Predicting early return to the operating room in early-onset scoliosis patients using machine learning techniques.

PURPOSE: Surgical treatment of early-onset scoliosis (EOS) is associated with high rates of complications, often requiring unplanned return to the operating room (UPROR). The aim of this study was to create and validate a machine learning model to predict which EOS patients will go on to require an UPROR during their treatment course. METHODS: A retrospective review was performed of all surgical EOS patients with at least 2 years follow-up. Patients were stratified based on whether they had experienced an UPROR. Ten machine learning algorithms were trained using tenfold cross-validation on an independent training set of patients. Model performance was evaluated on a separate testing set via their area under the receiver operating characteristic curve (AUC). Relative feature importance was calculated for the top-performing model. RESULTS: 257 patients were included in the study. 146 patients experienced at least one UPROR (57%). Five factors were identified as significant and included in model training: age at initial surgery, EOS etiology, initial construct type, and weight and height at initial surgery. The Gaussian naïve Bayes model demonstrated the best performance on the testing set (AUC: 0.79). Significant protective factors against experiencing an UPROR were weight at initial surgery, idiopathic etiology, initial definitive fusion construct, and height at initial surgery. CONCLUSIONS: The Gaussian naïve Bayes machine learning algorithm demonstrated the best performance for predicting UPROR in EOS patients. Heavier, taller, idiopathic patients with initial definitive fusion constructs experienced UPROR less frequently. This model can be used to better quantify risk, optimize patient factors, and choose surgical constructs. LEVEL OF EVIDENCE: Prognostic: III.

Incidence of Tether Breakage in Anterior Vertebral Body Tethering.

BACKGROUND: Thoracic anterior vertebral body tethering (TAVBT) is an emerging treatment for adolescent idiopathic scoliosis. Tether breakage is a known complication of TAVBT with incompletely known incidence. We aim to define the incidence of tether breakage in patients with adolescent idiopathic scoliosis who undergo TAVBT. The incidence of tether breakage in TAVBT is hypothesized to be high and increase with time postoperatively. METHODS: All patients with right-sided, thoracic curves who underwent TAVBT with at least 2 and up to 3 years of radiographic follow-up were included. Tether breakage between 2 vertebrae was defined a priori as any increase in adjacent screw angle >5 degrees from the minimum over the follow-up period. The presence and timing of tether breakage were noted for each patient. A Kaplan-Meier survival analysis was performed to calculate expected tether breakage up to 36 months. χ 2 analysis was performed to examine the relationship between tether breakage and reoperations. Independent t test was used to compare the average final Cobb angle between cohorts. RESULTS: In total, 208 patients from 10 centers were included in our review. Radiographically identified tether breakage occurred in 75 patients (36%). The initial break occurred at or beyond 24 months in 66 patients (88%). Kaplan-Meier survival analysis estimated the cumulative rate of expected tether breakage to be 19% at 24 months, increasing to 50% at 36 months. Twenty-one patients (28%) with a radiographically identified tether breakage went on to require reoperation, with 9 patients (12%) requiring conversion to posterior spinal fusion. Patients with a radiographically identified tether breakage went on to require conversion to posterior spinal fusion more often than those patients without identified tether breakage (12% vs. 2%; P =0.004). The average major coronal curve angle at final follow-up was significantly larger for patients with radiographically identified tether breakage than for those without tether breakage (31 deg±12 deg vs. 26 deg±12 deg; P =0.002). CONCLUSIONS: The incidence of tether breakage in TAVBT is high, and it is expected to occur in 50% of patients by 36 months postoperatively. LEVEL OF EVIDENCE: Level IV.

Split Tibialis Anterior Tendon Transfer to The Peroneus Brevis or Tertius for the Treatment of Varus Foot Deformities in Children with Static Encephalopathy: A retrospective case series.

