Utilizing a comprehensive machine learning approach to identify patients at high risk for extended length of stay following spinal deformity surgery in pediatric patients with early onset scoliosis.

PURPOSE: Early onset scoliosis (EOS) patient diversity makes outcome prediction challenging. Machine learning offers an innovative approach to analyze patient data and predict results, including LOS in pediatric spinal deformity surgery. METHODS: Children under 10 with EOS were chosen from the American College of Surgeon's NSQIP database. Extended LOS, defined as over 5 days, was predicted using feature selection and machine learning in Python. The best model, determined by the area under the curve (AUC), was optimized and used to create a risk calculator for prolonged LOS. RESULTS: The study included 1587 patients, mostly young (average age: 6.94 ± 2.58 years), with 33.1% experiencing prolonged LOS (n = 526). Most patients were female (59.2%, n = 940), with an average BMI of 17.0 ± 8.7. Factors influencing LOS were operative time, age, BMI, ASA class, levels operated on, etiology, nutritional support, pulmonary and neurologic comorbidities. The gradient boosting model performed best with a test accuracy of 0.723, AUC of 0.630, and a Brier score of 0.189, leading to a patient-specific risk calculator for prolonged LOS. CONCLUSIONS: Machine learning algorithms accurately predict extended LOS across a national patient cohort and characterize key preoperative drivers of increased LOS after PSIF in pediatric patients with EOS.

Characteristics of early complications in isolated primary anterior cruciate ligament reconstruction surgery.

PURPOSE: The early complications of isolated anterior cruciate ligament reconstruction surgery (ACLR) have not been well characterized using large databases. This study aims to characterize incidence, impact, and risk factors for short-term operative complications following elective, isolated ACLR surgery. We hypothesize that demographic and perioperative factors may predict 30-day complications after isolated ACLR. METHODS: This case-control analysis of the American College of Surgeons National Surgical Quality Improvement Program Database (2005-2017) used Current Procedural Terminology codes to identify elective, isolated ACLR patients. Patients undergoing concomitant procedures were excluded. Complications were analyzed using bivariate analysis against demographic variables. Multiple stepwise logistic regression was used to identify independent risk factors for morbidity after ACLR. RESULTS: A total 12,790 patients (37.0% female, p = 0.674) were included with a mean age of 32.2 years old (SD 10.7 years, p < 0.001). Mean BMI was 27.8 kg/m2 (6.5) where 28.9% of patients had a BMI > 30 (p = 0.064). The most common complications were wound-related (0.57%). In cases with complications, there were higher rates of (1.3% vs 0.8%, p = 0.004) prolonged operation (> 1.5 h), higher rate (2.9% vs 1.8%, p = 0.004) of extended length of stay (≥ 1 day), unplanned reoperation (15.8% vs 0.3%, p < 0.001), and unplanned readmission (17.5% vs 0.3%, p < 0.001). Multivariate analysis showed prolonged operative time (p = 0.001), dyspnea (p = 0.008), and non-ambulatory surgery (p = 0.034) to be predictive of any complication. Dependent functional status (p = 0.091), mFI-5 > 0.2 (= 0.173), female sex (p = 0.191), obesity (p = 0.101), and smoking (p = 0.113) were not risk factors for complications. CONCLUSION: ACLR is associated with low rates of morbidity and readmissions. The most common comorbidities, complications, and predictors of morbidities were identified to aid surgeons in further reducing adverse outcomes of ACLR. Operative time > 1.5 h, dyspnea, and non-ambulatory surgery are predictive of complications.

Perioperative outcomes in isolated versus multiligamentous anterior cruciate ligament reconstruction: a retrospective cohort analysis.

