Toward online modeling for lesion visualization and monitoring in cardiac ablation therapy

Despite extensive efforts to enhance catheter navigation, limited research has been done to visualize and monitor the tissue lesions created during ablation in the attempt to provide feedback for effective therapy. We propose a technique to visualize the temperature distribution and extent of induced tissue injury via an image-based model that uses physiological tissue parameters and relies on heat transfer principles to characterize lesion progression in near real time. The model was evaluated both numerically and experimentally using ex vivo bovine muscle samples while emulating a clinically relevant ablation protocol. Results show agreement to within 5 degreeC between the model-predicted and experimentally measured end-ablation tissue temperatures, as well as comparable predicted and observed lesion characteristics. The model yields temperature and lesion updates in near real-time, thus providing reasonably accurate and sufficiently fast monitoring for effective therapy.

Centerline tracking for quantification of reverse structural remodeling of the pulmonary veins following cardiac ablation therapy

RATIONALE AND OBJECTIVES

Patients with atrial fibrillation undergo structural remodeling resulting in increased pulmonary vein sizes. Studies have demonstrated that these changes are reversible following successful ablation therapy. To date, analyses of pulmonary vein structure have focused on measurements at the pulmonary vein ostia, and the full extent of reverse remodeling along the length of the pulmonary veins has not yet been fully characterized.

MATERIALS AND METHODS

An automated, three-dimensional method is proposed that quantifies pulmonary vein geometry starting at the ostia and extending several centimeters into the veins. A centerline is tracked along the length of the pulmonary vein, and orthogonal planes are computed along the curve. The method was validated against manual measurements on each of the four pulmonary veins for 10 subjects. The proposed methodology was used to analyze the pulmonary veins in 21 patients undergoing cardiac ablation therapy with preoperative and postoperative computed tomographic scans.

RESULTS

Validation results demonstrated that the automated measurements closely followed the manual measurements, with an overall mean difference of 11.50 mm(2). Significant differences in cross-sectional area at the two time points were observed at all pulmonary vein ostia and extending for 2.0 cm (excluding the 0.5-cm interval) into the left inferior pulmonary vein, 3.5 cm into the left superior pulmonary vein, and 2.0 cm into the right superior pulmonary vein.

CONCLUSIONS

Quantitative analysis along the length of the pulmonary veins can be accomplished using centerline tracking and measurements from orthogonal planes along the curve. The patient study demonstrated that reverse structural remodeling following ablation therapy occurs not only at the ostia but for several centimeters extending into the pulmonary veins.

Calibration and Evaluation of a Magnetically Tracked ICE Probe for Guidance of Left Atrial Ablation Therapy

The novel prototype system for advanced visualization for image-guided left atrial ablation therapy developed in our laboratory permits ready integration of multiple imaging modalities, surgical instrument tracking, interventional devices and electro-physiologic data. This technology allows subject-specific procedure planning and guidance using 3D dynamic, patient-specific models of the patient's heart, augmented with real-time intracardiac echocardiography (ICE). In order for the 2D ICE images to provide intuitive visualization for accurate catheter to surgical target navigation, the transducer must be tracked, so that the acquired images can be appropriately presented with respect to the patient-specific anatomy. Here we present the implementation of a previously developed ultrasound calibration technique for a magnetically tracked ICE transducer, along with a series of evaluation methods to ensure accurate imaging and faithful representation of the imaged structures. Using an engineering-designed phantom, target localization accuracy is assessed by comparing known target locations with their transformed locations inferred from the tracked US images. In addition, the 3D volume reconstruction accuracy is also estimated by comparing a truth volume to that reconstructed from sequential 2D US images. Clinically emulating validation studies are conducted using a patient-specific left atrial phantom. Target localization error of clinically-relevant surgical targets represented by nylon fiducials implanted within the endocardial wall of the phantom was assessed. Our studies have demonstrated 2.4 ± 0.8 mm target localization error in the engineering-designed evaluation phantoms, 94.8 ± 4.6 % volume reconstruction accuracy, and 3.1 ± 1.2 mm target localization error in the left atrial-mimicking phantom. These results are consistent with those disseminated in the literature and also with the accuracy constraints imposed by the employed technology and the clinical application.

