Nuclear magnetic resonance (NMR) is a means by which the magnetic properties of certain atomic nuclei can be observed. These nuclei are ‘polarized’ when placed in a uniform magnetic field and can then be studied through their interaction with low energy radiowaves. The signals (often referred to as resonances) collected from the sample have frequencies that depend on the applied magnetic field and are characteristic of the different nuclei; for example, hydrogen (¹H) produces a different signal from phosphorus (³¹P). In magnetic resonance imaging (MRI) the signals from the hydrogen atoms contained in tissue water and fat are spatially encoded by the use of magnetic field gradients, and are reconstructed to produce the anatomical images that have such widespread practical clinical applications. For each type of nucleus, the precise frequency of the resonance is also governed by the way in which the electrons surround the nucleus; that is, the chemical properties of a molecule containing that particular atom. We can therefore identify and quantitatively measure specific molecules in a sample. The data are collected in the form of high resolution NMR spectra, which record the frequency dependence of the resonances. The term NMR spectroscopy is still widely used in the scientific community but in the clinical context it is most commonly referred to as magnetic resonance spectroscopy (MRS).

MRS is used to study aspects of the biochemistry of selected regions of the human body by detecting and quantitatively measuring specific metabolites. For example, using phosphorus (³¹P), MRS signals from adenosine triphosphate, adenosine diphosphate, phosphocreatine, and inorganic phosphate are observable (Fig. 1) 198. In addition, the inorganic phosphate resonance gives a direct measurement of intracellular cytoplasmic pH. These parameters can be used to evaluate the relationship between energy supply and demand in tissues and organs, often referred to as ‘energy metabolism’.

Applications of MRS in surgery must still be largely considered as research. Knowledge of alterations in the cellular metabolism of diseased tissue not only increases our understanding of the disease but may be of clinical benefit. For example, a quantitative and objective measure of tissue ischaemia, when clinical examination and routine tests are equivocal, would help in deciding when tissue damage is irrevocable and may, therefore, indicate when reconstructive surgery is futile. Many diseases have been studied by MRS.

When patients with advanced peripheral vascular disease causing rest pain are studied by ³¹P MRS, the foot muscle (extensor digitorum brevis) shows marked abnormalities in phosphate-containing metabolites, the most consistent being relatively high levels of inorganic phosphate and low phosphocreatine concentrations. The changes are usually rapidly reversed by surgical restoration of blood supply. MRS measurements on this muscle provide a sensitive index of disease and a means of quantifying ischaemia. They relate more closely to the patient's symptoms than do conventional measurements of ankle pressure. In conditions that produce pain similar to that caused by vascular disease, such as arthritis of the spine, MRS clearly defines the condition causing the pain.

Patients with intermittent claudication suffer from pain and cramp in their calf muscle during exercise because blood flow is limited. Changes in calf muscle metabolism during rest and during and after exercise (plantar flexion) have shown that patients with claudication use up more muscle phosphocreatine than do control subjects. At the same time, the muscle cells of patients become more acid due to lactate production, indicating impaired oxygen supply. The rate of metabolic recovery after exercise is slower in patients with claudication, and this appears to provide the most sensitive index for objectively assessing the severity of the disease. The measurements are relatively simple, provided that the expensive instrumentation is available, can be repeated easily and, in principle could be used as part of the routine clinical evaluation prior to surgical treatment (Fig. 2) 199.

The metabolic viability of isolated, cold-preserved organs can be conveniently examined by ³¹P MRS and this has provided the basis of much work on kidney and heart preservation in animal models.

Several groups have examined the energy state of isolated human kidneys during cold ischaemia, prior to transplantation, and have shown that while the signal from adenosine triphosphate gradually decreases with time, that due to inorganic phosphate increases. Intracellular acidification is dependent on the nature of the preservation medium used. Initial function of the kidney after transplantation appears to reflect the adenosine triphosphate content of the preserved organ. While the method has not been explored further, largely because organ preservation is not a major problem in renal transplantation as practised today, it is likely that similar investigations in cardiac transplantation may be more helpful.

Preliminary reports on the MRS examination of transplanted human hearts indicate that phosphocreatine adenosine triphosphate ratio in vivo decreases during rejection. Such measurement may provide an alternative and non-invasive approach for the early detection of rejection. If these studies are confirmed, the need for regular cardiac biopsies could be eliminated.

