RESUMO
Introduction: An artificial intelligence as a medical device (AIaMD), built on convolutional neural networks, has demonstrated high sensitivity for melanoma. To be of clinical value, it needs to safely reduce referral rates. The primary objective of this study was to demonstrate that the AIaMD had a higher rate of correctly classifying lesions that did not need to be referred for biopsy or urgent face-to-face dermatologist review, compared to teledermatology standard of care (SoC), while achieving the same sensitivity to detect malignancy. Secondary endpoints included the sensitivity, specificity, positive and negative predictive values, and number needed to biopsy to identify one case of melanoma or squamous cell carcinoma (SCC) by both the AIaMD and SoC. Methods: This prospective, single-centre, single-arm, masked, non-inferiority, adaptive, group sequential design trial recruited patients referred to a teledermatology cancer pathway (clinicaltrials.gov NCT04123678). Additional dermoscopic images of each suspicious lesion were taken using a smartphone with a dermoscopic lens attachment. The images were assessed independently by a consultant dermatologist and the AIaMD. The outputs were compared with the final histological or clinical diagnosis. Results: A total of 700 patients with 867 lesions were recruited, of which 622 participants with 789 lesions were included in the per-protocol (PP) population. In total, 63.3% of PP participants were female; 89.0% identified as white, and the median age was 51 (range 18-95); and all Fitzpatrick skin types were represented including 25/622 (4.0%) type IV-VI skin. A total of 67 malignant lesions were identified, including 8 diagnosed as melanoma. The AIaMD sensitivity was set at 91 and 92.5%, to match the literature-defined clinician sensitivity (91.46%) as closely as possible. In both settings, the AIaMD identified had a significantly higher rate of identifying lesions that did not need a biopsy or urgent referral compared to SoC (p-value = 0.001) with comparable sensitivity for skin cancer. Discussion: The AIaMD identified significantly more lesions that did not need to be referred for biopsy or urgent face-to-face dermatologist review, compared to teledermatologists. This has the potential to reduce the burden of unnecessary referrals when used as part of a teledermatology service.
RESUMO
Artificial Intelligence (AI) systems could improve system efficiency by supporting clinicians in making appropriate referrals. However, they are imperfect by nature and misdiagnoses, if not correctly identified, can have consequences for patient care. In this paper, findings from an online survey are presented to understand the aptitude of GPs (n = 50) in appropriately trusting or not trusting the output of a fictitious AI-based decision support tool when assessing skin lesions, and to identify which individual characteristics could make GPs less prone to adhere to erroneous diagnostics results. The findings suggest that, when the AI was correct, the GPs' ability to correctly diagnose a skin lesion significantly improved after receiving correct AI information, from 73.6% to 86.8% (X2 (1, N = 50) = 21.787, p < 0.001), with significant effects for both the benign (X2 (1, N = 50) = 21, p < 0.001) and malignant cases (X2 (1, N = 50) = 4.654, p = 0.031). However, when the AI provided erroneous information, only 10% of the GPs were able to correctly disagree with the indication of the AI in terms of diagnosis (d-AIW M: 0.12, SD: 0.37), and only 14% of participants were able to correctly decide the management plan despite the AI insights (d-AIW M:0.12, SD: 0.32). The analysis of the difference between groups in terms of individual characteristics suggested that GPs with domain knowledge in dermatology were better at rejecting the wrong insights from AI.
RESUMO
AI virtual assistants have significant potential to alleviate the pressure on overly burdened healthcare systems by enabling patients to self-assess their symptoms and to seek further care when appropriate. For these systems to make a meaningful contribution to healthcare globally, they must be trusted by patients and healthcare professionals alike, and service the needs of patients in diverse regions and segments of the population. We developed an AI virtual assistant which provides patients with triage and diagnostic information. Crucially, the system is based on a generative model, which allows for relatively straightforward re-parameterization to reflect local disease and risk factor burden in diverse regions and population segments. This is an appealing property, particularly when considering the potential of AI systems to improve the provision of healthcare on a global scale in many regions and for both developing and developed countries. We performed a prospective validation study of the accuracy and safety of the AI system and human doctors. Importantly, we assessed the accuracy and safety of both the AI and human doctors independently against identical clinical cases and, unlike previous studies, also accounted for the information gathering process of both agents. Overall, we found that the AI system is able to provide patients with triage and diagnostic information with a level of clinical accuracy and safety comparable to that of human doctors. Through this approach and study, we hope to start building trust in AI-powered systems by directly comparing their performance to human doctors, who do not always agree with each other on the cause of patients' symptoms or the most appropriate triage recommendation.
