Quantitative SPECT/CT imaging of actinium-225 for targeted alpha therapy of glioblastomas

BACKGROUND

A new, alternative option for patients with recurrent glioblastoma is targeted alpha therapy (TAT), in the form of a local administration of substance P (neurokinin type 1 receptor ligand, NK-1) labelled with 225Ac. The purpose of the study was to confirm the feasibility of quantitative SPECT imaging of 225Ac, in a model reproducing specific conditions of TAT. In particular, to present the SPECT calibration methodology used, as well as the results of validation measurements and their accuracy. Additionally, to discuss the specific problems related to high noise in the presented case.

MATERIALS AND METHODS

All SPECT/CT scans were conducted using the Symbia T6 equipped with HE collimators, and acquired with multiple energy windows (three main windows: 440 keV, 218 keV, and 78 keV, with three lower scatter energy windows). A Jaszczak phantom with fillable cylindrical sources of various sizes was used to investigate quantitative SPECT/CT imaging characteristics. The planar sensitivity of the camera, an imaging calibration factor, and recovery coefficients were determined. Additionally, the 3D printed model of the glioblastoma tumour was developed and imaged to evaluate the accuracy of the proposed protocol.

RESULTS

Using the imaging calibration factor and recovery coefficients obtained with the Jaszczak phantom, we were able to quantify the activity in a 3D-printed model of a glioblastoma tumour with uncertainty of no more than 10% and satisfying accuracy.

CONCLUSIONS

It is feasible to perform quantitative 225Ac SPECT/CT imaging. However, there are still many more challenges that should be considered for further research on this topic (among others: accurate determination of ICF in the case of high background noise, better method of background estimation for recovery coefficient calculations, other methods for scatter correction than the dual-energy window scatter-compensation method used in this study).

Use of superpixels for improvement of inter-rater and intra-rater reliability during annotation of medical images

In the context of automatic medical image segmentation based on statistical learning, raters' variability of ground truth segmentations in training datasets is a widely recognized issue. Indeed, the reference information is provided by experts but bias due to their knowledge may affect the quality of the ground truth data, thus hindering creation of robust and reliable datasets employed in segmentation, classification or detection tasks. In such a framework, automatic medical image segmentation would significantly benefit from utilizing some form of presegmentation during training data preparation process, which could lower the impact of experts' knowledge and reduce time-consuming labeling efforts. The present manuscript proposes a superpixels-driven procedure for annotating medical images. Three different superpixeling methods with two different number of superpixels were evaluated on three different medical segmentation tasks and compared with manual annotations. Within the superpixels-based annotation procedure medical experts interactively select superpixels of interest, apply manual corrections, when necessary, and then the accuracy of the annotations, the time needed to prepare them, and the number of manual corrections are assessed. In this study, it is proven that the proposed procedure reduces inter- and intra-rater variability leading to more reliable annotations datasets which, in turn, may be beneficial for the development of more robust classification or segmentation models. In addition, the proposed approach reduces time needed to prepare the annotations.

Body composition radiomic features as a predictor of survival in patients with non-small cellular lung carcinoma: A multicenter retrospective study

OBJECTIVES

This study combined two novel approaches in oncology patient outcome predictions-body composition and radiomic features analysis. The aim of this study was to validate whether automatically extracted muscle and adipose tissue radiomic features could be used as a predictor of survival in patients with non-small cell lung cancer.

METHODS

The study included 178 patients with non-small cell lung cancer receiving concurrent platinum-based chemoradiotherapy. Abdominal imaging was conducted as a part of whole-body positron emission tomography/computed tomography performed before therapy. Methods used included automated assessment of the volume of interest using densely connected convolutional network classification model - DenseNet121, automated muscle and adipose tissue segmentation using U-net architecture implemented in nnUnet framework, and radiomic features extraction. Acquired body composition radiomic features and clinical data were used for overall and 1-y survival prediction using machine learning classification algorithms.

