HOME / PUBLICATIONS /

Publications by Year: 2014

2014
Ross JC, Diaz AA, Okajima Y, Wassermann D, Washko GR, Dy J, San Jose Estépar R. Airway labeling using a Hidden Markov Tree Model. In: Biomedical Imaging (ISBI), 2014 IEEE 11th International Symposium onBiomedical Imaging (ISBI), 2014 IEEE 11th International Symposium on. 2014 p. 554-558.Abstract
We present a novel airway labeling algorithm based on a Hidden Markov Tree Model (HMTM). We obtain a collection of discrete points along the segmented airway tree using particles sampling [1] and establish topology using Kruskal's minimum spanning tree algorithm. Following this, our HMTM algorithm probabilistically assigns labels to each point. While alternative methods label airway branches out to the segmental level, we describe a general method and demonstrate its performance out to the subsubsegmental level (two generations further than previously published approaches). We present results on a collection of 25 computed tomography (CT) datasets taken from a Chronic Obstructive Pulmonary Disease (COPD) study.
Wasserman D, Ross JC, Washko GR, Wells WM, San Jose Estépar R. Deformable Registration of Feature-Endowed Point Sets Based on Tensor Fields. In: Computer Vision and Pattern Recognition (CVPR), 2014 IEEE Conference onComputer Vision and Pattern Recognition (CVPR), 2014 IEEE Conference on. 2014 p. 2729-2735.Abstract
The main contribution of this work is a framework to register anatomical structures characterized as a point set where each point has an associated symmetric matrix. These matrices can represent problem-dependent characteristics of the registered structure. For example, in airways, matrices can represent the orientation and thickness of the structure. Our framework relies on a dense tensor field representation which we implement sparsely as a kernel mixture of tensor fields. We equip the space of tensor fields with a norm that serves as a similarity measure. To calculate the optimal transformation between two structures we minimize this measure using an analytical gradient for the similarity measure and the deformation field, which we restrict to be a diffeomorphism. We illustrate the value of our tensor field model by comparing our results with scalar and vector field based models. Finally, we evaluate our registration algorithm on synthetic data sets and validate our approach on manually annotated airway trees.
San Jose Estépar R, Vosburgh KG. Multimodality Guidance in Endoscopic and Laparoscopic Abdominal Procedures. In: Jolesz FA Intraoperative Imaging and Image-Guided Therapy. New York, NY: Springer New York; 2014 p. 767-778-778.Abstract
Interventions in the body increase markedly in difficulty when the instrument body and distal tip cannot be seen from an outside vantage point. Thus most endoscopic applications, including cystoscopy, neuroendoscopy, bronchoscopy, upper GI endoscopy, and colonoscopy, require long training periods to attain proficiency, and even experts may find themselves disoriented during a procedure. As well, retroperitoneal laparoscopy, where the distal tip of the laparoscope is not seen directly from the pneumoperitoneal space, faces similar challenges. In the following, we will discuss “endoscopy” with the understanding that the techniques may also be applied to laparoscopy.
Kurugol S, Washko GR, San Jose Estépar R. Ranking and classification of monotonic emphysema patterns with a multi-class hierarchical approach. In: Biomedical Imaging (ISBI), 2014 IEEE 11th International Symposium onBiomedical Imaging (ISBI), 2014 IEEE 11th International Symposium on. 2014 p. 1031-1034.Abstract
Emphysema has distinct and well-defined visually apparent CT patterns called centrilobular and panlobular emphysema. Existing studies concentrated on the classification of these patterns but they have not looked at the complete evolution of this disease as the destruction of lung parenchyma progresses from normal lung tissue to mild, moderate, and severe disease with complete effacement of the lung architecture. In this paper, we discretize this continuous process into five classes of increasing disease severity and construct a training set of 1161 CT patches. We exploit three solutions to this monotonic multi-class classification problem: a global rankSVM for ranking, hierarchical SVM for classification and a combination of these two, which we call a hierarchical rankSVM. Results showed that both hierarchical approaches were computationally efficient. The classification accuracies were slightly better for hierarchical SVM. However, in addition to classification, ranking approaches also provided a ranking of patterns, which can be utilized as a continuous disease progression score. In terms of the classification accuracy and ratio of pair-wise constraints satisfied, hierarchical rankSVM outperformed the global rankSVM.