Relationship of femoral neck areal bone mineral density to volumetric bone mineral density, bone size, and femoral strength in men and women

Using combined dual-energy X-ray absorptiometry (DXA) and quantitative computed tomography, we demonstrate that men matched with women for femoral neck (FN) areal bone mineral density (aBMD) have lower volumetric BMD (vBMD), higher bone cross-sectional area, and relatively similar values for finite element (FE)-derived bone strength.

INTRODUCTION

aBMD by DXA is widely used to identify patients at risk for osteoporotic fractures. aBMD is influenced by bone size (i.e., matched for vBMD, larger bones have higher aBMD), and increasing evidence indicates that absolute aBMD predicts a similar risk of fracture in men and women. Thus, we sought to define the relationships between FN aBMD (assessed by DXA) and vBMD, bone size, and FE-derived femoral strength obtained from quantitative computed tomography scans in men versus women.

METHODS

We studied men and women aged 40 to 90 years and not on osteoporosis medications.

RESULTS

In 114 men and 114 women matched for FN aBMD, FN total cross-sectional area was 38% higher (P < 0.0001) and vBMD was 16% lower (P < 0.0001) in the men. FE models constructed in a subset of 28 women and 28 men matched for FN aBMD showed relatively similar values for bone strength and the load-to-strength ratio in the two groups.

CONCLUSIONS

In this cohort of young and old men and women from Rochester, MN, USA who are matched by FN aBMD, because of the offsetting effects of bone size and vBMD, femoral strength and the load-to-strength ratio tended to be relatively similar across the sexes.

Augmented environments for minimally invasive therapy: implementation barriers from technology to practice

Augmented environments for medical applications have been explored and developed in an effort to enhance the clinician's view of anatomy and facilitate the performance of minimally invasive procedures. These environments must faithfully represent the real surgical field and require seamless integration of pre- and intra-operative imaging, surgical instrument tracking and display technology into a common framework centered around the patient. However, few image guidance environments have been successfully translated into clinical use. Several challenges that contribute to the slow progress of integrating such environments into clinical practice are discussed here in terms of both technical and clinical limitations.

Detail-on-demand visualization for lean understanding of lung abnormalities

In some respects, the lung is an anatomical bog - having limited referential landmarks. Nonetheless, precise understanding of the abnormalities that inflict this organ is crucial to effective clinical diagnosis and treatment. However, wading interactively through a three-dimensional scan of the lung poses a visual quagmire to the radiologist, resulting in significant interpretive differences due to inter and intra observer variation. Despite the continuing progress in quantitative imaging, lack of unambiguous visualization with accurately, relevant cues severely hinders the clinical adoption of many computational tools. We address this unmet need through a lean visualization paradigm wherein information is presented hierarchically to provide an interactive macro-to-micro view of lung pathologies. At the macro level, the structural and functional information is summarized into a synoptic glyph that is readily interpreted and correlated to a priori known disease states. The glyphs are "patho-spatio-temporally" tagged to facilitate navigation through the level-of-detail scales, down to the micro level values in the image voxels, providing quantitative interpretation of tissue type and the confidence level in the quantitation. A novel volume compositing scheme is proposed to specify and guide to the optimal site for surgical lung biopsy. This intuitive, interactive interface for rapid and unambiguous navigation towards the clinical endpoint harnesses the power of bio-informatics technology to provide an efficient, clinically relevant and comprehensive summary of pulmonary disease, including precise location, spatial extent and intrinsic character.

Accuracy considerations in image-guided cardiac interventions: experience and lessons learned

MOTIVATION

Medical imaging and its application in interventional guidance has revolutionized the development of minimally invasive surgical procedures leading to reduced patient trauma, fewer risks, and shorter recovery times. However, a frequently posed question with regard to an image guidance system is "how accurate is it?" On one hand, the accuracy challenge can be posed in terms of the tolerable clinical error associated with the procedure; on the other hand, accuracy is bound by the limitations of the system's components, including modeling, patient registration, and surgical instrument tracking, all of which ultimately impact the overall targeting capabilities of the system.

METHODS

While these processes are not unique to any interventional specialty, this paper discusses them in the context of two different cardiac image guidance platforms: a model-enhanced ultrasound platform for intracardiac interventions and a prototype system for advanced visualization in image-guided cardiac ablation therapy.

RESULTS

Pre-operative modeling techniques involving manual, semi-automatic and registration-based segmentation are discussed. The performance and limitations of clinically feasible approaches for patient registration evaluated both in the laboratory and in the operating room are presented. Our experience with two different magnetic tracking systems for instrument and ultrasound transducer localization is reported. Ultimately, the overall accuracy of the systems is discussed based on both in vitro and preliminary in vivo experience.

