In silico analysis of nuclei in glioblastoma using large-scale microscopy images improves prediction of treatment response

Annu Int Conf IEEE Eng Med Biol Soc. 2011:2011:87-90. doi: 10.1109/IEMBS.2011.6089903.

Abstract

In this paper, we present a complete and novel workflow for quantitative nuclear feature analysis of glioblastoma using high-throughput whole-slide microscopy image processing as it relates to treatment response and patient survival. With a complete suite of computer algorithms, large numbers of micro-anatomical structures, in this case nuclei, are analyzed and represented efficiently from whole-slide digitized images with numerical features. With regard to endpoints of treatment response, the computerized analysis presents a better discrimination than traditional neuropathologic review. As a result, this analysis method shows potential to facilitate a better understanding of disease progression and patients' response to therapy for glioblastoma.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Antineoplastic Agents / therapeutic use
  • Brain Neoplasms / drug therapy*
  • Brain Neoplasms / pathology*
  • Cell Nucleus / pathology*
  • Glioblastoma / drug therapy*
  • Glioblastoma / pathology*
  • Humans
  • Image Interpretation, Computer-Assisted / methods*
  • Microscopy / methods*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Treatment Outcome

Substances

  • Antineoplastic Agents