Medical Image Processing

The Problem

Early detection and diagnosis of pulmonary nodules is a problem attracting great interest in the fight against lung cancer. Lung nodules are potentially cancerous lesions of approximately ellipsoidal shape.  With high-resolution computed tomography (CT) imagery it is possible not only to detect nodules, but also make useful measurements regarding the volume, shape, density, etc.  Unlike some medical imaging modalities, CT provides numerous slices of image data that can be time consuming and potentially fatiguing for radiologists to review.   If lung screening becomes widespread, handling these vast amounts of data becomes an even more challenging problem.  

To assist radiologists and improve lung screening, we are developing state-of-the-art computer aided detection (CAD) systems for automatically identifying pulmonary nodules on CT and chest radiographs.  We also have new algorithms for automated nodule segmentation and volume estimation.  We refer to our CT CAD system as FlyerScan CT.  We also have FlyerScan CXR for detecting lung nodules in chest x-ray images.  FlyerScan CT has demonstrated world-class performance in the ANODE 2009 CAD competition.

FlyerScan CT graphical user interface.  A radiologist confirmed nodule that has been detected by our system is shown.  See how this CAD system, “FlyerScan” performed in the automatic nodule detection 2009 competition here:  FlyerScan ANODE 2009 CAD competition results.

Flyerscan CXR Output.  A true nodule is shown in red.  The green detections are false positives.

FlyerScan CXR uses active shape models to segment the lungs based on the method developed by Dr. Bram Van Ginneken.

New FlyerScan CT lung nodule segmentation algorithm result.

Selected References

1. Barath Narayanan Narayanan, Redha Ali, Russell C. Hardie, "Performance analysis of machine learning and deep learning architectures for malaria detection on cell images," Proc. SPIE 11139, Applications of Machine Learning, 111390W (6 September 2019); https://doi.org/10.1117/12.2524681

   1. B. N. Narayanan, R. C. Hardie and T.M. Kebede, "Performance Analysis of a Computer Aided Detection System for Lung Nodules in CT at Different Slice Thicknesses", SPIE Journal of Medical Imaging, 5(1) 014504 (2018). doi:10.1117/1.JMI.5.1.014504

   2. B. N. Narayanan, R. C. Hardie, T.M. Kebede and M.J. Sprague, "Optimized Feature Selection based Clustering Approach for Computer Aided Detection of Lung Nodules in Different Modalities", Pattern Analysis and Application Journal (2017). doi 10.1007/s10044-017-0653-4.

   3. T. Messay, R. C. Hardie, and T. R. Tuinstra, "Segmentation of pulmonary nodules in computed tomography using a regression neural network approach and its application to the Lung Image Database Consortium and Image Database Resource Initiative dataset", Medical Image Analysis, Volume 22, Issue 1, May 2015.

   4. T. Messay, R. C. Hardie and S. K. Rogers, “A New Computationally Efficient CAD System for Nodule Detection in CT Imagery,” Medical Image Analysis, 14(3):390-406, June 2010.

      • See how this CAD system, “FlyerScan” performed in the automatic nodule detection 2009 competition here:  FlyerScan ANODE 2009 CAD competitionresults.

   5. R. C. Hardie, S. K. Rogers, T. Wilson, and A. Rogers, “Performance analysis of a new computer aided detection system for identifying lung nodules on chest radiographs,” Medical Image Analysis, Vol. 12, Issue 3, pp 240-258, June 2008.  doi:10.1016/j.media.2007.10.004

   6. R. D. Ernst, R. C. Hardie, M. N. Gurcan, A. Oto, S. K. Rogers, J. W. Hoffmeister,  “CAD Performance Analysis for Pulmonary Nodule Detection: Comparison of Thick- and Thin-Slice Helical CT Scans,” Radiology Society of North America (RSNA), Nov. 28 – Dec. 3, 2004, Chicago, IL (presentation).

   7. M. N. Gurcan, K. Younis, D. Wormanns, R. C. Hardie, J. W. Hoffmeister, S. K. Rogers,  “Contrast Enhancement Analysis for the Classification of Solitary Pulmonary Nodules in Computed Tomography: Automated Assessment,” Radiology Society of North America (RSNA), Nov. 28 – Dec. 3, 2004, Chicago, IL (presentation).

   8. M. N. Gurcan, R. C. Hardie, et al, “Accurate Nodule Volume Estimation from Helical CT Images: Effect of Reconstruction Filter, Slice Thickness, and Volume Estimation Method,” Radiology Society of North America (RSNA), Nov. 30 – Dec. 5, 2003, Chicago, IL (presentation).

   9. R. D. Ernst, R. C. Hardie, M. N. Gurcan, A. Oto, E. M. Walser, B. H. Allen, et al “CAD Performance Analysis for Pulmonary Nodule Detection in Standard-Dose Thick-Slice Helical CT Images,” Radiology Society of North America (RSNA), Nov. 30 – Dec. 5, 2003, Chicago, IL (presentation).

   10. M. N. Gurcan, R. C. Hardie, S. K. Rogers, D. E. Dozer, B. H. Allen, R. V. Burns, J. W. Hoffmeister, “Automated Global Matching of Temporal Thoracic Helical CT Studies: Feasibility Study,” Computer Assisted Radiology and Surgery (CARS), June 25-28, 2003 London QEII Conference Centre, United Kingdom.

   11. M. N. Gurcan, R. C. Hardie, B. H. Allen, S. K. Rogers, D. E. Dozer, R. V. Burns, J. W. Hoffmeister, “Accurate Nodule Volume Estimation from Helical CT Images: Comparison of Slice-Based and Volume-Based Methods,” Radiology Society of North America (RSNA), Dec. 2-6, 2002, Chicago, IL (Poster).

Patents

Patent No.: US 7,486,812 B2, Feb. 3, 2009.  Shape Estimation and Temporal Registration of Lesions and Nodules.  Metin N. Gurcan, Russell C. Hardie, Steven K. Rogers.  iCAD, Inc. Beavercreek, OH.