Fusion of Images Improves Cancer Diagnosis and Treatment

Researchers at Ohio State University have developed a technique thatfuses the images produced by nuclear medicine scans with images producedby CT scans.

The composite image combines the ability of nuclear medicine to identifythe presence of certain tumors with the ability of CT to produce detailedimages of the tissue in which the tumor resides, thereby increasing theaccuracy of both techniques.

The combination technique promises to improve the treatment of certaintumors by providing surgeons with a more precise idea of a tumor's location,size, and extent before they begin to operate.

"Both of these imaging methods are powerful tools for cancer diagnosis,but both have deficiencies," said William Bennett, assistant professorof radiology at the Arthur G. James Cancer Hospital and Research Institute.The deficiencies are reduced when the images are combined.

Combination Technique Most Useful in Problem Cases

"Nuclear medicine scans can suggest the presence of a tumor, anda CT scan can show where the tumor is located." For this reason, hesaid, "the fusion technique is most useful in problem cases wherethere's confusion about whether the nuclear medicine study is truly positive,or if there's a question about the location of the tumor."

Nuclear medicine scans can often detect malignant tumors as small as1 cm (less than 1/2 in), about the size of an individual lymph node. Objectsthat small would be difficult to recognize as cancerous by using a CT scanalone.

In a nuclear medicine scan, a tumor shows up as a bright spot on anx-ray-like image. The surrounding organs, however, are indistinct or notvisible, which is one drawback to a nuclear medicine scan when used alone.

Also, the radioactive compounds used to label tumors in nuclear medicinestudies are sometimes taken up by tissues that are not cancerous. Thisleads to false-positive test results in 10% to 15% of cases.

A CT scan of the same area of the body will show the internal organsin sharp detail, but the area of the tumor, while often visible, may notbe large enough to be considered cancerous. As a result, tumor tissue cansometimes be missed when viewing a CT scan alone, producing a false-negativetest result. False-negative CT scans occur in about 20% to 30% of cases,depending on the type of tumor and region of the body.

"But when the computer is used to color the image of the tumorin a nuclear medicine scan, and that image is combined with a CT scan ofthe same area of the body, they reveal exactly where the tumor is located,"said Bennett.

New Technique Tested in Patients With Neuroendocrine Tumors

Bennett presented the technique at the December 1996 meeting of theRadiological Society of America. He and a team of researchers tested themethod using nuclear medicine and CT scans from 35 patients who had alreadycompleted treatment. All the patients had neuroendocrine tumors, a relativelyrare type of tumor that can occur in areas of the body such as the intestines,adrenals, pancreas, and lung.

In one case involved in the study, the fused image revealed a cancerouslymph node that had been missed by the patient's original CT scan. In thisway, the new method could reduce the number of falsely positive nuclearmedicine scans and falsely negative CT scans.

"This technique might have changed the therapy in about 25% to30% of cases used for this study," said Bennett.

The method does have limitations, however. It requires that a patientreceive both a CT scan and a nuclear medicine scan. It also takes additionaltime and expertise to fuse the two images. Moreover, both scans must includean identical landmark that the computer can use to match the two imagescorrectly. The agent used for the nuclear medicine scan in this study wastaken up by the kidneys, liver, and spleen, which are also imaged by theCT scan.

"We told the computer to use the kidneys, liver, or spleen as landmarksin both sets of scans," said Bennett. The computer then figured outhow everything in the two images relates to the landmarks."

This work was done by researchers at Arthur G. James Cancer Center inconjunction with software developers at Picker International, Cleveland,Ohio, who produce nuclear medicine imagers and CT scanners.