Nuclear Fractal Dimensions as a Tool for Prognostication of Oral Squamous Cell Carcinoma

Abstract

Background: Carcinogenesis follows complex molecular alterations, which are triggered by subtle chromatin architectural changes that are imperceptible to the human eye. As the treatment decisions in Oral Squamous Cell Carcinoma (OSCC) are hindered by the imprecise clinical stage determination and inter-observer variability in histological grading, focus in recent years has shifted to discovering identifiers related to neoplastic cell morphology studied through computer-aided image analysis. One such approach is the assessment of fractal geometry, a technique first described by Mandelbrot, which aids in precise assessment of architecture of natural objects. Assessment and quantification of degree of complexity of these fractal objects (self-similarities in structural complexity at different magnifying scales) is described as fractal dimension (FD). Aim: To evaluate the nuclear fractal dimension (NFD) in OSCC using computer-aided image analysis. Materials and Methods: Histological sections of 14 selected cases of Oral Squamous Cell Carcinoma (OSCC) and 6 samples of normal buccal mucosa (as control) were stained with Haematoxylin-Eosin and Feulgen stain for histopathological examination and evaluation of nuclear complexity respectively. Fifteen HPF at Invasive Tumour Front (ITF) and Tumour Proper (TP) of Feulgen-stained sections were selected and photographed in test and control samples. At ITF, TP and normal buccal mucosa 200 nuclei each were selected and analyzed using Image J software to quantify FD. The test and control groups were compared statistically using Independent sample t-test and One-way ANOVA. Results: Nuclear FD increased progressively towards worst tumour staging as compared to normal buccal mucosa. Conclusion: Nuclear FD can be considered for quantification of nuclear architectural changes as a prognostic indicator in OSCC.

Keywords

• image analysis
• prognostic indicator
• sarkar box counting method