Page 103 - Read Online

Martinez et al. Art Int Surg. 2025;5:16-23 https://dx.doi.org/10.20517/ais.2024.73 Page 20

spinal surgery to help with preoperative planning and surgical training [45,48] . The approach and positioning
of pedicle screws, foraminotomies, percutaneous interventions, and biopsies can all be achieved more safely,
[49]
with less margin of error, under the guidance of AR . AR also permits spine surgeons to view dissection
[50]
planes and tumor volumes with microscopic virtual mapping for performing osteotomies . Ma et al.
describe an ultrasound methodology to superimpose surgical planning in situ by incorporating CT images
[51]
with 3D anatomic landmarks .
Spine surgery can be challenging at baseline, as it is not uncommon to lack direct exposure or visualization
of the intricate, densely organized vessels and nerves along the axial skeleton. By the very nature of the field,
spino-plastics aims to treat an even more challenging subsect of patients. The distortion of native anatomy
in complex cases, whether caused by revision surgery or the mass effect of tumor bulk, presents additional
obstacles to intraoperative identification of neurovasculature. AR can aid surgeons in this difficult task,
employing visual information from MRI and CT scans to build surgical maps and chart paths around key
anatomic structures [45,46] . In spino-plastics cases, once the spinal instrumentation and fusion are complete,
the surgeon may harvest the VBG utilizing the standard arthrodesis instruments that are already on the
sterile field. If stealth guidance or AR is already being utilized for arthrodesis, it would be wise to consider
keeping the system available to assist the surgeon in harvesting and ensuring adequate fixation of the VBG.
Better spatial conceptualization of the instrumentation might reduce any risk of damaging nearby structures
in the vertebral column or retroperitoneum.

LOOKING TO THE FUTURE
Notably, the ultimate boundaries of AI have yet to be uncovered. AI has already contributed to our
[23]
understanding of driver mutations behind spinal cord tumors . This incredible technology will continue to
improve basic science research and treatment modalities to address the needs of spino-plastic patients from
many different perspectives. Despite the tremendous promise and exponential rise in these technological
advancements, there is much work to be done before clinicians may be completely comfortable about
incorporating this new technology into their workflow. Because ML is a powerful tool that is not fully
understood, caution must be exercised regarding the input of information to avoid the perpetuation of
misinformation and social biases. Overall, ML and AI currently lack transparency, which creates a “black
box” that may be difficult for surgeons to trust when comparing results to well-published algorithms that
have a more easily understood basis. However, there are methods currently being utilized to validate their
efficacy in clinical practice. This includes the results of case studies and trials - where technologies such as
imaging guidance can differentiate tumors from healthy tissue - comparative studies, and live integration
[52]
with surgical teams that provide constant feedback to enhance the safety and predictive power of AI
[3]
algorithms. Many metrics were used in these various studies to compare the performance of AI algorithms
to traditional models, such as the area under the curve, accuracy, and the receiver operating characteristic
curve. Furthermore, there is an upfront investment of time and resources essential for the development of
novel algorithms bearing any clinical significance. In other words, there is a significant lag time between
technological advancements and gaining necessary approvals for clinical application through the proper
avenues, including national supervisory organizations and individual hospital systems . In this stage of
[53]
conceptualization, there are limited existing data on AI in spino-plastic surgery and further long-term data
collection is required.

Despite the harvest and fixation of VBGs not requiring any additional tools that are traditionally used in
spinal fusion, the field of spino-plastics is in its nascent stages. Due to resource limitations or surgeon-
specific comfort levels with working in the spine and retroperitoneum, not all institutions have access to
plastic surgeons capable of performing this procedure. Developing strategies to implement novel AI