Enhanced Diagnosis of Generalised Pustular Psoriasis: Multi-Reader Multi-Case Evaluation of the Legit.Health Device as a Diagnosis Support Tool
ABSTRACT
Background:
Generalized pustular psoriasis (GPP), is a rare, chronic, systemic inflammatory disease with an unpredictable and heterogeneous clinical course characterized by chronic symptoms and periods of flaring. GPP presents diagnostic challenges due to its rarity and high similarity with other dermatologic diseases.
Objective:
To assess the performance of Legit.Health, a medical device powered by artificial intelligence software, in assisting healthcare practitioners (HCPs) to identify GPP.
Methods:
An algorithm was developed based on thousands of images of over 200 skin conditions (classified per the International Classification of Diseases 11th revision). The sensitivity and specificity of the algorithm for the differential diagnosis of GPP were assessed using a deep neural network for skin disease recognition. Due to the scarcity of GPP-related images, the medical device was fine-tuned using a dataset that included 4397 GPP images. Thereafter, 15 HCPs (11 primary care practitioners and 4 dermatologists) prospectively reviewed a total of 100 images of 15 visually similar skin conditions, virtually, in a clinical setting. After their diagnostic prediction, Legit.Health provided a prompt giving the top-5 possible skin conditions to assist them with their choice. Performance goals were to achieve a sensitivity of >75% and a specificity >80% for the top-3, and top-5 diagnoses results.
Results:
Legit.Health demonstrated high accuracy in identifying GPP, with top-1, top-3, and top-5 sensitivity and specificity of 80.3%, 86.3%, and 90.0% and 99.8%, 99.6%, and 96.1%, respectively. Results showed a notable increase in the diagnostic accuracy of the HCPs with assistance from Legit.Health, with a relative increase in GPP diagnostic accuracy of 96.9% overall, 120% for primary care practitioners and 58.4% for dermatologists.
Conclusions:
This improvement highlights the potential of Legit.Health in assisting HCPs in diagnosing rare diseases such as GPP, particularly in primary care settings where expertise may be limited, thereby improving patient outcomes. Clinical Trial: ClinicalTrials.gov NCT03782792; https://clinicaltrials.gov/study/NCT03782792 and NCT04399837; https://clinicaltrials.gov/study/NCT04399837
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.