RESUMO
BACKGROUND AND AIMS: Capsule endoscopy is a central element in the management of patients with suspected or known Crohn's disease. In 2017, PillCam™ Crohn's Capsule was introduced and demonstrated to have greater accuracy in the evaluation of extension of disease in these patients. Artificial intelligence [AI] is expected to enhance the diagnostic accuracy of capsule endoscopy. This study aimed to develop an AI algorithm for the automatic detection of ulcers and erosions of the small intestine and colon in PillCam™ Crohn's Capsule images. METHODS: A total of 8085 PillCam™ Crohn's Capsule images were extracted between 2017 and 2020, comprising 2855 images of ulcers and 1975 erosions; the remaining images showed normal enteric and colonic mucosa. This pool of images was subsequently split into training and validation datasets. The performance of the network was subsequently assessed in an independent test set. RESULTS: The model had an overall sensitivity and specificity of 90.0% and 96.0%, respectively. The precision and accuracy of this model were 97.1% and 92.4%, respectively. In particular, the algorithm detected ulcers with a sensitivity of 83% and specificity of 98%, and erosions with sensitivity and specificity of 91% and 93%, respectively. CONCLUSION: A deep learning model capable of automatically detecting ulcers and erosions in PillCam™ Crohn's Capsule images was developed for the first time. These findings pave the way for the development of automatic systems for detection of clinically significant lesions, optimizing the diagnostic performance and efficiency of monitoring Crohn's disease activity.
Assuntos
Endoscopia por Cápsula , Doença de Crohn/patologia , Redes Neurais de Computação , Colo/patologia , Humanos , Mucosa Intestinal/patologia , Intestino Delgado/patologia , Projetos Piloto , Sensibilidade e Especificidade , Úlcera/patologiaRESUMO
Background: Anatomopathological staging is the primary method to determine the prognosis of patients with colorectal carcinoma (CRC). However, new tools have been developed that can complement it, such as the analysis of the elevation of systemic inflammatory markers. Objective: To evaluate the impact of the elevation of scores based on inflammatory markers (the neutrophil-to-lymphocyte ratio [NLR], the Glasgow Prognostic Score [GPS], and isolated C-reactive protein [CRP]) in the prognosis of patients diagnosed with CRC and submitted to potentially curative surgery in Hospital de Braga, Portugal, between January 1st, 2005, and December 31st, 2010. Methods: A retrospective analysis of the data of 426 patients was performed, with a collection of several clinico-pathological variables, as well as the levels of lymphocytes, neutrophils, albumin and CRP, in the pre- and postoperative periods, to apply the different scores to the sample. Results: From the analysis of the survival curves, we concluded that patients with increased NLR in the pre- and postoperative periods present a lower cancer-related survival than patients with normal NLR (preoperative period: 93.7 versus 122 months; p<0.001; postoperative period: 112 versus 131 months; p=0.002). Patients with increased NLR in the pre- and postoperative periods also had a lower disease-free survival (preoperative period: 88.0 versus 122 months; p<0.001; postoperative period: 111 versus 132 months; p=0.002). In addition, increased pre- and postoperative NLR was associatedwith a higher risk of death due to CRC (preoperatively: hazard ratio [HR]=2.25; p<0.001; postoperatively: HR=2.18; p=0.003). However, the multivariate analysis shows that only postoperative NLR (ajusted HR =2.66; p=0.002) does so independently of the remaining variables. Conclusion: Regarding the scores applied to the sample, the NLR was the one that most consistently related to the prognosis of the patients. However, it would be useful to develop a prospective study that could confirm this relationship. (AU)