Diaz AA, Hardin ME, Come CE, San Jose Estépar R, Ross JC, Kurugol S, Okajima Y, Han MLK, Kim V, Ramsdell J, Silverman EK, Crapo JD, Lynch DA, Make B, Barr GR, Hersh CP, Washko GR. Childhood-Onset Asthma in Smokers. Association between CT Measures of Airway Size, Lung Function, and Chronic Airflow Obstruction. Annals ATSAnnals ATS 2014;11:1371-1378.Abstract
Rationale and Objectives: Asthma is associated with chronic airflow obstruction (CAO). Our goal was to assess the association of computed tomographic (CT) measures of airway wall volume (WV) and lumen volume (LV) with the forced expiratory volume in one second (FEV1) and CAO in smokers with childhood-onset asthma (CA). Methods: We analyzed clinical, lung function, and volumetric CT airway volumes data from 7,266 smokers including 590 with CA. Small WV and small LV of segmental airways were defined as measures 1 SD below the mean. We assessed the association between small WV, small LV, FEV1 and CAO (post-bronchodilator FEV1/forced vital capacity ratio <0.7) using linear and logistic models. Measurements and Main Results: CA subjects had smaller WV and LV than those without CA (mean ± SD, 371.3 ± 92.5mm3 vs. 388.7 ± 94.0mm3, P<0.0001; 230.8 ± 76.9 mm3 vs. 257.4 ± 80.9mm3, P<0.0001). Among CA subjects, those with the smallest WV and LV had the lowest FEV1 and greatest odds of CAO. A similar tendency was seen in those without CA. When comparing these 2 groups, small WV and small LV were more strongly associated with FEV1 (for WV, β [95%CI] -331ml [-488 - -174] vs. -244ml [-294 - -194]; for LV, -534ml [-669 - -398] vs. -435ml [-485 - -386] and CAO (Odds ratio 2.10 (1.27-3.46) vs. 2.02 [1.72-2.36]; 4.32 [2.64-7.08] vs. 3.55 [3.00-4.19], respectively) among CA subjects in multivariate models. Conclusion: In smokers with CA smaller airways are associated with reduced lung function and chronic airflow obstruction.
Hobbs BD, Foreman MG, Bowler R, Jacobson F, Make BJ, Castaldi PJ, San Jose Estépar R, Silverman EK, Hersh CP. Pneumothorax Risk Factors in Smokers with and without Chronic Obstructive Pulmonary Disease. Annals ATSAnnals ATS 2014;11:1387-1394.Abstract
Rationale: The demographic, physiologic, and computed tomography (CT) features associated with pneumothorax in smokers with and without chronic obstructive pulmonary disease (COPD) are not clearly defined. Objectives: We evaluated the hypothesis that pneumothorax in smokers is associated with male gender, tall and thin stature, airflow obstruction, and increased total and sub-pleural emphysema. Methods: The study included smokers with and without COPD from the COPDGene study, with quantitative chest CT analysis. Pleural-based emphysema was assessed using local histogram measures of emphysema. Pneumothorax history was defined by subject self-report. Measurements and Main Results: Pneumothorax was reported in 286 (3.2%) of 9,062 participants. In all participants, risk of prior pneumothorax was significantly higher in men (OR 1.55, 95% CI 1.08 - 2.22) and non-Hispanic whites (OR 1.90, 95% CI 1.34 - 2.69). Risk of prior pneumothorax was associated with increased percent CT emphysema in all participants and participants with COPD (OR 1.04 for each 1% increase in emphysema, 95% CI 1.03 - 1.06). Increased pleural-based emphysema was independently associated with risk of past pneumothorax in all participants (OR 1.05 for each 1% increase, 95% CI 1.01 - 1.10). In smokers with normal spirometry, risk of past pneumothorax was associated with non-Hispanic white race and lifetime smoking intensity (OR 1.20 for every 10 pack-years, 95% CI 1.09 - 1.33). Conclusions: Among smokers, pneumothorax is associated with male gender, non-Hispanic white race, and increased percentage of total and sub-pleural CT emphysema. Pneumothorax was not independently associated with height or lung function, even in participants with COPD. ClinicalTrials.gov Identifier: NCT00608764.
Diaz AA, Zhou L, Young TP, McDonald M-L, Harmouche R, Ross JC, San Jose Estépar R, Wouters EFM, Coxson HO, MacNee W, Rennard S, Maltais F, Kinney GL, Hokanson JE, Washko GR. Chest CT Measures of Muscle and Adipose Tissue in COPD: Gender-based Differences in Content and in Relationships with Blood Biomarkers. Academic RadiologyAcademic Radiology 2014;21:1255-1261.Abstract

