Path B: Radiation Detectors


The first three-dimensional (3D) tomographic images were nuclear medicine images  made in 1964 by David Kuhl  using a very special dual detector NaI scanner.

Kuhl’s scanner produced cross sectional and longitudinal tomographic images nine years before Hounsfield produced his first 3D X-ray images (1973). Both obtained data by rotating a scanning imaging system around the surface of the body region of interest, imaging  emissions and transmissions respectively . Computer-based mathematical algorithms were needed to reconstruct the sensed activity distributions. Kuhl obtained images by rotating a dual opposed collimated NaI crystal scanner about the subject, collecting digital data using in-home designed electronics. The ability to image the depth dimension was revolutionary and provided information previously obtained only at operation or autopsy.
The Anger camera (first commercialized in 1962) produced planar projection images recorded on film. Anger first demonstrated visual depth cues by rotating the patient in front of the camera. Longitudinal tomography images were first produced by Anger in 1968 Anger using a system called the Pho Con.   Two opposed  gamma cameras (each with a focused collimator) made a rectilinear scan over the body. Photographic images of activity were recorded at 6 different levels in the body. This was immediately found to be useful for whole body bone and cancer imaging studies using 99mTc and 67Ga respectively. Years later, we and others connected the device to a computers and displayed multi angular views,  in addition to transverse and longitudinal section images.

Ron Jazsczak and Gird Muhlennher came from industry (Nuclear Chicago), and introduced many innovations in single photon emission as they moved to Duke and U. Penn respectively. They developed and tested new imaging geometries, image analysis and display processing methods. Industry became increasingly active in nuclear medicine after finding that the sale of expensive X-ray CT systems proved that expensive devices that did something important could be marketed profitably. (click here for Jazsczak SPECT video interview).

Semi conductor detectors were developed in the 1960s and used in the National Labs, and within 5 years they were being tested for use in nuclear medicine. Our interest followed Randy’s visit to an IEEE Semi Conductor Symposium in 1967. In 1968, Gene Johnston and Randy met with Harold Glass at the Hammersmith Hospital, London who was working with germanium detectors. On the way  back from an IAEA meeting in Salzburg where Gene presented a paper we visited Harold Glass, and Brian Pullan (the latter at Imperial College). Our primary interest was in keeping up with new imaging receptor options.  Ge (Li) detectors had the ability to reduce  scatter  and improve image quantification. Single elements were not position sensitive,  but orthogonal strip detectors were already being tested by Roy Parker, and Victor McCready for use as gamma cameras.  Bob Beck and Harold Glass were testing Ge (Li) detectors in an attempt to improve the clarity of nuclear medical images.  Brian Pullan was investigating position sensitive gas detectors  and Gene visited to look into that option.  Shortly thereafter, we attended a meeting  on “Semiconductor Detectors in the future of nuclear medicine”,  in 1970 (published later by the Society of Nuclear Medicine) which contains  early papers describing ongoing work.  It contains  two of our papers describing  imaging, dosimetry and activation analysis studies conducted using Ge (Li) detectors.

→ Detector Applications: Our first systems used NaI probes for emission and transmission measurements. As commercial solid-state detectors became available we tested their utility in collaboration with ORTEC, for use in emission, and transmission imaging, including spectroscopy applications.  Their superior energy resolution made it possible to measure elemental content, and distribution of radioactive elements with overlapping energies in tissues, including stable elements by X-ray fluorescent induced radiations.

In a collaboration with RCA, we tested novel integrated NaI probes they developed with digital electronics built into each tube base. Bill Dunn’s thesis project used RCA new research PMTs with fast plastic scintillators to test their timing resolution for TOF PET. Jim Patton ‘s tomographic research used standard NaI probes to test new cylindrical geometry tomography systems. He also evaluated a series of Si(Li) and Ge(Li) detectors, and a modular High Purity Ge array for emission and fluorescent applications. David Pickens designed and evaluated a novel orthogonal array of NaI probes for brain tomographic imaging that used standard NaI probes, when the RCA probes were not available.

