Surgeons Enlist Computer’s Help in Solving the Three-Dimensional Puzzles of Broken Bones

Thaddeus P. Thomas, Donald D. Anderson, Andrew R. Willis, Matthew C. Frank,
J. Lawrence Marsh, and Thomas D. Brown

Comminuted (badly fragmented) bone fractures can be crippling. Putting the pieces back together presents a great challenge for orthopaedic surgeons. A new technique, three-dimensional puzzle solving, has the promise of allowing computers to help make it easier for surgeons to put badly fragment bones back together.

Comminuted fractures almost always occur as a result of high-energy trauma, whether from vehicular accidents, falls from a height, or battlefield events. After acute stabilization of the injured limb, surgeons attempt to restore bone structure and function. The special demands presented by comminuted articular (joint) fractures are an aspect of treating these injuries that has received relatively little attention. Here, an essential concern is avoiding post-traumatic osteoarthritis (OA). The physical impairment caused by OA of a single lower extremity joint is equivalent to that caused by end-stage kidney disease and congestive heart failure.1 Improved treatment of these severe injuries offers major benefits, including earlier return to activity, and reduced lifelong morbidity and disability.

Given the high risk of post-traumatic OA and the severe impairment when it occurs, forestalling joint degeneration is a major concern in managing high-energy articular fractures. Even with satisfactory wound healing and osseous union, joint degeneration is a frequent and disabling complication. As a point of reference, for axial compressive pilon fractures of the distal tibial articular surface, the incidence of post-traumatic OA of the ankle is in the range of 60 to 80%.2 OA development following an articular fracture has been attributed to the initial injury to the joint 3 and to elevated cartilage stresses from residual surface incongruity.4 The majority opinion (by far) is that elevated articular contact stresses associated with residual surface incongruity most strongly predispose to joint degeneration.

For these reasons better methods of precisely restoring articular surface congruity and bone structure are clearly justified. Unfortunately, such effort is hindered by the complexities of the highly disrupted anatomy, and by challenges associated with the surgical reconstruction. Once the acute soft tissue injury has stabilized, some form of definitive surgical reduction is required in most of these cases. Precise anatomical restoration unfortunately often comes at the expense of additional insult to already damaged soft tissues. Given the massive tissue bed insult from the original trauma, aggressive surgical exposure to facilitate precise bony reconstruction carries the risk of limb-threatening devitalization.6 Less invasive surgery is preferred, but with current technology, suitable fracture reduction is difficult to attain. In practice, perfectly congruous reduction is seldom achieved, particularly so for higher-energy fractures with appreciable comminution and fragment dispersion.

In orthopaedic trauma surgery, computer-aided fracture reduction has been pursued, with most efforts to date aimed at surgical navigation.7,8 Pre-operative planning has received less attention. Most planning is done in a two-dimensional radiographic environment. Existing pre-operative planning platforms for more complex bone fractures are site-specific, and they rely almost exclusively upon the surgeon, rather than upon any computational approach, to perform the time-consuming virtual reconstruction. Notably lacking to date has been any objective assessment of, or guidance in crafting, a fracture reduction/fixation plan.   

Three-dimensional (3D)
puzzle solving

Relevant new technologies have recently become available for planning articular fracture reduction. We term the first of these technologies computational 3D puzzle solving (Figure 1). 9-13 Interactive computer algorithms are used to solve the geometric problem of assembling arbitrary numbers of fragments from broken objects. The computational puzzle solver takes as its input the 3D geometry of each fragment. It then reconstructs the bone by aligning fracture surfaces and fragment boundaries. This approach requires minimal user interaction to carry out the alignments to an intact template, while allowing focused user interaction at critical decision points later in the reconstruction planning process.


Figure 1. The 3D puzzle solving software provides a virtual environment to pre-operatively plan comminuted fracture reconstruction by matching to an intact template of the fractured bone.


In partnership with investigators at the University of North Carolina at Charlotte, a team of investigators at the University of Iowa Department of Orthopaedics and Rehabilitation have adapted computational 3D puzzle solving methods for use with comminuted bone fragments.11-13 The approach aligns CT-derived surface models of fragments to an intact template, generated from an uninjured limb, to successively piece the bone back together (Figure 2). The original anatomy is restored by matching fragment native (periosteal and subchondral) bone surfaces. We have taken this technology to the point of computationally obtaining clinical fracture cases (Figure 3). The puzzle solving algorithm can accommodate incomplete fragment sets and bone compaction, common in high-energy extremity fractures. These capabilities open the door to target any remaining defects in the bone for augmentation with bone graft or other substituting materials (Figure 4).

Figure 2. 3D puzzle solving approach shown for a tibial plafond fracture case. Fragments were aligned to an intact template, derived from the contralateral limb.

fig3 fig4 

Figure 3. Ten high energy tibial plafond fracture cases, visualized by 3D CT, are shown before and after virtually reconstruction using puzzle solving methods.

Figure 4. Four 3D puzzle solutions are compared to what was achieved surgically. Bone compaction and defects can result in collapse and shortening.

