Automatic discrimination of actinic keratoses from clinical photographs

BACKGROUND AND OBJECTIVE Actinic keratoses (AK) are common premalignant skin lesions that can progress to invasive skin squamous cell carcinoma (sSCC). The subtle accumulation of multiple AK in aging individuals increases the risk of sSCC development, and this underscores the need for efficient treatment and patient follow-up. Our objectives were to develop a method based on color texture analysis of standard clinical photographs for the discrimination of AK from healthy skin and subsequently to test the developed approach in the quantification of field-directed treatment interventions. METHODS AK and healthy skin in clinical photographs of 22 patients were demarcated by experts and regions of interest (ROIs) of 50 × 50 pixels were cropped. The data set comprised 6010 and 13915 ROIs from AK and healthy skin, respectively. Color texture features were extracted using local binary patterns (LBP) or texton frequency histograms and evaluated employing a support vector machine (SVM) classifier. Classifier evaluation was performed using a leave-one-patient-out scheme in RGB, YIQ and CIE-Lab color spaces. The best configuration of the SVM model was tested using 157 AK and 216 healthy skin rectangular regions of arbitrary size. AK treatment outcome was evaluated in an additional group of eight patients with 32 skin lesions. RESULTS The best configuration of the discrimination model was achieved by employing LBP color texture descriptors estimated from the Y and I components of the YIQ color space. The sensitivity and specificity of the SVM model were 80.1% and 81.1% at ROI level and 89.8% and 91.7% at region level, respectively. Based on the classifier results the quantitative AK reduction was 83.6%. CONCLUSIONS It is important that patients with AK seek evaluation for treatment to reduce the risk of disease progression. Efficient patient follow-up and treatment evaluation require cost-effective and easy to use approaches. The proposed SVM discrimination model based on LBP color texture analysis renders clinical photography a practical, widely available and cost-effective tool for the evaluation of AK burden and treatment efficacy.

[1]  Hyeran Byun,et al.  Applications of Support Vector Machines for Pattern Recognition: A Survey , 2002, SVM.

[2]  Panagiota Spyridonos,et al.  Spatial fuzzy c-means algorithm with adaptive fuzzy exponent selection for robust vermilion border detection in healthy and diseased lower lips , 2014, Comput. Methods Programs Biomed..

[3]  Gurkirat Kaur,et al.  Automatic Detection and Segmentation of Skin Melanoma Images- An Introduction , 2015 .

[4]  John Platt,et al.  Probabilistic Outputs for Support vector Machines and Comparisons to Regularized Likelihood Methods , 1999 .

[5]  Paul F. Whelan,et al.  Experiments in colour texture analysis , 2001, Pattern Recognit. Lett..

[6]  Taeyoon Son,et al.  Dermatological feasibility of multimodal facial color imaging modality for cross‐evaluation of facial actinic keratosis , 2011, Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging.

[7]  Tarl W. Prow,et al.  Automated Detection of Actinic Keratoses in Clinical Photographs , 2015, PloS one.

[8]  Yong Man Ro,et al.  Using colour local binary pattern features for face recognition , 2010, 2010 IEEE International Conference on Image Processing.

[9]  Mark F. Naylor,et al.  Actinic keratoses , 2009, Cancer.

[10]  Robert B. Fisher,et al.  Using 3D information for classification of non-melanoma skin lesions , 2008 .

[11]  R. Lew,et al.  Improvement in precision of counting actinic keratoses , 2014, The British journal of dermatology.

[12]  S. Shumack,et al.  Daylight photodynamic therapy with methyl aminolevulinate cream as a convenient, similarly effective, nearly painless alternative to conventional photodynamic therapy in actinic keratosis treatment: a randomized controlled trial , 2014, The British journal of dermatology.

[13]  H. Wulf,et al.  Continuous activation of PpIX by daylight is as effective as and less painful than conventional photodynamic therapy for actinic keratoses; a randomized, controlled, single‐blinded study , 2008, The British journal of dermatology.

[14]  Javier M. Moguerza,et al.  Support Vector Machines with Applications , 2006, math/0612817.

[15]  M Aickin,et al.  Expression of p53 protein in actinic keratosis, adjacent, normal-appearing, and non-sun-exposed human skin. , 1997, Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology.

[16]  M. Thissen,et al.  Actinic keratosis: how to differentiate the good from the bad ones? , 2006, European journal of dermatology : EJD.

[17]  W. Chang,et al.  Computer-Aided Diagnosis of Skin Lesions Using Conventional Digital Photography: A Reliability and Feasibility Study , 2013, PloS one.

[18]  H. A. Miot,et al.  The reliability of counting actinic keratosis , 2013, Archives of Dermatological Research.

[19]  Rafael García,et al.  Computerized analysis of pigmented skin lesions: A review , 2012, Artif. Intell. Medicine.

[20]  María Huerta-Brogeras,et al.  Validation of dermoscopy as a real-time noninvasive diagnostic imaging technique for actinic keratosis. , 2012, Archives of dermatology.

[21]  J. Malvehy,et al.  A new vision of actinic keratosis beyond visible clinical lesions , 2015, Journal of the European Academy of Dermatology and Venereology : JEADV.

[22]  R. Hofmann-Wellenhof,et al.  Discrimination of Actinic Keratoses from Normal Skin with Reflectance Mode Confocal Microscopy , 2008, Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.].

[23]  B. Rao,et al.  Uses of non‐invasive imaging in the diagnosis of skin cancer: an overview of the currently available modalities , 2013, International journal of dermatology.

