Breast CT for Much Earlier Detection of Breast Cancer

Institution: University of California, Davis
Investigator(s): John Boone, Ph.D. - Karen Lindfors, M.D. -
Award Cycle: 2001 (Cycle VII) Grant #: 7EB-0075 Award: $500,000
Award Type: TRC Full Research Award
Research Priorities
Detection, Prognosis and Treatment>Imaging, Biomarkers, and Molecular Pathology: improving detection and diagnosis



Initial Award Abstract (2001)
Mammography is currently used to screen asymptomatic women for breast cancer, but there is a great need to improve sensitivity and the comfort associated with this method. We are developing a breast computed tomography (CT) scanner, which we think will allow the detection of breast cancer lesions in the 3-5 mm range. This is about 1/2 to 1/3 smaller than can be seem by conventional mammography. This increased sensitivity would allow detection almost a year earlier in the course of the disease. In addition, the breast CT scanner will not require compression as mammography does, and a commercial system would require about 2 seconds per breast. Importantly, we calculate that radiation dose will be comparable or less than existing mammography examinations.

Computed tomography (CT) is used every day for scanning heads, lungs, abdomens and pelvises, however it has not been seriously considered for breast cancer screening because of long-held misconceptions that very high radiation doses would be needed. We are questioning this dogma, and propose a scanner design that will image only breast tissue. We studied breast dose using computer simulations and performed experimental measurements on cadavers, and found that radiation levels comparable to or less than mammography could be used to produce high quality CT images of the breast. We feel that breast CT has the potential of much earlier detection of breast cancer.

This project is a collaboration between a medical physicist (Boone) and a radiologist, physician (Lindfors). We propose to build a breast screening-specific CT scanner. Women will lay face down on a table, with one breast hanging through a hole in the table. An x-ray tube and detector will produce data that can be used by a computer to generate a series of CT images. The breast CT scanner will require less time for screening because each breast would be scanned in only one position, as opposed to two for mammography. No compression will be used, but some non-painful breast elongation will probably be required. The radiation dose will be up to 30% less for women with larger breasts, and will be about the same as mammography for others. Once tested for safety, dose, and image quality, the breast CT system will be used to evaluate about 50 women who (based on screening mammography) are likely to have breast cancer. This is a necessary step towards using the breast CT system for a clinical trial on a larger group of women, which is an FDA requirement for breast CT to become a clinical technique. Finally, an important aspect of our research will be to incorporate breast cancer advocates assisting us in the design and use of the instrument for maximum comfort.


Final Report (2006)
Introduction. Over the course of this grant, we have designed, fabricated, and tested a computed tomography (CT) scanner custom-tailored to imaging the female breast for the earlier detection of breast cancer. In addition to having the potential to overcome the well known super position problem of mammography (where normal breast anatomy can obscure the visualization of a tumor), the breast CT scanner allows the patient to lie comfortably on her tummy during the scan, with her breast placed in a hole on the tabletop. Importantly, breast CT does not require compression of any kind, and therefore the scan is far more comfortable than mammography.

Topic addressed. Considerable effort at designing a breast CT system which is capable of producing high resolution tomogaphy images of the breast. While conventional mammography produces only two images of the breast (the so-called cranial-caudal view and the media lateral oblique view), the data acquired by our breast CT scanner is reconstructed in a computer to provide approximately 300 tomographic images– virtual slices through the breast– with each slice only 0.5 mm thick. By imaging the breast in a series of slices, we believe that the overlying or underlying normal anatomy will be removed from the image to allow the radiologist to better visualize the presence of a breast cancer tumor, if present.

Progress towards Specific Aims. In this grant, we achieved all the aims that involved the engineering aspects and construction of our dedicated breast CT scanner. However, our plans to scan of 50 women with likely breast cancer we only partially completed, but will be achieved using subsequent grant support from the NIH. However, the support from the CBCRP was critical in allowing us to pursue the Phase II clinical trial as originally proposed. Finally, our goal to compare both qualitatively and quantitatively mammography images with the breast CT images on a patient-by-patient basis is also only partially completed. A full quantitative comparison between mammography and breast CT will require the acquisition of more patient data, and this must be delayed until 50 women are first scanned using the breast CT scanner. The Phase II trial involving scanning women who have a BlRADS 4 or 5 diagnosis based upon mammography will begin in January 2005.

