This Article Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Koh, D-M Articles by Padhani, A R Search for Related Content PubMed PubMed Citation Articles by Koh, D-M Articles by Padhani, A R

Diffusion-weighted MRI: a new functional clinical technique for tumour imaging

¹Department of Radiology, Royal Marsden Hospital, Downs Road, Sutton, Surrey SM2 9PT, ²Mount Vernon Hospital, The Paul Strickland Scanner Centre, Rickmansworth Road, Northwood, Middlesex HA6 2RN, UK

One of the key aims of oncological imaging is to differentiate between malignant and non-malignant tissues at all stages of the patient's cancer care. Accurate staging and precise delineation of the extent of malignancy influences therapeutic decisions, therapy outcomes and, ultimately, patient prognosis.

Conventional imaging using ultrasound, CT or MRI detects cancer by identifying anatomical distortion or altered tissue appearances. Tumour tissue conspicuity may be increased after the administration of intravenous contrast medium, thus enhancing detection and delineation. However, identification of small volume active tumour, either at presentation or at early disease relapse remains challenging because small volume disease may not result in detectable structural or morphological change on conventional imaging. Furthermore, the effects of therapy and complications thereof may obscure or mimic recurrent disease.

Functional imaging techniques using CT, MRI and positron emission tomography (PET) are increasingly being applied to the evaluation of tumours. These techniques exploit as their contrast mechanism unique pathophysiological changes that occur within tumours; such as altered blood flow, increased glucose metabolism, hypoxia and cellularity. Such functional techniques are increasingly used for tumour detection, for the monitoring of treatment response and to detect relapsed disease. Clinical experience has shown that functional techniques have their own unique strengths and limitations.

A new, emerging functional technique that is now finding a role in cancer imaging is diffusion-weighted MRI (DWI or DW-MRI), which produces information about tissue cellularity and the integrity of cellular membranes. This technique may not be well appreciated by general radiologists. DWI can be performed on most modern MRI machines with relative ease, in a short period of time and without the need for contrast medium administration. The potential for this technique to evaluate the larynx for tumour recurrence after prior radiotherapy is demonstrated in a short communication from Vandecaveye et al in this issue [1].

At a fundamental level, DWI provides information on the random (Brownian) motion of water molecules in tissues. The Brownian displacements of millions of water molecules over time are normally distributed with a mean final value of zero for all time periods measured, but with a standard deviation that is proportional to the diffusion coefficient and time measured. This was the basis for Einstein's diffusion equation published in 1905, which subsequently helped to earn him the 1921 Physics Nobel Prize.

In tissues, DWI probes the movement of water molecules, which occurs largely in the extracellular space. However, the movement of water molecules in the extracellular space is not entirely free, but is modified by interactions with hydrophobic cellular membranes and macromolecules. Hence, diffusion in biological tissue is often referred to as "apparent diffusion". By comparing differences in the apparent diffusion between tissues, tissue characterization becomes possible. For example, a tumour would exhibit more restricted apparent diffusion compared with a cyst because intact cellular membranes in a tumour would hinder the free movement of water molecules.

One of the simplest methods of obtaining DWI images is to apply pairs of opposing and balanced magnetic field gradients (but of differing durations and amplitudes) around a spin-echo refocusing pulse of a T₂ weighted sequence. Stationary water molecules are unaffected by the paired gradients, and thus retain their signal. Non-stationary water molecules acquire phase information from the first gradient, but are not rephased by the second gradient, leading to an overall loss of the MR signal. The signal reduction on the DWI image is proportional to the amount of diffusion water motion occurring during the pulse sequence. Hence, on DWI, there is usually less signal attenuation (i.e. higher signal intensity) of tumour compared with normal tissue due to the restricted diffusion of water molecules in tumours, which is presumed to be due to an increased cellular density.

At some anatomical locations (e.g. the brain), DWI is usually performed in three or more gradient directions because of the unequal limitations to diffusion in some directions imposed by tissue organization (e.g. white matter tracts). This phenomenon is termed anisotropy and may be also observed in some visceral organs. For example, in the normal prostate gland, diffusion is greater along the line of the ducts than across the ducts. However, anisotropy is usually not seen in tumours since cancers typically grow in a disorganized fashion.

The degree of diffusion-weighting applied is indicated by the b-value (measured in s mm^–2), which indicates the magnitude and duration of the applied gradients and time between the paired gradients. By varying the amplitudes, lengths and intervals between the diffusion gradients, the sensitivity to the degree of diffusion motion can be altered and the data processed to provide information about actual diffusion distances. Hence, DWI using a larger b-value (e.g. b = 500 s mm^–2) is more sensitive to the slower motion of water molecules and smaller diffusion distances, whereas the converse is true with a smaller b-value (e.g. b = 50 s mm^–2). It is important to remember that the phenomenon of water molecule movement detected by DWI occurs at a length scale (typically micrometres) that is significantly larger than intracellular distances, but significantly smaller than the pixel dimensions of typical MR images (typically millimetres).

When performing DWI cancer studies, images are typically acquired using different b-values (typically 0–1000 s mm^–2). The images obtained at different b-values allow the calculation of the apparent diffusion coefficient (ADC; unit mm² s^–1), which is usually presented as a quantitative parametric map. From the discussions above, it is not surprising that on ADC maps, tumours usually demonstrate low ADC values and appear as low signal intensity area compared with normal tissue. This appearance is the inverse of that observed on "raw" DWI images obtained at high b-values. The quantitative ADC values can aid in lesion characterization, and can also be applied to evaluate the treatment response of tumours.