Rationale and Objectives Computed tomography (CT) of the chest can be used to assess pectoralis muscle area (PMA) and subcutaneous adipose tissue (SAT) area. Adipose tissue content is associated with inflammatory mediators in chronic obstructive pulmonary disease (COPD) subjects. Based on gender differences in body composition, we aimed to assess the hypothesis that in subjects with COPD, the relationships between PMA, SAT, and blood biomarkers of inflammation differ by gender. Materials and Methods We compared chest CT measures of PMA and SAT on a single slice at aortic arch and supraesternal notch levels from 73 subjects (28 women) with COPD between genders. The relationships of PMA and SAT area to biomarkers were assessed using within-gender regression models. Results Women had a lesser PMA and a greater SAT area than men (difference range for PMA, 13.3–22.8 cm2; for SAT, 11.8–12.4 cm2; P &lt; .05 for all comparisons) at both anatomic levels. These differences in PMA and SAT remained significant after adjustment for age and body mass index. Within-gender regression models adjusted for age showed that SAT was directly associated with C-reactive protein (for aortic arch level, P = .04) and fibrinogen (for both anatomic locations, P = .003) only in women, whereas PMA was not associated with any biomarkers in either gender. Conclusions It appears that in subjects with COPD, there are gender-based differences in the relationships between subcutaneous adipose tissue and inflammatory biomarkers.

acad_radiol_2014_diaz.pdf
Rudyanto RD, Kerkstra S, van Rikxoort EM, Fetita C, Brillet P-Y, Lefevre C, Xue W, Zhu X, Liang J, Oksüz I, Unay D, Kadipaşaogˇlu K, San Jose Estépar R, Ross JC, Washko GR, Prieto J-C, Hoyos MH, Orkisz M, Meine H, Hüllebrand M, Stöcker C, Mir FL, Naranjo V, Villanueva E, Staring M, Xiao C, Stoel BC, Fabijanska A, Smistad E, Elster AC, Lindseth F, Foruzan AH, Kiros R, Popuri K, Cobzas D, Jimenez-Carretero D, Santos A, Ledesma-Carbayo MJ, Helmberger M, Urschler M, Pienn M, Bosboom DGH, Campo A, Prokop M, de Jong PA, Ortiz-de-Solorzano C, Muñoz-Barrutia A, van Ginneken B. Comparing algorithms for automated vessel segmentation in computed tomography scans of the lung: the VESSEL12 study. Medical Image AnalysisMedical Image Analysis 2014;18:1217-1232.Abstract

Abstract The VESSEL12 (VESsel SEgmentation in the Lung) challenge objectively compares the performance of different algorithms to identify vessels in thoracic computed tomography (CT) scans. Vessel segmentation is fundamental in computer aided processing of data generated by 3D imaging modalities. As manual vessel segmentation is prohibitively time consuming, any real world application requires some form of automation. Several approaches exist for automated vessel segmentation, but judging their relative merits is difficult due to a lack of standardized evaluation. We present an annotated reference dataset containing 20 CT scans and propose nine categories to perform a comprehensive evaluation of vessel segmentation algorithms from both academia and industry. Twenty algorithms participated in the VESSEL12 challenge, held at International Symposium on Biomedical Imaging (ISBI) 2012. All results have been published at the VESSEL12 website http://vessel12.grand-challenge.org. The challenge remains ongoing and open to new participants. Our three contributions are: (1) an annotated reference dataset available online for evaluation of new algorithms; (2) a quantitative scoring system for objective comparison of algorithms; and (3) performance analysis of the strengths and weaknesses of the various vessel segmentation methods in the presence of various lung diseases.

