for the identification and grading of hepatomegaly from automatic volumetric liver assessment normalized to body surface area (H scores) are introduced. H scores match well with clinical interpretations for hepatomegaly and may improve hepatomegaly detection compared with height measurements or visual inspection, commonly used in current clinical practice.

Effect of testosterone supplementation with and without a dual 5α-reductase inhibitor on fat-free mass in men with suppressed testosterone production: a randomized controlled trial

CONTEXT

Steroid 5α-reductase inhibitors are used to treat benign prostatic hyperplasia and androgenic alopecia, but the role of 5α-dihydrotestosterone (DHT) in mediating testosterone's effects on muscle, sexual function, erythropoiesis, and other androgen-dependent processes remains poorly understood.

OBJECTIVE

To determine whether testosterone's effects on muscle mass, strength, sexual function, hematocrit level, prostate volume, sebum production, and lipid levels are attenuated when its conversion to DHT is blocked by dutasteride (an inhibitor of 5α-reductase type 1 and 2).

DESIGN, SETTING, AND PATIENTS

The 5α-Reductase Trial was a randomized controlled trial of healthy men aged 18 to 50 years comparing placebo plus testosterone enthanate with dutasteride plus testosterone enanthate from May 2005 through June 2010.

INTERVENTIONS

Eight treatment groups received 50, 125, 300, or 600 mg/wk of testosterone enanthate for 20 weeks plus placebo (4 groups) or 2.5 mg/d of dutasteride (4 groups).

MAIN OUTCOME MEASURES

The primary outcome was change in fat-free mass; secondary outcomes: changes in fat mass, muscle strength, sexual function, prostate volume, sebum production, and hematocrit and lipid levels.

RESULTS

A total of 139 men were randomized; 102 completed the 20-week intervention. Men assigned to dutasteride were similar at baseline to those assigned to placebo. The mean fat-free mass gained by the dutasteride groups was 0.6 kg (95% CI, -0.1 to 1.2 kg) when receiving 50 mg/wk of testosterone enanthate, 2.6 kg (95% CI, 0.9 to 4.3 kg) for 125 mg/wk, 5.8 kg (95% CI, 4.8 to 6.9 kg) for 300 mg/wk, and 7.1 kg (95% CI, 6.0 to 8.2 kg) for 600 mg/wk. The mean fat-free mass gained by the placebo groups was 0.8 kg (95% CI, -0.1 to 1.7 kg) when receiving 50 mg/wk of testosterone enanthate, 3.5 kg (95% CI, 2.1 to 4.8 kg) for 125 mg/wk, 5.7 kg (95% CI, 4.8 to 6.5 kg) for 300 mg/wk, and 8.1 kg (95% CI, 6.7 to 9.5 kg) for 600 mg/wk. The dose-adjusted differences between the dutasteride and placebo groups for fat-free mass were not significant (P = .18). Changes in fat mass, muscle strength, sexual function, prostate volume, sebum production, and hematocrit and lipid levels did not differ between groups.

CONCLUSION

Changes in fat-free mass in response to graded testosterone doses did not differ in men in whom DHT was suppressed by dutasteride from those treated with placebo, indicating that conversion of testosterone to DHT is not essential for mediating its anabolic effects on muscle.

TRIAL REGISTRATION

clinicaltrials.gov Identifier: NCT00493987.

Improved T2 mapping accuracy with dual-echo turbo spin echo: effect of phase encoding profile orders

Turbo spin echo (TSE) pulse sequences have been applied to estimate T(2) relaxation times in clinically feasible scan times. However, T(2) estimations using TSE pulse sequences has been shown to differ considerable from reference standard sequences due to several sources of error. The purpose of this work was to apply voxel-sensitivity formalism to correct for one such source of error introduced by differing phase encoding profile orders with dual-echo TSE pulse sequences. The American College of Radiology phantom and the brains of two healthy volunteers were imaged using dual-echo TSE as well as 32-echo spin-echo acquisitions and T(2) estimations from uncorrected and voxel-sensitivity formalism-corrected dual-echo TSE and 32-echo acquisitions were compared. In all regions of the brain and the majority of the analyses of the American College of Radiology phantom, voxel-sensitivity formalism correction resulted in considerable improvements in dual-echo TSE T(2) estimation compared with the 32-echo acquisition, with improvements in T(2) value accuracy ranging from 5.2% to 18.6%.