PURPOSE: The outcomes of anterior cruciate ligament reconstruction in the setting of multiligamentous knee injury (M-ACLR) have not been well characterized compared to isolated ACLR (I-ACLR). This study aims to characterize and compare short-term outcomes between I-ACLR and M-ACLR. METHODS: This is a retrospective cohort analysis of the American College of Surgeons National Surgical Quality Improvement Program database from 2005 to 2017. Current Procedural Terminology codes were used to identify and compare elective I- and M-ACLR patients, excluding patients undergoing concomitant meniscal or chondral procedures. Patient demographics and outcomes after I- and M-ACLR were compared using bivariate analysis. Multiple logistic regression analyzed if multiligamentous ACLR was an independent risk factor for adverse outcomes. RESULTS: There was a total of 13,131 ACLR cases, of which 341 were multiligamentous cases. The modified fragility index-5 was higher in multiligamentous ACLR (p < 0.001). Multiligamentous ACLR had worse perioperative outcomes, with higher rate of all complications (3.8%, p = 0.013), operative time > 1.5 h (p < 0.001), length of stay (LOS) ≥ 1 day (p < 0.001), wound complication (2.1%, p = 0.001), and intra- or post-op transfusions (p < 0.001). In multiple logistic regression, multiligamentous ACLR was an independent risk factor for LOS ≥ 1 (odds ratio [OR] 5.8), and intra-/post-op transfusion (OR 215.1) and wound complications (OR 2.4). M-ACLR was not an independent risk factor for any complication, reoperation at 30 days, readmission, urinary tract infection (UTI), or venous thromboembolism (VTE). CONCLUSION: M-ACLR generally had worse outcomes than I-ACLR, including longer LOS, need for perioperative transfusions, and wound complications.

Increased age and modified fragility index increases risk of short-term complications after anterior cruciate ligament reconstruction surgery.

BACKGROUND: Large data analysis of anterior cruciate ligament reconstruction (ACLR) short-term complications on age will help surgeons stratify and counsel at-risk patients. The purpose of this study is to assess if older patients are at greater risk for short-term complications after ACLR. METHODS: This retrospective cohort study included patients who underwent elective ACLR with or without concomitant meniscal procedures in the National Surgical Quality Improvement Program from 2005 to 2017. Patients were divided into age groups 16-30, 31-45, and > 45. Modified fragility index-5 (mFI-5), demographics and short-term outcomes were examined with bivariate and multivariate analysis to determine if age was a risk factor for complications. RESULTS: A total of 23,581 patients (35.4% female) were included in this analysis. Mean age was 32.1 ± 10.8 years. Older patients had higher mFI-5 scores (p < 0.001), shorter operative times (p < 0.001), lower use of only general anesthesia (p < 0.001). The oldest patients had similar rates of complications as the two younger groups. Older age was an independent risk factor for VTE, but decreased risk of prolonged operations. A mFI-5 > 0 increased risk factors for readmission (Odds ratio 2.2, P = 0.006). Infection was the most common cause 30-day readmissions (40/135, 29.6%). CONCLUSION: In the early postoperative period, older age is an independent risk factor for VTE and younger age is a significant factor for prolonged surgeries. Having an mFI-5 > 0 increased risk factors for readmission.

The Posterior Cranial Vertical Line: A Novel Radiographic Marker for Classifying Global Sagittal Alignment.

OBJECTIVE: To define a novel radiographic measurement, the posterior cranial vertical line (PCVL), in an asymptomatic adult population to better understand global sagittal alignment. METHODS: We performed a multicenter retrospective review of prospectively collected radiographic data on asymptomatic volunteers aged 20-79. The PCVL is a vertical plumb line drawn from the posterior-most aspect of the occiput. The horizontal distances of the PCVL to the thoracic apex (TA), posterior sagittal vertical line (PSVL, posterosuperior endplate of S1), femoral head center, and tibial plafond were measured. Classification was either grade 1 (PCVL posterior to TA and PSVL), grade 2 (PCVL anterior to TA and posterior to PSVL), or grade 3 (PCVL anterior to TA and PSVL). RESULTS: Three hundred thirty-four asymptomatic patients were evaluated with a mean age of 41 years. Eighty-three percent of subjects were PCVL grade 1, 15% were grade 2, and 3% were grade 3. Increasing PCVL grade was associated with increased age (p < 0.001), C7-S1 sagittal vertical axis (SVA) (p < 0.001), C2-7 SVA (p < 0.001). Additionally, it was associated with decreased SS (p = 0.045), increased PT (p < 0.001), and increased knee flexion (p < 0.001). CONCLUSION: The PCVL is a radiographic marker of global sagittal alignment that is simple to implement and interpret. Increasing PCVL grade was significantly associated with expected changes and compensatory mechanisms in the aging population. Most importantly, it incorporates cervical alignment parameters such as C2-7 SVA. The PCVL defines global sagittal alignment in adult volunteers and naturally distributes into 3 grades, with only 3% being grade 3 where the PCVL lies anterior to the TA and PSVL.

