Does digital image enhancement facilitate semi-automated fracture assessment? CT scout-based inter- and intra-rater reliability study.

This poster is published under an open license. Please read the disclaimer for further details.

Congress:

ECR 2015

Poster Number:

C-0713

Type:

Scientific Exhibit

Keywords:

Bones, Musculoskeletal spine, Computer applications, CT-Quantitative, CAD, Image manipulation / Reconstruction, Acceptance testing, Computer Applications-Detection, diagnosis, Observer performance, Image registration, Image verification, Osteoporosis

Authors:

J. Narloch, W. Glinkowski; Warsaw/PL

DOI:

10.1594/ecr2015/C-0713

DOI-Link:

https://dx.doi.org/10.1594/ecr2015/C-0713

Fig. 1: Scanogram obtained directly from CT

Fig. 2: Enhanced image of the Scanogram with "Sharpen" filter of ImageJ software.

Fig. 3: Combined image of 2 screenshots of results and CT scout view in...

Aims and objectives

Underreporting of vertebral fractures remains a world-wide problem in clinical practice, resulting in under treatment of patients suffering from osteoporotic fractures [6,9,10]. Lateral CT scout views can be useful in clinical research settings to assess and classify vertebral fractures based on a reliable semi-quantitative method [1,2,4,5,8-10]. Digital image enhancement was reported in dental research [3,7]. This study was aimed to determine whether image quality digital enhancement can improve vertebral fracture detection using semi-automated quantitative morphometry.

Methods and materials

Two researchers independently assessed vertebral morphometry based on CT lateral scout views in 250 subjects (49 men, 201 women) diagnosed and treated at the Orthopaedic Trauma facility. One thousand and eighty five vertebral bodies were evaluated twice with shape-based statistical modeling software (SpineAnalyzer, Optasia Medical, Cheadle, UK) before and after image enhancement. Digital image enhancement was processed with the use of ImageJ 1.47 (NIH, Bethesda, MD, USA). Intra- and inter-rater reliability was performed using Interrater Agreement Coefficients and kappa (k) statistics.

Results

Wedge, biconcave and crush vertebral fractures were identified by Raters A and B. Analysis performed after image processing revealed decreased number of fractures of all types, for both raters and solely automated rating. Intra- and inter-rater AC for wedge ratio ranged from 0.527 to 0.534,; 0.514-0.636 for biconcave and 0.533-0.592 for crush fractures. Reliability measured with AC showed substantial to almost perfect agreement (AC=0.859-0.964 and 0.877-0.963, respectively).

Conclusion

This study confirmed that CT scout-based semi-automated quantitative vertebral morphometry is a reliable tool to identify vertebral compression fractures. Digital enhancement tool failed to improve vertebral fracture detection of its reliability.

Personal information

Acknowledgement: This study is supported by project funded by National Science Center N404 695940

References

1. Bazzocchi A, Fuzzi F, Garzillo G, Diano D, Rimondi E, Merlino B, Moio A, Albisinni U, Battista G, Guglielmi G. Reliability and accuracy of scout CT in the detection of vertebral fractures. Br J Radiol. 2013;86(1032):20130373. 2. Bazzocchi A, Spinnato P, Albisinni U, Battista G, Rossi C, Guglielmi G. A careful evaluation of scout CT lateral radiograph may prevent unreported vertebral fractures. Eur J Radiol. 2012;81(9):2353-7. 3. Belém MD, Ambrosano GM, Tabchoury CP, Ferreira-Santos RI, Haiter-Neto F. Performance of digital radiography with enhancement filters...