IMAGE ANALYSIS 111 caused by the presence of tiniea versicolor on the subject's back. The clinical expert grader was able to discern this condition and interpret accordingly. The IA system, however, confused the dermatological problem with horny impactions. The above prob- lems can be addressed by a brief inspection of the slides under a microscope prior to IA processing. Finally, irritation of the stratum corneum may cause problems when evaluating FB slides. In our study, sample C (acetylated lanolin alcohol) caused an increase in uplifted scaly material on the stratum corneum surface. When evaluated by IA, this sample gave higher than expected numbers for number of objects and object area, approximately double those of the other samples. The IA task we used is more sensitive to this type of background noise and may also be helpful in detecting irritation during a human comedogenicity study. CONCLUSION Image analysis has been used to quantify follicular biopsy slides obtained during a human comedogenicity study. Results indicate that there is a linear correlation between IA data and the interpretive score. Image analysis data is reproducible over time, in that the system is non-biased. This method of evaluation is more sensitive to irritation on samples than previous methods. IA can reduce errors during the evaluation because it is semiautomated. Central data interpretation of studies conducted at multiple investiga- tor sites can also be achieved through the use of digital image analysis. The comput- erized nature of this method allows for mathematical analysis of results and easy data storage. REFERENCES (1) O. H. Mills and A.M. Kligman, A human model for assessing comedogenic substances, Arch. Dermatol., 118, 903-905 (1982). (2) O. H. Mills and A.M. Kligman, Follicular biopsy, Dermatologica, 167, 57-63 (1980). (3) O. H. Mills et al., "Assessing Comedogenicity: Current and Future Trends," in Clinical Safety and Efficacy Testing of Cosmetics, W. C. Wagginer, Ed. (Marcel Dekker, New York, 1990), pp. 83-91. (4) A. M. Kligman and T. Kuong, An improved rabbit ear model for assessing comedogenic substances, Br. J. Dermatol., 100, 699-702 (1979). (5) A. Awajan, D. Rondot, and J. Mignot, Quick method of measuring the furrows distribution on skin surface replicas, Med. Biol. Eng. Comput., 379-389 (July 1989). (6) G. L. Grove, M. J. Grove, and J. J. Reydon, Optical profilemetry: An objective method for quan- tiffcation of facial wrinkles, J. Am. Acad. Dermatol., 213, 631-637 (1989). (7) R. Marks, Methods for the assessment of the effects of topical retinoic acid in photo-aging and actinic keratosis, J. Int. Med. Res., 18, 29C-34C (1990). (8) M. Monto and R. Caputo, Chemical efficacy and patient tolerance of topical trentinoin therapy in photo-aging, J. Int. Med. Res., 18, 35C-40C (1990). (9) J. Bhawan et al., Effects of trentinoin on photodamaged skin, Arch. Dermatol., 127, 666-672 (1991). (10) G. L. Grove et al., Skin replica analysis of photodamaged skin after therapy with trentinoin emollient cream, J. Am. Acad. Dermatol., 25, 231-237 (1991). (11) T. J. Flotto et al., A computerized image analysis method for measuring elastic tissue, J. Invest. Dermatol., 93, 358-362 (1989).

112 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS (12) (13) (14) (15) (16) (17) D. T. Woodley et al., Treatment of photoaged skin with topical tretinoin increases epidermal-dermal anchoring fibrils, J.A.M.A., 263, 3057-3059 (1990). C. R. Roquet and M. Kermici, A method for measuring the various constituents of the human hair follicle, J. Microsc., 156, 115-123 (1981). V. A. Moss, D. M. Jenkinson, and H. Y. Elde, Automated image segmentation and serial section reconstruction in microscopy, J. Microsc., 158, 187-196 (1990). G. Sauermann, B. Ebens, and U. Hoppe, Analysis of facial comedones by porphyrin fluorescence and image analysis, J. Toxicol.-Cut. Ocular Toxicol., 8, 369-385 (1989/1990). R. R. Anderson, Polarized light examination and photography of the skin, Arch. Dermatol., 127 (July 1991). D. Halliday and R. Resnick, Physics (John Wiley & Sons, New York, 1978), pp. 1069-1073.