Current Evidence for Hybrid Constructs: Simultaneous ACDF/Arthroplasty and Arthroplasty Adjacent to Previous ACDF.

It is not surprising that the utilization of hybrid constructs, combining cervical disc arthroplasty with anterior cervical disc arthroplasty, has steadily increased over the last decade. Known limitations exist with multi-level anterior cervical disc arthroplasty and cervical disc arthroplasty procedures. Hybrid surgery offers the possibility to address patient-specific pathology in a more tailored manner by restoring functional mobility and promoting fusion where appropriate. This review discusses the current evidence, both biomechanical and clinical, of hybrid surgery for 2-level and 3-level cervical disease.

Artificial Intelligence and Machine Learning Applications in Spine Surgery.

The complexity of patients with spine pathology and high rates of complications has driven extensive research directed toward optimizing outcomes and reducing complications. Traditional statistical analysis has been limited both in validity and in the number of predictor variables considered. Over the past decade, artificial intelligence and machine learning have taken center stage as the possible solution to creating more accurate and applicable patient-centered predictive models in spine surgery. This review discusses the current published machine learning applications on preoperative optimization, risk stratification, and predictive modeling for the cervical, lumbar, and adult spinal deformity populations.

Predicting postoperative coronal alignment for adult spinal deformity: do lower-extremity factors matter?

OBJECTIVE: The objective was to describe an intraoperative method that accurately predicts postoperative coronal alignment for up to 2 years of follow-up. The authors hypothesized that the intraoperative coronal target for adult spinal deformity (ASD) surgery should account for lower-extremity parameters, including pelvic obliquity (PO), leg length discrepancy (LLD), lower-extremity mechanical axis difference (MAD), and asymmetrical knee bending. METHODS: Two lines were drawn on intraoperative prone radiographs: the central sacral pelvic line (CSPL) (the line bisecting the sacrum and perpendicular to the line touching the acetabular sourcil of both hips) and the intraoperative central sacral vertical line (iCSVL) (which is drawn relative to CSPL based on the preoperative erect PO). The distance from the C7 spinous process to CSPL (C7-CSPL) and the distance from the C7 spinous process to iCSVL (iCVA) were compared with immediate and 2-year postoperative CVA. To account for LLD and preoperative lower-extremity compensation, patients were categorized into four preoperative groups: type 1, no LLD (< 1 cm) and no lower-extremity compensation; type 2, no LLD with lower-extremity compensation (PO > 1°, asymmetrical knee bending, and MAD > 2°); type 3, LLD and no lower-extremity compensation; and type 4, LLD with lower-extremity compensation (asymmetrical knee bending and MAD > 4°). A retrospective review of a consecutively collected cohort with ASD who underwent minimum 6-level fusion with pelvic fixation was performed for validation. RESULTS: In total, 108 patients (mean ± SD age 57.7 ± 13.7 years, 14.0 ± 3.9 levels fused) were reviewed. Mean preoperative/2-year postoperative CVA was 5.0 ± 2.0/2.2 ± 1.8 cm. For patients with type 1, both C7-CSPL and iCVA had similar error margins for immediate postoperative CVA (0.5 ± 0.6 vs 0.5 ± 0.6 cm, p = 0.900) and 2-year postoperative CVA (0.3 ± 0.4 vs 0.4 ± 0.5 cm, p = 0.185). For patients with type 2, C7-CSPL was more accurate for immediate postoperative CVA (0.8 ± 1.2 vs 1.7 ± 1.8 cm, p = 0.006) and 2-year postoperative CVA (0.7 ± 1.1 vs 2.1 ± 2.2 cm, p < 0.001). For patients with type 3, iCVA was more accurate for immediate postoperative CVA (0.3 ± 0.4 vs 1.7 ± 0.8 cm, p < 0.001) and 2-year postoperative CVA (0.3 ± 0.2 vs 1.9 ± 0.8 cm, p < 0.001). For patients with type 4, iCVA was more accurate for immediate postoperative CVA (0.6 ± 0.7 vs 3.0 ± 1.3 cm, p < 0.001) and 2-year postoperative CVA (0.5 ± 0.6 vs 3.0 ± 1.6 cm, p < 0.001). CONCLUSIONS: This system, which accounted for lower-extremity factors, provided an intraoperative guide to determine both immediate and 2-year postoperative CVA with high accuracy. For patients with type 1 and 2 (no LLD, with or without lower-extremity compensation), C7-intraoperative CSPL accurately predicted postoperative CVA up to 2-year follow-up (mean error 0.5 cm). For patients with type 3 and 4 (LLD, with or without lower-extremity compensation), iCVA accurately predicted postoperative CVA up to 2-year follow-up (mean error 0.4 cm).