CONCLUSION

While clinical accuracy is specific to a particular patient and procedure and vastly dependent on the surgeon's experience, the system's engineering limitations are critical to determine whether the clinical requirements can be met.

A piecewise patch-to-model matching method for image-guided cardiac catheter ablation

Accurate and fast fusion and display of real-time images of anatomy and associated data is critical for effective use in image guided procedures, including image guided cardiac catheter ablation. We have developed a piecewise patch-to-model matching method, a modification of the contractive projection point technique, for accurate and rapid matching between an intra-operative cardiac surface patch and a pre-operative cardiac surface model. Our method addresses the problems of fusing multi-modality images and using non-rigid deformation between a surface patch and a surface model. A projection lookup table, K-nearest neighborhood search, and a final iteration of point-to-projection are used to reliably find the surface correspondence. Experimental results demonstrate that the method is fast, accurate and robust for real-time matching of intra-operative surface patches to pre-operative 3D surface models of the left atrium.

Fast adaptation of pre-operative patient specific models to real-time intra-operative volumetric data streams

Image-guided catheter ablation therapy is becoming an increasingly popular treatment option for atrial fibrillation. Successful treatment relies on accurate guidance of the treatment catheter. Integration of high-resolution, pre-operative data with electrophysiology data and positional data from tracked catheters improves targeting, but lacks the means to monitor changes in the atrial wall. Intra-operative ultrasound provides a method for imaging the atrial wall, but the real-time, dynamic nature of the data makes it difficult to seamlessly integrate with the static pre-operative patient-specific model. In this work, we propose a technique which uses a self-organizing map (SOM) for dynamically adapting a pre-operative model to surface patch data. The surface patch would be derived from a segmentation of the anatomy in a real-time, intra-operative ultrasound data stream. The method is demonstrated on two regular geometric shapes as well as data simulated from a real, patient computed tomography dataset.

Relation of vertebral deformities to bone density, structure, and strength

Because they are not reliably discriminated by areal bone mineral density (aBMD) measurements, it is unclear whether minimal vertebral deformities represent early osteoporotic fractures. To address this, we compared 90 postmenopausal women with no deformity (controls) with 142 women with one or more semiquantitative grade 1 (mild) deformities and 51 women with any grade 2-3 (moderate/severe) deformities. aBMD was measured by dual-energy X-ray absorptiometry (DXA), lumbar spine volumetric bone mineral density (vBMD) and geometry by quantitative computed tomography (QCT), bone microstructure by high-resolution peripheral QCT at the radius (HRpQCT), and vertebral compressive strength and load-to-strength ratio by finite-element analysis (FEA) of lumbar spine QCT images. Compared with controls, women with grade 1 deformities had significantly worse values for many bone density, structure, and strength parameters, although deficits all were much worse for the women with grade 2-3 deformities. Likewise, these skeletal parameters were more strongly associated with moderate to severe than with mild deformities by age-adjusted logistic regression. Nonetheless, grade 1 vertebral deformities were significantly associated with four of the five main variable categories assessed: bone density (lumbar spine vBMD), bone geometry (vertebral apparent cortical thickness), bone strength (overall vertebral compressive strength by FEA), and load-to-strength ratio (45-degree forward bending ÷ vertebral compressive strength). Thus significantly impaired bone density, structure, and strength compared with controls indicate that many grade 1 deformities do represent early osteoporotic fractures, with corresponding implications for clinical decision making.

Seed localization and TRUS-fluoroscopy fusion for intraoperative prostate brachytherapy dosimetry

OBJECTIVE

To develop and evaluate an integrated approach to intra-operative dosimetry for permanent prostate brachytherapy (PPB) by combining a fluoroscopy-based seed localization routine with a transrectal ultrasound (TRUS)-to-fluoroscopy fusion technique.

MATERIALS AND METHODS

Three-dimensional seed coordinates are reconstructed based on the two-dimensional seed locations identified from three fluoroscopic images acquired at different angles. A seed-based registration approach was examined in both simulation and phantom studies to register the seed locations identified from the fluoroscopic images to the TRUS images. Dose parameters were then evaluated and compared to CT-based dosimetry from a patient dataset.

