Unplanned return to operation room (OR) following growing spinal constructs (GSCs) in early onset scoliosis (EOS)-a multi-centric study

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.

Younger age at spinal cord detethering is potentially associated with a reduced risk of curve progression in children with early onset scoliosis

PURPOSE

In children with early onset scoliosis (EOS) who have tethered spinal cord (TSC), spinal cord detethering is commonly performed prior to spinal deformity correction (SDC). The purpose of this study was to investigate whether age or curve magnitude at the time of detethering is associated with curve progression at a follow-up of at least 2 years. It was hypothesized that patients who undergo detethering at a younger age, or those with a smaller curve magnitude, would experience a reduced rate of curve progression when compared with those who are older or with larger curves.

METHODS

Patients with EOS who underwent detethering at least 2 years prior to SDC were identified in a multicenter international registry. Radiographs were assessed just prior to the detethering procedure (pre-detether) and at the most recent visit prior to SDC (most recent post-detether). The rate of curve progression > 10° was examined. Owing to unequal follow-up in individual patients, Cox regression was used to investigate associations between primary variables (age and magnitude of major coronal curve) and rate of curve progression.

RESULTS

37 patients met inclusion criteria and 18 (mean age: 3.7 ± 2.9 years, 66.7% female, mean follow-up: 3.4 ± 1.3 years) had radiographic data available for analysis. Pre-detether and most recent post-detether major coronal curves were 44.8° ± 18.5° and 47.6° ± 23.9°, respectively. 5 (27.8%) patients had curve progression > 10° at a follow-up of 3.2 ± 1.2 years. Patients with progression > 10° were older at the time of detethering when compared with those without (5.6 ± 2.8 vs. 3 ± 2.7 years, p = 0.084). Regression analysis demonstrated that as age at detethering increased by 1 year, the rate of curve progression > 10° increased by 28.6% [95% confidence interval (CI) 0.899; 1.839, p = 0.169]. There was no evidence of an association between pre-detethering curve magnitude and rate of curve progression > 10° [HR: 1.027, 95% CI 0.977; 1.079, p = 0.297].

CONCLUSION

In a small multicenter cohort of EOS patients with TSC, younger age, but not curve size, at the time of detethering was associated with a lower rate of scoliosis progression. Although these results indicate a potential role for early spinal cord detethering in the EOS population, they require further prospective investigation with a larger number of patients.

LEVEL OF EVIDENCE

Level II.

Characteristics and therapeutic profile of TBI patients who underwent bilateral decompressive craniectomy: experience with 151 cases

Background

Decompressive craniectomy (DC) and intracranial pressure (ICP) monitoring are common approaches to reduce the death rate of Traumatic brain injury (TBI) patients, but the outcomes of these patients are unfavorable, particularly those who receive bilateral DC. The authors discuss their experience using ICP and other potential methods to improve the outcomes of TBI patients who receive bilateral DC.

Methods

Data from TBI patients receiving bilateral DC from Jan. 2008 to Jan. 2022 were collected via a retrospective chart review. Included patients who received unplanned contralateral DC after initial surgery were identified as unplanned secondary surgery (USS) patients. Patients’ demographics and baseline medical status; pre-, intra-, and postoperative events; and follow-up visit outcome data were analyzed.

Results

A total of 151 TBI patients were included. Patients who underwent USS experienced more severe outcomes as assessed using the 3-month modified Rankin Scale score (P = 0.024). In bilateral DC TBI patients, USS were associated with worsen outcomes, moreover, ICP monitoring was able to lower their death rate and was associated with a lower USS incidence. In USS patients, ICP monitoring was not associated with improved outcomes but was able to lower their mortality rate (2/19, 10.5%, vs. 10/25, 40.0%; P = 0.042).

Conclusion

The avoidance of USS may be associated with improved outcomes of TBI patients who underwent bilateral DC. ICP monitoring was a potential approach to lower USS rate in TBI patients, but its specific benefits were uncertain.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13049-022-01046-w.

Current benchtop protocols are not appropriate for the evaluation of distraction-based growing rods: a literature review to justify a new protocol and its development

PURPOSE

Although distraction-based growing rods (GR) are the gold standard for the treatment of early onset scoliosis, they suffer from high failure rates. We have (1) performed a literature search to understand the deficiencies of the current protocols, (2) in vitro evaluation of GRs using our proposed protocol and performed a finite element (FE) model validation, and (3) identified key features which should be considered in mechanical testing setups.

METHODS

PubMed, Embase, and Web of Science databases were searched for articles published on (a) in vivo animal, in vitro cadaveric, and biomechanical studies analyzing the use of GRs as well as (b) failure mechanisms and risk factors for GRs. Both FE and benchtop models of a proposed TGR test construct were developed and evaluated for two cases, long tandem connectors (LT), and side-by-side connectors (SBS). The test construct consisted of five polymer blocks representing vertebral bodies, joined with springs to simulate spinal stiffness. The superior and inferior blocks accepted the pedicle screw anchors, while the three middle blocks were floating. After the pedicle screws, rods, and connectors were assembled onto this construct, distraction was performed, mimicking scoliosis surgery. The resulting distracted constructs were then subjected to static compression-bending loading. Yield load and stiffness were calculated and used to verify/validate the FE results.

RESULTS

From the literature search, key features identified as significant were axial and transverse connectors, contoured rods, and distraction, distraction being the most challenging feature to incorporate in testing. The in silico analyses, once they are validated, can be used as a complementing technique to investigate other anatomical features which are not possible in the mechanical setup (like growth/scoliosis curvature). Based on our experiment, the LT constructs showed higher stiffness and yield load compared to SBS (78.85 N/mm vs. 59.68 N/mm and 838.84 N vs. 623.3 N). The FE predictions were in agreement with the experimental outcomes (within 10% difference). The maximum von Mises stresses were predicted adjacent to the distraction site, consistent with the location of observed failures in vivo.

CONCLUSION

The two-way approach presented in this study can lead to a robust prediction of the contributing factors to the in vivo failure.