One of the most obvious applications of MRS is to investigate the nature of solid tumours and their response to therapy. From the point of view of surgery, methods for determining tumour type, grading, and potential response to chemotherapy are important in clinical management. ³¹P MRS has provided much new information about tumours, particularly in the brain: the bioenergetic characteristics of menangiomas for example, are significantly different from those of gliomas. However, because of the large diversity of metabolic patterns in human tumours, the greatest application of ³¹P MRS may not be in characterizing tumour respiratory activity in general terms, but in identifying the intrinsic biochemical characteristics of tumour due to its particular environment and physiological state. This may help clinical decisions in specific cases. In a more general sense it appears that there are some characteristic changes in the spectroscopic signals associated with phospholipid metabolism and structures. In particular it has been suggested that the levels of the metabolites phosphoethanolamine and phosphocholine may reflect cellular growth and proliferation in some complex way. Nevertheless, up until now no clear relationships between tumour grades and metabolism have been observed.

In some recent empirical studies, proton (¹H) spectroscopy has been used to correlate tumour grade with spectral patterns. In ¹H spectroscopy of the brain the main signals are from choline-containing compounds, creatine and phosphocreatine and N-acetyl aspartate. High lactate concentrations are observed in some tumours. It remains to be established whether any of these signals could be used for tumour grading or for distinguishing between malignant and benign tumours.

Once an unequivocal diagnosis of cancer is made, pathological and clinical staging will determine decisions about appropriate treatment. No single system of staging is applicable to all cancers but several non-invasive staging procedures are routinely used. Hepatic spread is not uncommonly the initial manifestation of cancer elsewhere, yet diagnosis of hepatic metastases is often difficult in patients without advanced disease. The sensitivity of CT scanning and ultrasound for diagnosing diffuse liver involvement is poor.

Some recent studies have shown that hepatic spread of lymphoma produces biochemical changes that can be detected by ³¹P MRS of the liver: elevated phosphomonoester levels (likely to be from phosphoethanolamine) have been found in patients with hepatic infiltration. This increase in the phosphomonoester signal approximately relates to clinical stage and in several patients there was some indication that the spectroscopic measurement was more sensitive to infiltration than conventional techniques. In addition, the monoester signal appears to provide a useful marker of the response of lymphoma cells to chemotherapy and of recurrent progressive disease.

Although MRI gives direct anatomical guidance to the surgeon, MRS has unique contributions to make in patient management and therapeutic decision making prior to surgery.

Cadoux-Hudson TA, Blackledge MJ, Rajagopalan B, Taylor DJ, Radda GK. Human primary brain tumour metabolism in vivo: a phosphorus magnetic resonance spectroscopy study. Br J Cancer 1989; 60: 430–6.

den Hollander JA, et al. Potentials of quantitative image-localized human ³¹P nuclear magnetic resonance spectroscopy in the clinical evaluation of intracranial tumors. Magn Res Q, 1989; 5: 152–68.

Dixon RM, Angus PW, Rajagopalan B, Radda GK. Abnormal phosphomonoester signals in ³¹P MR spectra from patients with hepatic lymphoma. A possible marker of liver infiltration and response to chemotherapy. Br J Cancer, 1991; 63: 953–8.

Hands LJ, Bore PJ, Galloway G, Morris PJ, Radda GK. Muscle metabolism in patients with claudication investigated by P-31 NMR spectroscopy. Clin Sci, 1986; 71: 283–90.

Hands LJ, Sharif MH, Payne GS, Morris PJ, Radda GK. Muscle ischaemia in peripheral vascular disease studied by ³¹P-magnetic resonance spectroscopy. Eur J Vasc Surg, 1990; 4: 637–42.

Radda GK, Rajagopalan B, Taylor DJ. Biochemistry in vivo: an appraisal of clinical magnetic resonance spectroscopy. Magn Res Q, 1989; 5: 122–51.

Segebarth CM, Baleriaux DF, Luyten PR, den Hollandere JA. Detection of metabolic heterogeneity of human intracranial tumours in vivo by ¹H NMR spectroscopic imaging. Magn Res Med, 1990; 13: 62–76.