INTRODUCTION: The study purpose was to determine the safety/efficacy of a split anterior tibialis tendon transfer (SPLATT) to the peroneus tertius or brevis in children with static encephalopathy and varus feet. METHODS: A retrospective review of short- and long-term complications, change in ankle range of motion, strength, and gait kinematics. Predictors of postoperative varus or valgus were examined. RESULTS: One hundred thirty-three patients were included (average age [SD] 10.3 [3.7]), with an average follow-up of 3.9 (3.4) years. Forefoot/hindfoot eversion range of motion improved (P ≤ 0.05), dorsiflexor strength was maintained or improved in 76.9% of patients, and dorsiflexion in swing phase was maintained. Complications occurred in 6 of 133 patients (4.5%) and included 1 transfer failure, 1 wound dehiscence, and four pressure areas from casts. Successful correction was achieved in 77% of patients. Later onset of recurrent varus (14.4%, 10.6% requiring revision surgery) and pes valgus (8.7%, 4.8% requiring revision surgery) occurred. The length of the follow-up predicted the development of the pes valgus (odds ratio 1.28, 95% CI 1.0 to 1.6). DISCUSSION: SPLATT to the peroneus tertius or brevis is effective, and complications are rare. Subsequent valgus or recurrent varus deformities may occur, possibly requiring repeat surgery.

Predictors of Walking Activity in Children and Adolescents With Myelomeningocele.

OBJECTIVE: To assess the relationship between real-world walking activity in children and adolescents with myelomeningocele (MMC) and gross measures of lower extremity strength, range of motion, demographics, and medical history. DESIGN: Prospective study. SETTING: Participants recruited in outpatient clinics; data collected in a hospital-based motion analysis laboratory and in the community. PARTICIPANTS: Children and adolescents (N=52) with daily step count data available from a larger study of ambulatory children and adolescents with MMC. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURE: Univariate and multivariate regression was used to assess which participant characteristics and clinical factors were related to average number of steps per day. RESULTS: Univariate analysis showed a lower number of steps per day correlated with older age, male sex, higher body mass index, higher lesion level, use of assistive devices for ambulation, history of shunt placement, more television (TV) watched per week, lower hip extension and abduction strength, knee flexion strength, and ankle dorsiflexion and plantarflexion strength, and decreased knee and hip range of motion. Only assistive device usage and hours of TV watched per week remained in the final multivariate model predicting number of steps per day. CONCLUSIONS: Walking activity in children and adolescents with MMC was best predicted by assistive device use and amount of sedentary activity. Other predictors of walking activity from univariate analysis were related to assistive device use. This information can help tailor rehabilitation efforts and educate patients and families. Interventions targeting early prevention of strength loss and contractures may be important to retain or increase walking activity in children and adolescents with MMC.

Effect of physiological loads on cortical and traditional pedicle screw fixation.

STUDY DESIGN: Human cadaveric biomechanical study. OBJECTIVE: To determine the fixation strength of laterally directed, cortical pedicle screws under physiological loads. SUMMARY OF BACKGROUND DATA: Lateral trajectory cortical pedicle screws have been described as a means of obtaining improved fixation while minimizing soft-tissue dissection during lumbar instrumentation. Biomechanical data have demonstrated equivalent strength in a quasi-static model; however, no biomechanical information is available comparing the fixation of cortical with traditional pedicle screws under cyclic physiological loads. METHODS: Seventeen vertebral levels (T11-L5) underwent quantitative computed tomography. On 1 side, a laterally directed, cortical pedicle screw was inserted with a traditional, medially directed pedicle screw placed on the contralateral side. With the specimen constrained in a testing apparatus, each screw underwent cyclic craniocaudal toggling under incrementally increasing physiological loads until 2 mm of head displacement occurred. Next, uniaxial pullout of each toggled screw was performed. The number of craniocaudal toggle cycles and load (N) required to achieve pedicle screw movement as well as axial pullout resistance (N) were compared between the 2 techniques. RESULTS: The mean trabecular bone mineral density of the specimens was 202 K2HPO4 mg/cm. Cortical pedicle screws demonstrated significantly improved resistance to toggle testing, requiring 184 cycles to reach 2 mm of displacement compared with 102 cycles for the traditional pedicle screws (P=0.002). The force necessary to displace the screws was also significantly greater for the cortical versus the traditional screws (398 N vs. 300 N, P=0.004). There was no statistical difference in axial pullout strength between the previously toggled cortical and traditional pedicle screws (1722 N vs. 1741 N, P=0.837). CONCLUSION: Laterally directed cortical pedicle screws have superior resistance to craniocaudal toggling compared with traditional pedicle screws. LEVEL OF EVIDENCE: N/A.