The Lumbosacral Fractional Curve vs Maximum Coronal Cobb Angle in Adult Spinal Deformity Patients with Coronal Malalignment: Which Matters More?

STUDY DESIGN: Retrospective cohort study. OBJECTIVES: In patients undergoing adult spinal deformity (ASD) surgery we sought to: 1) report preoperative and postoperative lumbosacral fractional (LSF) curve and maximum coronal Cobb angles and 2) determine their impact on radiographic, clinical, and patient-reported outcomes (PROs). METHODS: A single-institution cohort study was undertaken. The LSF curve was the cobb angle between the sacrum and most tilted lower lumbar vertebra. Coronal/sagittal vertical axis (CVA/SVA) were collected. Patients were compared between 4 groups: 1) Neutral Alignment (NA); 2) coronal malalignment only (CM); 3) Sagittal malalignment only (SM); and 4) Combined-Coronal-Sagittal-Malalignment (CCSM). Outcomes including postoperative CM, postoperative coronal vertical axis, complications, readmissions, reoperation, and PROs. RESULTS: A total of 243 patients underwent ASD surgery with mean total instrumented levels of 13.5. Mean LSF curve was 12.1±9.9°(0.2-62.3) and mean max Cobb angle was 43.0±26.5° (0.0-134.3). The largest mean LSF curves were seen in patients with CM (14.6°) and CCSM (13.1°) compared to NA (12.1°) and SM (9.5°) (p=0.100). A higher LSF curve was seen in patients with fusion to the sacrum and instrumentation to the pelvis (p=0.009), and a higher LSF curve was associated with more TLIFs (p=0.031). Postoperatively, more TLIFs were associated with greater amount of LSF curve correction (p<0.001). Comparing the LSF and the max Cob angle among Qiu types, the highest mean max Cobb angle was in Qiu Type B patients (p=0.025), whereas the highest mean LSF curve was in Qiu Type C patients (p=0.037). Moreover, 82.7% of patients had a LSF curve opposite the max Cobb angle. The LSF curve was larger than the max Cobb angle in 22/243 (9.1%) patients, and most of these 22 patients were Qiu Type A (59.1%). Regarding correction, the max Cobb angle achieved more correction than the LSF curve, judged by the percent improved from preop (54.5% Cobb vs. 46.5% LSF, p=0.025) in patients with max cobb>20° and LSF curve >5°. The LSF curve underwent greater correction in Qiu Type C patients (9.2°) compared to Type A (5.7°) and Type B (5.1°) (p=0.023); however, the max Cobb angle was similarly corrected among Qiu Types: Type A 21.8°, Type B 24.6°, and Type C 25.4° (p=0.602). Minimal differences were seen comparing the preop/postop/change in LSF curve and max Cobb angle regarding postop CM, postop CVA, complications, readmissions, reoperation, and PROs. CONCLUSIONS: The LSF curve was highest in patients with CM, CCSM, and Qiu Type C curves. Most patients had a LSF curve opposite the max Cobb angle. The max Cobb angle was more often corrected than the LSF curve. The LSF curve underwent greater correction among Qiu Type C patients, whereas the max Cobb angle was similarly corrected among all Qiu Types. No clear trend was seen regarding postoperative complications and PROs between the LSF curve and max Cobb angle.