medical_image_analysis_2014_rudyanto.pdf
Castaldi PJ, Cho MH, San Jose Estépar R, McDonald M-LN, Laird N, Beaty TH, Washko G, Crapo JD, Silverman EK. Genome-Wide Association Identifies Regulatory Loci Associated with Distinct Local Histogram Emphysema Patterns. American journal of respiratory and critical care medicineAmerican journal of respiratory and critical care medicine 2014;190:399-409.Abstract

RATIONALE:Emphysema is a heritable trait that occurs in smokers with and without chronic obstructive pulmonary disease. Emphysema occurs in distinct pathologic patterns, but the genetic determinants of these patterns are unknown.OBJECTIVES:To identify genetic loci associated with distinct patterns of emphysema in smokers and investigate the regulatory function of these loci.METHODS:Quantitative measures of distinct emphysema patterns were generated from computed tomography scans from smokers in the COPDGene Study using the local histogram emphysema quantification method. Genome-wide association studies (GWAS) were performed in 9,614 subjects for five emphysema patterns, and the results were referenced against enhancer and DNase I hypersensitive regions from ENCODE and Roadmap Epigenomics cell lines.MEASUREMENTS AND MAIN RESULTS:Genome-wide significant associations were identified for seven loci. Two are novel associations (top single-nucleotide polymorphism rs379123 in MYO1D and rs9590614 in VMA8) located within genes that function in cell-cell signaling and cell migration, and five are in loci previously associated with chronic obstructive pulmonary disease susceptibility (HHIP, IREB2/CHRNA3, CYP2A6/ADCK, TGFB2, and MMP12). Five of these seven loci lay within enhancer or DNase I hypersensitivity regions in lung fibroblasts or small airway epithelial cells, respectively. Enhancer enrichment analysis for top GWAS associations (single-nucleotide polymorphisms associated at P < 5 × 10(-6)) identified multiple cell lines with significant enhancer enrichment among top GWAS loci, including lung fibroblasts.CONCLUSIONS:This study demonstrates for the first time genetic associations with distinct patterns of pulmonary emphysema quantified by computed tomography scan. Enhancer regions are significantly enriched among these GWAS results, with pulmonary fibroblasts among the cell types showing the strongest enrichment.

Paper
Kinsey MC, San Jose Estépar R, Zhao Y, Yu X, Diao N, Heist RS, Wain JC, Mark EJ, Washko G, Christiani DC. Invasive adenocarcinoma of the lung is associated with the upper lung regions. Lung cancerLung cancer 2014;84:145-150.Abstract

OBJECTIVES:We postulated that ventilation-perfusion (V/Q) relationships within the lung might influence where lung cancer occurs. To address this hypothesis we evaluated the location of lung adenocarcinoma, by both tumor lobe and superior-inferior regional distribution, and associated variables such as emphysema.MATERIALS AND METHODS:One hundred fifty-nine cases of invasive adenocarcinoma and adenocarcinoma with lepidic features were visually evaluated to identify lobar or regional tumor location. Regions were determined by automated division of the lungs into three equal volumes: (upper region, middle region, or lower region). Automated densitometry was used to measure radiographic emphysema.RESULTS:The majority of invasive adenocarcinomas occurred in the upper lobes (69%), with 94% of upper lobe adenocarcinomas occurring in the upper region of the lung. The distribution of adenocarcinoma, when classified as upper or lower lobe, was not different between invasive adenocarcinoma and adenocarcinoma with lepidic features (formerly bronchioloalveolar cell carcinoma, P=0.08). Regional distribution of tumor was significantly different between invasive adenocarcinoma and adenocarcinoma with lepidic features (P=0.001). Logistic regression analysis with the outcome of invasive adenocarcinoma histology was used to adjust for confounders. Tumor region continued to be a significant predictor (OR 8.5, P=0.008, compared to lower region), whereas lobar location of tumor was not (P=0.09). In stratified analysis, smoking was not associated with region of invasive adenocarcinoma occurrence (P=0.089). There was no difference in total emphysema scores between invasive adenocarcinoma cases occurring in each of the three regions (P=0.155). There was also no difference in the distribution of region of adenocarcinoma occurrence between quartiles of emphysema (P=0.217).CONCLUSION:Invasive adenocarcinoma of the lung is highly associated with the upper lung regions. This association is not related to smoking, history of COPD, or total emphysema. The regional distribution of invasive adenocarcinoma may be due to V/Q relationships or other local factors.

