Divining Death's Cause
Pathologists are embracing new no-touch "virtual autopsies," with 3-D images that reveal valuable clues
The mysterious death of a 3-week-old Swedish infant in early 2004 could have been an episode of CSI. The baby girl died while a defect in her vocal cords was being examined with a laryngoscope, and an autopsy failed to explain her death. But finely detailed 3-D images from a "virtual autopsy"-a CT scan performed prior to the usual kind-held the answer. The images showed a tiny perforation in the baby's voice box where a needle on the laryngoscope had gone astray. Air had rushed into her chest, collapsing her left lung. As doctors tried to relieve the pressure on the lung, the needle punctured the sac around her heart. Air leaked in. Her heart stopped.
The Pentagon has employed this emerging technology since late 2004 on the remains of military members in Iraq and Afghanistan, to determine how helmets and body armor can better protect against bullets and bombs. Forensic investigators in Sweden and Switzerland have put virtual autopsy to use in cases such as the one above, and specialists are flocking to conferences in Australia and Europe to learn the technique. Medical examiners in the United States are also beginning to show interest. A new mortuary for the Maryland medical examiner is slated to include the ability to perform virtual autopsies.
Peer power. Virtual autopsy exploits the ability of a modern CT scanner to create images of a body, intact or battered, without physically invading it. (Sometimes other forms of scanning, such as magnetic resonance imaging, are combined with CT.) In as little as 10 minutes, data representing thin X-ray slices of the body are reconstructed by a powerful computer into crisp, detailed images of bone and tissue. Pathologists and radiologists can zero in on a fractured skull like the one above, deciphering the pattern to determine how the blow arrived and exactly how death occurred. Images can be sized up and down and turned at various angles, providing instant flexibility not afforded by conventional autopsy. Nor can a physical autopsy show the path of a bullet at a glance, as virtual autopsy can.
Virtual autopsy is now automatic in LinkÃÂöping, where the baby girl was examined. That is also where the technique is being refined at the Center for Medical Image Science and Visualization at LinkÃÂöping University. It and the Center for Forensic Imaging and Virtopsy at the University of Bern are the two focal points of the latest work. Last week, CMIV Director Anders Persson described a case he had reviewed with police only a few minutes before. A 14-year-old girl had been murdered. Virtual autopsy showed that she had been struck in the head with great force. The fracture pattern indicated that the weapon had been both large and heavy, which let investigators quickly eliminate other possibilities for the murder weapon. "In a normal case they would have had the results only after 24 hours," says Persson.
The Pentagon is mum about what it has learned from performing virtual autopsies on some 1,700 bodies that have passed through the Dover Air Force Base mortuary in Delaware, the receiving point for all military casualties re-entering the country. "One of the things we are doing is providing information about head injuries," says Armed Forces Medical Examiner Craig Mallak, the ranking military pathologist. He lauds the new tool for what it reveals and also for improving the efficiency of physical autopsies, which are still performed. "Previously we could spend an hour or more looking for a projectile fragment," he says. "Now the 3-D view shows us exactly where to look."
In a decade or two, hands-on autopsies will be gone-replaced by the virtual version, says CFIV Director Michael Thali. At least in Europe, it is beginning to be integrated into the teaching curriculum. Sweden's CMIV has a 15-meter, high-resolution screen for that purpose. "We have stopped using ordinary autopsy as a training tool," says Persson, because the clear, precise images on the huge screen are far more instructive.
Autopsy comeback? The technology's boosters think it also might substitute, if not perfectly, for hospital autopsies, which have all but disappeared. One reason for the dive in autopsy rates from about 50 percent in the 1960s to an average of 5 percent now is that since 1970, hospitals no longer have to perform a minimum rate of autopsies to remain accredited. Another is physicians' widespread belief that autopsies tell them nothing about diagnosing and treating patients that sophisticated monitoring, imaging, and lab tests don't provide. And at a cost of several thousand dollars per autopsy, few hospitals are eager to do lots of them.
Accreditation and expense aside, autopsies have repeatedly demonstrated that doctors are often wrong about the reason a patient died. One pathology journal reported in 2002 that half the autopsy reports reviewed at one teaching hospital reflected at least one misdiagnosis, and the majority of the errors were so bad that at least some of the patients might have lived had they been diagnosed correctly. "If autopsy findings were used broadly across multiple doctors and specialties, one could learn from others' problems," says George Lundberg, former editor of the Journal of the American Medical Association and now editor-in-chief of the professional site eMedicine.com. A pathologist who frequently and vigorously laments the demise of the autopsy, Lundberg thinks virtual autopsies could be useful. "A good pathologist," he says, "has always used other modalities besides cutting and feeling."
Even if lessons learned from autopsies do not always prevent deaths, "families need to know," says Kim Collins, director of medical and forensic autopsy at the Medical University of South Carolina in Charleston. "There are a lot of diseases you die with, but not necessarily from-heart disease and cancer are two. The autopsy findings could foretell a family's medical future." Yet doctors often don't ask families for permission to perform an autopsy, and most families don't know they can request one.
The benefits can extend beyond a few doctors and families, says Collins. Now 42 years old, she was diagnosed with high-grade breast cancer in 2003, and had a bilateral mastectomy followed by chemotherapy and radiation. "So far I am doing great," she says. "But if I die of breast cancer, I would like for my doctors to be able to study the cancer cells." The findings could be used by researchers, and adding virtual autopsy, says Collins, "would be super."
This story appears in the September 11, 2006 print edition of U.S. News & World Report.