METHODS: In this paper, we propose a novel approach to distinguish colonic polyps by integrating several techniques, including a modified deep residual network, principal component analysis and AdaBoost ensemble learning. A powerful deep residual network architecture, ResNet-50, was investigated to reduce the computational time by altering its architecture. To keep the interference to a minimum, median filter, image thresholding, contrast enhancement, and normalisation techniques were exploited on the endoscopic images to train the classification model. Three publicly available datasets, i.e., Kvasir, ETIS-LaribPolypDB, and CVC-ClinicDB, were merged to train the model, which included images with and without polyps.
RESULTS: The proposed approach trained with a combination of three datasets achieved Matthews Correlation Coefficient (MCC) of 0.9819 with accuracy, sensitivity, precision, and specificity of 99.10%, 98.82%, 99.37%, and 99.38%, respectively.
CONCLUSIONS: These results show that our method could repeatedly classify endoscopic images automatically and could be used to effectively develop computer-aided diagnostic tools for early CRC detection.
OBJECTIVE: Our objective was to create a framework that can guide future implementation and research on the use of eHealth tools to support patients with growth disorders who require growth hormone therapy.
METHODS: A total of 12 pediatric endocrinologists with experience in eHealth, from a wide geographical distribution, participated in a series of online discussions. We summarized the discussions of 3 workshops, conducted during 2020, on the use of eHealth in the management of growth disorders, which were structured to provide insights on existing challenges, opportunities, and solutions for the implementation of eHealth tools across the patient journey, from referral to the end of pediatric therapy.
RESULTS: A total of 815 responses were collected from 2 questionnaire-based activities covering referral and diagnosis of growth disorders, and subsequent growth hormone therapy stages of the patient pathway, relating to physicians, nurses, and patients, parents, or caregivers. We mapped the feedback from those discussions into a framework that we developed as a guide to integration of eHealth tools across the patient journey. Responses focused on improved clinical management, such as growth monitoring and automation of referral for early detection of growth disorders, which could trigger rapid evaluation and diagnosis. Patient support included the use of eHealth for enhanced patient and caregiver communication, better access to educational opportunities, and enhanced medical and psychological support during growth hormone therapy management. Given the potential availability of patient data from connected devices, artificial intelligence can be used to predict adherence and personalize patient support. Providing evidence to demonstrate the value and utility of eHealth tools will ensure that these tools are widely accepted, trusted, and used in clinical practice, but implementation issues (eg, adaptation to specific clinical settings) must be addressed.
CONCLUSIONS: The use of eHealth in growth hormone therapy has major potential to improve the management of growth disorders along the patient journey. Combining objective clinical information and patient adherence data is vital in supporting decision-making and the development of new eHealth tools. Involvement of clinicians and patients in the process of integrating such technologies into clinical practice is essential for implementation and developing evidence that eHealth tools can provide value across the patient pathway.