Paper
Rahaghi FN, Lazea D, Dihya S, San Jose Estépar R, Bueno R, Sugarbaker D, Frendl G, Washko GR. Preoperative Pulmonary Vascular Morphology and Its Relationship to Postpneumonectomy Hemodynamics. Academic RadiologyAcademic Radiology 2014;21:704-710.Abstract

Pulmonary edema and pulmonary hypertension are postsurgical complications of pneumonectomy that may represent the remaining pulmonary vasculature's inability to accommodate the entirety of the cardiac output. Quantification of the aggregate pulmonary vascular cross-sectional area (CSA) has been used to study the development of pulmonary vascular disease in smokers. In this study, we applied this technique to demonstrate the potential utility of pulmonary vascular quantification in surgical risk assessment. Our hypothesis was that those subjects with the lowest aggregate vascular CSA in the nonoperative lung would be most likely to have elevated pulmonary vascular pressures in the postoperative period.A total of 61 subjects with postoperative hemodynamics and adequate imaging were identified from 159 patients undergoing pneumonectomies for mesothelioma. The total CSA of blood vessels perpendicular to the plane of computed tomographic (CT) scan slices was computed for blood vessels <5 mm(2) (CSA 5 mm). This measurement expressed as a percentage of lung parenchyma area (CSA 5%) was compared to postoperative hemodynamic measurements obtained by right heart catheterization.In patients where a contrasted CT scan was used (n = 26), CSA 5% was correlated with postoperative day 0 minimum cardiac index (R = 0.37, P = .03) but not with the maximum pulmonary arterial pressures. In patients with noncontrast CT scans (n = 35), CSA 5% was inversely correlated with postoperative day 0 maximum pulmonary arterial pressures (R = 0.43, P = .03) but not with the minimum cardiac index. The preoperative perfusion fraction of the nonsurgical lung did not correlate with postoperative hemodynamics.CSA of pulmonary vasculature with an area ≤5 mm(2) has potential in estimating the ability of pulmonary vascular bed to accommodate postsurgical changes in pneumonectomy.Copyright © 2014 AUR. Published by Elsevier Inc. All rights reserved.

Paper
McDonald M-LN, Diaz AA, Ross JC, San Jose Estépar R, Zhou L, Regan EA, Eckbo E, Muralidhar N, Come CE, Cho MH, Hersh CP, Lange C, Wouters E, Casaburi RH, Coxson HO, MacNee W, Rennard SI, Lomas DA, Agusti A, Celli BR, Black-Shinn JL, Kinney GL, Lutz SM, Hokanson JE, Silverman EK, Washko GR. Quantitative Computed Tomography Measures of Pectoralis Muscle Area and Disease Severity in Chronic Obstructive Pulmonary Disease. A Cross-Sectional Study. Annals ATSAnnals ATS 2014;11:326-334.Abstract