Living with a C2-Sacrum Spinal Fusion: Surgical Outcomes and Quality of Life in Patients Fused from C2 to the Sacrum.

STUDY DESIGN: Single center, retrospective cohort study. OBJECTIVES: Little is known about the surgical outcomes and quality of life in patients with C2-sacrum posterior spinal fusion (PSF). Though it is thought to be a "final" construct, it remains unknown how patients fare postoperatively. We sought to evaluate the surgical outcomes and quality of life of patients after C2-sacrum PSF. METHODS: Consecutive patients undergoing C2-Sacrum PSF from 2015-2020 by 4 surgeons at a single institution were included. The study time period for each patient began after their index operation that led to the C2-sacrum fusion. Dates of surgery, complications, reoperations, patient reported outcomes (PROs) including EuroQol 5 Dimensions (EQ-5D), Oswestry Disability Index (ODI), Scoliosis Research Society (SRS) questionnaires, and activities of daily living (ADL) questions were collected and analyzed. Descriptive statistics, paired t-tests, student t-tests, and linear regression were used. RESULTS: Of the 23 patients who underwent C2-sacrum PSF, 6 patients (26%) required a total of 10 reoperations after a mean of 1.5 years (range 0-4 years) after C2-sacrum PSF. Five reoperations were for mechanical failure; 3 for wound complications/infection; and 2 for instrumentation and spinous process prominence. PROs were collected on 18 patients with mean follow-up of 2.4 years (range .5-4.5) after their C2-sacrum PSF. At 6-months, both SRS-22 and ODI scores improved significantly after C2-sacrum PSF (SRS: 57.5 to 76.3, P = .0014; ODI: 47.0 to 31.7, P = .013). Similarly, at a mean 2.4 years postoperatively, mean ODI improved significantly (47.0 to 30.4, P = .0032). Six patients (33%) had minimal symptoms (ODI <20). The median postoperative EQ-5D score was .74 (range .19 to 1.0), which compares favorably to patients with hip/knee osteoarthritis (EQ-5D .63) and diabetes mellitus (DM) (EQ-5D .69) and hypertension (HTN). In terms of activities of daily living (ADL), 10 patients (56%) exercised regularly-a mean 4.5 days/week. 11 (61%) could do light aerobic activity (e.g. stationary bike). 10 (55%) were able to play with children/grandchildren as desired. Eight patients (44%) hiked, and 2 (11%) drove independently. 11 (61%) could tolerate short air-travel comfortably. Of the 17 patients who could toilet and perform basic hygiene preoperatively, 16 (94%) were able to do so postoperatively. CONCLUSION: Though C2-sacrum PSF is thought to be a "final" construct, approximately 1 in 4 patients require subsequent operations. However, C2-sacrum PSF patients had a significant improvement in SRS and ODI scores by 6 months postop. Over 60% of patients were regularly performing light aerobic activity 2 years after their C2-sacrum PSF. EQ-5D suggests that this population fares better than those with degenerative hip/knee arthritis and similarly to those with common chronic conditions like DM and HTN.

