Healthc Inform Res.  2022 Jul;28(3):222-230. 10.4258/hir.2022.28.3.222.

Quantification of Efflorescences in Pustular Psoriasis Using Deep Learning

Affiliations
  • 1Department of Biomedical Engineering, University of Basel, Basel, Switzerland
  • 2Department of Dermatology, University Hospital Zurich, Zurich, Switzerland
  • 3Department of Health Sciences and Technology, Swiss Federal Institute of Technology, Zurich, Switzerland
  • 4Lucerne School of Computer Science and Information Technology, Lucerne University of Applied Sciences and Arts, Lucerne, Switzerland
  • 5Department of Dermatology, University Hospital of Basel, Basel, Switzerland

Abstract


Objectives
Pustular psoriasis (PP) is one of the most severe and chronic skin conditions. Its treatment is difficult, and measurements of its severity are highly dependent on clinicians’ experience. Pustules and brown spots are the main efflorescences of the disease and directly correlate with its activity. We propose an automated deep learning model (DLM) to quantify lesions in terms of count and surface percentage from patient photographs.
Methods
In this retrospective study, two dermatologists and a student labeled 151 photographs of PP patients for pustules and brown spots. The DLM was trained and validated with 121 photographs, keeping 30 photographs as a test set to assess the DLM performance on unseen data. We also evaluated our DLM on 213 unstandardized, out-of-distribution photographs of various pustular disorders (referred to as the pustular set), which were ranked from 0 (no disease) to 4 (very severe) by one dermatologist for disease severity. The agreement between the DLM predictions and experts’ labels was evaluated with the intraclass correlation coefficient (ICC) for the test set and Spearman correlation (SC) coefficient for the pustular set.
Results
On the test set, the DLM achieved an ICC of 0.97 (95% confidence interval [CI], 0.97–0.98) for count and 0.93 (95% CI, 0.92–0.94) for surface percentage. On the pustular set, the DLM reached a SC coefficient of 0.66 (95% CI, 0.60–0.74) for count and 0.80 (95% CI, 0.75–0.83) for surface percentage.
Conclusions
The proposed method quantifies efflorescences from PP photographs reliably and automatically, enabling a precise and objective evaluation of disease activity.

Keyword

Psoriasis; Dermatology; Computer-Assisted Diagnosis; Machine Learning; Deep Learning

Figure

  • Figure 1 Sample image (A) with expert labels (B) and the DLM prediction (C). This picture came from the test set used to evaluate the DLM and was not used in the training process. The original image is shown in (A), while (B) shows the image overlaid with expert labels and (C) the image overlaid with the DLM predictions. The pustules are colored in yellow, the brown spots in red, the patient’s skin in blue, and the background in violet. DLM: deep learning model.

  • Figure 2 Agreement of DLM lesion count predictions with expert labels. The figure shows the Bland-Altman plots of the predicted count for pustules (A), spots (C), and combined lesions (E). The plots for pustules (B), spots (D), and both lesions (F) show the third quartile of the mean difference and the mean absolute difference of the predicted count for patches with up to the number of lesions specified on the horizontal axis value. DLM: deep learning model.

  • Figure 3 Agreement of DLM lesion surface predictions with expert labels. The figure shows the Bland-Altman plots of the predicted surface percentage for pustules (A), spots (C) and combined lesions (E). The plots for pustules (B), spots (D), and both lesions (F) show the third quartile of the mean difference and the mean absolute difference of the predicted surface percentage for patches with up to the lesion surface specified on the horizontal axis value. DLM: deep learning model.