RESULTS

The volume of interest detection model achieved the following metric scores: 0.98 accuracy, 0.89 precision, 0.96 recall, and 0.92 F1 score. Automated segmentation achieved a median dice coefficient >0.99 in all segmented regions. We extracted 330 body composition radiomic features for every patient. For overall survival prediction using clinical and radiomic data, the best-performing feature selection and prediction method achieved areas under the curve-receiver operating characteristic (AUC-ROC) of 0.73 (P < 0.05); for 1-y survival prediction AUC-ROC was 0.74 (P < 0.05).

CONCLUSION

Automatically extracted muscle and adipose tissue radiomic features could be used as a predictor of survival in patients with non-small cell lung cancer.

Fully automated 3D body composition analysis and its association with overall survival in head and neck squamous cell carcinoma patients

Objectives

We developed a method for a fully automated deep-learning segmentation of tissues to investigate if 3D body composition measurements are significant for survival of Head and Neck Squamous Cell Carcinoma (HNSCC) patients.

Methods

3D segmentation of tissues including spine, spine muscles, abdominal muscles, subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), and internal organs within volumetric region limited by L1 and L5 levels was accomplished using deep convolutional segmentation architecture - U-net implemented in a nnUnet framework. It was trained on separate dataset of 560 single-channel CT slices and used for 3D segmentation of pre-radiotherapy (Pre-RT) and post-radiotherapy (Post-RT) whole body PET/CT or abdominal CT scans of 215 HNSCC patients. Percentages of tissues were used for overall survival analysis using Cox proportional hazard (PH) model.

Results

Our deep learning model successfully segmented all mentioned tissues with Dice's coefficient exceeding 0.95. The 3D measurements including difference between Pre-RT and post-RT abdomen and spine muscles percentage, difference between Pre-RT and post-RT VAT percentage and sum of Pre-RT abdomen and spine muscles percentage together with BMI and Cancer Site were selected and significant at the level of 5% for the overall survival. Aside from Cancer Site, the lowest hazard ratio (HR) value (HR, 0.7527; 95% CI, 0.6487-0.8735; p = 0.000183) was observed for the difference between Pre-RT and post-RT abdomen and spine muscles percentage.

Conclusion

Fully automated 3D quantitative measurements of body composition are significant for overall survival in Head and Neck Squamous Cell Carcinoma patients.

Performance of Fully Automated Algorithm Detecting Bone Marrow Edema in Sacroiliac Joints

This study evaluates the performance of a fully automated algorithm to detect active inflammation in the form of bone marrow edema (BME) in iliac and sacral bones, depending on the quality of the coronal oblique plane in patients with axial spondyloarthritis (axSpA). The results were assessed based on the technical correctness of MRI examination of the sacroiliac joints (SIJs). A total of 173 patients with suspected axSpA were included in the study. In order to verify the correctness of the MRI, a deviation angle was measured on the slice acquired in the sagittal plane in the T2-weighted sequence. This angle was located between the line drawn between the posterior edges of S1 and S2 vertebrae and the line that marks the actual plane in which the slices were acquired in T1 and STIR sequences. All examinations were divided into quartiles according to the deviation angle measured in degrees as follows: 1st group [0; 2.2], 2nd group (2.2; 5.7], 3rd group (5.7; 10] and 4th group (10; 29.2]. Segmentations of the sacral and iliac bones were acquired manually and automatically using the fully automated algorithm on the T1 sequence. The Dice coefficient for automated bone segmentations with respect to reference manual segmentations was 0.9820 (95% CI [0.9804, 0.9835]). Examinations of BME lesions were assessed using the SPARCC scale (in 68 cases SPARCC > 0). Manual and automatic segmentations of the lesions were performed on STIR sequences and compared. The sensitivity of detection of BME ranged from 0.58 (group 1) to 0.83 (group 2) versus 0.76 (total), while the specificity was equal to 0.97 in each group. The study indicates that the performance of the algorithm is satisfactory regardless of the deviation angle.

