Interdisciplinary Initiatives Program Round 4 – 2008

Paul Keall, no longer at Stanford
Jonathan Berger, Music
Billy Loo, Radiation Oncology

4D thoracic CT is an imaging modality invented in 2003 and has since gained widespread use in radiation treatments for lung, breast and abdominal cancer patients. Despite the widespread use, a recent review of 50 consecutive 4D CT studies at Stanford shows artifacts of 4mm or greater in 46/50 patients (Yamamoto et al. IJROBP 2008)‐ clearly an alarmingly high and unacceptable failure rate in an area where millimeter accuracy is desired in tumor delineation and guided radiation delivery. Despite not being optimized from an audio, visual or patient perspective, previous independent studies have demonstrated the efficacy of simple audiovisual devices to improve respiratory regularity. Furthermore, there is a clear link between respiratory irregularity and artifacts on 4D CT images.

In this study we have developed a next generation audio‐visual biofeedback device using a representative waveform, musical guides with a beat period corresponding to the respiratory waveform period and performance feedback to engage the patient in the training process during the imaging and treatment procedures. The prototype has been built, tested and refined on volunteers. A quality assurance program has been developed and is in place. The next step is to introduce this for patient care and test the hypothesis Audiovisual biofeedback significantly improves 4D PET/CT image quality. This hypothesis test will be performed in year 2 of the study. In addition to auditory biofeedback as a means of self‐regulation, the project contributes to the broader study of the potential of auditory feedback for data representation particularly with high dimensional and noisy data with potential application in a wide range of the biological sciences. Understanding the perceptual and psychophysical nature of auditory feedback contributes both to our knowledge of human perception and cognition as well as to refining methods of numerical and data representation useful in biology research.