Prognostic value of automatically extracted nuclear morphometric features in whole slide images of male breast cancer

摘要: Numerous studies have shown the prognostic significance of nuclear morphometry in breast cancer patients. Wide acceptance morphometric methods has, however, been hampered by tedious and time consuming nature manual segmentation nuclei lack equipment for high throughput digitization slides. Recently, whole slide imaging became more affordable widely available, making fully digital pathology archives feasible. In this study, we employ an automatic algorithm to extract features related size analyze their value male cancer. The study population comprised 101 patients whom survival data was available (median follow-up 5.7 years). Automatic performed on digitized tissue microarray slides, each patient, mean area standard deviation were calculated. univariate analysis, a significant difference found between with low (P=0.022), while atypia score did not provide value. Cox regression, had independent additional (P=0.032) tumor tubule formation. conclusion, present method its application prognosis. automatically extracted proved be indicator. With increasing availability scanning labs, these kinds quantitative approaches can easily integrated workflow routine practice.