Deep Learning Algorithm for Differentiating Patients with a Healthy Liver from Patients with Liver Lesions Based on MR Images

The problems in diagnosing the state of a vital organ such as the liver are complex and remain unresolved. These problems are underscored by frequently published studies on this issue. At the same time, demand for imaging diagnostics, preferably using a method that can detect the disease at the earliest possible stage, is constantly increasing. In this paper, we present liver diseases in the context of diagnosis, diagnostic problems, and possible elimination. We discuss the dataset and methods and present the stages of the pipeline we developed, leading to multiclass segmentation of the liver in multiparametric MR image into lesions and normal tissue. Finally, based on the processing results, each case is classified as either a healthy liver or a liver with lesions. For the training set, the AUC ROC is 0.925 (standard error 0.013 and a p-value less than 0.001), and for the test set, the AUC ROC is 0.852 (standard error 0.039 and a p-value less than 0.001). Further refinements to the proposed pipeline are also discussed. The proposed approach could be used in the detection of focal lesions in the liver and the description of liver tumors. Practical application of the developed multi-class segmentation method represents a key step toward standardizing the medical evaluation of focal lesions in the liver.

Measurement Systems for Use in the Navigation of the Cannula-Guide Assembly within the Deep Regions of the Bronchial Tree

BACKGROUND

The purpose of this paper is to present the spatial navigation system prototype for localizing the distal tip of the cannula-guide assembly. This assembly is shifted through the channel of a bronchoscope, which is fixed in relation to the patient. The navigation is carried out in the bronchial tree, based on maneuvers of the aforementioned assembly.

METHODS

The system consists of three devices mounted on the guide handle and at the entrance to the bronchoscope working channel. The devices record the following values: cannula displacement, rotation of the guide handle, and displacement of the handle ring associated with the bending of the distal tip of the guide.

RESULTS

In laboratory experiments, we demonstrate that the cannula displacement can be monitored with an accuracy of 2 mm, and the angles of rotation and bending of the guide tip with an accuracy of 10 and 20 degrees, respectively, which outperforms the accuracy of currently used methods of bronchoscopy support.

CONCLUSIONS

This accuracy is crucial to ensure that we collect the material for histopathological examination from a precisely defined place. It makes it possible to reach cancer cells at their very early stage.

DeepBeam: a machine learning framework for tuning the primary electron beam of the PRIMO Monte Carlo software

Background

Any Monte Carlo simulation of dose delivery using medical accelerator-generated megavolt photon beams begins by simulating electrons of the primary electron beam interacting with a target. Because the electron beam characteristics of any single accelerator are unique and generally unknown, an appropriate model of an electron beam must be assumed before MC simulations can be run. The purpose of the present study is to develop a flexible framework with suitable regression models for estimating parameters of the model of primary electron beam in simulators of medical linear accelerators using real reference dose profiles measured in a water phantom.

Methods

All simulations were run using PRIMO MC simulator. Two regression models for estimating the parameters of the simulated primary electron beam, both based on machine learning, were developed. The first model applies Principal Component Analysis to measured dose profiles in order to extract principal features of the shapes of the these profiles. The PCA-obtained features are then used by Support Vector Regressors to estimate the parameters of the model of the electron beam. The second model, based on deep learning, consists of a set of encoders processing measured dose profiles, followed by a sequence of fully connected layers acting together, which solve the regression problem of estimating values of the electron beam parameters directly from the measured dose profiles. Results of the regression are then used to reconstruct the dose profiles based on the PCA model. Agreement between the measured and reconstructed profiles can be further improved by an optimization procedure resulting in the final estimates of the parameters of the model of the primary electron beam. These final estimates are then used to determine dose profiles in MC simulations.

Results

Analysed were a set of actually measured (real) dose profiles of 6 MV beams from a real Varian 2300 C/D accelerator, a set of simulated training profiles, and a separate set of simulated testing profiles, both generated for a range of parameters of the primary electron beam of the Varian 2300 C/D PRIMO simulator. Application of the two-stage procedure based on regression followed by reconstruction-based minimization of the difference between measured (real) and reconstructed profiles resulted in achieving consistent estimates of electron beam parameters and in a very good agreement between the measured and simulated photon beam profiles.

Conclusions

The proposed framework is a readily applicable and customizable tool which may be applied in tuning virtual primary electron beams of Monte Carlo simulators of linear accelerators. The codes, training and test data, together with readout procedures, are freely available at the site: https://github.com/taborzbislaw/DeepBeam.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13014-021-01847-w.

