Deep Neural Frameworks Improve the Accuracy of General Practitioners in the Classification of Pigmented Skin Lesions.

Maximiliano Lucius,Martín Belvisi,José Antonio De All,Carlos M Galmarini,Victoria Lorenzatti,Jorge De All,Marisa Lanfranconi,Luciana Radizza

doi:10.3390/diagnostics10110969

Maximiliano Lucius, Martín Belvisi + Show 6 more

Open Access

https://doi.org/10.3390/diagnostics10110969

Copy DOI

Abstract

This study evaluated whether deep learning frameworks trained in large datasets can help non-dermatologist physicians improve their accuracy in categorizing the seven most common pigmented skin lesions. Open-source skin images were downloaded from the International Skin Imaging Collaboration (ISIC) archive. Different deep neural networks (DNNs) (n = 8) were trained based on a random dataset constituted of 8015 images. A test set of 2003 images was used to assess the classifiers’ performance at low (300 × 224 RGB) and high (600 × 450 RGB) image resolution and aggregated data (age, sex and lesion localization). We also organized two different contests to compare the DNN performance to that of general practitioners by means of unassisted image observation. Both at low and high image resolution, the DNN framework differentiated dermatological images with appreciable performance. In all cases, the accuracy was improved when adding clinical data to the framework. Finally, the least accurate DNN outperformed general practitioners. The physician’s accuracy was statistically improved when allowed to use the output of this algorithmic framework as guidance. DNNs are proven to be high performers as skin lesion classifiers and can improve general practitioner diagnosis accuracy in a routine clinical scenario.

Highlights

Diagnosis in dermatology is largely based on the visual inspection of a lesion on the suspicious skin area
This study aimed to evaluate whether deep learning frameworks trained in large datasets can help non-dermatologist physicians improve their accuracy in categorizing the seven most common pigmented skin lesions representing more than 90% of the pigmented skin lesions
Almost all deep neural networks (DNNs) showed the highest true positive rate (TPR) for melanocytic nevi classification when compared to the other pigmented lesions, with observed geometric mean values equal or lower than 0.65; interestingly, VGG16, VGG19 and MobileNet showed high TPR for vascular lesion classification

Summary

Introduction

Diagnosis in dermatology is largely based on the visual inspection of a lesion on the suspicious skin area. If the lesion is suspicious, the visual inspection is supplemented with different diagnostic tools (e.g., dermoscopy, confocal microscopy or optical coherence tomography) providing the ability to explore the skin in vivo, in depth and at a higher resolution [5,6]. Access to these instruments remains limited due to time, logistical and cost concerns. Even when this technical support is feasible, dermatologists rarely achieve average rates greater than 85% [7,8]. The situation is even worse if we consider that there is a shortage of dermatologists whilst the diagnostic accuracy of non-expert clinicians is sensibly below than what is observed with dermatologists, reaching estimate rates between 20 and 40% [9,10,11,12,13]

Objectives

Methods

Results

Discussion

Conclusion