A phase and frequency alignment protocol for 1H MRSI data of the prostate

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Spectroscopy

Background:

• Proton Magnetic Resonance Spectroscopic Imaging (1H MRSI) identifies prostate cancer by analyzing metabolite levels.
• Current analysis methods struggle with data artifacts like phase and frequency variations, impacting accuracy.
• Citrate resonance frequency is particularly sensitive to physiological conditions, posing challenges for peak alignment.

Purpose of the Study:

• To develop and validate a frequency and phase correction algorithm for automatic alignment of prostate MRSI spectra.
• To improve the reliability of statistical pattern recognition for prostate cancer detection.
• To enhance the accurate localization of prostate tumors using MRSI.

Main Methods:

• Developed a novel frequency and phase correction algorithm for MRSI data.
• Validated the algorithm using simulated datasets, assessing phase and frequency standard deviations for citrate resonances.
• Evaluated alignment improvement in patient data using principal component variance, signal-to-noise ratio, and cross-correlation metrics.

Main Results:

• The algorithm achieved precise alignment, with phase standard deviation of 0.095 rad and frequency standard deviation of 0.68 Hz for citrate resonances in simulated data.
• Analysis of five patient datasets showed significant improvements in spectral similarity and data quality post-alignment.
• Key parameters indicated enhanced spectral consistency, demonstrating effective phase and frequency correction.

Conclusions:

• The developed peak alignment algorithm effectively corrects phase and frequency variations in prostate MRSI spectra.
• Improved spectral alignment is expected to significantly enhance pattern recognition capabilities for prostate cancer detection.
• This advancement holds promise for more accurate and reliable localization of prostate cancer using MRSI.