Sep 17, 2013
Neutrons identify kidney cancer
Renal cell carcinoma (RCC) can be distinguished from benign kidney tumours using a non-invasive 3D imaging modality called neutron stimulated emission computed tomography (NSECT), according to simulations by researchers in North Carolina (Phys. Med. Biol. 58 5867).
The technique detects characteristic gamma rays emitted by trace elements specific to different types of cancer upon irradiation of the patient or a tissue sample with a neutron beam. Still in the pre-clinical stages of development, NSECT has the potential to detect disease earlier than conventional, anatomical imaging approaches.
"Some of these elements are suspected to undergo changes in concentration in the earliest stages of tumour growth," said Anuj Kapadia, senior author and faculty member at Duke University in Durham, NC. "Our technique aims to measure these elements and detect cancer in the body earlier than other clinical imaging methods in use today."
While conventional imaging techniques like MRI or CT can detect renal tumours, they cannot always differentiate between benign and malignant disease. The only definitive way to do so is by biopsy, an invasive procedure that carries complication risks, motivating the researchers to investigate whether NSECT could non-invasively identify RCC in vivo.
Monte Carlo simulations
The researchers used a mathematical human phantom and the MCNP5 Monte Carlo code to simulate NSECT scans. Organs were assigned elemental compositions using data from the literature, including one kidney containing a RCC tumour, which was assigned a composition based on samples taken from 70 patients with the disease.
The modelled scanner comprised two semi-circular detector rings positioned side by side and a 5 MeV collimated neutron source. Each detector ring comprised five high-purity germanium radiation detectors. The beam energy was selected to induce gamma emission in the elements specifically associated with RCC.
During imaging simulation, the detectors and neutron source were rotated around the phantom, measuring gamma emission spectra at 10 angles and at 60 discrete source positions at each angle. The process was repeated for 10 longitudinal slices.
Characteristic spectral peaks resulting from the stable isotopes of six elements were separated out and used to reconstruct tomographic images for each, using the expectation-maximization algorithm, where the pixel value is proportional to the concentration of the element. The elements included phosphorus, present in RCC in higher concentrations; sodium and potassium, which are present in healthy kidneys in higher concentrations than other organs; and sulphur, a background element present throughout the body.
Calculations revealed an effective dose to the patient of 3.86 mSv for the scan, less than half of that for an abdominal CT scan. Images of phosphorus, potassium and sodium distributions, each superimposed on a sulphur image to provide an outline of the body, demonstrated qualitative differences between the diseased and healthy kidney. The images demonstrated higher levels of phosphorus and lower levels of potassium and sodium in the diseased kidney, consistent with the modelled compositions, and correctly mapped the distribution of healthy and diseased tissue.
The researchers used the data from the three images to isolate the kidneys from the surrounding anatomy, enabling a quantitative comparison of the two. For each of the three elements, the differences in pixel values between the healthy and diseased kidney were statistically significant, though not as large as the differences in element concentration predicted. The shortfalls were attributed to the low concentrations of the elements that were close to the detection sensitivity limit of the modelled NSECT system. For example, on average, potassium makes up only 0.005% of an RCC tumour.
"Due to this fact, we were able to identify the difference between element concentrations across tissues, but we couldn't perform an accurate absolute measure," said Rodrigo Viana, first author and PhD candidate at the Nuclear Energy Research Institute in São Paulo, Brazil, who was a research fellow at Duke University at the time of the study. The group have several ongoing lines of research, including the design of a clinical specification scanner that will improve detector sensitivity with the aim of routine absolute concentration measurement in patients.
• Related articles in PMB
3D element imaging using NSECT for the detection of renal cancer: a simulation study in MCNP
R S Viana et al Phys. Med. Biol. 58 5867
Sensitivity analysis for liver iron measurement through neutron stimulated emission computed tomography: a Monte Carlo study in GEANT4
G A Agasthya et al Phys. Med. Biol. 57 113
Experimental detection of iron overload in liver through neutron stimulated emission spectroscopy
A J Kapadia et al Phys. Med. Biol. 53 2633
About the author
Jude Dineley is a freelance science writer and former medical physicist based in Sydney, Australia.