Is There a Difference in Screw Accuracy, Robot Time Per Screw, Robot Abandonment, and Radiation Exposure Between the Mazor X and the Renaissance? A Propensity-Matched Analysis of 1179 Robot-Assisted Screws.

STUDY DESIGN: Prospective single-cohort analysis. OBJECTIVES: To compare the outcomes/complications of 2 robotic systems for spine surgery. METHODS: Adult patients (≥18-years-old) who underwent robot-assisted spine surgery from 2016-2019 were assessed. A propensity score matching (PSM) algorithm was used to match Mazor X to Renaissance cases. Preoperative CT scan for planning and an intraoperative O-arm for screw evaluation were preformed. Outcomes included screw accuracy, robot time/screw, robot abandonment, and radiation. Screw accuracy was measured using Vitrea Core software by 2 orthopedic surgeons. Screw breach was measured according to the Gertzbein/Robbins classification. RESULTS: After PSA, a total of 65 patients (Renaissance: 22 vs. X: 43) were included. Patient/operative factors were similar between robot systems (P > .05). The pedicle screw accuracy was similar between robots (Renaissance: 1.1%% vs. X: 1.3%, P = .786); however, the S2AI screw breach rate was significantly lower for the X (Renaissance: 9.5% vs. X: 1.2%, P = .025). Robot time per screw was not statistically different (Renaissance: 4.6 minutes vs. X: 3.9 minutes, P = .246). The X was more reliable with an abandonment rate of 2.3% vs. Renaissance:22.7%, P = .007. Radiation exposure were not different between robot systems. Non-robot related complications including dural tear, loss of motor/sensory function, and blood transfusion were similar between robot systems. CONCLUSION: This is the first comparative analyses of screw accuracy, robot time/screw, robot abandonment, and radiation exposure between the Mazor X and Renaissance systems. There are substantial improvements in the X robot, particularly in the perioperative planning processes, which likely contribute to the X's superiority in S2AI screw accuracy by nearly 8-fold and robot reliability by nearly 10-fold.

Do Adult Spinal Deformity Patients Undergoing Surgery Continue to Improve From 1-Year to 2-Years Postoperative?

OBJECTIVE: Evaluate clinical improvement as measured by patient-reported outcomes (PROs) during the 1 to 2-year interval. STUDY DESIGN: Retrospective Cohort. METHODS: A single-institution registry of ASD patients undergoing surgery was queried for patients with ≥6 level fusions. Demographics and radiographic variables were collected. PROs collected were the ODI and SRS-22r scores at: preoperative, 1-year and 2-years. Outcome measures of clinical improvement during the 1-2 year time interval were: 1) group medians, 2) percent minimum clinically important difference (MCID), and 3) percent minimal symptom scale (MSS)(ODI < 20 or SRS-pain + function >8). Wilcoxon rank-sum tests, chi-squared tests, Kruskal-Wallis tests, and logistic regression were performed. RESULTS: 157 patients undergoing ASD surgery with minimum of 1-year follow-up were included. Mean age was 53.2 and mean instrumented levels was 13.1. Preoperative alignment was: Neutral Alignment (NA) 49%, Coronal Malalignment (CM) 17%, Sagittal Malalignment (SM 17%), and Combined Coronal/Sagittal Malalignment (CCSM) 18%. Preoperative to 1-year, and preoperative to 2-years, all ODI/SRS-22r significantly improved (P < .001). In all patients, the only significant improvement in PROs between 1-and 2-year postoperative were those reaching ODI MCID (69% 1-year vs. 84% 2-years; P < .001). Subgroup analysis: ≥55 years had an improved median ODI (18 vs. 8; P = .047) and an improved percent achieving ODI MCID (73% vs. 84%, P = .048). CCSM patients experienced significant improvement in SRS-appearance score (75% vs. 100%; P = .050), along with those with severe preoperative SM >7.5 cm (73% vs. 100%; P = .032). CONCLUSIONS: Most ASD patients experience the majority of PRO improvement by 1-year postoperative. However, subsets of patients that may continue to improve up to 2-years postoperative include patients ≥55 years, combined coronal/sagittal malalignment, and those with severe sagittal malalignment ≥7.5 cm.

