Accuracy of computer diagnosis of melanoma: a quantitative meta-analysis.

BACKGROUND Recent developments in computer technology have raised expectations that fully automated diagnostic instruments will become available to diagnose cutaneous melanoma without the need of human expertise. OBJECTIVES To critically review the contemporary literature on computer diagnosis of melanoma, evaluate the accuracy of such computer diagnosis, analyze the influence of study characteristics, and compare the accuracy of computer diagnosis of melanoma with human diagnosis. METHODS Quantitative meta-analysis of published reports. DATA SOURCES Eligible studies were identified by a MEDLINE search covering the period from January 1991 to March 2002, by manual searches of the reference lists of retrieved articles, and by direct communication with experts. RESULTS Thirty studies with substantial differences in methodological quality were deemed eligible for meta-analysis. Five of these complied with the predetermined list of "good quality" requirements, but none met all methodological quality requirements. Ten of these studies compared the performance of computer diagnosis with human diagnosis. The diagnostic accuracy achieved with computer diagnosis was statistically not different from that of human diagnosis (log odds ratios, 3.36 vs 3.51; P =.80). The diagnostic performance of the computer diagnosis was better for studies that used dermoscopic images than for studies that used clinical images (log odds ratios, 4.2 vs 3.4; P =.08). Other study characteristics did not significantly influence the accuracy of the computer diagnosis. CONCLUSIONS The computer diagnosis of melanoma is accurate under experimental conditions, but the practical value of automated diagnostic instruments under real-world conditions is currently unknown. We suggest minimum requirements for methodological quality in future experimental studies or, ideally, randomized controlled trials.

[1]  C. Grin,et al.  Accuracy in the clinical diagnosis of malignant melanoma. , 1990, Archives of dermatology.

[2]  A. Green,et al.  Computer image analysis of pigmented skin lesions , 1991, Melanoma research.

[3]  E Claridge,et al.  Shape analysis for classification of malignant melanoma. , 1992, Journal of biomedical engineering.

[4]  C Bartoli,et al.  Results obtained by using a computerized image analysis system designed as an aid to diagnosis of cutaneous melanoma , 1992, Melanoma research.

[5]  T. Schindewolf,et al.  Classification of melanocytic lesions with color and texture analysis using digital image processing. , 1993, Analytical and quantitative cytology and histology.

[6]  L E Moses,et al.  Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. , 1993, Statistics in medicine.

[7]  L E Moses,et al.  Estimating Diagnostic Accuracy from Multiple Conflicting Reports , 1993, Medical decision making : an international journal of the Society for Medical Decision Making.

[8]  R. H. Moss,et al.  Neural network diagnosis of malignant melanoma from color images , 1994, IEEE Transactions on Biomedical Engineering.

[9]  W Abmayr,et al.  Evaluation of different image acquisition techniques for a computer vision system in the diagnosis of malignant melanoma. , 1994, Journal of the American Academy of Dermatology.

[10]  A. Green,et al.  Computer image analysis in the diagnosis of melanoma. , 1994, Journal of the American Academy of Dermatology.

[11]  R Marchesini,et al.  The invisible colours of melanoma. A telespectrophotometric diagnostic approach on pigmented skin lesions. , 1996, European journal of cancer.

[12]  W Abmayr,et al.  Improving early recognition of malignant melanomas by digital image analysis in dermatoscopy. , 1997, Studies in health technology and informatics.

[13]  P Glasziou,et al.  Comparing dichotomous screening tests when individuals negative on both tests are not verified. , 1997, Journal of clinical epidemiology.

[14]  Hugues Talbot,et al.  Automated Instrumentation for the Diagnosis of Invasive Melanoma: Image Analysis of Oil Epiluminescence Microscopy , 1997 .

[15]  P Bauer,et al.  Diagnosis of cutaneous melanoma: accuracy of a computerized image analysis system (Skin View) , 1997, 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.

[16]  H. Kittler,et al.  Epiluminescence microscopy-based classification of pigmented skin lesions using computerized image analysis and an artificial neural network , 1998, Melanoma research.

[17]  G Pellacani,et al.  Digital videomicroscopy improves diagnostic accuracy for melanoma. , 1998, Journal of the American Academy of Dermatology.

[18]  P Barbini,et al.  Digital dermoscopy analysis for the differentiation of atypical nevi and early melanoma: a new quantitative semiology. , 1999, Archives of dermatology.

[19]  S. Menzies,et al.  Automated epiluminescence microscopy: human vs machine in the diagnosis of melanoma. , 1999, Archives of dermatology.

[20]  G Pellacani,et al.  Digital videomicroscopy and image analysis with automatic classification for detection of thin melanomas. , 1999, Melanoma research.

[21]  K Wolff,et al.  Computer-aided epiluminescence microscopy of pigmented skin lesions: the value of clinical data for the classification process , 2000, Melanoma research.

[22]  Eric Lefevre,et al.  Knowledge modeling methods in the framework of evidence theory: an experimental comparison for melanoma detection , 2000, Smc 2000 conference proceedings. 2000 ieee international conference on systems, man and cybernetics. 'cybernetics evolving to systems, humans, organizations, and their complex interactions' (cat. no.0.

[23]  A Ingber,et al.  Improving screening for melanoma by measuring similarity to pre‐classified images , 2000, Melanoma research.

[24]  H. Kittler,et al.  Frequency and characteristics of enlarging common melanocytic nevi. , 2000, Archives of Dermatology.

[25]  H. Kittler,et al.  Follow-up of melanocytic skin lesions with digital epiluminescence microscopy: patterns of modifications observed in early melanoma, atypical nevi, and common nevi. , 2000, Journal of the American Academy of Dermatology.

[26]  P Bauer,et al.  Digital epiluminescence microscopy: usefulness in the differential diagnosis of cutaneous pigmentary lesions. A statistical comparison between visual and computer inspection , 2000, Melanoma research.

[27]  R Marchesini,et al.  Multispectral imaging approach in the diagnosis of cutaneous melanoma: potentiality and limits. , 2000, Physics in medicine and biology.

[28]  Philippe Schmid-Saugeon,et al.  Symmetry axis computation for almost-symmetrical and asymmetrical objects: Application to pigmented skin lesions , 2000, Medical Image Anal..

[29]  S. Menzies,et al.  Short-term digital surface microscopic monitoring of atypical or changing melanocytic lesions. , 2001, Archives of dermatology.

[30]  Harald Ganster,et al.  Automated Melanoma Recognition , 2001, IEEE Trans. Medical Imaging.

[31]  M. Oliviero,et al.  Automatic differentiation of melanoma from melanocytic nevi with multispectral digital dermoscopy: a feasibility study. , 2001, Journal of the American Academy of Dermatology.

[32]  H. Kittler,et al.  Risks and benefits of sequential imaging of melanocytic skin lesions in patients with multiple atypical nevi. , 2001, Archives of dermatology.

[33]  G. Cevenini,et al.  Differentiation between pigmented Spitz naevus and melanoma by digital dermoscopy and stepwise logistic discriminant analysis , 2001, Melanoma research.

[34]  R Hofmann-Wellenhof,et al.  Tissue counter analysis of dermatoscopic images of melanocytic skin tumours: preliminary findings , 2002, Melanoma research.

[35]  H. Kittler,et al.  Diagnostic accuracy of dermoscopy. , 2002, The Lancet. Oncology.