POS0958 RESPONSIVENESS OF CONVENTIONAL, SEMI-AUTOMATIC AND FULL-AUTOMATIC METHODS TO QUANTIFY MARROW BONE EDEMA LESIONS IN MRI OF AXIAL SPONDYLOARTHRITIS PATIENTS: A PILOT STUDY

Background:

The presence or absence of marrow bone edema (MBE) in the sacroiliac joints (SIJ) is very important in the diagnosis of axial Spondyloarthritis (axSpA). The quantification of this lesion and its extension may be important to analyze responsiveness of the treatment. Several scoring systems have been proposed for MRI images of SIJ, some of them being observer dependent (Berlin, SPARCC). Others, works in a semi-automatically way, such as the s-SCAISS[1], which makes it possible to quantify the size of the lesion, based on the indication of the expert. Recently, methods like the KITs4R[2], based on a deep patch-based classification network. allow a fully automated detection and quantification of the MBE lesions.

Objectives:

To analyze responsiveness of several scores (observer independent, semiautomatic and full automatic) for quantify MBE in SIJ of axSpA patients.

Methods:

Two rheumatologists independently quantified SIJ images from axSpA patients by visual inspection methods (Berlin and SPARCC indexes) and a semiautomatic system (s-SCAISS) on a single semi-coronal MRI slide (STIR). Full semi-coronal MRI images (15 to 18 slices) were used for an automatic detection algorithm (KITs4R), where total MBE was calculated as sum of areas of MBE in each slice. Patients were assessed before TNF-α therapy (PRE) and 3 months later (POST). Spearman correlations was used to analyze relationship between variables, Wilcoxon signed-rank test for significant differences and Cohen’s d for calculating the effect size of improvement. Figure 1 shows processing of the MRI images: A) Area of MBE selected by the SCAISS, when the rheumatologist “click” on each MBE; B) Lesions detected automatically by KITs4R; C) Automated deep segmentation of bones and subchondral regions (with split into quadrants with central axes of joints also shown. D) Superposed A, B and C images.

Results:

12 axSpA patients were recruited from the CASTRO cohort (42% female, age 46±11 years, disease duration 16±13 years, BMI 28±5). Results PRE and POST are shown in Table 1: mean values (sd), statistical significance (NS, not significant; *, p<0.05; **, p<0.01), and effect size. Activity indexes and CRP were lower. ASDAS and CRP shown significant differences and a large effect size. All MRI scores showed good responsiveness (ES medium-large, p<0.01), specially KITs4R. Agreement between all MRI scores were high (r>0.8;p<0.01). Between semiautomatic and automatic methods, this agreement was also excellent (r=0.92;p<0.001). Correlation in improvements (reductions in scores) were also significant between all MRI scores (r>0.7;p<0.05).

PREPOSTSignESBASDAI6.12 (2.45)4.96 (2.74)N.S.0.44-SmallASDAS3.61 (1.04)2.58 (1.25)*0.89-LargeCRP11.26 (8.93)4.81 (4.08)**0.84-LargeBERLIN2.58 (1.98)1.17 (1.85)**0.74-MediumSPARCC3.92 (3.42)1.58 (2.43)**0.69-MediumSCAISS295 (332)95 (163)**0.68-MediumKITs4R1671 (1596)258 (421)**1.04-Large

NS, not significant; *, p<0.05; **, p<0.01

Conclusion:

All MRI scores have good level of agreement between them and good responsiveness. Berlin and SPARCC are observer dependent, and do not quantify the extension of the MBE area. s-SCAIS helps to this quantification. KITs4R is not observer dependent but clinimetric validation, analizing agreement level with human expert, is necessary. New advanced tools are improving quantitative and objective measurement of BME which is important to analyze responsiveness.

References:

[1]Development and validation of SCAISS, a tool for semi-automated quantification of sacroilitis by magnetic resonance in spondyloarthritis. Rheumatol Int. 2018 Oct;38(10):1919-1926.

[2]The semi-automated algorithm for the detection of bone marrow oedema lesions in patients with axial spondyloarthritis. Rheumatol Int. 2020 Apr;40(4):625-633.

Disclosure of Interests:

None declared.

