Multi-modal 3D Simulation Makes the Impossible Possible.

B.Y. was born full term after a large vertex encephalocele was diagnosed prenatally. The unique challenge to repairing B.Y.'s encephalocele was a microcephalic skull and large proportion of likely functional extracranial brain tissue, which would need to be preserved. At Boston Children's Hospital, a simulation-based collaborative presurgical planning and rehearsal process, using both digital and 3D printed models, enabled successful technical completion and outcome of an otherwise inoperable case.

B.Y. was prenatally diagnosed with a large vertex encephalocele. When he was born at full term, the complexity of the encephalocele was considered too high to operate. Although he was sent home with hospice care, he continued to thrive. His parents reported that other than his unwieldy head shape, which kept him immobile on his back, he behaved just like their older son.

B.Y.’s pediatrician recommended another opinion at Boston Children’s Hospital, where he presented for evaluation at 5 months of age, weighing 6.85 kg. He had limited vision, but otherwise no seizures or other medical issues. On examination, he had a large encephalocele with good quality overlying skin and no leakage of cerebrospinal fluid. His head circumference was under the second percentile, with 20–30% of the brain contents and a large amount of fluid herniated into the encephalocele. Magnetic resonance imaging showed a complex lesion, containing what appeared to be functional brain tissue.

The microcephaly and large amount of extracranial brain posed a unique challenge to reconstruction. Skull expansion would be required to accommodate the possibly functional brain tissue. Neurosurgery and plastic surgery expertise were combined with an on-site hospital-based simulator program—including simulation methodology, imaging, segmentation, mechanical engineering, 3D printing and design—to enable detailed surgical preplanning necessary to design, test, and practice multiple options without posing any harm to the patient. Although 3D models have been valuable tools in craniofacial procedures for decades,[1-15] the unique collaboration between surgeons and a simulator program that could produce individualized 3D renderings provided a novel iterative process of physical and digital modeling to help make an otherwise impossible case possible.

SIMULATION METHODOLOGY

Models Created

An initial set of 3D digital models were created from segmentation of the head computed tomography study using Materialise Mimics (Materialise, Leuven Belgium) software. These models included B.Y.’s skull, brain parenchyma, and dura, and were printed using polymer-jet 3D printers to facilitate simulation of the surgical procedure (Fig. 1). Using volumetric analysis of the 3D model, the intracranial volume was estimated at 531 cc, and the extracranial brain volume was 105 cc; thus, an additional 100 cc of volume would need to be created to reposition the brain back into the skull (Fig. 1).

Fig. 1.

Preoperative 3D models showing fluid and solid components of the encephalocele (A) and just the extracranial brain (B).

Initial Presurgical Revision (Physical Model)

Vertical osteotomies arranged over the parietal-occipital region were marked and cut to expand the volume (Fig. 2). The cranial defect at the apex was expected to fill in with time, as has been demonstrated in previous pediatric cranial reconstruction.[16-20] After making these cuts on the model, the segments were “out-fractured,” carefully bending them outward in a radial pattern.

Fig. 2.

3D model for surgical preplanning with proposed osteotomies.

Digital Revision and Volumetric Calculation

After the revised models were rescanned using CT, the resulting volumetric imaging was resegmented and digital volumetric models produced (Fig. 3). The out-fractured parietal-occipital segments were isolated from the anterior skull and positioned to estimate the resulting intracranial volume. To increase volume further, 2 more osteotomies were created virtually along the parietal bones. Several other reconstructions were created in software, to roughly correlate rotational position of the barrel staves with resulting increases in intracranial volume.

Fig. 3.

Digital manipulation of model with additional osteotomies laterally (A) and volumetric analysis of digital model (B).

INTRAOPERATIVE DETAILS

The preoperative planning process facilitated an efficient approach in the operating room. Following anesthetic preparation, he was repositioned safely into the prone position. The encephalocele was drained of approximately 300 ml of fluid. Through a coronal incision, the encephalocele was dissected free and the preplanned posterior osteotomies were created. A segment of brain tissue emanating from the right occipital area was independent of other attachments and therefore this completely disorganized nonfunctional brain tissue was excised. The remainder of the brain tissue was reducible into the expanded cranium. Due to the lack of normal brain architecture, there were no definable ventricles into which to place an external ventricular drain or shunt.

The out-fractured posterior segments demonstrated a tendency to collapse back together, thus 2 of the small parietal segments were removed and secured over the encephalocele defect to prevent reherniation.

As a rare and complex case, there is no average procedure time for comparison, but the 5-hour operative time was a perceived reduction in time, attributed to the presurgical planning and rehearsal. Blood loss was also minimized by the efficient approach; B.Y. was hemodynamically stable throughout, had a calculated blood loss of 125 ml and received 75 ml of packed red blood cells during the procedure. He required no further transfusion postoperatively. He was extubated and recovered in the intensive care unit.

POSTOPERATIVE COURSE

B.Y. recovered well initially, but after 48 hours, he became lethargic. Repeat imaging showed hydrocephalus, and he was taken back to the operating room for insertion of an external ventricular drain, which was later converted to a permanent shunt. His neurologic status improved back to baseline. He continues to take levetiracetam for seizure prophylaxis. B.Y. has been followed for 1 year without further postoperative complications. The potential risks of seizure, stroke, devastating neurologic injury, or death that had been discussed preoperatively were avoided (Fig. 4).

Fig. 4.

The patient at 2 months postoperative.

LIMITATIONS AND FURTHER FRONTIERS

Though informative, computer/screen-based digital planning remains limited by lack of physicality, highlighting the importance of a complimentary 3D physical model. In our hands, the 3D print replica adds “feel,” haptic and mechanical information. For instance, the initially designed osteotomies were technically reasonable and accessible within digital design but also required inspection of the resulting 3D print, to confirm they were actually safe and technically feasible in the actual patient. Additionally, unanticipated forces, such as supine positioning, yielded motion of the posterior osteotomized segments, which was otherwise unpredictable by a virtual model alone. As an additional “view,” future procedures may incorporate finite element analysis to predict movement and pressure effects from patient positioning.

CONCLUSIONS

New suites of simulation-based approaches offer novel paradigms of surgical preplanning allowing for risk-free discovery, rehearsal, and preparation of innovative procedures without peril to patients. Collaboration of surgeons with simulator experts can convert cases once deemed too risky to endeavor into novel, life-saving procedures offered to new populations of patients.