RESULTS

Less than 0.2% error in the D90 value was observed using the TRUS-fluoroscopy image-fusion-based method relative to the CT-based post-implantation dosimetry. In the phantom study, an average distance of 3 mm was observed between the seeds identified from TRUS and the reconstructed seeds at registration. Isodose contours were displayed superimposed on the TRUS images.

CONCLUSIONS

Promising results were observed in this preliminary study of a TRUS-fluoroscopy fusion-based brachytherapy dosimetry analysis method, suggesting that the method is highly sensitive and calculates clinically relevant dosimetry, including the prostate D90. Further validation of the method is required for eventual clinical application.

An event-driven distributed processing architecture for image-guided cardiac ablation therapy

Medical imaging data is becoming increasing valuable in interventional medicine, not only for preoperative planning, but also for real-time guidance during clinical procedures. Three key components necessary for image-guided intervention are real-time tracking of the surgical instrument, aligning the real-world patient space with image-space, and creating a meaningful display that integrates the tracked instrument and patient data. Issues to consider when developing image-guided intervention systems include the communication scheme, the ability to distribute CPU intensive tasks, and flexibility to allow for new technologies. In this work, we have designed a communication architecture for use in image-guided catheter ablation therapy. Communication between the system components is through a database which contains an event queue and auxiliary data tables. The communication scheme is unique in that each system component is responsible for querying and responding to relevant events from the centralized database queue. An advantage of the architecture is the flexibility to add new system components without affecting existing software code. In addition, the architecture is intrinsically distributed, in that components can run on different CPU boxes, and even different operating systems. We refer to this Framework for Image-Guided Navigation using a Distributed Event-Driven Database in Real-Time as the FINDER architecture. This architecture has been implemented for the specific application of image-guided cardiac ablation therapy. We describe our prototype image-guidance system and demonstrate its functionality by emulating a cardiac ablation procedure with a patient-specific phantom. The proposed architecture, designed to be modular, flexible, and intuitive, is a key step towards our goal of developing a complete system for visualization and targeting in image-guided cardiac ablation procedures.

Virtual pelvic anatomy and surgery simulator: an innovative tool for teaching pelvic surgical anatomy

Pelvic three-dimensional (3-D) anatomy is complex and can be difficult to visualize. We have developed an alpha-version virtual anatomic model that will facilitate a) student learning of anatomy b) resident preparedness for operation, c) surgical planning for complex, multidisciplinary procedures. This educational module is composed of the virtual anorectum, an anatomically accurate, interactive model of the anorectum and pelvic viscera based on co-registered human computed tomographic (CT) and magnetic resonance (MR) images. It also consists of a physical model fabricated from the same data-set using stereolithography. Usability and functionality was assessed in a sample of surgical trainees. Feedback demonstrated that the module is an efficient adjunct to medical education and surgical planning. Ultimately, we envision a deformable virtual model that allows residents and surgeons to rehearse complex surgical procedures by performing virtual surgery, thereby improving preparedness for real operations. Such a model will be integrated into a colorectal surgery educational curriculum, and in the future may represent a unique method of competency testing.

Application of magnetic resonance imaging to measure fasting and postprandial volumes in humans

Our aims were to measure the gastric volume response in excess of ingested meal volume (i.e. gastric accommodation), contribution of swallowed air to this excess, day-to-day variability of gastric volumes measured by MRI and their relationship to volumes measured by single-photon-emission computed tomography (SPECT). In 20 healthy volunteers, fasting and postprandial gastric volumes were measured after technetium(99m)-pertechnetate labeling of the gastric mucosa by SPECT and separately by MRI, using 3D gradient echo and 2D half-Fourier acquisition single-shot turbo spin echo (HASTE) sequences. Ten of these subjects had a second MRI exam to assess intra-individual variation. Thereafter, another 10 subjects had two MRI studies during which they ingested the nutrient in 30 or 150 mL aliquots. During MRI, the postprandial gastric volume change exceeded the ingested meal volume by 106 +/- 12 mL (Mean +/- SEM). The HASTE and gradient echo sequences distinguished air from fluid under fasting and postprandial conditions respectively. This postprandial excess mainly comprised air (61 +/- 5 mL), which was not significantly different when ingested as 30 or 150 mL aliquots. Fasting and postprandial gastric volumes measured by MRI were generally reproducible within subjects. During SPECT, postprandial volumes increased by 158 +/- 18 mL; gastric volumes measured by SPECT were higher than MRI. MRI measures gastric volumes with acceptable performance characteristics; the postprandial excess primarily consists of air, which is not affected by the mode of ingestion. Gastric volumes are technique specific and differ between MRI and SPECT.

