Mr. Ronald A. Karwoski Profile

Ronald A. Karwoski

at Mayo Clinic College of Medicine

SPIE Involvement:

Author

Publications (10)

SPIE Journal Paper | 13 September 2022

Individualized and generalized models for predicting observer performance on liver metastasis detection using CT

Parvathy Sudhir Pillai, David Holmes, Rickey Carter, Akitoshi Inoue, David Cook, Ron Karwoski, Jeff Fidler, Joel Fletcher, Shuai Leng, Lifeng Yu, Cynthia McCollough, Scott Hsieh

JMI, Vol. 9, Issue 05, 055501, (September 2022) https://doi.org/10.1117/12.10.1117/1.JMI.9.5.055501

KEYWORDS: Performance modeling, Liver, Computed tomography, Feature extraction, Data modeling, 3D modeling, Radiology, RGB color model, Reconstruction algorithms, Convolutional neural networks

Read Abstract +

Purpose: Radiologists exhibit wide inter-reader variability in diagnostic performance. This work aimed to compare different feature sets to predict if a radiologist could detect a specific liver metastasis in contrast-enhanced computed tomography (CT) images and to evaluate possible improvements in individualizing models to specific radiologists.Approach: Abdominal CT images from 102 patients, including 124 liver metastases in 51 patients were reconstructed at five different kernels/doses using projection domain noise insertion to yield 510 image sets. Ten abdominal radiologists marked suspected metastases in all image sets. Potentially salient features predicting metastasis detection were identified in three ways: (i) logistic regression based on human annotations (semantic), (ii) random forests based on radiologic features (radiomic), and (iii) inductive derivation using convolutional neural networks (CNN). For all three approaches, generalized models were trained using metastases that were detected by at least two radiologists. Conversely, individualized models were trained using each radiologist’s markings to predict reader-specific metastases detection.Results: In fivefold cross-validation, both individualized and generalized CNN models achieved higher area under the receiver operating characteristic curves (AUCs) than semantic and radiomic models in predicting reader-specific metastases detection ability (p < 0.001). The individualized CNN with an AUC of mean (SD) 0.85(0.04) outperformed the generalized one [AUC = 0.78 ( 0.06 ) , p = 0.004]. The individualized semantic [AUC = 0.70 ( 0.05 ) ] and radiomic models [AUC = 0.68 ( 0.06 ) ] outperformed the respective generalized versions [semantic AUC = 0.66 ( 0.03 ) , p = 0.009; radiomic AUC = 0.64 ( 0.06 ) , p = 0.03].Conclusions: Individualized models slightly outperformed generalized models for all three feature sets. Inductive CNNs were better at predicting metastases detection than semantic or radiomic features. Generalized models have implementation advantages when individualized data are unavailable.

Proceedings Article | 18 March 2013 Paper

Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns

Sushravya Raghunath, Srinivasan Rajagopalan, Ronald Karwoski, Brian Bartholmai, Richard Robb

Proceedings Volume 8670, 867037 (2013) https://doi.org/10.1117/12.2008110

KEYWORDS: Lung, Machine learning, Emphysema, Computed tomography, Glasses, Visualization, Expectation maximization algorithms, Pathology, Chest, Data modeling

Read Abstract +

Proceedings Article | 16 April 2012 Paper

Validation of geometric measurements of the left atrium and pulmonary veins for analysis of reverse structural remodeling following ablation therapy

M. Rettmann, D. Holmes, M. Gunawan, X. Ge, R. Karwoski, J. Breen, D. Packer, R. Robb

Proceedings Volume 8317, 831724 (2012) https://doi.org/10.1117/12.912801

KEYWORDS: Veins, Image segmentation, Computed tomography, 3D metrology, Volume rendering, 3D imaging standards, 3D image processing, Atrial fibrillation, Visualization, Image visualization

Read Abstract +

Proceedings Article | 23 February 2012 Paper

Effect of denoising on supervised lung parenchymal clusters

Padmapriya Jayamani, Sushravya Raghunath, Srinivasan Rajagopalan, Ronald Karwoski, Brian Bartholmai, Richard Robb

Proceedings Volume 8315, 83152X (2012) https://doi.org/10.1117/12.911650

KEYWORDS: Denoising, Lung, Glasses, Computed tomography, Emphysema, Quality measurement, Medical imaging, Signal to noise ratio, Quantitative analysis, Sensors

Read Abstract +

Proceedings Article | 23 February 2012 Paper

Automating the expert consensus paradigm for robust lung tissue classification

Srinivasan Rajagopalan, Ronald Karwoski, Sushravya Raghunath, Brian Bartholmai, Richard Robb

Proceedings Volume 8315, 831530 (2012) https://doi.org/10.1117/12.912009

KEYWORDS: Lung, Visualization, Tissues, Computed tomography, Glasses, Emphysema, Resolution enhancement technologies, Reliability, Sensors, Chest