Future direction. Because we have received subsequent and significant funding to continue the breast tomography project at UC Davis, we hope to accrue possibly up to 190 patients. We have almost three more years of secured funding to carry out this task.

Impact. The success or failure of breast CT is largely dependent upon how it compares with mammography, and therefore we are eagerly pursuing this objective to better understand the potential of breast CT. While we make no specific scientific claims about the superiority of breast CT until such time that the science is ready, we remain very enthusiastic and hopeful that breast CT will provide a superior methodology for breast cancer screening, and will ultimately lead to the earlier detection of breast cancer, with better prognosis for those affected by the disease.


Symposium Abstract (2003)
The use of computed tomography (CT) for breast cancer screening has for decades not been considered feasible due to unacceptably high radiation doses. To overcome this barrier, we are developing a custom-designed CT scanner for breast imaging. In our design, the woman lays flat on a table with her breast hanging comfortably through a hole in the table. This geometry requires less radiation to produce excellent images, because the women’s thorax is not exposed, as it would using a standard commercial CT scanner. In addition, the geometry of our breast CT scanner avoids compression, a common complaint with mammography.

The mechanical design of a dedicated breast CT scanner has been completed, and a scanner is under construction in our laboratory. We have evaluated a number of performance criteria; including scan position reproducibility, x-ray spectral properties, and radiation output levels. Specific to this study, we optimized the radiation dose levels that are required and compared them against doses typically employed in mammography. We have made extensive measurements of the breast diameters typical of the screening environment, and we used computer simulation to estimate the radiation dose that women would experience in breast CT. Radiation dose levels are related to image quality in CT, and thus acceptable radiation levels need to produce a diagnostically useful image quality. Using measurements of image noise performance on a clinical CT system, we have demonstrated that good image quality can be produced using radiation dose levels comparable to, or lower than, that of currently practiced two-view mammography.

Using cadaver breasts imaged on a commercial CT system, we demonstrated that our dedicated breast instrument will likely produce far more detailed information compared to mammography. Although the clinical testing of the breast CT system in our laboratory is still not complete, our initial evaluations suggest that the system has excellent clinical potential. The lack of compression and the need to scan each breast only once suggest that breast CT may be well received by the patient population. We remain enthusiastic that breast CT may provide a comfortable technique that is capable of delivering earlier breast cancer detection with radiation dose levels similar to that of currently practiced mammography.

Development and Monte Carlo Analysis of Antiscatter Grids for Mammography
Periodical:Technology in Cancer Research and Treatment
Index Medicus:
Authors: Boone JM, Makarova OV, Zyryanov VN, Tang, CM, Mancini DC, Moldovan N, Divan R
Yr: 2002 Vol: 1 Nbr: Abs: Pg:441-447

Small-animal X-ray dose from micro-CT.
Periodical:Molecular Imaging
Index Medicus: Mol Imaging
Authors: Boone JM, Velazquez O, Cherry SR.
Yr: 2004 Vol: 3 Nbr: Abs: Pg:149-158

Cassette-based digital mammography.
Periodical:Technology in Cancer Research and Treatment
Index Medicus:
Authors: Seibert JA, Boone JM, Cooper VN, Lindfors KK.
Yr: 2004 Vol: 3 Nbr: 5 Abs: Pg:413-427

Normal organ volume assessment from abdominal CT.
Periodical:Journal of Abdominal Imaging
Index Medicus: Abdom Imaging
Authors: Geraghty EM, Boone JM, McGahan JP, Jain K.
Yr: 2004 Vol: 29 Nbr: 4 Abs: Pg:482-490

A comprehensive analysis of DgN(CT) coefficients for pendant-geometry cone-beam breast computed tomography.
Periodical:Medical Physics
Index Medicus: Med Phys
Authors: Boone JM, Shah N, Nelson TR
Yr: 2004 Vol: 31 Nbr: 2 Abs: Pg:226-235

Normalized glandular dose (DgN) coefficients for arbitrary X-ray spectra in mammography: computer-fit values of Monte Carlo derived data.
Periodical:Medical Physics
Index Medicus: Med Phys
Authors: Boone JM
Yr: 2002 Vol: 29 Nbr: 5 Abs: Pg:869-875