Computer program that enhances the diagnostic accuracy of positron emission tomography (PET)

Jun 20 2005

German scientists developed a computer program that enhances the diagnostic accuracy of positron emission tomography (PET) scans with Alzheimer's patients, opening the door for earlier treatment of this progressive brain disorder. Details of their findings were provided at the Society of Nuclear Medicine's 52nd Annual Meeting in Toronto.

"Positron emission tomography (PET) with FDG (fluorodeoxyglucose) in the hands of the experienced investigator is the most reliable noninvasive test to detect Alzheimer's disease in the early stage; however, reading these scans depends on the experience of the interpreter," said Peter Bartenstein, department of nuclear medicine, University Mainz, Mainz, Germany. "Our expert system enables reliable diagnosis of Alzheimer's with PET even for inexperienced physicians. For the experienced reader, it is also very helpful in difficult cases to establish the diagnosis," he added.

Alzheimer's disease gradually destroys one's memory and ability to learn, reason, make judgments, communicate and carry out daily activities. More than 4.5 million Americans suffer from Alzheimer's disease, including former U.S. President Ronald Reagan, who endured the disease for years before his death last June. An early, accurate diagnosis of Alzheimer's helps patients and their families plan for the future, and it gives them time to discuss care while the patient can still take part in making decisions.

Recent studies indicate that positron emission tomography (PET), a powerful medical imaging procedure, can supply important diagnostic information and confirm an Alzheimer's diagnosis. Conventionally, the confirmation of Alzheimer's is a long process of elimination that averages between two and three years of diagnostic and cognitive testing. Early diagnosis provides a patient access to therapies, which are more effective earlier in the disease, and allows for future planning before loss of mental capacity. "Early and reliable diagnosis of Alzheimer's disease is of paramount importance, especially with determining treatment options," said the co-author of "Use of an Expert System for the Diagnosis of Suspected Alzheimer's Disease (AD) With FDG PET."

PET uses noninvasive special imaging systems and radioactive tracers to produce pictures of the function and metabolism of the cells in the body. A radiopharmaceutical, such as FDG, which includes both sugar (glucose) and a radionuclide (a radioactive element) that gives off signals, is injected into the patient; its emissions are measured by a PET scanner.

"Our system, which was tested in 150 cases, did as well as experts in identifying Alzheimer's patients, even in the very early stages, and will help enhance the diagnostic accuracy of FDG PET," said Bartenstein. "This automated, observer-independent analytical approach for therapeutic applications will be an important tool for quality assurance in molecular/nuclear imaging," he added.

"The knowledge of several experts in the field was utilized in the development of the computer program by defining several rules used by these experts for diagnosing and excluding Alzheimer's," explained Bartenstein. Based on this knowledge, the program identifies abnormalities in the PET images and detects Alzheimer's typical changes, he said. "Besides determining the probability that a patient has Alzheimer's, it is able to identify various other dementia disorders (like frontal lobe dementia or Lewy body disease) that also show typical, but different abnormalities in the FDG-PET image," he added. This prototype research tool is not yet designed for commercial application.

Authors of "Use of an Expert System for the Diagnosis of Suspected Alzheimer's Disease (AD) With FDG PET" are Thomas Siessmeier, department of nuclear medicine, University Mainz, Mainz, Germany; Sebastian Oehm, CeBiTec/BRF, University Bielefeld, Bielefeld, Germany; Hans-Georg Buchholz, department of nuclear medicine, University Mainz, Mainz, Germany; Alexander Drzezga, department of nuclear medicine, Technical University Munich, Munich, Germany; Andreas Fellgiebel, department of psychiatry, University Mainz, Mainz, Germany; Mathias Schreckenberger, department of nuclear medicine, University Mainz, Mainz, Germany; Thomas Uthman, faculty of mathematics, University Mainz, Mainz, Germany; and Peter Bartenstein, department of nuclear medicine, University Mainz, Mainz, Germany.