Rationale: Muscle wasting in chronic obstructive pulmonary disease (COPD) is associated with a poor prognosis and is not readily assessed by measures of body mass index (BMI). BMI does not discriminate between relative proportions of adipose tissue and lean muscle and may be insensitive to early pathologic changes in body composition. Computed tomography (CT)-based assessments of the pectoralis muscles may provide insight into the clinical significance of skeletal muscles in smokers. Objectives: We hypothesized that objective assessment of the pectoralis muscle area on chest CT scans provides information that is clinically relevant and independent of BMI. Methods: Data from the ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints) Study (n = 73) were used to assess the relationship between pectoralis muscle area and fat-free mass. We then used data in a subset (n = 966) of a larger cohort, the COPDGene (COPD Genetic Epidemiology) Study, to explore the relationship between pectoralis muscle area and COPD-related traits. Measurements and Main Results: We first investigated the correlation between pectoralis muscle area and fat-free mass, using data from a subset of participants in the ECLIPSE Study. We then further investigated pectoralis muscle area in COPDGene Study participants and found that higher pectoralis muscle area values were associated with greater height, male sex, and younger age. On subsequent clinical correlation, compared with BMI, pectoralis muscle area was more significantly associated with COPD-related traits, including spirometric measures, dyspnea, and 6-minute-walk distance (6MWD). For example, on average, each 10-cm(2) increase in pectoralis muscle area was associated with a 0.8-unit decrease in the BODE (Body mass index, Obstruction, Dyspnea, Exercise) index (95% confidence interval, -1.0 to -0.6; P < 0.001). Furthermore, statistically significant associations between pectoralis muscle area and COPD-related traits remained even after adjustment for BMI. Conclusions: CT-derived pectoralis muscle area provides relevant indices of COPD morbidity that may be more predictive of important COPD-related traits than BMI. However, the relationship with clinically relevant outcomes such as hospitalization and death requires additional investigation. Pectoralis muscle area is a convenient measure that can be collected in the clinical setting in addition to BMI.

Paper
San Jose Estépar R, Reeves AP, Yankelevitz D, Henschke CI, Mendelson DS, de la Hoz RE. A Quantitative Method for Mosaic Gas Trapping based on Residual Mass. International journal of computer assisted radiology and surgeryInternational journal of computer assisted radiology and surgery 2014;9 Suppl 1:9-10.Abstract

Purpose

Air trapping is a pathophysiologic condition indicating the retention of excess gas in all or part of the lung at any stage of expiration. Expiratory CT has been used to reveal air trapping in patients with airway diseases such as pulmonary emphysema, chronic bronchitis, asthma, and small-airway disease. Air trapping can be classified as extensive and mosaic. Extensive gas trapping is characterized by well-defined low attenuation areas in expiratory scans. Mosaic air trapping is characterized by the appearance of inhomogeneous air emptying on expiratory compared to normal inspiratory chest CT scans. A number of quantitative chest CT methods have been pro- posed to assess extensive gas trapping [1, 2], however those methods fail to detect the diffuse nature of mosaic gas trapping. We have developed an approach to tackle this problem and we have validated the technique in a population of World Trade Center workers with abnormal lung emptying reported by a radiologist.

int_j_comput_assist_radiol_surg_2014_san_jose_estepar.pdf
Perez F, Vera S, Fernandez-Esparrach G, Córdova H, San Jose Estépar R, Jover JH, González Ballester MÁ. Surgical Workflow Analysis, Design and Development of an Image-Based Navigation System for Endoscopic Interventions. In: Erdt M, Linguraru MG, Oyarzun Laura C, Shekhar R, Wesarg S, González Ballester MÁ, Drechsler K Lecture Notes in Computer Science. Cham: Springer International Publishing; 2014 p. 91-98-98.Abstract

Abstract Endoscopic interventions in the abdominal and thoracic cavity are often hampered by the difficulty to orient in the endoscopic view. This is due to the small field of view and the inhomogeneous illumination, but also because abdominal organs are highly deformable ...