Reference

References

1. Gooderham MJ, Van Voorhees AS, Lebwohl MG. An update on generalized pustular psoriasis. Expert Rev Clin Immunol. 2019; 15(9):907–19. https://doi.org/10.1080/1744666X.2019.1648209.
Article
2. Puzenat E, Bronsard V, Prey S, Gourraud PA, Aractingi S, Bagot M, et al. What are the best outcome measures for assessing plaque psoriasis severity? A systematic review of the literature. J Eur Acad Dermatol Venereol. 2010; 24(Suppl 2):10–6. https://doi.org/10.1111/j.1468-3083.2009.03562.x.
Article
3. Bhushan M, Burden AD, McElhone K, James R, Vanhoutte FP, Griffiths CE. Oral liarozole in the treatment of palmoplantar pustular psoriasis: a randomized, double-blind, placebo-controlled study. Br J Dermatol. 2001; 145(4):546–53. https://doi.org/10.1046/j.1365-2133.2001.04411.x.
Article
4. Kolios AG, French LE, Navarini AA. Detection of small changes in psoriasis intensity with PrecisePASI. Dermatology. 2015; 230(4):314–7. https://doi.org/10.1159/000371811.
Article
5. Youn SW, Choi CW, Kim BR, Chae JB. Reduction of inter-rater and intra-rater variability in psoriasis area and severity index assessment by photographic training. Ann Dermatol. 2015; 27(5):557–62. https://doi.org/10.5021/ad.2015.27.5.557.
Article
6. Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017; 542(7639):115–8. https://doi.org/10.1038/nature21056.
Article
7. Meienberger N, Anzengruber F, Amruthalingam L, Christen R, Koller T, Maul JT, et al. Observer-independent assessment of psoriasis-affected area using machine learning. J Eur Acad Dermatol Venereol. 2020; 34(6):1362–8. https://doi.org/10.1111/jdv.16002.
Article
8. Andermatt S, Horvath A, Pezold S, Cattin P. Pathology segmentation using distributional differences to images of healthy origin. Crimi A, Bakas S, Kuijf H, Keyvan F, Reyes M, van Walsum T, editors. Brainlesion: glioma, multiple sclerosis, stroke and traumatic brain injuries. Cham, Switzerland: Springer;2018. p. 228–38. https://doi.org/10.1007/978-3-030-11723-8_23.
Article
9. Furger F, Amruthalingam L, Navarini A, Pouly M. Applications of generative adversarial networks to dermatologic imaging. Schilling FP, Stadelmann T, editors. Artificial neural networks in pattern recognition. Cham, Switzerland: Springer;2020. p. 187–99. https://doi.org/10.1007/978-3-030-58309-5_15.
Article
10. Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation. Navab N, Hornegger J, Wells W, Frangi A, editors. Medical image computing and computer-assisted intervention – MICCAI 2015. Cham, Switzerland: Springer;2015. p. 234–41. https://doi.org/10.1007/978-3-319-24574-4_28.
Article
11. He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. In : Proceedings of the IEEE conference on computer vision and pattern recognition; 2016 Jun 26–Jul 1; Las Vegas, NV. p. 770–8.
Article
12. Deng J, Dong W, Socher R, Li LJ, Li K, Fei-Fei L. ImageNet: a large-scale hierarchical image database. In : Proceedings of 2009 IEEE Conference on Computer Vision and Pattern Recognition; 2009 Jun 20–25; Miami, FL. p. 248–55. https://doi.org/10.1109/CVPR.2009.5206848.
Article
13. Smith LN. A disciplined approach to neural network hyper-parameters: Part 1--learning rate, batch size, momentum, and weight decay [Internet]. Ithaca (NY): arXiv.org;2018. [cited at 2022 Jul 20]. Available from: https://arxiv.org/abs/1803.09820.
14. Yeung M, Sala E, Schonlieb CB, Rundo L. Unified focal loss: generalising dice and cross entropy-based losses to handle class imbalanced medical image segmentation. Comput Med Imaging Graph. 2022; 95:102026. https://doi.org/10.1016/j.compmedimag.2021.102026.
Article
15. Lin TY, Goyal P, Girshick R, He K, Dollar P. Focal loss for dense object detection. IEEE Trans Pattern Anal Mach Intell. 2020; 42(2):318–27. https://doi.org/10.1109/TPAMI.2018.2858826.
Article