A bone structural basis for fracture risk in diabetes

CONTEXT

Elevated areal bone mineral density (aBMD) in type 2 diabetes mellitus is inconsistent with increased fracture risk at some skeletal sites.

OBJECTIVES

Because aBMD is an imperfect surrogate for bone strength, we assessed bone structure and strength more directly using quantitative computed tomography.

DESIGN

Diabetic and nondiabetic subjects were evaluated in a cross-sectional study.

SETTING

Subjects were recruited from a random sample of the Rochester, MN, population.

PARTICIPANTS

Forty-nine subjects (28 women and 21 men) with type 2 diabetes were compared with age- and sex-matched nondiabetic controls.

MAIN OUTCOME MEASUREMENTS

We measured bone geometry, strength, and volumetric BMD (vBMD) at the hip, spine, and wrist, along with hip aBMD, using central and peripheral quantitative computed tomography and estimated bone load to bone strength ratios at each site.

RESULTS

Adjusted for differences in body mass index between cases and controls (29.8 vs. 27.6), hip aBMD was greater in diabetic subjects, but this was accounted for by greater trabecular vBMD. Cortical vBMD was similar in the two groups, as was bone cross-sectional area and cortical thickness. Bone strength measures were generally better in diabetic subjects, but bone loads were higher from their greater weight. Consequently, load to strength ratios (i.e. factor-of-risk) were similar.

CONCLUSIONS

Patients with type 2 diabetes enjoy little benefit from elevated aBMD in terms of improved bone load to strength ratios. With no deficit in bone density, the rationale for antiresorptive therapy in diabetic patients is uncertain, but potential adverse effects of diabetes on bone quality need more study.

Anatomy for biomedical engineers

There is a perceived need for anatomy instruction for graduate students enrolled in a biomedical engineering program. This appeared especially important for students interested in and using medical images. These students typically did not have a strong background in biology. The authors arranged for students to dissect regions of the body that were of particular interest to them. Following completion of all the dissections, the students presented what they had learned to the entire class in the anatomy laboratory. This course has fulfilled an important need for our students.

Patient specific physical anatomy models

The advent of small footprint stereo-lithographic printers and the ready availability of segmentation and surface modeling software provide a unique opportunity to create patient-specific physical models of anatomy, validation of image guided intervention applications against phantoms that exhibit naturally occurring anatomic variation. Because these models can incorporate all structures relevant to a procedure, this allows validation to occur under realistic conditions using the same or similar techniques as would be used in a clinical application. This in turn reduces the number of trials and time spent performing in-vivo validation experiments. In this paper, we describe our general approach for the creation of both non-tissue and tissue-mimicking patient-specific models as part of a general-purpose patient emulation system used to validate image guided intervention applications.

Structural determinants of vertebral fracture risk

Vertebral fractures are more strongly associated with specific bone density, structure, and strength parameters than with areal BMD, but all of these variables are correlated.

INTRODUCTION

It is unclear whether the association of areal BMD (aBMD) with vertebral fracture risk depends on bone density per se, bone macro- or microstructure, overall bone strength, or spine load/bone strength ratios.

MATERIALS AND METHODS

From an age-stratified sample of Rochester, MN, women, we identified 40 with a clinically diagnosed vertebral fracture (confirmed semiquantitatively) caused by moderate trauma (cases; mean age, 78.6 +/- 9.0 yr) and compared them with 40 controls with no osteoporotic fracture (mean age, 70.9 +/- 6.8 yr). Lumbar spine volumetric BMD (vBMD) and geometry were assessed by central QCT, whereas microstructure was evaluated by high-resolution pQCT at the ultradistal radius. Vertebral failure load ( approximately strength) was estimated from voxel-based finite element models, and the factor-of-risk (phi) was determined as the ratio of applied spine loads to failure load.

