Purpose: Conventional endoscopy is limited by human capability to recognize and to differentiate pathology. Fractal analysis of blood vessels has been used in other organs, such as the retina, but never before to supplement colonoscopy. The aim of this study was to assess whether it is possible to differentiate between healthy and pathologic rectal mucosa using fractal analysis of the mucosal microvascular architecture seen during colonoscopic examination (Supplemental Digital Content 1, http://links.lww.com/SLE/A145).
Methods: A total of 300 consecutive patients, 133 females and 167 males with a mean age of 49.1 (±11.3) years, undergoing endoscopy were included in the prospective cohort study. Colonoscopy of the sigmorectal region was performed, and then analyzed with computer-assisted image fractal analysis.
Results: Fractal analysis of mucosal vasculature allowed for differentiation between healthy and pathologic rectal mucosa, as well as different pathologies (P<0.0001). The sensitivity of fractal analysis to diagnose rectal neoplasia was 92.8% to 96.4%, while the specificity was 91.9% to 98.5% depending on the fractal parameter. The sensitivity of fractal analysis to diagnose rectal colitis was 84.2% to 92.1%, while the specificity was 95.0% to 96.0%, depending on the fractal parameter.
Conclusions: Computer-assisted fractal analysis allows for differentiation between healthy and pathologic rectal mucosa, as well as between different mucosal pathologies, seen during colonoscopy. Fractal analysis improves the sensitivity and specificity of colonoscopy to aid in the diagnosis of neoplasia or colitis, and should be included in the screening and surveillance of these pathologies.
*First Department of General, Oncological and Gastrointestinal Surgery
†Department of Anatomy, Jagiellonian University Medical College
‡Institute of Computer Science, Jagiellonian University
§Chair of Computational Mathematics, Jagiellonian University
∥Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical Engineering, AGH University of Science and Technology, Krakow, Poland
A.G. and K.A.T. contributed to this article equally.
A.G., M.M., and M.Z. are the creators and owners of COLON software. However, fractal analysis, such as described and used in this study, can also be performed using software programs other than COLON. This study has been funded using statutory funds of the Jagiellonian University Medical College. K.A.T. was supported by the Foundation for Polish Science (FNP). The remaining authors declare no conflicts of interest.
Reprints: Krzysztof A. Tomaszewski, MD, PhD, Department of Anatomy, Jagiellonian University Medical College, 12 Kopernika Street, 31-034 Krakow, Poland (e-mail: firstname.lastname@example.org).
Received February 6, 2016
Accepted August 11, 2016