Warning! OUTDATED BROWSER DETECTED!   Please update your browser immediately for a better experience on this website. Learn More
  • 3-D Scoliotic Spine Model Aids Pre-Surgical Planning in 8-Year-Old Girl

    Radiologists were part of a team at Mt. Sinai Hospital who used a 3-D printed model to plan and guide surgery on an 8-year-old girl. By Elizabeth Gardner

    April 1, 2017

    A 3-D model of an 8-year-old girl’s scoliotic spine proved so helpful in pre-surgical planning that surgeons used it in the operating room to help guide a complex — and ultimately successful — multi-stage procedure.

    The patient, an orphan from Armenia — a country lacking such advanced medical care — had a severe rotatory kyphoscoliosis, multi-level malsegmentation of the vertebrae and ribs and Type I diastematomyelia, or “split cord syndrome.”

    While routine 3-D reconstruction could not adequately display all of the anomalies, diagnostic radiologist Javin Schefflein, MD, outlined production methods for 3-D printed models created at New York’s Mount Sinai Hospital during an RSNA 2016 presentation. Researchers worked closely with Mount Sinai’s Neurosurgical Simulation Core.

    “We contacted the neurosurgery team who were excited at the prospect of generating a precise physical model to help visualize the pathology and plan surgery,” - Dr. Schefflein said.

    Onsite 3-D printing can be a boon for numerous medical applications, but producing complex models needs to be a group effort among radiologists, engineers, surgeons and computer scientists.

    “The collaborative nature of this endeavor cannot be overstated,” Dr. Schefflein said. “Each team member contributes to every pre-operative 3-D printing project we work on. The uses for this technology are boundless, and every time we have added a different discipline to our modeling collective, a new purpose has emerged.”

    The model was used to plan a two-stage surgery involving T12-L2 laminectomy, resection of the midline bony spur at L1, intradural exploration to de-tether the spinal cord, asymmetric pedicle subtraction osteotomy at T1-L1 to straighten the curvatures, and long-segment posterior fusion with instrumentation from T2-L5. The model was also used during the surgery to help surgeons visualize steps in the procedure.

    In Mount Sinai’s onsite dedicated 3-D printing lab, the first step was to obtain CT images of the full spinal column and proximal ribs. Initial seeding for the segmentation was completed via high-contrast thresholding of the image. A connected component growth model with origins from the seed mask completed the rough mask of all the bony components.

    The model was refined using a low-propagation level-set model. Geometry-preserving Taubin smoothing followed by quadratic edge collapse decimation yielded the final model, which was printed life size with a gypsum powder-based 3-D printer. The finished model exhibited weight and texture very similar to bone.

    Final Model a Joint Effort

    Even with advanced software, some human intervention was needed to tweak the instructions for producing the final model, Dr. Schefflein said.

    “Our neuroradiologists worked hand-in-hand with the neurosurgery department to define what should be included in the print, which was then explained to the computer engineering arm of the modeling group,” he said.

    Surgeons planned the procedure by physically rotating the printed model to see the interconnections among fused ribs, fused vertebrae and anterior and posterior attachments of the bone spur, as well as the relationships of all the spinal curves to the plane of the pelvis.

    As the child underwent the two-stage surgery, a member of the surgical staff held up and manipulated the model so the surgeon could “visualize” the portions of the spinal anatomy that weren’t visible at a given point in the procedure.

    The operation, according to Dr. Schefflein, was a complete success. In terms of creating the model itself, the process took more than 10 hours including scanning (10 minutes), segmenting (three hours), printing (five hours) and drying/hardening time (two to three hours), at a cost of about $710.

    “The materials and labor were cheaper than we expected, though the start-up cost for accurate modeling can be daunting,” Dr. Schefflein said, adding that Mount Sinai’s printer alone cost about $60,000.

    Finding a workable payment policy is key, Dr. Schefflein said. So far, surgery teams pay for the models generated at Mount Sinai, which is not a sustainable option.

    Paradoxically, there are already billing codes covering models produced by outside contractors. Dr. Schefflein urges radiologists to press for a code for in-house models. “It is not a drastic change,” he said.


    Pre-operative computer-generated 3-D segmentation of severe scoliosis created by the Mount Sinai Hospital Neurosurgery Simulation Core.

    Post-operative computer-generated 3-D segmentation following two-stage corrective surgery, depicting spinal alignment of long-segment posterior fusion.Pre- and post-operative images courtesy of the Mount Sinai Hospital Department of Radiology and Dr. Anthony Costa, head of the Mount Sinai Neurosurgery Simulation Core.

    3-D printed severe thoracolumbar scoliosis with multilevel malsegmentations and Type I diastematomyelia for use in pre-surgical planning and intraoperative visualization. Image courtesy of the Mount Sinai Hospital Department of Radiology and the Rapid Prototyping Center.