A simulation-based method for evaluating geometric tests of a linac c-arm in quality control in radiotherapy

Purpose

Assessment of the accuracy of geometric tests of a linac used in external beam therapy is crucial for ensuring precise dose delivery. In this paper, a new simulation‐based method for assessing accuracy of such geometric tests is proposed and evaluated on a set of testing procedures.

Methods

Linac geometry testing methods used in this study are based on an established design of a two‐module phantom. Electronic portal imaging device (EPID) images of fiducial balls contained in these modules can be used to automatically reconstruct linac geometry. The projection of the phantom modules fiducial balls onto the EPID detector plane is simulated for assumed nominal geometry of a linac. Then, random errors are added to the coordinates of the projections of the centers of the fiducial balls and the linac geometry is reconstructed from these data.

Results

Reconstruction is performed for a set of geometric test designs and it is shown how the dispersion of the reconstructed values of geometric parameters depends on the design of a geometric test. Assuming realistic accuracy of EPID image analysis, it is shown that for selected testing plans the reconstruction accuracy of geometric parameters can be significantly better than commonly used action thresholds for these parameters.

Conclusions

Proposed solution has the potential to improve geometric testing design and practice. It is an important part of a fully automated geometric testing solution.

Automated measurement of parameters related to the deformities of lower limbs based on x-rays images

Measurement of the deformation of the lower limbs in the current standard full-limb X-rays images presents significant challenges to radiologists and orthopedists. The precision of these measurements is deteriorated because of inexact positioning of the leg during image acquisition, problems with selecting reliable anatomical landmarks in projective X-ray images, and inevitable errors of manual measurements. The influence of the random errors resulting from the last two factors on the precision of the measurement can be reduced if an automated measurement method is used instead of a manual one. In the paper a framework for an automated measurement of various metric and angular quantities used in the description of the lower extremity deformation in full-limb frontal X-ray images is described. The results of automated measurements are compared with manual measurements. These results demonstrate that an automated method can be a valuable alternative to the manual measurements.

Automated assessment of synovitis in 0.2T magnetic resonance images of the wrist

According to the current recommendations in diagnosis of rheumatoid arthritis (RA), Magnetic Resonance (MR) images of wrist joints are used to evaluate three main types of lesions: synovitis, bone edema and bone erosions. In the clinical practice, the RA-related lesions seen in MR images are assessed manually with the semi-quantitative RAMRIS scoring system. In this paper we present an automated method for inflamed synovial membrane volume determination, based on the analysis of pre- and post-contrast MR images and segmentation of wrist bones seen in MR images. We found that the correlation between the automatically quantified volume of synovitis and RAMRIS scores was in the range from 0.76 to 0.87 for the total RAMRIS synovitis score. This can be compared with the correlation between the manually quantified volume of synovitis and RAMRIS scores which was in the range from 0.75 to 0.81 for the total synovitis score. The results of the study demonstrate that computer assisted methods for assessment of synovitis have great potential for clinical applications.

A microcomputed tomography-based comparison of root canal filling quality following different instrumentation and obturation techniques

OBJECTIVE

With a microcomputed tomography (microCT) imaging device, we aimed to quantitatively evaluate root canal fillings after commonly used endodontic procedures and also tested the suitability of microCT for this purpose.

MATERIALS AND METHODS

Eighty single roots were instrumented and obturated with gutta-percha and Tubli-Seal. They were divided into 4 groups of 20. The Hand groups were instrumented with hand files and filled with thermoplastic (Th) compaction and cold lateral (CL) condensation, i.e. Hand-Th and Hand-CL, respectively. The Rot groups, i.e. Rot-Th and Rot-CL, were instrumented with a rotary ProFile system and filled as above. The roots were scanned and 3-dimensional (3D) visualization was obtained. The number, size, percentage of volume and distribution of voids at the filling/dentine interface (i-voids) and voids surrounded by filling material (s-voids) were measured.

RESULTS

Canal fillings differed significantly with regard to the size of both types of voids and the average number of i-voids. All canals presented a low volume of voids. The highest percentage (0.69%) was found for i-voids in the Hand-CL group, while the lowest volume (0.11% for s-voids and 0.14% for i-voids) was in the Hand-Th canals. Apically, in the last 3 mm, i-voids were observed mainly in the Th groups, and s-voids occurred mostly in the coronal part of the canal filling in all cases.