Our first clinical use of a lithium-drifted germanium (Ge (Li) semiconductor detector was to measure stable iodine in the thyroid gland. With it we mapped and quantified the distribution of iodine in the thyroid by measuring the Am-241 induced fluorescent x-rays emitted from the neck. The Ge (Li) detector performed well in in-vivo emission and transmission imaging and in-vitro spectral analysis applications, i.e. analysis of complex isotope mixtures, including radioactive isotopes, and stable elements induced by XRF, and NAA (neutron activation). We also tested lithium drifted silicon Si (Li) detectors for thyroid XRF (as their energy resolution at low energies was better than obtained with Ge (Li) detectors. Our interest in X-ray fluorescence (XRF) followed work by Hoffer and Beck at the Argonne Cancer Research Hospital (ACRH). They used a Si (Li) solid-state detector and an X-ray tube excitation source  to excite and image stable iodine in the thyroid by imaging the fluorescent x-rays from iodine in the neck. We collaborated with EG&G ORTEC (an Oak Ridge Company) that produced and allowed us to test their solid state detectors. In this, project we used 241Am(as the exciting source instead of an X-ray beam). ORTEC ultimately commercialized the 241Am based system.

→ Clinical Applications: There are several valid medical, industrial and environmental applications of XRF. We used XRF to distinguish benign from malignant thyroid nodules based on the low iodine content of malignant nodules. Low levels in thyroid cancer reflect  the decreased ability of cancer tissues to trap and retain iodine (<5%) of that trapped by normal thyroid tissues. TIC also is decreased in acute stages of thyroiditis. Many patients treated with 131I become hypothyroid following 131I therapy. GS Lee found that low TIC was an early predictor of hypothyroidism 131I treated hyperthyroid patients. Evaluation of goiters in patients from endemic goiter regions can also be aided by XRF measured low thyroid iodine content (TIC). Elemental analysis of environmental samples using Si (Li) with anti coincidence guard rings provides enhanced spectral sensitivity and specificity.
A major advantage of XRF, over radioisotope imaging of the thyroid, is the extremely low dose to the patient (< 60 mrad) limited to the thyroid region. Thus, XRF can be used to evaluate thyroid function in radiosensitive populations (children and pregnant women) without the need to use radioactive iodine. The methods and findings in our studies are in a link to clinical studies. We found a diagnostically useful correlation between TIC in malignant and benign nodules with XRF buttressed by low uptake emission scan uptake when planning for surgery. For general use, we preferred the Ge (Li) because of it wider range of applications. (XRF System Details, Clinical XRF Findings)

→ Other Unexpected Potential Fluorescent Therapeutic Applications: We realized later, that one could destroy tumors if one could get sufficient 125I linked to DNA . [iodo deoxy uridine (IUDR)] into the tumor cell nuclei. Lethal doses could b delivered by the fluorescent radiations emitted when the tumor is irradiated with x-rays in the 30-35 keV region, High local energy is deposited by the short ejected range low energy x-rays and Auger electrons. This became apparent when we analyzed the relative efficiency of radiation sensitization when bromo deoxy uridine (BUDR) or iodo deoxy uridine (IUDR) are used to enhance absorbed dose in conjunction with radiation therapy. The need is for means of getting the low energy exciting beam to deep tumors The development of encapsulated x-ray sources for this purpose is an area being used in breast cancer therapy and is a topic of continued interest.

TRANSMISSION IMAGING. Differential attenuation of bone, soft tissue, and air is the basis for radiographic planar and tomographic x-ray imaging. Current studies use one or more polychromatic X-ray sources. The attenuation information derived from these studies provides both diagnostic images, and data needed to correct for attenuation and quantify tomographic x-ray images, and also to correct for attenuation in the quantification of isotope tracer images. Jim Sorenson developed the conjugate imaging method of quantifying nuclear images. He was the first to use the transmission of monochromatic beams of low energy x- and gamma rays from collimated isotope sources to measure bone mineral content in the wrist and other bones. These provide images and numerical data on regional bone mineral content based upon differential transmission of the emitted energies.