RESULTS

Spine loading (axial compressive force on L3) was similar in vertebral fracture cases and controls (e.g., for 90 degrees forward flexion, 2639 versus 2706 N; age-adjusted p = 0.173). However, fracture cases had inferior values for most bone density and structure variables. Bone strength measures were also reduced, and the factor-of-risk (phi) was 35-37% greater (worse) among women with a vertebral fracture. By age-adjusted logistic regression, relative risks for the strongest fracture predictor in each of the five main variable categories were bone density (total lumbar spine vBMD: OR per SD change, 2.2; 95% CI, 1.1-4.3), bone geometry (vertebral apparent cortical thickness: OR, 2.1; 95% CI, 1.1-4.1), bone microstructure (none significant); bone strength ("cortical" [outer 2 mm] compressive strength: OR, 2.5; 95% CI, 1.3-4.8), and factor-of-risk (phi for 90 degrees forward flexion/overall vertebral compressive strength: OR, 3.2; 95% CI, 1.4-7.5). These variables were correlated with spine aBMD (partial r, -0.32 to 0.75), but each was a stronger predictor of fracture in the logistic regression analyses.

CONCLUSIONS

The association of aBMD with vertebral fracture risk is explained by its correlation with more specific bone density, structure, and strength parameters. These may allow deeper insights into fracture pathogenesis.

Prostate brachytherapy seed reconstruction using an adaptive grouping technique

Fluoroscopy-based three-dimensional seed localization as a component of intraoperative dosimetry for prostate brachytherapy is an active area of research. A novel adaptive-grouping-based reconstruction approach is developed. This approach can recover overlapped seeds that are not detected from the fluoroscopic images. Two versions of the adaptive-grouping-based reconstruction approach are implemented and compared to an epipolar geometry-based seed reconstruction technique. Simulations based on nine patient datasets are used to validate the algorithms. A total of 2259 reconstructions is performed in which different types of error such as random noise in seed image locations and ambiguities in projection geometry are incorporated. Among those reconstructions, nine of the cases with overlapping seeds and the different types of error are performed. It is demonstrated that the adaptive-grouping-based reconstruction method is more accurate than the epipolar geometry method and allows faster reconstruction. At a random noise level of 0.6 mm, the mean distance error in reconstructed seed locations is approximately 1.0 mm for one of the relevant cases examined in detail. The best adaptive-grouping-based approach successfully recovered overlapped seeds in the majority of simulated cases (89%), with the remainder of cases generating one false positive seed. Phantom validation is also performed, and overlapped seeds are successfully recovered with all 92 seeds correctly localized and reconstructed. The mean distance error between segmented seed images and projected seeds is 0.5 mm in the phantom study.

High resolution multidetector CT-aided tissue analysis and quantification of lung fibrosis

RATIONALE AND OBJECTIVES

Volumetric high-resolution scans can be acquired of the lungs with multidetector CT (MDCT). Such scans have potential to facilitate useful visualization, characterization, and quantification of the extent of diffuse lung diseases, such as usual interstitial pneumonitis or idiopathic pulmonary fibrosis (UIP/IPF). There is a need to objectify, standardize, and improve the accuracy and repeatability of pulmonary disease characterization and quantification from such scans. This article presents a novel texture analysis approach toward classification and quantification of various pathologies present in lungs with UIP/IPF. The approach integrates a texture matching method with histogram feature analysis.

MATERIALS AND METHODS

Patients with moderate UIP/IPF were scanned on a Lightspeed 8-detector GE CT scanner (140 kVp, 250 mAs). Images were reconstructed with 1.25-mm slice thickness in a high-frequency sparing algorithm (BONE) with 50% overlap and a 512 x 512 axial matrix, (0.625 mm(3) voxels). Eighteen scans were used in this study. Each dataset is preprocessed and includes segmentation of the lungs and the bronchovascular trees. Two types of analysis were performed, first an analysis of independent volume of interests (VOIs) and second an analysis of whole-lung datasets. 1) Fourteen of the 18 scans were used to create a database of independent 15 x 15 x 15 cubic voxel VOIs. The VOIs were selected by experts as having greater than 70% of the defined class. The database was composed of: honeycombing (number of VOIs 337), reticular (130), ground glass (148), normal (240), and emphysema (54). This database was used to develop our algorithm. Three progressively challenging classification experiments were designed to test our algorithm. All three experiments were performed using a 10-fold cross-validation method for error estimation. Experiment 1 consisted of a two-class discrimination: normal and abnormal. Experiment 2 consisted of a four-class discrimination: normal, reticular, honeycombing, and emphysema. Experiment 3 consisted of a five-class discrimination: normal, ground glass, reticular, honeycombing, and emphysema. 2) The remaining four scans were used to further test the algorithm on new data in the context of a whole lung analysis. Each of the four datasets was manually segmented by three experts. These datasets included normal, reticular and honeycombing regions and did not include ground glass or emphysema. The accuracy of the classification algorithm was then compared with results from experts.