Paper
Gazourian L, Rogers AJ, Ibanga R, Weinhouse GL, Pinto-Plata V, Ritz J, Soiffer RJ, Antin JH, Washko GR, Baron RM, Ho VT. Factors associated with bronchiolitis obliterans syndrome and chronic graft-versus-host disease after allogeneic hematopoietic cell transplantation. Am J Hematol 2014;89(4):404-9.Abstract
Bronchiolitis obliterans syndrome (BOS) is a form of chronic graft vs. host disease (cGVHD) and a highly morbid pulmonary complication after allogeneic hematopoietic stem cell transplantation (HSCT). We assessed the prevalence and risk factors for BOS and cGVHD in a cohort of HSCT recipients, including those who received reduced intensity conditioning (RIC) HSCT. Between January 1, 2000 and June 30, 2010, all patients who underwent allogeneic HSCT at our institution (n = 1854) were retrospectively screened for the development of BOS by PFT criteria. We matched the BOS cases with two groups of control patients: (1) patients who had concurrent cGVHD without BOS and (2) those who developed neither cGVHD nor BOS. Comparisons between BOS patients and controls were conducted using t-test or Fisher's exact tests. Multivariate regression analysis was performed to examine factors associated with BOS diagnosis. All statistical analyses were performed using SAS 9.2. We identified 89 patients (4.8%) meeting diagnostic criteria for BOS at a median time of 491 days (range: 48-2067) after HSCT. Eighty-six (97%) of our BOS cohort had extra-pulmonary cGVHD. In multivariate analysis compared to patients without cGVHD, patients who received busulfan-based conditioning, had unrelated donors, and had female donors were significantly more likely to develop BOS, while ATG administration was associated with a lower risk of BOS. Our novel results suggest that busulfan conditioning, even in RIC transplantation, could be an important risk factor for BOS and cGVHD.
Ross JC, Díaz AA, Okajima Y, Wassermann D, Washko GR, Dy J, San José Estépar R. AIRWAY LABELING USING A HIDDEN MARKOV TREE MODEL. Proc IEEE Int Symp Biomed Imaging 2014;2014:554-558.Abstract
We present a novel airway labeling algorithm based on a Hidden Markov Tree Model (HMTM). We obtain a collection of discrete points along the segmented airway tree using particles sampling [1] and establish topology using Kruskal's minimum spanning tree algorithm. Following this, our HMTM algorithm probabilistically assigns labels to each point. While alternative methods label airway branches out to the segmental level, we describe a general method and demonstrate its performance out to the subsubsegmental level (two generations further than previously published approaches). We present results on a collection of 25 computed tomography (CT) datasets taken from a Chronic Obstructive Pulmonary Disease (COPD) study.
biomedical_imaging_isbi_2014_ieee_11th_international_symposium_on_2014_ross.pdf
Diaz AA, Zhou L, Young TP, McDonald M-L, Harmouche R, Ross JC, Estepar RSJ, Wouters EFM, Coxson HO, MacNee W, Rennard S, Maltais F, Kinney GL, Hokanson JE, Washko GR. Chest CT measures of muscle and adipose tissue in COPD: gender-based differences in content and in relationships with blood biomarkers. Acad Radiol 2014;21(10):1255-61.Abstract
RATIONALE AND OBJECTIVES: Computed tomography (CT) of the chest can be used to assess pectoralis muscle area (PMA) and subcutaneous adipose tissue (SAT) area. Adipose tissue content is associated with inflammatory mediators in chronic obstructive pulmonary disease (COPD) subjects. Based on gender differences in body composition, we aimed to assess the hypothesis that in subjects with COPD, the relationships between PMA, SAT, and blood biomarkers of inflammation differ by gender. MATERIALS AND METHODS: We compared chest CT measures of PMA and SAT on a single slice at aortic arch and supraesternal notch levels from 73 subjects (28 women) with COPD between genders. The relationships of PMA and SAT area to biomarkers were assessed using within-gender regression models. RESULTS: Women had a lesser PMA and a greater SAT area than men (difference range for PMA, 13.3-22.8 cm²; for SAT, 11.8-12.4 cm²; P < .05 for all comparisons) at both anatomic levels. These differences in PMA and SAT remained significant after adjustment for age and body mass index. Within-gender regression models adjusted for age showed that SAT was directly associated with C-reactive protein (for aortic arch level, P = .04) and fibrinogen (for both anatomic locations, P = .003) only in women, whereas PMA was not associated with any biomarkers in either gender. CONCLUSIONS: It appears that in subjects with COPD, there are gender-based differences in the relationships between subcutaneous adipose tissue and inflammatory biomarkers.
Wassermann D, Ross J, Washko G, Wells WM, San Jose-Estepar R. Deformable Registration of Feature-Endowed Point Sets Based on Tensor Fields. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit 2014;2014:2729-2735.Abstract
The main contribution of this work is a framework to register anatomical structures characterized as a point set where each point has an associated symmetric matrix. These matrices can represent problem-dependent characteristics of the registered structure. For example, in airways, matrices can represent the orientation and thickness of the structure. Our framework relies on a dense tensor field representation which we implement sparsely as a kernel mixture of tensor fields. We equip the space of tensor fields with a norm that serves as a similarity measure. To calculate the optimal transformation between two structures we minimize this measure using an analytical gradient for the similarity measure and the deformation field, which we restrict to be a diffeomorphism. We illustrate the value of our tensor field model by comparing our results with scalar and vector field based models. Finally, we evaluate our registration algorithm on synthetic data sets and validate our approach on manually annotated airway trees.
computer_vision_and_pattern_recognition_cvpr_2014_ieee_conference_on_2014_wasserman.pdf
Ross JC, Castaldi PJ, Cho MH, Dy JG. Dual Beta Process Priors for Latent Cluster Discovery in Chronic Obstructive Pulmonary Disease [Internet]. In: Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, NY, USA: ACM; 2014 p. 155–162. Publisher's VersionAbstract

