BJU international · 2026
The PENIMA project: artificial intelligence decision-support tool for penile lesions - protocol paper and invitation to participate.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Mølsted Janni, Als Anne Birgitte, Jensen Jørgen Bjerggaard, Aagaard Mikael, Lethan Lisa Anni Larsen, Kristiansen Sinja, Moen Christian Arvei, Gerdtsson Axel, Kibsgaard Line, Jespersen Christina Gade, Sørensen Frederikke Eichner, Rossau Ann-Kathrine, Rangen Jens Kramer, Jakobsen Jakob Kristian
Аннотация
OBJECTIVES: To present and describe the design, data collection framework, and interim cohort characteristics of the PENIMA project, a multicentre initiative establishing a clinically annotated penile lesion image repository to support future artificial intelligence (AI)-driven decision-support development - and to encourage urology and dermatology colleagues working with penile lesions to participate in the project.
PATIENTS AND METHODS: Prospective participant inclusion commenced in March 2023 across Scandinavian urological and dermatological departments. Patients with penile anatomy undergoing clinically indicated genital examination are invited to participate and provide oral and written consent as defined by local authority regulations. Clinical photographs and structured metadata are collected. Lesion classification is based on a specialist evaluation, clinical follow-up and histopathology when available. Images are pseudonymised and annotated using a secure, containerised Computer Vision Annotation Tool (CVAT) environment. Supervised machine-learning model development using convolutional neural networks is planned with patient-level dataset partitioning and evaluation using standard diagnostic performance metrics. The study was approved by relevant local authorities, and the approval number is 1-10-72-150-22.
RESULTS: As of February 2026, 1125 participants and approximately 7000 images have been included. In all, 40% represented normal anatomical variants. Among lesional cases, 33% are clinically assessed as benign, 13% premalignant, and 8% malignant; 22% of all cases have histopathological verification. Concordance between clinical suspicion and histopathology is highest for penile carcinoma (84.6%, 95% confidence interval 75.0-91.0%).
CONCLUSION: The PENIMA project includes a multicentre, clinically grounded image repository reflecting real-world penile lesion presentations. The feasible and successful infrastructure provides a foundation for future development and validation of image-based clinical decision-support tools.
PATIENTS AND METHODS: Prospective participant inclusion commenced in March 2023 across Scandinavian urological and dermatological departments. Patients with penile anatomy undergoing clinically indicated genital examination are invited to participate and provide oral and written consent as defined by local authority regulations. Clinical photographs and structured metadata are collected. Lesion classification is based on a specialist evaluation, clinical follow-up and histopathology when available. Images are pseudonymised and annotated using a secure, containerised Computer Vision Annotation Tool (CVAT) environment. Supervised machine-learning model development using convolutional neural networks is planned with patient-level dataset partitioning and evaluation using standard diagnostic performance metrics. The study was approved by relevant local authorities, and the approval number is 1-10-72-150-22.
RESULTS: As of February 2026, 1125 participants and approximately 7000 images have been included. In all, 40% represented normal anatomical variants. Among lesional cases, 33% are clinically assessed as benign, 13% premalignant, and 8% malignant; 22% of all cases have histopathological verification. Concordance between clinical suspicion and histopathology is highest for penile carcinoma (84.6%, 95% confidence interval 75.0-91.0%).
CONCLUSION: The PENIMA project includes a multicentre, clinically grounded image repository reflecting real-world penile lesion presentations. The feasible and successful infrastructure provides a foundation for future development and validation of image-based clinical decision-support tools.