Objective Assessment of Psoriasis Lesion Thickness for PASI Scoring using 3D Digital Imaging

Psoriasis is a chronic inflammatory skin condition which affects 2-3% of population around the world. Psoriasis Area and Severity Index (PASI) is a gold standard to assess psoriasis severity as well as the treatment efficacy. Although a gold standard, PASI is rarely used because it is tedious and complex. In practice, PASI score is determined subjectively by dermatologists, therefore inter and intra variations of assessment are possible to happen even among expert dermatologists. This research develops an algorithm to assess psoriasis lesion for PASI scoring objectively. Focus of this research is thickness assessment as one of PASI four parameters beside area, erythema and scaliness. Psoriasis lesion thickness is measured by averaging the total elevation from lesion base to lesion surface. Thickness values of 122 3D images taken from 39 patients are grouped into 4 PASI thickness score using K-means clustering. Validation on lesion base construction is performed using twelve body curvature models and show good result with coefficient of determinant (R2) is equal to 1.




References:
[1] Peter van de Kerkhof, Textbook of Psoriasis, 2003, Blackwell
Publishing: Massachussetts
[2] The Psoriasis Association, What is Psoriasis?, 2008, The Psoriasis
Association: UK
[3] Lionel Fry, An Atlas of Psoriasis, 2005, Taylor&Francis: London
[4] T. Frederiksson, U. Pettersson, Severe Psoriasis: Oral Therapy with a
New Retinoid, Dermatologica, 1978, 157(4), pp: 238-44
[5] M. Alper, A. Kavak, A.H. Parlak, R. Demirici, I. Belenli, N. Yesildal,
Measurement of Epidermal Thickness in a Patient with Psoriasis by
Computer Supported Image Analyisis, Brazilian Journal of Medical
and Biological Research, 2004, 37, pp: 111-117.
[6] Harold Alexander, D.L. Miller, Determining Skin Thickness with
Pulsed Ultra Sound. The Journal of Investigative Dermatology, Vol 72,
pp: 17-19. 1979.
[7] Serup, J.: Non-invasive quantification of psoriasis plaques-
measurement of skin thickness with 15 MHz pulsed ultrasound. Journal
of Clinical and Experimental Dermatology, Volume 9 Issue 5, 502 --
508 (2006)
[8] Konica Minolta Vivid 910 Non Contact 3D Digitizer Manual
Handbook, Japan (2001)
[9] Bryan F. Jones, Peter Plassman, An Instrument to Measure the
Dimension of Skin Wounds, IEEE Trancastion on Biomedical
Engineering, Vol. 42, No.5 1995, pp: 464 - 470
[10] Zhilin Li, Qing Zhu, Christopher Gold, Digital Terrain Modeling:
Principle and Methodology,2005, CRC Press: Florida
[11] J. P. Luntama, S. Koponen, M. Hallikainen, Analysis of Sea Ice
Thickness and Mass Estimation with a Spaceborne Laser Altimerer,
Geosciense and Remote Sensing, 1997, Volume 3. pp: 1314-1316
[12] Frederic Gibou and Ronald Fedkiw. "A fast hybrid k-means level set
algorithm for segmentation". In 4th Annual Hawaii International
Conference on Statistics and Mathematics, pages 281-291, 2005.
[13] R. Herwig, A.J. Poustka, C. Muller, C. Bull, H. Lehrach, and J O-Brien.
Large-scale clustering of cdna-fingerprinting data. Genome Research,
9:1093-1105, 1999.
[14] Paul J. Besl, Ramesh C. Jain, Three-Dimensional Object Recognition,
Annals of Discrete Mathematics-ACM, Vol. 17, 1985, pp: 75-145