What Is the Impact of Surgical Approach in the Treatment of Degenerative Cervical Myelopathy in Patients With OPLL? A Propensity-Score Matched, Multi-Center Analysis on Inpatient and Post-Discharge 90-Day Outcomes.

STUDY DESIGN: Retrospective cohort. OBJECTIVE: Provide a comparison of surgical approach in the treatment of degenerative cervical myelopathy in patients with OPLL. METHODS: A national database was queried to identify adult (≥18 years) patients with OPLL, who underwent at least a 2-level cervical decompression and fusion for cervical myelopathy from 2012-2014. A propensity-score-matching algorithm was employed to compare outcomes by surgical approach. RESULTS: After propensity-score matching, 627 patients remained. An anterior approach was found to be an independent predictor for higher inpatient surgical complications(OR 5.9), which included dysphagia:14%[anterior]vs.1.1%[posterior] P-value < 0.001, wound hematoma:1.7%[anterior]vs.0%[posterior] P-value = 0.02, and dural tear:9.4%[anterior]vs.3.2%[posterior] P-value = 0.001. A posterior approach was an predictor for longer hospital length of stay by nearly 3 days(OR 3.4; 6.8 days[posterior]vs.4.0 days[anterior] P-value < 0.001). The reasons for readmission/reoperation did not vary by approach for 2-3-level fusions; however, for >3-level fusions, patients with an anterior approach more often had respiratory complications requiring mechanical ventilation(P-value = 0.038) and required revision fusion surgery(P-value = 0.015). CONCLUSIONS: The national estimates for inpatient complications(25%), readmissions(9.9%), and reoperations(3.5%) are substantial after the surgical treatment of multi-level OPLL. An anterior approach resulted in significantly higher inpatient surgical complications, but this did not result in a longer hospital length of stay and the overall 90-day complication rates requiring readmission or reoperation was similar to those seen after a posterior approach. For patients requiring >3-level fusion, an anterior approach is associated with significantly higher risk for respiratory complications requiring mechanical ventilation and revision fusion surgery. Precise neurological complications and functional outcomes were not included in this database, and should be further assessed in future studies.

Can Natural Language Processing and Artificial Intelligence Automate The Generation of Billing Codes From Operative Note Dictations?

STUDY DESIGN: Retrospective Cohort Study. OBJECTIVES: Using natural language processing (NLP) in combination with machine learning on standard operative notes may allow for efficient billing, maximization of collections, and minimization of coder error. This study was conducted as a pilot study to determine if a machine learning algorithm can accurately identify billing Current Procedural Terminology (CPT) codes on patient operative notes. METHODS: This was a retrospective analysis of operative notes from patients who underwent elective spine surgery by a single senior surgeon from 9/2015 to 1/2020. Algorithm performance was measured by performing receiver operating characteristic (ROC) analysis, calculating the area under the ROC curve (AUC) and the area under the precision-recall curve (AUPRC). A deep learning NLP algorithm and a Random Forest algorithm were both trained and tested on operative notes to predict CPT codes. CPT codes generated by the billing department were compared to those generated by our model. RESULTS: The random forest machine learning model had an AUC of .94 and an AUPRC of .85. The deep learning model had a final AUC of .72 and an AUPRC of .44. The random forest model had a weighted average, class-by-class accuracy of 87%. The LSTM deep learning model had a weighted average, class-by-class accuracy 0f 59%. CONCLUSIONS: Combining natural language processing with machine learning is a valid approach for automatic generation of CPT billing codes. The random forest machine learning model outperformed the LSTM deep learning model in this case. These models can be used by orthopedic or neurosurgery departments to allow for efficient billing.

How common is acute pelvic fixation failure after adult spine surgery? A single-center study of 358 patients.