Our first transmission imaging system used a NaI detector to measure bone mineral content (BMC) in the wrist using a collimated 125I source. A number of students worked with us on this project including the first black medical student at Vanderbilt, Levi Watkins. Levi worked with Gene Johnston during the summer between his first and second year. Later, Levi went on to become Professor of Thoracic Surgery and Dean at Johns Hopkins. Obviously, a good start in research is good for one’s career. Later students compared the performance of different semi conductor detectors for this application. Later, Ron Price used the large NaI detector in the scanning whole body counter to measure the distribution of bone mineral content in the whole body using 153Gd. This predated the higher resolution images subsequently provided by commercial whole body bone mineral counting systems.

For information on BMC and Transmission Imaging, click here.


Emission imaging of tracers with energies well above 30 keV is done with either NaI or Ge (Li). The sensitivity of Si (Li) detectors is too low to be useful at energies above 30 keV. Ge (Li) detectors on the other hand can be used at high and low energies. At low energies, it can image fluorescent x-rays from iodine, and for BMC transmission imaging using Gd medium, energy sources. The Ge (Li) detectors were the only semiconductor detectors that were then capable of imaging higher energy gamma rays used in patient imaging. Later, Cadmium zinc telluride (CZT) and Cadmium telluride CdTe) detectors found to compete well with Germanium for medium and high-energy gamma imaging.

Detector size and sensitivity are strong determinants of potential utility of solid-state detectors (typically small) and NaI detectors (typically large). A comparison of 6 different tracers and detectors (Ge (Li) and NaI) was performed in dogs with spontaneous lymphoma. The different tumor-seeking radiopharmaceuticals were injected, and several days later, autopsy samples were counted to determine   absolute and relative tumor uptake in many distinct enlarged lymph nodes. The Ge (Li) detector easily resolved the different energies. However, the accuracy of its quantitation of the different amounts of the individual tracers (computer-based spectral stripping) in the tumors, was better with NaI based on NaI’s higher stopping power (greater sensitivity) and higher S/N, Note: the data on sample composition from the NaI detector required special computer spectrum stripping analyses, not required with Ge(Li). The study is further discussed in the Tumor imaging Section.

→ Comparative Studies: The efficacy of emission imaging using Ge (Li) and NaI detectors was studied and results published in 1971 (Semi Conductor Symposium (Brill, Patton, Johnston, Dyer, Baglan, Erickson, and Rolfes). The small size, < 3 cm diameter Ge (LI) detectors limited their use for human imaging where sensitivity was critical. They were very effective for in vitro applications, like activation analysis where long counting times were possible. In an attempt to overcome the size/sensitivity penalty, Fred Goulding and the Lawrence Berkeley Lab (LBL). designed an approx. 9-fold larger detector, consisting of a 3 x 3 array of 2” diameter 1 cm thick Ge (Li) detectors, housed in a common cryostat. Jim Patton was the PI on the NIH grant that funded the work. The device was built at LBL, and tested at Vanderbilt. Its stability and performance was state of the art, but even the 9-fold larger surface area of the detector, sensitivity was still too low. Phantom and patient studies favored the larger area Anger camera and we could not demonstrate a benefit for the Ge device. An additional problem with patient studies is the difficulty in correcting for patient motion when using a scanning detector. With camera type stationary imaging systems, gating can be used to correct for external motions based on continuous interspersed measurements of body surface-fixed markers. Depending on different collimation options, the device produced planar and/or longitudinal tomography images.

For more details on the 3×3 Device, VU Tomoscanner and VU Orthoscanner.

NextArrowBackArrow Back to Path B: Computing at Vanderbilt
Next to Path B: Research Funding