Quantitative MR imaging: physical principles and sequence design in abdominal imaging

Quantitative magnetic resonance (MR) imaging seeks to quantify fundamental biologic and MR-inducible tissue properties, in contrast to the routine application of MR imaging in the clinic, in which differences in MR parameters are used to generate contrast for subsequent subjective image analysis. Fundamental parameters that are commonly quantified by using MR imaging include proton density, diffusion, T1 relaxation, T2 and T2* relaxation, and magnetization transfer. Applications of these MR imaging-quantifiable parameters to abdominal imaging include oncologic imaging, evaluation of diffuse liver disease, and assessment of splenic, renal, and pancreatic disease. An understanding of the inherent physical principles underlying the basic quantitative parameters as well as the commonly used pulse sequences requisite to their derivation is critical, as this field is rapidly growing and its use will likely continue to expand in the clinic. The full potential of quantitative MR imaging applied to abdominal imaging has yet to be realized, but the myriad applications reported to date will undoubtedly continue to grow.

Effect of disease progression on liver apparent diffusion coefficient values in a murine model of NASH at 11.7 Tesla MRI

PURPOSE

To evaluate the apparent diffusion coefficient (ADC) values of liver in a murine model of non-alcoholic steatohepatitis using 11.7 Tesla (T) MRI.

MATERIALS AND METHODS

This animal study was IACUC approved. Seventeen male C57BL/6 mice were divided into control (n = 3) and experimental groups (n = 14) fed a methionine-deficient choline-deficient (MCD) diet to induce steatohepatitis. Livers underwent ex vivo diffusion-weighted MR imaging and ADC maps were calculated. A pathologist determined subjective scores of steatosis, classified from 0 to 3. Digital image analysis was used to determine percentage areas of steatosis. Graphs comparing ADC to subjective and digital image analysis (DIA) determinations of steatosis were plotted.

RESULTS

Subjective assessments of steatosis ranged up to values of 3 and DIA determined areas of steatosis to range up to approximately 16%. ADC values approximated 800 × 10(-6) mm(2) /s (range, 749-811 × 10(-6) mm(2) /s, mean 786 × 10(-6) mm(2) /s) in controls and 500 × 10(-6) mm(2) /s (range, 478-733 × 10(-6) mm(2) /s, mean 625 × 10(-6) mm(2) /s) in experimental mice. Moderate correlation between ADC and subjective scores of steatosis (R = -0.56) was observed. Strong correlation between ADC values and percentage areas of steatosis was between ADC values and percentage areas of steatosis was observed greater (R = -0.81) and very strong correlation was observed with the exclusion of a single outlying data point (R = -0.91).

CONCLUSION

Based on the comparison of ADC values and steatosis determinations by DIA, increasing degrees of steatosis are seen to result in decreased hepatic ADC values.

Freehand liver volumetry by using an electromagnetic pen tablet: accuracy, precision, and rapidity

The purpose of this study is to assess the accuracy, precision, and rapidity of liver volumes calculated by using a freehand electromagnetic pen tablet contourtracing method as compared with the volumes calculated by using the standard optical mouse contourtracing method. The imaging data used as input for accuracy and precision testing were computed by software developed in our institution. This computer software can generate models of solid organs and allows both standard mouse-based and electromagnetic pen-driven segmentation (number of data sets, n = 70). The images used as input for rapidity testing was partly computed by modeling software (n = 70) and partly selected from contrast-enhanced computed tomography (CT) examinations (n = 12). Mean volumes and time required to perform the segmentation, along with standard deviation and range values with both techniques, were calculated. Student's t test was used to assess significance regarding mean volumes and time calculated by using both segmentation techniques on phantom and CT data sets. P value was also calculated. The mean volume difference was significantly lower with the use of the freehand electromagnetic pen as compared with the optical mouse (0.2% vs. 1.8%; P < .001). The mean segmentation time per patient was significantly shorter with the use of the freehand electromagnetic pen contourtracing method (354.5 vs. 499.1 s on phantoms; 457.4 vs. 610.0 s on CT images; P < .001). Freehand electromagnetic pen-based volumetric technique represents a technologic advancement over manual mouse-based contourtracing because of the superior statistical accuracy and sensibly shorter time required. Further studies focused on intra- and interobserver variability of the technique need to be performed before its introduction in clinical application.

Comparison of semiautomated and manual measurements for simulated hypo- and hyper-attenuating hepatic tumors on MDCT: effect of slice thickness and reconstruction increment on their accuracy

RATIONALE AND OBJECTIVES

The aims of this study were to compare accuracy between semiautomated and manual measurements of the longest diameter and volume of simulated hepatic tumors in phantoms and to evaluate the effects of slice thickness (ST) and reconstruction increment (RI) on accuracy.