RESULTS

Independent VOIs: 1) two-class discrimination problem (sensitivity, specificity): normal versus abnormal (92.96%, 93.78%). 2) Four-class discrimination problem: normal (92%, 95%), reticular (86%, 87%), honeycombing (74%, 98%), and emphysema (93%, 98%). 3) Five-class discrimination problem: normal (92%, 95%), ground glass (75%, 89%), reticular (22%, 92%), honeycombing (74%, 91%), and emphysema (94%, 98%). Whole-lung datasets: 1) William's index shows that algorithm classification of lungs agrees with the experts as well as the experts agree with themselves. 2) Student t-test between overlap measures of algorithm and expert (AE) and expert and expert (EE): normal (t = -1.20, P = .230), Reticular (t = -1.44, P = .155), Honeycombing (t = -3.15, P = .003). 3) Lung volumes intraclass correlation: dataset 1 (ICC = 0.9984, F = 0.0007); dataset 2 (ICC = 0.9559, F = 0); dataset 3 (ICC = 0.8623, F= 0.0015); dataset 4 (ICC = 0.7807, F = 0.0136).

CONCLUSIONS

We have demonstrated that our novel method is computationally efficient and produces results comparable to expert radiologic judgment. It is effective in the classification of normal versus abnormal tissue and performs as well as the experts in distinguishing among typical pathologies present in lungs with UIP/IPF. The continuing development of quantitative metrics will improve quantification of disease and provide objective measures of disease progression.

Dosimetry accuracy as a function of seed localization uncertainty in permanent prostate brachytherapy: increased seed number correlates with less variability in prostate dosimetry

The variation of permanent prostate brachytherapy dosimetry as a function of seed localization uncertainty was investigated for I-125 implants with seed activities commonly employed in contemporary practice. Post-implant imaging and radiation dosimetry data from nine patients who underwent permanent prostate brachytherapy served as the source of clinical data for this simulation study. Gaussian noise with standard deviations ranging from 0.5 to 10 mm was applied to the seed coordinates for each patient dataset and 1000 simulations were performed at each noise level. Dose parameters, including D90, were computed for each case and compared with the actual dosimetry data. A total of 81 000 complete sets of post-brachytherapy dose volume statistics were computed. The results demonstrated that less than 5% deviation of prostate D90 can be expected when the seed localization uncertainty is 2 mm, whereas a seed localization uncertainty of 10 mm yielded an average decrease in D90 of 33 Gy. The mean normalized decrement in the prostate V100 was 10% at 5 mm uncertainty. Implants with greater seed number and larger prostate volume correlated with less sensitivity of D90 and V100 to seed localization uncertainty. Estimated target volume dose parameters tended to decrease with increasing seed localization uncertainty. The bladder V100 varied more significantly both in mean and standard deviation as compared to the urethra V100. A larger number of implanted seeds also correlated to less sensitivity of the bladder V100 to seed localization uncertainty. In contrast, the deviation of urethra V100 did not correlate with the number of implanted seeds or prostate volume.

Computer aided classification of cell nuclei in the gastrointestinal tract by volume and principal axis

Normal function of the gastrointestinal tract involves the coordinated activity of several cell types Human disorders of motor function of the gastrointestinal tract are often associated with changes in the number of these cells. For example, in diabetic patients, abnormalities in gastrointestinal transit are associated with changes in nerves and interstitial cells of Cajal (ICC), two key cells that generate and regulate motility. ICC are cells of mesenchymal origin that function as pacemakers and amplify neuronal signals in the gastrointestinal tract. Quantifying the changes in number of specific cell types in tissues from patients with motility disorders is challenging and requires immunolabeling for specific antigens. The shape of nuclei differs between the cell types in the wall of the gastrointestinal tract. Therefore the objective of this study was to determine whether cell nuclei can be classified by analyzing the 3D morphology of the nuclei. Furthermore, the orientation of the long axis of nuclei changes within and between the muscle layers. These features can be used to classify and differentially label the nuclei in confocal volume images of the tissue by computing the principal axis of the coordinates of the set of voxels forming each nucleus and thereby to identify cells by their nuclear morphology. Using this approach, we were able to separate and quantify nuclei in the smooth muscle layers of the tissue. Therefore we conclude that computer-aided classification of cell nuclei can be used to identify changes in the cell types expressed in gastrointestinal smooth muscle.