Results obtained by using a computerized image analysis system designed as an aid to diagnosis of cutaneous melanoma

N. Cascinelli, M. Ferrario, R. Bufalino, S. Zurrida, V. Galimberti, L. Mascheroni, C. Bartoli, C. Clemente

Research output: Contribution to journalArticlepeer-review

Abstract

Results obtained using a computerized image analysis system as an aid to clinical diagnosis of melanoma are reported. The system comprises a colour television camera connected through a digitizing board to a 386 personal computer. By means of original algorithms able to measure the shape, the colours and texture of a pigmented lesion of the skin, the system provides eight on/off indicators that are matched with the histological diagnosis to identify benign and malignant pigmented lesions. The chances that a given lesion is malignant increase with the increasing number of positive indicators. The training field of the system was constituted of images and data of 169 cutaneous lesions in 165 patients. Taking two positive indicators as the threshold between pigmented benign and malignant lesions, the efficiency of the system is 0.98, the positive predictive value is 0.45 and the negative predictive value is 0.95. These values were confirmed in a series of 44 pigmented lesions, 10 of which were melanoma, that constitute the present test series. The authors conclude that this computerized image analysis system should be regarded as a useful aid to diagnosis for a non-expert clinician. The system limit is transformation within a naevus.

Original languageEnglish
Pages (from-to)163-170
Number of pages8
JournalMelanoma Research
Volume2
Issue number3
Publication statusPublished - 1992

Keywords

  • Computerized image analysis
  • Cutaneous pigmented lesions
  • Diagnosis
  • Melanoma

ASJC Scopus subject areas

  • Cancer Research
  • Dermatology
  • Oncology

Fingerprint

Dive into the research topics of 'Results obtained by using a computerized image analysis system designed as an aid to diagnosis of cutaneous melanoma'. Together they form a unique fingerprint.

Cite this