Chronic obstructive pulmonary disease (COPD) is a lung disease characterized by airflow limitation usually associated with an inflammatory response to noxious particles, such as cigarette smoke. COPD is currently the third leading cause of death in the United States and is the only leading cause of death that is increasing in prevalence. It also represents an enormous financial burden to society, costing tens of billions of dollars annually in the U.S. It is widely accepted by the medical community that COPD is a heterogeneous disease, with substantial evidence indicating that genetic variation contributes to varying levels of disease susceptibility. This heterogeneity makes it difficult to predict health decline and develop targeted treatments for better patient care. Although researchers have made several attempts to discover disease subtypes, results have been inconclusive, in part because standard clustering methods have not properly dealt with disease manifestations that may worsen with increased exposure. In this paper we introduce a transformative way of looking at the COPD subtyping task. Specifically, we model the relationship between risk factors (such as age and smoke exposure) and manifestations of disease severity using Gaussian Processes, which allow us to represent so-called "disease trajectories". We also posit that individuals can be associated with multiple disease types (latent clusters), which we assume are influenced by genetics. Furthermore, we predict that only subsets of the numerous disease-related quantitative features are useful for describing each latent subtype. We model these associations using two separate beta process priors, and we describe a variational inference approach to discover the most probable latent cluster assignments. Results are validated with associations to genetic markers.

Nava R, Escalante-Ramírez B, Cristóbal G, San José Estépar R. Extended Gabor approach applied to classification of emphysematous patterns in computed tomography. Med Biol Eng Comput 2014;52(4):393-403.Abstract
Chronic obstructive pulmonary disease (COPD) is a progressive and irreversible lung condition typically related to emphysema. It hinders air from passing through airpaths and causes that alveolar sacs lose their elastic quality. Findings of COPD may be manifested in a variety of computed tomography (CT) studies. Nevertheless, visual assessment of CT images is time-consuming and depends on trained observers. Hence, a reliable computer-aided diagnosis system would be useful to reduce time and inter-evaluator variability. In this paper, we propose a new emphysema classification framework based on complex Gabor filters and local binary patterns. This approach simultaneously encodes global characteristics and local information to describe emphysema morphology in CT images. Kernel Fisher analysis was used to reduce dimensionality and to find the most discriminant nonlinear boundaries among classes. Finally, classification was performed using the k-nearest neighbor classifier. The results have shown the effectiveness of our approach for quantifying lesions due to emphysema and that the combination of descriptors yields to a better classification performance.
medical_biological_engineering_computing_2014_nava.pdf

Pages