MATERIALS AND METHODS

Liver phantoms with 45 hypoattenuating and 45 hyperattenuating lesions of different sizes (diameter, 13.3-50.7 mm; volume, 0.4-54.0 mm(3)) and shapes (spherical or elliptical) were scanned using a 64-row multidetector computed tomographic scanner. Images were reconstructed with ST and RI settings of 0.75 and 0.7 mm, 1.0 and 0.7 mm, 1.5 and 1.0 mm, 3.0 and 2.0 mm, 3.0 and 3.0 mm, and 5.0 and 5.0 mm. The longest diameter and volume of each lesion were measured both manually and semiautomatically. To assess accuracy, measurements were compared to reference values by calculating absolute percentage error. Comparisons of absolute percentage error between methods and between ST and RI settings were performed using paired t tests. The degree of correlation between each measurement and a reference value was also assessed.

RESULTS

The semiautomated method showed significantly higher accuracy than the manual method in volume for most ST and RI settings (0.75 and 0.7 mm, 1.0 and 0.7 mm, and 1.5 and 1.0 mm in hypoattenuating lesions and all settings in hyperattenuating lesions; P < .05) and showed similar accuracy in diameter for all ST and RI settings regardless of lesion attenuation (P > .05). Semiautomated measurements also demonstrated higher correlation with reference values than the manual method for both diameter and volume. The absolute percentage error tended to be increased as ST and RI increased for both methods, and acceptable maximum ST and RI in semiautomated method were 1.5 and 1.0 mm.

CONCLUSIONS

Semiautomated computed tomographic measurement showed higher accuracy and correlation than the manual method in measuring the diameter and volume of hepatic lesions. The accuracy of both methods was highly dependent on z-axis resolution.

Liver volume variations as a parameter to assess therapy response in advanced metastatic liver disease

BACKGROUND

The aim of this study was to evaluate liver volume variations (LVV) as a parameter to assess the therapy response in patients with advanced liver metastasis (aLM).

METHODS

Patients with colorectal cancer and consecutive computed tomography (CT) scans were divided into a group with aLM receiving palliative chemotherapy (n = 24) and a control group (n = 21) being followed after curative therapy. Liver volumetry was performed manually. The therapy response was assessed according to the response evaluation criteria in solid tumors (RECIST, n = 69). LVV were compared between groups and to variations in the sum-of-longest-diameter (SLDV). Using receiver operating characteristic (ROC) analysis, LVV were evaluated for distinguishing between progressive disease (PD) and stable disease (SD)/partial remission (PR).

RESULTS

Median LVV between patients with aLM (10.0%) and the control group (4.0%) differed significantly (p < 0.01). PD led to a larger median LVV (26.8%) than PR/SD (5.0%, p < 0.01). LVV in aLM patients correlated positively with SLDV (r = 0.71, p < 0.01). A cut-off value of 9.5% allowed distinguishing between PD and SD/PR (sensitivity: 86%, specificity: 88%, p < 0.01).

CONCLUSION

LVV are helpful to assess PD in patients with aLM.

Adrenal gland volume measurement in septic shock and control patients: a pilot study

OBJECTIVES

To compare adrenal gland volume in septic shock patients and control patients by using semi-automated volumetry.

METHODS

Adrenal gland volume and its inter-observer variability were measured with tomodensitometry using semi-automated software in 104 septic shock patients and in 40 control patients. The volumes of control and septic shock patients were compared and the relationship between volume and outcome in intensive care was studied.

RESULTS

The mean total volume of both adrenal glands was 7.2 ± 2.0 cm(3) in control subjects and 13.3 ± 4.7 cm(3) for total adrenal gland volume in septic shock patients (p < 0.0001). Measurement reproducibility was excellent with a concordance correlation coefficient value of 0.87. The increasing adrenal gland volume was associated with a higher rate of survival in intensive care.

CONCLUSION

The present study reports that with semi-automated software, adrenal gland volume can be measured easily and reproducibly. Adrenal gland volume was found to be nearly double in sepsis compared with control patients. The absence of increased volume during sepsis would appear to be associated with a higher rate of mortality and may represent a prognosis factor which may help the clinician to guide their strategy.

Automated segmentation and quantification of liver and spleen from CT images using normalized probabilistic atlases and enhancement estimation

PURPOSE

To investigate the potential of the normalized probabilistic atlases and computer-aided medical image analysis to automatically segment and quantify livers and spleens for extracting imaging biomarkers (volume and height).

