Purpose

The aim of this research is to study the potential of a wavelet-based method for the analysis of nuclear magnetic Resonance Spectroscopic (MRS) signals obtained from patients with tumors.

 The Problem of Tumor Characterization Using Current Methodology

There are two basic questions which must be answered once a tumor is found:

1. Is the tumor benign or malignant?
2. Can it be categorized?

Clearly, an accurate and timely answer to these questions is crucial. If the tumor is on the surface of the human body, its diagnosis is easier.

However, for interior tumors the situation is more difficult. The first step is to detect the tumor via X-ray, CAT or MRI examination. If the tumors are large, and have distinct geometrical boundaries, then the trained radiologist can draw some conclusions. However, if the tumors are small and not clearly visible, and are subject to psychophysical perception interpretations, a sure diagnosis is difficult. Furthermore, no biochemical analysis is possible from such examinations.

Unfortunately, in almost all cases, the patient goes through the procedure of biopsy. Biopsy is an invasive surgical technique, which will provide the most reliable biochemical composition of the tumor. There are two kinds of biopsies. The first one is a biopsy involving surgery and the second is known as needle aspiration. Both kinds will create trauma and bleeding. Note that bleeding, in the case of a malignant tumor can be risky in the sense that it may propagate cancerous cells in other parts of the body, and could also favor the creation of blood clots. In the case of the surgical biopsy the samples are usually large and provide adequate data for an accurate biochemical analysis, but this is not true in the case of needle aspiration. In the needle aspiration case the samples are usually small and in many cases inadequate for an accurate and reliable diagnosis. Furthermore, due to the tumor’s location, the degrees of freedom of getting the desired sample are limited. Therefore, the need for a better method is clear.

Proposition

In this study, the analysis of FID signals is done using a powerful mathematical tool known as WAVELETS. Wavelets will provide high resolution spectra containing not only the chemical markers obtained from the currently used Fourier analysis, but also a host of other potential chemical markers that need to be identified. These new markers could be useful for diagnosis purposes and  for the evaluation of medical treatments as well. However, such study can be clinically useful only if proper tumor calibration is done first.

Note: From 1995 until now, Dr. Nick Panagiotacopulos (Dr. P.) has developed a software system which is doing exactly what we discussed earlier. Initially, the purpose of the system was to analyze FID signals related to cancer, but the system can be used also for the analysis of other medical signals such as electromyograms, electrocardiograms, encephalograms, and  cases involving Parkinson, Alzheimer’s and dementia. For example, measure dopamine levels in the brain.

Dr. P. is donating this system to FKW, and he is inviting medical research groups from anywhere in the world, to collaborate with him in the tumor calibration aspects of the study and subsequently in the FID signal analysis.

Significance

The wavelet approach it can be (for certain cases) an alternative to biopsy, and is suitable for the biochemical analysis of tumors. Furthermore, the method is non-invasive (non-surgical) and therefore convenient, safe, fast, cheap and it will provide reliable (accurate, and reproducible) results.