the Judgement Comparison between Radiologists and Computer Aided Diagnosis Systems in Polyps Diagnosis based on Colorectal CT Images

Abstract

Introduction: Colorectal cancer is one of the most important types of cancer in the world that often leads to death. Computer-aided diagnostic systems can play an important role in the early diagnosis of this disease. The aim of this study was to investigate the effect of the proposed computer-aided diagnostic system on the accuracy of radiologist diagnosis in colorectal CT images.

Methods: A computer-aided diagnostic system based on MLP artificial neural network was developed to diagnose polyps in colorectal CT images. To validate the developed model, colorectal CT images of 35 patients of Loghman Hakim Hospital in Tehran and Al-Zahra Hospital in Isfahan were collected. These patients underwent colonography on the recommendation of their doctor and then underwent conventional colonoscopy due to the possibility of polyps. The results of the CTC-CAD method and the radiologist according to the Colonoscopy specialist opinion, participated in a Binominal statistical test to check the accuracy of diagnosis.

Results: The result of p-value obtained from the binomial test comparing the correct diagnoses of CTC-CAD method and radiologist about the presence of polyps from Loghman Hakim and Alzahra hospitals were 0.001 and 0.02, respectively. In the binomial test comparing correct diagnoses, CTC-CAD and radiologist methods for the absence of polyps from Loghman Hakim and Alzahra hospitals were 0.262 and 0.367, respectively.

Conclusion: According to the findings of this study, it can be concluded that CTC-CAD performed better in diagnosing the presence of polyps than the radiologist and in the absence of polyps compared to the radiologist.