Prediction of individual patient survival and GBM characteristics

• Research type

  Research Study

• Full title

  Prediction of individual patient survival and glioblastoma characteristics, using routinely acquired MRI scan and other NHS data, computational modelling and machine learning.

• IRAS ID

  309957

• Contact name

  Kismet Hossain-Ibrahim

• Contact email

  kzhossainibrahim@dundee.ac.uk

• Sponsor organisation

  Tayside Academic Science Centre (TASC)

• Duration of Study in the UK

  3 years, 3 months, 30 days

• Research summary

  Review existing published Texture analysis (TA) methods to determine the most applicable for routinely acquired data in Ninewells.
  Develop new multiscale computer code for TA and combine it with machine learning to determine accuracy, sensitivity and specificity for prediction of patient survival and tumour characteristics.
  Further develop multiscale tumour growth models to take account of repeated MRI scans over time, fitting these models to NHS scan data. Determine whether this added information on tumour growth enhances the accuracy of individual patient predictions.

• REC name

  North of Scotland Research Ethics Committee 1

• REC reference

  22/NS/0021

• Date of REC Opinion

  15 Feb 2022

• REC opinion

  Favourable Opinion