They come to the Johns Hopkins Medical Center from all over the world -- scared-looking children with misshapen heads and faces that make you want to avert your eyes at the same time you want to comfort them.
It is as if the hand of their Creator slipped in shaping their tiny skulls. In one picture, a little girl's skull balloons out on one side. In another, a little boy's bulging right eye seems to dribble down his cheek, inches lower than the left.
To most people, these children have horrible deformities. To Dr. Craig A. Vander Kolk and other plastic surgeons at Hopkins, they have "cranio-facial anomalies."
His job is to fix them -- to create a nose where none existed, to give a brain room to grow, to fashion a face that can go out in public without ridicule from other children.
Sculpting a child's face and skull is a delicate job because the artists' medium is flesh and bone -- growing constantly but less predictably than in luckier children. Surgery that would ideally be science is often educated guesswork.
To improve their odds of operating at the right time and in the right way, a small team of physicians, engineers and other scientists at Hopkins is enlisting the imaging and storage capabilities of powerful computers to create a vast data base of these cranio-facial anomalies.
The goal is to generate "maps" of these little skulls, not just as they are but as they likely will be in six months, a year, two years.
Dr. Vander Kolk's hope is that in the future he and other surgeons will be able to plan and practice complicated surgery before they ever touch a scalpel, using advanced computer simulations and three-dimensional imaging technology.
"If there were some way to understand which areas were growing and which areas weren't growing, we'd have a better idea of what we want to do," says Dr. Vander Kolk. The surgeon, who is also an associate professor at Hopkins medical school, is working with Dr. Ben Carson, director of pediatric neurosurgery, to coordinate the project.
The stakes are especially high when the patients come from some Third World countries, where a child with a malformed skull or facial tumor faces ostracism or worse. "In some places, they would even be killed because there is a genetic role," Dr. Vander Kolk says.
The doctor flashes a color slide across a projection screen. It shows a child of about 2 who had never been able to chew, talk or fully open his mouth because of a severely receded chin with a fused jaw joint.
The surgery to correct the condition involved cutting out the fused joint and moving the jaw into its correct position. Then the surgeon had to reconstruct the bones around the joint using a graft from the child's ribs, bolting the graft to the existing bone with titanium screws.
In this case, the operation was successful, letting the boy eat solid food and talk for the first time, Dr. Vander Kolk says. But such surgery could be more precise, and less reliant on educated guesswork, with three-dimensional simulations of future growth, he adds.
"If I can get more information at my fingertips before I go in, it can be shorter, it can be safer for the patient, and it can cost less," says Dr. Vander Kolk.
The person in charge of gathering the information to put at Dr. Vander Kolk's fingertips works out of a small, cluttered laboratory filled with human and animal skulls in the Hopkins physiology building.
Joan Richtsmaier is an anthropologist by training, and the skulls are not morbid decorations but the tools of her trade. Trying to explain a detail of anatomy to a visitor, she casually picks up a monkey skull and points to a groove in the bone where a nerve would run.
Dr. Richtsmaier, a professor of anatomy and plastic surgery but not a physician, hopes to draw on medical archives all over the country to find useful radiological images of deformed and normal children's skulls.
Her objective is to assemble a data base of images of children's skulls -- electronically sorted by race, sex, age and the type of anomaly the surgeon must deal with.
Expanding the data base
As of July, Hopkins had compiled 219 digitized CT scans of cranio-facial anomaly cases in its data bases, Dr. Richtsmaier says. By the end of the summer, she expects to have 500-600 cases stored in the system's memory banks.
Next year, she hopes to have enough data to generate the first surgically useful imaging based on cases in the data base. She hopes to have eventually tens of thousands of cases stored in electronic files. The more images of a particular anomaly that can included in the mathematical model, the more accurate the resulting "map" can be.
"We'd like to have hundreds of cases for everything," she says.
Converting the immense electronic storehouse of data Dr. Richtsmaier is accumulating into useful images is a monumental task, however. Hopkins has the mathematical models it needs, she says, but it doesn't have the visualization software to create three-dimensional images.
That's where Diana Hauser comes in. When she's not teaching mechanical engineering at Hopkins' Homewood campus, she is writing the computer programs that will turn ones and zeros in the data bank into vivid three-dimensional projections that can be turned and viewed from any angle. Since coming aboard the project in January, she's written 64,000 lines of software, with millions more to come.
Maintaining such a vast data base and generating 3-D images would be a pipe dream for Hopkins if it weren't for the vast increases in computer storage capacity and data-processing speed in recent years.
The technological firepower Hopkins' Center for Biomedical Visualization is devoting to the project is as close to state of the art as you can get in the fast-changing world of computing.
Most of Dr. Hauser's work is done on a $6,000 computer work station made by Silicon Graphics Inc. But for the heavy data-crunching, she uses a $150,000 Reality Engine II -- the same machine used to create the special effects in the movie "Jurassic Park."
Dr. Richtsmaier's ambitions for the data base go far beyond Hopkins.
'A national resource'
To increase the number of cases that can be cataloged, Hopkins and Bell Atlantic Corp. have installed a sophisticated fiber-optic phone link to University Hospital, which has agreed to transmit radiological images from its cases to the Hopkins data bank.
But Dr. Richtsmaier sees that as only the beginning. Eventually she would like to see other hospitals contributing radiological data and tapping into the bank to generate their own images through electronic connections.
"We hope this will be a national resource," she says.
The radiological data storage and imaging system Hopkins is developing could be applied to other fields of health care, researchers say. Dr. Hauser cited health care for the aged as a particularly promising application.
"The relationship we discover for growing tissue can also be applied to deteriorating tissue," she says. "Right now, it's traditional to do repetitive surgeries for any kind of deteriorating bone tissue. . . . This will reduce the number of operations."
Those possibilities fit in neatly with Dr. Hauser's long-term interest in designing "smart" prostheses that change with the body as it ages. For instance, she believes that the imaging technology she is developing today could be used to help create an artificial hip that "ages" along with its recipient and never needs replacement.
Dr. Hauser says she believes we are just entering an era in which medicine, engineering, mathematics and computer science converge to change how health care is practiced.
"I'm just very pleased to be at the gateway at Hopkins where all these things are coming together," she says.