16. Zhu W, Huang Y, Zeng L, Chen X, Liu Y, Qian Z, et al. AnatomyNet: deep learning for fast and fully automated whole-volume segmentation of head and neck anatomy. Med Phys. 2019; 46(2):576–89. https://doi.org/10.1002/mp.13300.
Article
17. El Jurdi R, Petitjean C, Honeine P, Cheplygina V, Abdallah F. High-level prior-based loss functions for medical image segmentation: a survey. Comput Vis Image Underst. 2021; 210:103248. https://doi.org/10.1016/j.cviu.2021.103248.
Article
18. Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, et al. PyTorch: an imperative style, high-performance deep learning library. Adv Neural Inf Process Syst. 2019; 32:8024–35.
19. Howard J, Gugger S. Fastai: a layered API for deep learning. Information. 2020; 11(2):108. https://doi.org/10.3390/info11020108.
Article
20. van Stralen KJ, Dekker FW, Zoccali C, Jager KJ. Measuring agreement, more complicated than it seems. Nephron Clin Pract. 2012; 120(3):c162–7. https://doi.org/10.1159/000337798.
Article
21. Schaap MJ, Cardozo NJ, Patel A, de Jong EM, van Ginneken B, Seyger MM. Image-based automated psoriasis area severity index scoring by convolutional neural networks. J Eur Acad Dermatol Venereol. 2022; 36(1):68–75. https://doi.org/10.1111/jdv.17711.
Article
22. Wu X, Yan Y, Zhao S, Kuang Y, Ge S, Wang K, et al. Automatic severity rating for improved psoriasis treatment. Medical image computing and computer assisted intervention – MICCAI 2021. Cham, Switzerland: Springer;2021. p. 185–94. https://doi.org/10.1007/978-3-030-87234-2_18.
Article
23. Pal A, Chaturvedi A, Garain U, Chandra A, Chatterjee R, Senapati S. Severity assessment of psoriatic plaques using deep CNN based ordinal classification. OR 20 Context-aware operating theaters, computer assisted robotic endoscopy, clinical image-based procedures, and skin image analysis. Cham, Switzerland: Springer;2018. p. 252–9. https://doi.org/10.1007/978-3-030-01201-4_27.
Article
24. Cazzolato MT, Ramos JS, Rodrigues LS, Scabora LC, Chino DY, Jorge AE, et al. The UTrack framework for segmenting and measuring dermatological ulcers through telemedicine. Comput Biol Med. 2021; 134:104489. https://doi.org/10.1016/j.compbiomed.2021.104489.
Article
25. Zhao C, Shuai R, Ma L, Liu W, Wu M. Segmentation of dermoscopy images based on deformable 3D convolution and ResU-NeXt +. Med Biol Eng Comput. 2021; 59(9):1815–32. https://doi.org/10.1007/s11517-021-02397-9.
Article
26. Goyal M, Oakley A, Bansal P, Dancey D, Yap MH. Skin lesion segmentation in dermoscopic images with ensemble deep learning methods. IEEE Access. 2019; 8:4171–81. https://doi.org/10.1109/ACCESS.2019.2960504.
Article
27. Schnurle S, Pouly M, vor der Bruck T, Navarini A, Koller T. On using support vector machines for the detection and quantification of hand eczema. In : Proceedings of the 9th International Conference on Agents and Artificial Intelligence (ICAART); 2017 Feb 24–26; Porto, Portugal. p. 75–84.
Article
28. Raj R, Londhe ND, Sonawane RS. Deep learning based multi-segmentation for automatic estimation of psoriasis area score. In : Proceedings of 2021, 8th International Conference on Signal Processing and Integrated Networks (SPIN); 2021 Aug 26–27; Noida, India. p. 1137–42. https://doi.org/10.1109/SPIN52536.2021.9566039.
Article
29. Liu Y, Sun P, Wergeles N, Shang Y. A survey and performance evaluation of deep learning methods for small object detection. Expert Syst Appl. 2021; 172:114602. https://doi.org/10.1016/j.eswa.2021.114602.
Article
30. Finnane A, Curiel-Lewandrowski C, Wimberley G, Caffery L, Katragadda C, Halpern A, et al. Proposed technical guidelines for the acquisition of clinical images of skin-related conditions. JAMA Dermatol. 2017; 153(5):453–7. https://doi.org/10.1001/jamadermatol.2016.6214.
Article
Full Text Links
  • HIR
Actions
Cited
CITED
export Copy
Close
Share
  • Twitter
  • Facebook
Similar articles
Copyright © 2024 by Korean Association of Medical Journal Editors. All rights reserved.     E-mail: koreamed@kamje.or.kr