Read Abstract +

Proceedings Article | 23 February 2012 Paper

Active relearning for robust supervised classification of pulmonary emphysema

Sushravya Raghunath, Srinivasan Rajagopalan, Ronald Karwoski, Brian Bartholmai, Richard Robb

Proceedings Volume 8315, 83152Q (2012) https://doi.org/10.1117/12.911648

KEYWORDS: Emphysema, Statistical modeling, Computed tomography, Radiology, Chest, Lung, Spatial resolution, Diffractive optical elements, Diagnostics, Image classification

Read Abstract +

Proceedings Article | 20 March 2008 Paper

Processing of CT images for analysis of diffuse lung disease in the lung tissue research consortium

Ronald Karwoski, Brian Bartholmai, Vanessa Zavaletta, David Holmes, Richard Robb

Proceedings Volume 6916, 691614 (2008) https://doi.org/10.1117/12.773159

KEYWORDS: Lung, Tissues, Image segmentation, Computed tomography, Emphysema, Databases, Chronic obstructive pulmonary disease, Glasses, Image processing, Image analysis

Read Abstract +

The goal of Lung Tissue Resource Consortium (LTRC) is to improve the management of diffuse lung diseases through a better understanding of the biology of Chronic Obstructive Pulmonary Disease (COPD) and fibrotic interstitial lung disease (ILD) including Idiopathic Pulmonary Fibrosis (IPF). Participants are subjected to a battery of tests including tissue biopsies, physiologic testing, clinical history reporting, and CT scanning of the chest. The LTRC is a repository from which investigators can request tissue specimens and test results as well as semi-quantitative radiology reports, pathology reports, and automated quantitative image analysis results from the CT scan data performed by the LTRC core laboratories. The LTRC Radiology Core Laboratory (RCL), in conjunction with the Biomedical Imaging Resource (BIR), has developed novel processing methods for comprehensive characterization of pulmonary processes on volumetric high-resolution CT scans to quantify how these diseases manifest in radiographic images. Specifically, the RCL has implemented a semi-automated method for segmenting the anatomical regions of the lungs and airways. In these anatomic regions, automated quantification of pathologic features of disease including emphysema volumes and tissue classification are performed using both threshold techniques and advanced texture measures to determine the extent and location of emphysema, ground glass opacities, "honeycombing" (HC) and "irregular linear" or "reticular" pulmonary infiltrates and normal lung. Wall thickness measurements of the trachea, and its branches to the 3^rd and limited 4^th order are also computed. The methods for processing, segmentation and quantification are described. The results are reviewed and verified by an expert radiologist following processing and stored in the public LTRC database for use by pulmonary researchers. To date, over 1200 CT scans have been processed by the RCL and the LTRC project is on target for recruitment of the 2200 patients with 1800 CT scans in the repository for the 5-year effort. Ongoing analysis of the results in the LTRC database by the LTRC participating institutions and outside investigators are underway to look at the clinical and physiological significance of the imaging features of these diseases and correlate these findings with quality of life and other important prognostic indicators of severity. In the future, the quantitative measures of disease may have greater utility by showing correlation with prognosis, disease severity and other physiological parameters. These imaging features may provide non-invasive alternative endpoints or surrogate markers to alleviate the need for tissue biopsy or provide an accurate means to monitor rate of disease progression or response to therapy.

Proceedings Article | 13 March 2006 Paper

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

Vanessa Zavaletta, Ronald Karwoski, Brian Bartholmai, Richard Robb

Proceedings Volume 6143, 61432Z (2006) https://doi.org/10.1117/12.653450

KEYWORDS: Lung, Tissues, Computed tomography, Image resolution, Visualization, Image segmentation, Image analysis, Error analysis, Pulmonary fibrosis, Pulmonary function tests

Read Abstract +

Proceedings Article | 12 April 2005 Paper

Automated segmentation of the lungs from high resolution CT images for quantitative study of chronic obstructive pulmonary diseases

Ishita Garg, Ronald Karwoski, Jon Camp, Brian Bartholmai, Richard Robb

Proceedings Volume 5744, (2005) https://doi.org/10.1117/12.595827

KEYWORDS: Image segmentation, Lung, Chronic obstructive pulmonary disease, Spherical lenses, Computed tomography, Diagnostics, Emphysema, Image resolution, Visualization, Image processing

Read Abstract +

Proceedings Article | 30 April 2004 Paper

System for the analysis of whole-bone strength from helical CT images

Jon Camp, Ronald Karwoski, Mahlon Stacy, Elizabeth Atkinson, Sundeep Khosla, L. Joseph Melton, B. Lawrence Riggs, Richard Robb

Proceedings Volume 5369, (2004) https://doi.org/10.1117/12.535745

KEYWORDS: Bone, Neck, Image segmentation, Minerals, Computed tomography, Databases, 3D image processing, Dual energy x-ray absorptiometry, Image processing, Spine

Read Abstract +

Showing 5 of 10 publications

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Keywords/Phrases

Search In:

Publication Years