Topodermatographic image analysis for melanoma screening and the quantitative assessment of tumor dimension parameters of the skin

Abstract

Background. The clinical need to identify and evaluate changes of cutaneous lesions in melanoma screening or follow‐up of patients with cancer is of paramount importance. Because skin‐lesion changes may be small and numerous, clinical assessment alone does not meet the requirements of quantitative assessment. Using the computer as a diagnostic tool for the image analysis of sequentially captured skin surface images has resulted in the technical problem of insufficient registration reproducibility. This paper describes the technical logistics, setup procedure, and clinical evaluation of the novel technique termed “topodermatography,” which performs the quantitative videographic image analysis of skin‐lesion changes over time. Methods. Digitized measurements of skin‐surface image parameters were performed using a high‐speed processor with an onboard coprocessor, a high‐resolution video camera, specifically designed image processing software, and a position framework for the adjustment of the patient's standing position. The topodermatographic image analysis was performed on 109 consecutive patients who were at risk for melanoma (N = 98), had lesions from Kaposi's sarcoma (N = 4), had metastatic skin deposits from melanoma (N = 3), and had breast cancer (N = 4). Results. Skin lesion changes over time could be identified reliably within a few millimeters of diametric enlargement. In this series, a 0.51% early melanoma detection rate was assessed in 19 of 98 patients followed for 12 months. By monitoring manifest neoplastic skin lesions, tumor growth kinetics were analyzed quantitatively to determine the total area of skin involvement, thus facilitating precise response assessment. Conclusions. Topodermatographic image analysis helps to optimize screening and follow‐up procedures for patients with melanoma and populations at risk for melanoma. In addition, metastatic tumor lesions on the skin can be monitored dynamically, facilitating the accurate evaluation of the impact of systemic therapy on multiple skin deposits from melanoma and nonmelanoma cancers. Cancer 1995;75:981‐8.

A Prelimary Approach for the Automated Recognition of Malignant Melanoma

“…Of course, shaving the hairs before imaging sessions may be a solution. However, this solution not only adds extra costs and time to the image session, but also is uncomfortable and impractical especially for multiple lesions or total-body nevus imaging (Voight and ClaBen, 1995). Hence, in spite of its algorithmic complexity, the preprocessing technique called DullRazor (Lee et al, 1997) appeared to us as the best solution, from a practical point of view, for thick hair removal.…”

Automated Detection of New or Evolving Melanocytic Lesions Using a 3D Body Model

“…Segmentation/detection approaches usually identify a set of lesion candidates using simple image processing methods [7,[9][10][11] and then filter the results using unsupervised [7,9] or supervised [10] classification. Matching lesions in images taken at different times is challenging and approaches take many forms.…”

“…Voigt and Classen [11] perform both segmentation and registration. Images of the patient's front and back torso are acquired with a single camera and a positioning framework for adjusting the patient's pose.…”

Counting Moles Automatically From Back Images

“…However, these methods were designed to work with images with exactly one mole in the middle of the image surrounded by skin background, and these methods cannot be applied without modifications to our study, which attempts to identify all moles in a back torso. Currently, only a few methods have been described to extract multiple moles from images of large anatomic sites [14], [15]. In [14], the moles in the front and the back torso are detected by thresholding the output of a Sobel operation, which highlights the border of the moles.…”

“…Currently, only a few methods have been described to extract multiple moles from images of large anatomic sites [14], [15]. In [14], the moles in the front and the back torso are detected by thresholding the output of a Sobel operation, which highlights the border of the moles. However, a global threshold value is difficult, if not impossible, to obtain.…”