Healthc Inform Res.  2013 Sep;19(3):196-204. 10.4258/hir.2013.19.3.196.

Quantitative Measurement Method for Possible Rib Fractures in Chest Radiographs

• Kim J
• Kim S
• Kim YJ
• Kim KG
• Park J

• KMID: 2166698
• DOI: http://doi.org/10.4258/hir.2013.19.3.196

Abstract

OBJECTIVES
This paper proposes a measurement method to quantify the abnormal characteristics of the broken parts of ribs using local texture and shape features in chest radiographs.
METHODS
Our measurement method comprises two steps: a measurement area assignment and sampling step using a spline curve and sampling lines orthogonal to the spline curve, and a fracture-ness measurement step with three measures, asymmetry and gray-level co-occurrence matrix based measures (contrast and homogeneity). They were designed to quantify the regional shape and texture features of ribs along the centerline. The discriminating ability of our method was evaluated through region of interest (ROI) analysis and rib fracture classification test using support vector machine.
RESULTS
The statistically significant difference was found between the measured values from fracture and normal ROIs; asymmetry (p < 0.0001), contrast (p < 0.001), and homogeneity (p = 0.022). The rib fracture classifier, trained with the measured values in ROI analysis, detected every rib fracture from chest radiographs used for ROI analysis, but it also classified some unbroken parts of ribs as abnormal parts (8 to 17 line sets; length of each line set, 2.998 +/- 2.652 mm; length of centerlines, 131.067 +/- 29.460 mm).
CONCLUSIONS
Our measurement method, which includes a flexible measurement technique for the curved shape of ribs and the proposed shape and texture measures, could discriminate the suspicious regions of ribs for possible rib fractures in chest radiographs.

Keyword

Rib Fractures; Radiography; Computer-Aided Radiographic Image Interpretation; Image Processing; Decision Support Techniques

MeSH Terms

Decision Support Techniques
Rib Fractures
Ribs
Thorax

Figure

• 

  Figure 1 (A) Measurement area assignment using a spline and a thickness value. (B) Fracture-ness measurement process on a single rib.

• 

  Figure 2 Various morphologic features of (A) fractured and (B) non-fractured (normal) ribs on which region of interests were drawn. The window-level of each image is adapted for print.

• 

  Figure 3 Measured values with various unit lengths for normal and fracture region of interests (ROIs). (A, B) Asymmetry, (C, D) Contrast , and (E, F) 10.0 × Homogeneity . "Centerline" means the length of the centerline for each ROI. The dotted vertical lines indicate 0.1 of the ratio, "Unit length/Centerline".

• 

  Figure 4 Box plots of measured values with a constant unit length (0.5 mm) for fracture region of interests (ROIs) and normal ROIs. (A) Asymmetry, (B) Contrast, and (C) 10.0 × Homogeneity.

• 

  Figure 5 Examples of rib fracture detection results. The training data of the rib fracture classifier were the asymmetry values of our measurement method. The dots indicate the end points of the sampling lines which are classified into fracture group by the rib fracture classifier. The black circles indicate the actual broken part of ribs. The lines with square dots are the centerlines of the ribs.

Reference

1. Bhavnagri SJ, Mohammed TL. When and how to image a suspected broken rib. Cleve Clin J Med. 2009; 76(5):309–314.

2. Dubinsky I, Low A. Non-life-threatening blunt chest trauma: appropriate investigation and treatment. Am J Emerg Med. 1997; 15(3):240–243.

3. Abe K, Katsuragawa S, Sasaki Y, Yanagisawa T. A fully automated adaptive unsharp masking technique in digital chest radiograph. Invest Radiol. 1992; 27(1):64–70.

4. Prokop M, Schaefer CM, Oestmann JW, Galanski M. Improved parameters for unsharp mask filtering of digital chest radiographs. Radiology. 1993; 187(2):521–526.

5. Stahl M, Aach T, Dippel S. Digital radiography enhancement by nonlinear multiscale processing. Med Phys. 2000; 27(1):56–65.

6. Donnelley M, Knowles G. Computer aided long bone fracture detection. In : Proceedings of the 8th International Symposium on Signal Processing and Its Applications; 2005 Aug 28-31; Sydney, Australia. p. 175–178.

7. Lum VL, Leow WK, Chen Y, Howe TS, Png MA. Combining classifiers for bone fracture detection in X-ray images. In : Proceedings of the 12th IEEE International Conference on Image Processing; 2005 Sep 11-14; Genoa, Italy. p. 1149–1152.

8. He JC, Leow WK, Howe TS. Hierarchical classifiers for detection of fractures in X-ray images. In : Kropatsch WG, Kampel M, Hanbury A, editors. Computer analysis of images and patterns. Heidelberg, Germany: Springer;2007. p. 962–969.

9. Linda CH, Jiji GW. Crack detection in X-ray images using fuzzy index measure. Appl Soft Comput. 2011; 11(4):3571–3579.

10. Kim TY, Son J, Kim KG. The recent progress in quantitative medical image analysis for computer aided diagnosis systems. Healthc Inform Res. 2011; 17(3):143–149.

11. Lange S, Walsh G. Radiology of chest diseases. 3rd ed. New York (NY): Thieme Medical Publishers;2007.

12. Costa LF, Cesar RM Jr. Shape analysis and classification: theory and practice. 2nd ed. Boca Raton (FL): CRC Press;2000.

13. Hall-Beyer M. The GLCM tutorial home page [Internet]. Calgary (AB): University of Calgary;c2007. cited at 2013 Sep 1. Available from: http://www.fp.ucalgary.ca/mhallbey/tutorial.htm.

14. Cortes C, Vapnik V. Support-vector networks. Mach Learn. 1995; 20(3):273–297.

15. Karatzoglou A, Meyer D, Hornik K. Support vector machines in R. J Stat Softw. 2006; 15(9):1–28.