RAI Labs Duke
NSECT | GSECT | Quant CT/Tomo | Dual Energy | Chest Tomo | Breast Tomo | Breast Density
Quant. Image | Emerg. Quant. Imaging | Perf. Metrology | Clinical Trials | Emerg. Clinical

Neutron imaging for element quantification in breast and liver


Fig 1: Schematic of the NSECT acquisition system showing major components.

Pioneered at Duke, Neutron Stimulated Emission Computed Tomography (NSECT) presents a novel non-invasive approach for in-vivo quantitative imaging of element distributions in the human body.

NSECT uses a beam of fast neutrons at low intensities to stimulate gamma emission from naturally occurring elements in the body in order to generate tomographic images of localized element concentrations in a specific organ. Such information can be used to diagnose cancer, liver disease, or a number of other abnormalities characterized by element-disorders.



Fig 2: Experimental NSECT acquisition. A collimated neutron beam enters through the gray opening in the wall and exposes the sample in the center. High-purity Germanium clover detectors detect the emitted gamma spectrum corresponding to the elemental composition of the sample.


Studies have indicated that changes in trace element concentrations in human tissue may be a precursor to malignancy in several organs such as the brain, prostate and breast. Our goal is to measure these concentration changes at a very early stage of tumor development in order to diagnose cancer and differentiate between malignant and benign tissue.

Liver disorders such as sickle cell anemia, thalassemia major, hemochromatosis and Wilson’s disease are characterized by changes in concentration of iron and copper. Through a single non-invasive tomographic scan NSECT has the potential to image localized element concentrations in regions of interest within the liver without the need for a liver biopsy.

Figure 1 shows a schematic of the NSECT acquisition system with its major components. Currently, the neutron source of choice is a Van-de-Graaff accelerator with a 20 MV tandem capable of generating neutrons with energies up to 23 MeV. This source provides a high-flux monochromatic neutron beam that can be collimated down to a few millimeters. High-purity Germanium (HPGe) detectors are preferred due to their superior energy-resolution compared with scintillation detectors.

Several experiments have been performed at Duke University to demonstrate the abilities of NSECT for diagnosis of different disorders. These experiments have all been performed at the Triangle Universities Nuclear Laboratory located at Duke University.

Figure 2 shows the acquisition system used in NSECT experiments. Results from some of these key experiments are presented below.


Fig 3: Spectrum from Benign Breast showing prominently identified elements.

Fig 4: Spectrum from Malignant Breast showing prominently identified elements.
Note: Both spectra are from same patient excised through biopsy.


  Fig 5: Table showing major elemental differences between benign and malignant breast tissue.

1. Breast Cancer Diagnosis

Figure 3 shows a spectrum obtained from an excised breast-tissue specimen that was confirmed benign through biopsy. Figure 4 shows a spectrum from a malignant tumor obtained from the same patient. A list of the elemental differences detected is shown in Figure 5. The largest differences were found in Al, Br, Cl, Co, Fe, K, Rb and Zn. (Results from Kapadia et. al., "Neutron Stimulated Emission Computed Tomography for Diagnosis of Breast Cancer," IEEE Transactions on Nuclear Science, vol. 55(1), pp .501-509, 2008.)

2. Liver Iron Overload
Figure 6 shows a spectrum from bovine liver tissue with artificially induced iron overload obtained with a 0.375 mSv scan. A strong gamma-line can be observed at 847 keV, corresponding to elevated liver iron. Using this NSECT measurement, it is possible to obtain a detection sensitivity of a few milligrams per gram. (Results from Kapadia et. al., “Experimental detection of iron overload in liver through neutron stimulated emission spectroscopy,” Physics in Medicine and Biology, vol 53, pp 2633-2649, 2008.)

3. Small-animal Imaging
Figure 7 shows a spectrum from a small-animal mouse specimen. Gamma lines are visible for12C from tissue, 40Ca from bone, and elements 39K, 37Cl, 68Zn and 25Mg from the specimen, and 158Gd, 160Gd from the fixing solution. Elements 27Al, 56Fe and several isotopes in Ge originate from detector components and mounting hardware. (Results from Kapadia et. al., "Neutron Stimulated Emission Computed Tomography for Diagnosis of Breast Cancer," IEEE Transactions on Nuclear Science, vol. 55(1), pp. 501-509, 2008.)

Fig 6: Spectrum from bovine liver tissue with iron overload. The gamma-line at 847 keV corresponds to an elevated iron concentration level.
Fig 7: Spectrum from small-animal mouse specimen showing prominently identified elements.


Fig 8: Reconstruction of "N" phantom showing faithful reproduction of phantom shape for each individual element (Fe and Cu).
Fig 9: Spectrum corresponding to the "N" phantom shown above. Gamma lines are clearly identified for several states in iron and copper.
  The outer, vertical bars are copper, while the inner diagonal bars are iron. Each bar measures 7 mm in width and 6 cm in height.

4. Tomographic Imaging
Figure 8 and Figure 9 show tomographic and spectral data from a solid metal phantom of natural iron and natural copper. The two vertical orange bars are copper, while the two diagonal gray bars are iron. The reconstructed image is a faithful representation of the locations of each individual element in the phantom. (Results from Floyd et. al., "Neutron Stimulated Emission Computed Tomography of a Multi-Element Phantom," Physics in Medicine and Biology, vol. 53, pp 2313-2326, 2008.)



Fig 10: Simulation of the NSECT acquisition system in GEANT4 for liver iron overload detection. The liver is visible as a pink volume inside an elliptical torso. The spinal cord is visible as a white cylinder. The concentration of iron in the liver can be increased to any desired amount. Six germanium detectors (green and gray cylinders) are used to detect emitted gamma photons.

5. Monte-Carlo Simulations in GEANT4
In addition to the experiments performed with the Van-de-Graaff source, we have developed Monte-Carlo simulations of the NSECT acquisition system in GEANT4. These simulations are used to analyze NSECT parameters such as SNR, detection sensitivity and detection accuracy, and optimize the NSECT system for minimum patient dose. Figure 10 shows a simulated NSECT system with six HPGe detectors around a human torso with liver iron overload used in NSECT

sensitivity analysis. Using this system, detection sensitivity was estimated to be approximately 3 mg/g (i.e. 3 mg iron per gram of liver tissue).

Figure 11 shows a reconstructed tomographic image from a simulated liver with localized iron deposits within an iron-overloaded liver. The reconstructed image is a faithful representation of the phantom, showing accurate locations and concentrations for the iron deposits.

These experiments and simulations demonstrate exciting possibilities and a promising future for NSECT. Currently work is under way to develop a portable prototype scanner to perform liver-iron quantification in a clinical environment, and demonstrating iron-overload detection in a live animal specimen.

Fig 11a: Simulated model of an adult human liver with localized iron deposits. The iron concentration in the red lesion is twice the concentration of the pink lesion. The spinal cord, seen as a white hollow cylinder, contains no iron.
Fig 11b: Tomographic reconstruction of the simulated liver phantom. The reconstructed image shows accurate locations and concentrations for the iron deposits.
Radiographic Tech. | Low-Cost Tomo | Optimiz. of Rad. Ther. | Personal CAD | Adapt. Educ.