Multimodal Brain Tumor Segmentation

Gliomas are the most frequent primary brain tumors in adults, presumably originating from glial cells and infiltrating the surrounding tissues [1]. Despite considerable advances in glioma research, patient diagnosis remains poor. The clinical population with the more aggressive form of the disease, classified as high-grade gliomas, have a median survival rate of two years or less and require immediate treatment [2], [3]. The slower growing low-grade variants, such as low-grade astrocytomas or oligodendrogliomas, come with a life expectancy of several years so aggressive treatment is often delayed as long as possible. For both groups, intensive neuroimaging protocols are used before and after treatment to evaluate the progression of the disease and the success of a chosen treatment strategy. In current clinical routine, as well as in clinical studies, the resulting images are evaluated either based on qualitative criteria only (indicating, for example, the presence of characteristic hyper-intense tissue appearance in contrast-enhanced T1-weighted MRI), or by relying on such rudimentary quantitative measures as the largest diameter visible from axial images of the lesion [4], [5].

By replacing the current basic assessments with highly accurate and reproducible measurements of the relevant tumor substructures, image processing routines that can automatically analyze brain tumor scans would be of enormous potential value for improved diagnosis, treatment planning, and follow-up of individual patients.  Segmenting brain tumours from multi-modal imaging data is one of the most challenging tasks in medical image analysis due to their unpredictable appearance and shape.

However, developing automated brain tumor segmentation techniques is technically challenging, because lesion areas are only defined through intensity changes that are relative to surrounding normal tissue, and even manual segmentations by expert raters show significant variations when intensity gradients between adjacent structures are smooth or obscured by partial voluming or bias field artifacts. Furthermore, tumor structures vary considerably across patients in terms of size, extension, and localization.


  • Students with a major in computer science, software engineering, image processing, pattern classification, artificial intelligence, or a related area in the final stage of master level studies are invited to contribute
  • Affinity with programming is required
  • Familiarity with Matlab or C++ / Java / C#, (OpenCV, ITK Toolkit)


  • Project duration: TBD
  • Supervision: Dr. Muhammad Moazam Fraz
  • Location: School of Electrical Engineering and Computer Science (SEECS), National University of Sciences and Technology (NUST), H-12, Islamabad, Pakistan.

For more information please contact moazam dot fraz at seecs dot edu dot pk



[1] E. C. Holland, “Progenitor cells and glioma formation,” Current Opinion in Neurology, vol. 14, pp. 683–688, 2001.
[2] H. Ohgaki and P. Kleihues, “Population-based studies on incidence, survival rates, and genetic alterations in astrocytic and oligodendroglial gliomas.” J Neuropathol Exp Neurol, vol. 64, no. 6, pp. 479–489, Jun 2005.
[3] D. H. Louis, H. Ohgaki, O. D. Wiestler, and W. K. Cavanee, “WHO classification of tumours of the central nervous system,” WHO/IARC.,
Lyon, France, Tech. Rep., 2007.
[4] E. Eisenhauer, P. Therasse, J. Bogaerts, L. Schwartz, D. Sargent, R. Ford, J. Dancey, S. Arbuck, S. Gwyther, M. Mooney, L. Rubinstein, L. Shankar, L. Dodd, R. Kaplan, D. Lacombe, and J. Verweij, “New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1),” European Journal of Cancer, vol. 45, no. 2, pp. 228–247, 2009.
[5] P. Y. Wen, D. R. Macdonald, D. a. Reardon, T. F. Cloughesy, a. G. Sorensen, E. Galanis, J. Degroot, W. Wick, M. R. Gilbert, A. B. Lassman, C. Tsien, T. Mikkelsen, E. T. Wong, M. C. Chamberlain, R. Stupp, K. R. Lamborn, M. a. Vogelbaum, M. J. van den Bent, and S. M. Chang, “Updated response assessment criteria for highgrade gliomas: response assessment in neuro-oncology working group.” Journal of clinical oncology : official journal of the American Society of Clinical Oncology, vol. 28, pp. 1963–72, 2010.

Posted in Medical Image Analysis, Projects.