Effect of Flip Angle Error for T 1

  • Slides: 15
Download presentation
Effect of Flip Angle Error for T 1 Bias Correction in MRI Fat Fraction

Effect of Flip Angle Error for T 1 Bias Correction in MRI Fat Fraction Estimation March 22, 2011 Issac Yang Supervisor: Charles Mc. Kenzie, Ph. D Department of Medical Biophysics University of Western Ontario

Background and Motivation • Non-Alcoholic Fatty Liver Disease (NAFLD) is becoming the most common

Background and Motivation • Non-Alcoholic Fatty Liver Disease (NAFLD) is becoming the most common chronic liver disease – Affects 30% of the Western population • Fat depositing in liver cells (steatosis) • Associated with metabolic syndromes – Obesity, type II diabetes Normal Fatty Liver

Background and Motivation • NAFLD is asymptomatic • Current procedure of detection is liver

Background and Motivation • NAFLD is asymptomatic • Current procedure of detection is liver biopsy – Invasive procedure • Complications involves minor bleeding, hospitalization, and (rarely) death – Steatosis distribution is inhomogeneous • Prone to sampling error • MRI offers a non-invasive alternative – Unaffected by sampling error

Obtaining Fat Fractions Fat + Water • An MRI technique called IDEAL was used

Obtaining Fat Fractions Fat + Water • An MRI technique called IDEAL was used – Allows production of separate fat and water signal images. IDEAL Fat Fraction = + Water Fat

Bias in Fat Fraction • Signals in each pixel is characterized by the equation

Bias in Fat Fraction • Signals in each pixel is characterized by the equation • T 1 values for fat is significantly different from T 1 of water – Typically: T 1 f=382 ms, T 1 w=809 ms at 3. 0 T in vivo • Difference in T 1 values cause unequal weighting of water and fat signals in fat fraction equation

Current Method • T 1 weighting in the signal increases with flip angle •

Current Method • T 1 weighting in the signal increases with flip angle • Current method of mitigating bias is to use a low flip angle. – Decreases signal intensity – Reduced Signal to Noise Ratio (SNR)

Bias Correction • Bias can be reduced by performing corrections to the image by

Bias Correction • Bias can be reduced by performing corrections to the image by using estimated T 1 values. • This method assumes known and uniform flip angle — Untrue for fields of 3. 0 T and above due to RF pulse inhomogeneity — Flip angle measurements should be performed to further reduce error

Method T 1 f Estimate Fat Image Bias Correction Corrected Fat Image Fat Fraction

Method T 1 f Estimate Fat Image Bias Correction Corrected Fat Image Fat Fraction Map T 1 w Estimate Water Image Flip Angle Map Bias Correction Corrected Water Image

Results • Flip angle measurements show true flip angle to be 9 degrees on

Results • Flip angle measurements show true flip angle to be 9 degrees on average – Prescribed 12 degrees – 25% lower than expected • T 1 measurements yielded: – T 1 w=274 ms, T 1 f=181 ms – Estimated values: T 1 w=350 ms, T 1 f=200 ms Flip Angle Map

Correction Results

Correction Results

Discussion • Measured fat fractions was brought closer to the true fat fraction value

Discussion • Measured fat fractions was brought closer to the true fat fraction value for phantoms

Conclusion • Experiments suggest that including flip angle mapping into imaging protocol would decrease

Conclusion • Experiments suggest that including flip angle mapping into imaging protocol would decrease fat fraction bias

Acknowledgement Mc. Kenzie Lab Dr. Charles Mc. Kenzie Dr. Trevor Wade Dr. Lanette Friesen-Waldner

Acknowledgement Mc. Kenzie Lab Dr. Charles Mc. Kenzie Dr. Trevor Wade Dr. Lanette Friesen-Waldner Bartha Lab Dr. Robert Bartha Dr. John Drozd Jacob Penner Curtis Wiens Bryan Addeman Yifan Cui Samantha Flood

Thank You For Your Attention Questions?

Thank You For Your Attention Questions?