Page 100 - Read Online

Page 17 Martinez et al. Art Int Surg. 2025;5:16-23 https://dx.doi.org/10.20517/ais.2024.73

under the definition of AI, including machine learning (ML), deep learning (DL), natural language
[2]
processing, and computer vision . Given the immeasurable quantity of patient health data and increasingly
advanced technologies capable of processing it, there are ample applications for AI in the spino-plastics
[1]
domain .

Spino-plastic surgery is one surgical subspecialty that combines the talents of interdisciplinary surgical
subspecialists within plastics, orthopedics, and neurosurgery to meet the needs of patients requiring
complex spinal reconstruction. In brief, spino-plastics utilizes vascularized bone grafts (VBGs) from the iliac
crest, spinous process, rib, scapula, clavicle, and occiput to augment the strength of spinal fusions
necessitated by pathologies such as trauma, degeneration, or tumor . VBGs are pedicled on muscle and
[4-9]
supplied by Sharpey’s fibers, which physically connect the muscle to bone and allow small unnamed
[10]
periosteal feeding vessels to vascularize Haversian canals . VBGs are increasingly indicated for the
treatment of pseudoarthrosis, as they increase osteogenesis, osteoconductivity, and osteoinductivity
compared to non-vascularized bone grafts (N-VBGs) . Rates of pseudoarthrosis following arthrodesis can
[10]
reach 60% or higher, leading to reoperations and significant morbidity that negatively impact quality of
life [11,12] . VBGs have been incorporated into the existing reconstructive algorithm that is divided into six
levels: allograft, bony substitution, autograft, N-VBG, VBG, pedicled vascularized bone flap, and free bone
[13]
[10]
flap . As VBGs have been found to enhance the strength of spinal fusion and decrease rates of
pseudoarthrosis, there is a need for an AI algorithm to identify those at risk for pseudoarthrosis who may
benefit from VBG. Key areas of research interest within spino-plastics include the identification of optimal
surgical candidates given the expanding indications for VBGs, as well as improving surgical techniques to
enhance patient outcomes.

In the literature, there is already evidence of AI algorithms developed to screen for vulnerable patient
populations and identify surgical candidates [1,3,14] . Furthermore, there are many existing AI algorithms with
similar functions of patient risk stratification. Within spine surgery, AI has already been applied to identify
surgical candidates and treatment options for anterior decompression and fusion for cervical spondylotic
[17]
myelopathy [15,16] , as well as to predict quality of life outcomes in adult spinal deformities . The future of
spine surgery may be guided by bioinformaticians, data engineers, and computer scientists who process big
[18]
data in a way that informs patient care and scientific discovery . In this article, we conducted a non-
systematic narrative review of the literature to better understand AI’s capability to transform the field of
spino-plastics through assessment of surgical candidacy and patient selection, imaging and virtual surgical
planning (VSP), and intraoperative instrument manipulation.

SURGICAL CANDIDACY AND PATIENT SELECTION
Disease classification systems are invaluable tools when applied appropriately within medical practice.
While a classification score does not solely drive available treatment options, it is a standardized entry point
and a piece to the overall puzzle in the care of patients with complex pathology. Unsupervised AI data
analysis can create new hierarchical clustering that accounts for patient frailty scores, functional status,
[19]
radiographic characteristics, and many demographic factors . Sophisticated pattern analysis incorporates
more data than could have been previously imagined, making surgeon education easier with elaborate risk-
benefit grids for various treatment pathways .
[19]
Predictive algorithms are an excellent way to identify high-risk patients more effectively, such as those who
are at a greater than average risk of pseudarthrosis, wound breakdown, or morbidity/mortality associated
with spinal fusion. In general, earlier identification of high-risk patients allows for earlier intervention with
proactive employment of strategies to mitigate the risks inherent to the patient or pathology itself. In spino-