Shape-machined structural bone defect fillers

Working with collaborators at Iowa State University, our group has been developing methods for the rapid manufacturing of patient-specific bone implants using rapid prototyping processes14-17. The geometry of segmental bone defects resulting from traumatic injury can be reverse-engineered from medical images (such as CT scans), and then accurate defect fillers can be automatically generated in advanced biomaterials and other bioactive/biocompatible materials. We have utilized a subtractive RP process in which defect fillers are machined from any of a variety of materials (up to and including allograft bone). This work uniquely enables the rapid manufacturing of implant fillers with several key characteristics, including suitable bio-compatible materials, pre-drilled fixation screw/wire hole geometries, and custom surface characteristics on specified patches of the filler geometry. Among the benefits of this approach is the ability to create accurate filler geometries that may improve initial fixation strength and stability through accurate mating geometry.

In a test of clinical practicality, a recent fracture case was puzzle solved prospectively for the first time. A 19 year old male presented with an open distal tibia fracture, sustained after jumping from a moving train (Figure 5). In addition to substantial articular fragmentation, the diaphysis was severely comminuted. For initial treatment, the fracture was spanned with an external fixator and antibiotic beads were implanted, followed by a CT scan. An accurate puzzle solution was obtained within a four hour time span, and it was given to the surgeon one week prior to definitive surgery. Successfully completing this process in a reasonable time period provided confidence in the feasibility of this approach.


Figure 5. These two cases show how 3D puzzle solving can be used to plan for a case, and where needed, devise strategies for dealing with segmental bone defects.


Figure 6.
Puzzle Solving Video


References Cited

  • Saltzman CL, Zimmerman MB, O'Rourke M, et al. Impact of comorbidities on the measurement of health in patients with ankle osteoarthritis. J Bone Joint Surg Am. 88(11):2366-2372, 2006.
  • Marsh JL, Weigel DP, Dirschl DR. Tibial plafond fractures. How do these ankles function over time? J Bone Joint Surg Am. 85-A(2):287-295, 2003.
  • Buckwalter JA, Brown TD. Joint injury, repair, and remodeling: roles in post-traumatic osteoarthritis. Clin Orthop Relat Res. 423:7-16, 2004.
  • Anderson DD, Van Hofwegen C, Marsh JL, Brown TD. Is elevated contact stress predictive of post-traumatic osteoarthritis for imprecisely reduced tibial plafond fractures? J Orthop Res. 2010 Jul 6. [Epub ahead of print].
  • Dirschl DR, Marsh JL, Buckwalter JA, et al. Articular fractures. J Am Acad Orthop Surg. 12:416-423, 2004.
  • Wyrsch B, McFerran MA, McAndrew M, et al. Operative treatment of fractures of the tibial plafond. A randomized, prospective study. J Bone Joint Surg Am. 78(11):1646-1657, 1996.
  • Hüfner T, Pohlemann T, Tarte S, et al. Computer-assisted fracture reduction of pelvic ring fractures: an in vitro study. Clin Orthop Relat Res. 399:231-239, 2002.
  • Schmucki D, Gebhard F, Grützner PA, et al. Computer aided reduction and imaging. Injury. 35 Suppl 1:S-A96-104, 2004.
  • Bhandarkar SM, Chowdhury AS, Tang Y, et al. Computer vision guided virtual craniofacial reconstruction. Comput Med Imaging Graph. 31(6):418-427, 2007.
  • Chowdhury AS, Bhandarkar SM, Robinson RW, Yu JC. Virtual multi-fracture craniofacial reconstruction using computer vision and graph matching. Comput Med Imaging Graph. 33(5):333-342, 2009.
  • Thomas TP, Anderson DD, Willis AR, et al. A computational/experimental platform for investigating three-dimensional puzzle solving of comminuted articular fractures. Comput Methods Biomech Biomed Engin. (accepted, in press), 2010.
  • Willis A, Anderson D, Thomas T, Brown T, Marsh JL. 3D reconstruction of highly fragmented bone fractures. In: Pluim JPW, Reinhardt JM, eds. Medical Imaging 2007: Image Processing.6512: 65121P1-6512P10; 2007.
  • Zhou B, Willis A, Sui Y, Anderson D, Thomas T, Brown T. Improving inter-fragmentary alignment for virtual 3D reconstruction of highly fragmented bone fractures. In: Pluim JPW, Dawant BM, eds. Medical Imaging 2009: Image Processing.  7259:7259341-7259349; 2009.
  • Frank MC. Subtractive rapid prototyping: Creating a completely automated process for rapid machining. In: Kamrani A, Abouel Nasr E, eds. Rapid Prototyping. Springer U.S. pp. 165-196; 2006.
  • Frank MC, Hunt CV, Anderson DD, et al. Rapid manufacturing in biomedical materials: Using subtractive rapid prototyping for bone replacement. Proceedings of the Solid Freeform Fabrication Symposium, 2008.
  • Anderson DD, Frank MC, McKinley TO, Brown TD. Fragment substitutes for anatomically-interfaced segmental bone defect repair. 54th Annual Meeting of the Orthopaedic Research Society. Paper #998, 2008.
  • Sutradhar A, Paulino GH, Miller MJ, Nguyen TH. Topological optimization for designing patient-specific large craniofacial segmental bone replacements. PNAS. EarlyEdition (, 2010.