METHODS

A clinical tool was developed to segment livers and spleen from 257 abdominal contrast-enhanced CT studies. There were 51 normal livers, 44 normal spleens, 128 splenomegaly, 59 hepatomegaly, and 23 partial hepatectomy cases. 20 more contrast-enhanced CT scans from a public site with manual segmentations of mainly pathological livers were used to test the method. Data were acquired on a variety of scanners from different manufacturers and at varying resolution. Probabilistic atlases of livers and spleens were created using manually segmented data from ten noncontrast CT scans (five male and five female). The organ locations were modeled in the physical space and normalized to the position of an anatomical landmark, the xiphoid. The construction and exploitation of liver and spleen atlases enabled the automated quantifications of liver/spleen volumes and heights (midhepatic liver height and cephalocaudal spleen height) from abdominal CT data. The quantification was improved incrementally by a geodesic active contour, patient specific contrast-enhancement characteristics passed to an adaptive convolution, and correction for shape and location errors.

RESULTS

The livers and spleens were robustly segmented from normal and pathological cases. For the liver, the Dice/Tanimoto volume overlaps were 96.2%/92.7%, the volume/height errors were 2.2%/2.8%, the root-mean-squared error (RMSE) was 2.3 mm, and the average surface distance (ASD) was 1.2 mm. The spleen quantification led to 95.2%/91% Dice/Tanimoto overlaps, 3.3%/ 1.7% volume/height errors, 1.1 mm RMSE, and 0.7 ASD. The correlations (R2) with clinical/manual height measurements were 0.97 and 0.93 for the spleen and liver, respectively (p < 0.0001). No significant difference (p > 0.2) was found comparing interobserver and automatic-manual volume/height errors for liver and spleen.

CONCLUSIONS

The algorithm is robust to segmenting normal and enlarged spleens and livers, and in the presence of tumors and large morphological changes due to partial hepatectomy. Imaging biomarkers of the liver and spleen from automated computer-assisted tools have the potential to assist the diagnosis of abdominal disorders from routine analysis of clinical data and guide clinical management.

Exact CT-based liver volume calculation including nonmetabolic liver tissue in three-dimensional liver reconstruction

Exact preoperative determination of the liver volume is of great importance prior to hepatobiliary surgery, especially in living donated liver transplantation (LDLT) and extended hepatic resections. Modern surgery-planning systems estimate these volumes from segmented image data. In an experimental porcine study, our aim was (1) to analyze and compare three volume measurement algorithms to predict total liver volume, and (2) to determine vessel tree volumes equivalent to nonmetabolic liver tissue. Twelve porcine livers were examined using a standardized three-phase computed tomography (CT) scan and liver volume was calculated computer-assisted with the three different algorithms. After hepatectomy, livers were weighed and their vascular system plasticized followed by CT scan, CT reconstruction and re-evaluation of total liver and vessel volumes with the three different algorithms. Blood volume determined by the plasticized model was at least 1.89 times higher than calculated by multislice CT scans (9.7% versus 21.36%, P=0.028). Analysis of 3D-CT-volumetry showed good correlation between the actual and the calculated liver volume in all tested algorithms with a high significant difference in estimating the liver volume between Heymsfield versus Heidelberg (P=0.0005) and literature versus Heidelberg (P=0.0060). The Heidelberg algorithm reduced the measuring error with deviations of only 1.2%. The present results suggest a safe and highly predictable use of 3D-volumetry in liver surgery for evaluating liver volumes. With a precise algorithm, the volume of remaining liver or single segments can be evaluated exactly and potential operative risks can therefore be better calculated. To our knowledge, this study implies for the first time a blood pool, which corresponds to nonmetabolic liver tissue, of more than 20% of the whole liver volume.

Atlas-based Automated Segmentation of Spleen and Liver using Adaptive Enhancement Estimation

The paper presents the automated segmentation of spleen and liver from contrast-enhanced CT images of normal and hepato/splenomegaly populations. The method used 4 steps: (i) a mean organ model was registered to the patient CT; (ii) the first estimates of the organs were improved by a geodesic active contour; (iii) the contrast enhancements of liver and spleen were estimated to adjust to patient image characteristics, and an adaptive convolution refined the segmentations; (iv) lastly, a normalized probabilistic atlas corrected for shape and location for the precise computation of each organ's volume and height (mid-hepatic liver height and cephalocaudal spleen height). Results from test data demonstrated the method's ability to accurately segment the spleen (RMS error = 1.09mm; DICE/Tanimoto overlaps = 95.2/91) and liver (RMS error = 2.3mm, and DICE/Tanimoto overlaps = 96.2/92.7). The correlations (R(2)) with clinical/manual height measurements were 0.97 and 0.93 for the spleen and liver respectively.