Diffusion-weighted MRI is an established tool for the evaluation of intracranial diseases. The technique has been applied successfully to detect early cerebral infarction and for the characterization of brain tumours. More recently, DWI has been used to demonstrate early response of brain tumours to radiation treatment. However, motion-related artefacts, which degrade image quality, have limited the clinical application of DWI to extracranial sites. These motion-related artefacts can now be substantially reduced by the use of parallel imaging, combined with breath-hold, single-shot, echo-planar MRI techniques. Such techniques are useful for the evaluation of the larynx, where respiratory and swallowing movement, together with susceptibility effects from air within the larynx, can significantly degrade image quality.

In this issue of BJR, Vandecaveye et al describe their experience of using DWI for the detection of post-treatment recurrent laryngeal tumours in a small number of patients [1]. All cases were validated by histopathology. Thin section (48 slices at 4 mm thickness) axial DWI of the larynx was performed using six b-values, which ranged from b = 0 s mm^–2 to 1000 s mm^–2. Although parallel imaging was not employed, each DWI study was completed in less than 6 min. The four examples presented elegantly demonstrate the potential role of DWI in distinguishing tumour from post-treatment change. Furthermore, the exquisite radiological-pathological comparison enhances our understanding of the pathological basis for their imaging appearances at DWI. In the study, two cases with recurrent tumour were detected at DWI as focal areas of restricted diffusion returning low ADC values. By comparison, two cases demonstrated focal asymmetry of the larynx due to inflammatory change. These areas returned high ADC values at DWI and were due to laryngeal necrosis and oedema at histopathology. Intriguingly, one case with inflammatory change was also evaluated using ¹⁸FDG-PET imaging, which revealed a moderately hypermetabolic focus in the larynx (false-positive). As the PET imaging was performed within 6 months of radiotherapy, this case emphasises the potential pitfall of using ¹⁸FDG-PET imaging to distinguish between tumour and inflammation in the early post-treatment period.

There are challenges to the use of DWI for tumour evaluation. First, the dichotomy of identifying tumour versus post-treatment change may not always be straightforward. There can be substantial overlap in the ADC values between malignant and non-malignant tissue making it difficult to determine disease status. For example, a predominantly necrotic tumour may potentially be confused with necrosis arising from radiation treatment since both would result in higher ADC values. Thus, as with all functional imaging techniques, DWI image information should be interpreted with information from conventional imaging to improve disease assessment. Second, the averaged ADC values derived from regions of interest drawn around tumours may not sufficiently characterize tumour heterogeneity. More sophisticated methods of analysis are needed to adequately account for regional variations, which is particularly important for the evaluation of treatment response in tumours. Third, the analysis of quantitative ADC on commercial platforms lacks standardization. Some commercial software does not allow drawing of free forms to encompass tumour regions. Most commercial software also does not allow image registration and noise filtration, which can significantly impact on the quality of the quantitative data. Clearly, further collaborative work in this area would be welcomed.

Despite some of the above limitations, DWI is emerging as a powerful, new diagnostic tool which will be increasingly applied to the evaluation of tumours, as has been demonstrated in the accompanying paper by Vandecaveye et al [1]. Potential applications include distinguishing tumour from non-tumour tissue, assessing of treatment response and for the prediction of treatment outcome. As the examination is quick and can be conveniently incorporated into existing protocols, assessment of its role in everyday clinical practice could be expediently achieved.

Received for publication March 30, 2006. Revision received May 19, 2006. Accepted for publication May 30, 2006.

	References

Top References

 

1. Vandecaveye V, De Keyzer F, Poorten VV, Deraedt K, Alaerts H, Landuyt W, et al. Evaluation of the larynx for tumour recurrence by diffusion-weighted MRI after radiotherapy: initial experience in four cases. Br J Radiol 2006;79:681–7.[Abstract/Free Full Text]

This article has been cited by other articles:

				S. H. Kim, J. M. Lee, S. H. Hong, G. H. Kim, J. Y. Lee, J. K. Han, and B. I. Choi Locally Advanced Rectal Cancer: Added Value of Diffusion-weighted MR Imaging in the Evaluation of Tumor Response to Neoadjuvant Chemo- and Radiation Therapy Radiology, October 1, 2009; 253(1): 116 - 125. [Abstract] [Full Text] [PDF]

				Y. Ohba, H. Nomori, T. Mori, K. Ikeda, H. Shibata, H. Kobayashi, S. Shiraishi, and K. Katahira Is diffusion-weighted magnetic resonance imaging superior to positron emission tomography with fludeoxyglucose F 18 in imaging non-small cell lung cancer? J. Thorac. Cardiovasc. Surg., August 1, 2009; 138(2): 439 - 445. [Abstract] [Full Text] [PDF]

				Y. Cui, X.-P. Zhang, Y.-S. Sun, L. Tang, and L. Shen Apparent Diffusion Coefficient: Potential Imaging Biomarker for Prediction and Early Detection of Response to Chemotherapy in Hepatic Metastases Radiology, September 1, 2008; 248(3): 894 - 900. [Abstract] [Full Text] [PDF]

				BJR review of the year - 2006 Br. J. Radiol., March 1, 2007; 80(951): 147 - 151. [Full Text] [PDF]

This Article Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Koh, D-M Articles by Padhani, A R Search for Related Content PubMed PubMed Citation Articles by Koh, D-M Articles by Padhani, A R