[24]  Georgios Gaitanis,et al.  Immunocryosurgery for Nonmelanoma Skin Cancer: Applications and Practical Tips , 2015 .

[25]  Ammara Masood,et al.  Computer Aided Diagnostic Support System for Skin Cancer: A Review of Techniques and Algorithms , 2013, Int. J. Biomed. Imaging.

[26]  Ilias Maglogiannis,et al.  Overview of Advanced Computer Vision Systems for Skin Lesions Characterization , 2009, IEEE Transactions on Information Technology in Biomedicine.

[27]  Randy H. Moss,et al.  Advances in skin cancer image analysis , 2011, Comput. Medical Imaging Graph..

[28]  J. McGregor,et al.  British Association of Dermatologists’ guidelines for the care of patients with actinic keratosis 2017 , 2017, The British journal of dermatology.

[29]  Andrew Zisserman,et al.  A Statistical Approach to Texture Classification from Single Images , 2004, International Journal of Computer Vision.

[30]  F. Gruijl,et al.  Early p53 alterations in mouse skin carcinogenesis by UVB radiation: immunohistochemical detection of mutant p53 protein in clusters of preneoplastic epidermal cells. , 1996, Proceedings of the National Academy of Sciences of the United States of America.

[31]  Leonard H Goldberg,et al.  Review ofactinic keratosis. Part I: etiology, epidemiology and clinical presentation. , 2010, Journal of drugs in dermatology : JDD.

[32]  Panagiota Spyridonos,et al.  Gray Hausdorff distance measure for medical image comparison in dermatology: Evaluation of treatment effectiveness by image similarity , 2013, Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging.

[33]  S Schreml,et al.  Clinical, histopathological and immunohistochemical assessment of human skin field cancerization before and after photodynamic therapy , 2012, The British journal of dermatology.

[34]  Mika Laiho,et al.  Towards Understanding the Formation of Uniform Local Binary Patterns , 2013 .

[35]  Robert B. Fisher,et al.  A Color and Texture Based Hierarchical K-NN Approach to the Classification of Non-melanoma Skin Lesions , 2013 .

[36]  Chengjun Liu,et al.  Color space normalization: Enhancing the discriminating power of color spaces for face recognition , 2010, Pattern Recognit..

[37]  E. Veledar,et al.  Reliability of quantification measures of actinic keratosis , 2013, The British journal of dermatology.

[38]  Paul Maruff,et al.  Amyloid-Related Memory Decline in Preclinical Alzheimer’s Disease Is Dependent on APOE ε4 and Is Detectable over 18-Months , 2015, PloS one.

[39]  Gerard de Haan,et al.  Automatic imaging sysem with decision support for inspection of pigmented skin lesions and melanoma diagnosis. , 2009 .

[40]  M Suppa,et al.  A new algorithm for the discrimination of actinic keratosis from normal skin and squamous cell carcinoma based on in vivo analysis of optical properties by high‐definition optical coherence tomography , 2016, Journal of the European Academy of Dermatology and Venereology : JEADV.

[41]  Matti Pietikäinen,et al.  Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[42]  Panagiota Spyridonos,et al.  Evaluation of vermillion border descriptors and relevance vector machines discrimination model for making probabilistic predictions of solar cheilosis on digital lip photographs , 2015, Comput. Biol. Medicine.

[43]  R M Hurwitz,et al.  Solar keratosis: an evolving squamous cell carcinoma. Benign or malignant? , 1995, Dermatologic surgery : official publication for American Society for Dermatologic Surgery [et al.].

[44]  Nikolaos G. Bourbakis,et al.  A survey of skin-color modeling and detection methods , 2007, Pattern Recognit..

[45]  S. Salasche,et al.  Epidemiology of actinic keratoses and squamous cell carcinoma. , 2000, Journal of the American Academy of Dermatology.

[46]  D. Rigel,et al.  The importance of early diagnosis and treatment of actinic keratosis. , 2013, Journal of the American Academy of Dermatology.

[47]  Andrew Zisserman,et al.  A Statistical Approach to Material Classification Using Image Patch Exemplars , 2009, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[48]  C. Cockerell,et al.  Histopathology of incipient intraepidermal squamous cell carcinoma ("actinic keratosis"). , 2000, Journal of the American Academy of Dermatology.

[49]  T. Jørgensen,et al.  Machine‐learning classification of non‐melanoma skin cancers from image features obtained by optical coherence tomography , 2008, Skin research and technology : official journal of International Society for Bioengineering and the Skin (ISBS) [and] International Society for Digital Imaging of Skin (ISDIS) [and] International Society for Skin Imaging.

[50]  R Marks,et al.  The role of treatment of actinic keratoses in the prevention of morbidity and mortality due to squamous cell carcinoma. , 1991, Archives of dermatology.

[51]  Lauge Sørensen,et al.  Multiple Classifier Systems in Texton-Based Approach for the Classification of CT Images of Lung , 2010, MCV.

[52]  Gerhard Goos,et al.  Computer Vision - ECCV 2000 , 2003, Lecture Notes in Computer Science.

[53]  Jacob Scharcanski,et al.  Automated prescreening of pigmented skin lesions using standard cameras , 2011, Comput. Medical Imaging Graph..

[54]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[55]  Matti Pietikäinen,et al.  Gray Scale and Rotation Invariant Texture Classification with Local Binary Patterns , 2000, ECCV.