CONCLUSION

MicroCT was a useful tool for 3D quantitative evaluations of these root canal fillings. None of the root canal instrumentation and filling methods ensured void-free obturation. CL condensation produced mainly i-voids. With Th compaction, internal s-voids were particularly common, but there were mainly i-voids in the apical part.

An automatic recognition of the frontal sinus in x-ray images of skull

In the present paper, a procedure is developed for fully automatic recognition of the frontal sinus in cranial radiographs. An X-ray image of a whole skull is required at the input of the procedure, which consists of three subsequent steps: the selection of a rectangular region of interest, containing the sinus; detection of the line of brow ridges; and fronto-nasal suture, detection of the borders of the frontal sinus. The recognition algorithm is based on a method of connectivity-preserving thresholding, introduced in the present study, and on watersheds from markers. Totally, 50 X-ray images have been analyzed. The frontal sinus borders were recognized correctly in 41 cases.

What the training of a neuronal network optimizes

In the study a model of training of neuronal networks built of integrate-and-fire neurons is investigated. Neurons are assembled into complex networks of Watts-Strogatz type. Every neuronal network contains a single receptor neuron. The receptor neuron, stimulated by an external signal, evokes spikes in equal time intervals. The spikes generated by the receptor neuron induce subsequent activity of a whole network. The depolarization signals, traveling the network, modify synaptic couplings according to a kick-and-delay rule, whose process is termed "training." It is shown that the training decreases the mean length of paths along which a depolarization signal is transmitted from the receptor neuron. Consequently, the training also decreases the reaction time and the energy expense necessary for the network to react to the external stimulus. It is shown that the initial distribution of synaptic couplings crucially determines the performance of trained networks.

Quantifying deterioration of bone tissue from grey-level images

Quantification of the severity of the deterioration in bone tissue is an important diagnostic problem involving the assessment of connectedness of the tissue from its grey-level images. In this study a fuzzy parameter was introduced for quantifying the severity of discontinuities of network-like structures. For each two pixels of an analyzed image, the fuzzy parameter was derived from the grey-level intensity along the brightest path connecting them. The performance of this parameter was tested on images of continuous and discontinuous samples of vertebral trabecular bone, matched for mean gray-level intensity. It was shown that the values of the parameter were significantly different (P-value<10(-6)) in both groups and equal to 0.02+/-0.11 and 0.46+/-0.22 for discontinuous and continuous samples, respectively. The decrease in the fuzzy parameter could be interpreted as the result of either an increasing gap size in discontinuous structures or a decreasing thickness of structure elements in continuous structures. A possible application of the method for use in quantifying severity of discontinuities within the vertebral cortical rim was described.

Quantifying anisotropy of trabecular bone from gray-level images

In this study, a gray-level image-based approach to quantifying structural anisotropy is described. The secant modulus was estimated for thirty L(3) vertebral bodies using nondestructive testing. The vertebral bodies were imaged with a clinical CT scanner. QCT measurements of BMD were also performed for trabecular regions. Structural anisotropy in trabecular regions was quantified from binarized images using the mean intercept length (MIL) method and from gray-level images using the gray-level structure tensor (GST) method. BMD alone explained 28% of the variation of the secant modulus. Multivariable regression models combining BMD and measures of anisotropy, as proposed by the relations formulated by Cowin, revealed significant improvement in the prediction of the secant modulus. Combining a principal value of the fabric tensor, as computed by either MIL or GST methods, with BMD resulted in increased correlation with the secant modulus. The highest correlation (R(2)=0.81) was achieved with a combination of BMD and the third principal value of the GST. Adding a term proportional to the minimal cross-sectional area of the vertebral body explained 86% of the variation of the secant modulus.

Estimating structural properties of trabecular bone from gray-level low-resolution images

In this paper the relationship between three-dimensional histomorphometric parameters derived from microCT and MRI images of distal radius trabecular bone samples is studied. microFE analysis of the trabecular samples is performed and Young's modulus for cranio-caudal direction is calculated. Most of the MRI and microCT parameters correlate significantly with, respectively, MRI and microCT estimates of bone volume fraction. For some of the parameters strong correlation between microCT and MRI results is also observed. However, in these cases there simultaneously exists correlation between: microCT parameter and microCT bone volume fraction; microCT and MRI bone volume fraction; MRI bone volume fraction and MRI parameter. It is found that, comparing to bone volume fraction, histomorphometric information derived from binarized MRI images does not improve estimation of the Young's modulus of trabecular bone samples (calculated for "gold standard" microCT data). Thus a novel method of "optimal paths" analysis of gray-level MRI images is introduced. "Optimal paths" parameters improve estimation of the Young's modulus of trabecular bone samples. They also provide surrogate, gray-level image-based measure of trabecular thickness.

Optimal response function in networks of excitatory elements

In this paper the problem of signal propagation in networks of excitatory elements is studied. It is found that the geometry of signal transmission paths depends crucially on how an excitatory element responds to a stimulus. Two types of responses are defined: fast and slow. In the slow response case the signal transmission paths are in the same universality class as optimal paths in the limit of strong disorder. The signal transmission paths formed in the fast response case constitute possibly a new universality class.

Optimal cut of trabecular network

It has been shown recently that failure of mechanically tested trabecular bone samples can be localized within a part of the volume of the samples. Bone volume fraction BV/TV of failure regions was found to be a better determinant of the mechanical competence of the specimens and was smaller than BV/TV of the whole samples. These results suggest that localization of a failure within a part of an inhomogeneous trabecular network can be related to the presence of a surface of minimal cut-a surface separating the analyzed trabecular sample into two disjoint parts in such a way, that the separation requires removal of minimal possible amount of bone material. Thus, to properly address the problem of mechanical competence of a trabecular bone sample, one must be able to detect and describe the surface of minimal cut. In this paper an algorithm localizing surfaces of minimal cut within 3D trabecular structures is introduced.

From simple to complex networks: inherent structures, barriers, and valleys in the context of spin glasses

Given discrete degrees of freedom (spins) on a graph interacting via an energy function, what can be said about the energy local minima and associated inherent structures? Using the lid algorithm in the context of a spin glass energy function, we investigate the properties of the energy landscape for a variety of graph topologies. First, we find that the multiplicity N(s) of the inherent structures generically has a log-normal distribution. In addition, the large volume limit of ln <N(s)>/<ln N(s)> differs from unity, except for the Sherrington-Kirkpatrick model. Second, we find simple scaling laws for the growth of the height of the energy barrier between the two degenerate ground states and the size of the associated valleys. For finite connectivity models, changing the topology of the underlying graph does not modify qualitatively the energy landscape, but at the quantitative level the models can differ substantially.

Novel algorithm detecting trabecular termini in muCT and MRI images

In this paper, a novel algorithm detecting trabecular termini in three-dimensional images of trabecular bone is introduced. The algorithm is applied to the analysis of muCT and MRI images of distal radius trabecular bone samples. In muCT images, the volume of the trabecular termini constitutes at most 2.1% of the bone volume fraction BV/TV and is typically smaller than 1% of BV/TV. Isolated trabeculae are not observed in the interior of the trabecular bone samples. Trabecular bone structure assessed with muCT appears thus highly optimized. The volume and the number of the trabecular termini do not correlate with BV/TV. These quantities do not correlate also with apparent Young's modulus of the samples. In contrast in MRI images, segmented with the dual reference limit method, the volume of the trabecular termini and the volume of isolated parts constitute even up to 14% of the apparent bone volume fraction App.BV/TV. For MRI images, the volume of the trabecular termini increases significantly with decreasing App.BV/TV. The volume and the number of the trabecular termini, derived from MRI images do not correlate with Young's modulus. There is also no correlation between the number and the volume of the trabecular termini, estimated from MRI and muCT images. The volume of the trabecular termini is overestimated 15 times in MRI images. App.BV/TV correlates strongly with BV/TV. Young's modulus derived from MRI images correlates strongly with Young's modulus found for muCT data. It is shown that the diagnostic significance of latter result is highly limited.