OBJECTIVE: There is a paucity of literature on pelvic fixation failure after adult spine surgery in the early postoperative period. The purpose of this study was to determine the incidence of acute pelvic fixation failure in a large single-center study and to describe the lessons learned. METHODS: The authors performed a retrospective review of adult (≥ 18 years old) patients who underwent spinal fusion with pelvic fixation (iliac, S2-alar-iliac [S2AI] screws) at a single academic medical center between 2015 and 2020. All patients had a minimum of 3 instrumented levels. The minimum follow-up was 6 months after the index spine surgery. Patients with prior pelvic fixation were excluded. Acute pelvic fixation failure was defined as revision of the pelvic screws within 6 months of the primary surgery. Patient demographics and operative, radiographic, and rod/screw parameters were collected. All rods were cobalt-chrome. All iliac and S2AI screws were closed-headed screws. RESULTS: In 358 patients, the mean age was 59.5 ± 13.6 years, and 64.0% (n = 229) were female. The mean number of instrumented levels was 11.5 ± 5.5, and 79.1% (n = 283) had ≥ 6 levels fused. Three-column osteotomies were performed in 14.2% (n = 51) of patients, and 74.6% (n = 267) had an L5-S1 interbody fusion. The mean diameter/length of pelvic screws was 8.5/86.6 mm. The mean number of pelvic screws was 2.2 ± 0.5, the mean rod diameter was 6.0 ± 0 mm, and 78.5% (n = 281) had > 2 rods crossing the lumbopelvic junction. Accessory rods extended to S1 (32.7%, n = 117) or S2/ilium (45.8%, n = 164). Acute pelvic fixation failure occurred in 1 patient (0.3%); this individual had a broken S2AI screw near the head-neck junction. This 76-year-old woman with degenerative lumbar scoliosis and chronic lumbosacral zone 1 fracture nonunion had undergone posterior instrumented fusion from T10 to pelvis with bilateral S2AI screws (8.5 × 90 mm); i.e., transforaminal lumbar interbody fusion L4-S1. The patient had persistent left buttock pain postoperatively, with radiographically confirmed breakage of the left S2AI screw 68 days after surgery. Revision included instrumentation removal at L2-pelvis and a total of 4 pelvic screws. CONCLUSIONS: The acute pelvic fixation failure rate was exceedingly low in adult spine surgery. This rate may be the result of multiple factors including the preference for multirod (> 2), closed-headed pelvic screw constructs in which large-diameter long screws are used. Increasing the number of rods and screws at the lumbopelvic junction may be important factors to consider, especially for patients with high risk for nonunion.

Postoperative coronal malalignment after adult spinal deformity surgery: incidence, risk factors, and impact on 2-year outcomes.

PURPOSE: To evaluate the incidence, risk factors, and patient-reported outcomes (PROs) of adult spinal deformity (ASD) patients with postoperative coronal malalignment. METHODS: A single-institution, retrospective cohort study of ASD patients undergoing ≥ 6 level fusions from 2015 to 2019 was undertaken. The primary outcome was postoperative coronal malalignment, defined as C7-coronal vertical axis (CVA) > 3 cm. Secondary outcomes included: complications, readmissions, reoperations, and 2-year PROs. RESULTS: A total of 243 ASD patients undergoing spinal surgery had preoperative and immediate postoperative measurements, and 174 patients (72%) had 2-year follow-up. Mean age was 49.3 ± 18.3yrs and mean instrumented levels was 13.5 ± 3.9. Mean preoperative CVA was 2.9 ± 2.7 cm, and 90 (37%) had preoperative coronal malalignment. Postoperative coronal malalignment occurred in 43 (18%) patients. Significant risk factors for postoperative coronal malalignment were: preoperative CVA (OR 1.21, p = 0.001), preoperative SVA (OR 1.05, p = 0.046), pelvic obliquity (OR 1.21; p = 0.008), Qiu B vs. A (OR 4.17; p = 0.003), Qiu C vs. A (OR 7.39; p < 0.001), lumbosacral fractional (LSF) curve (OR 2.31; p = 0.021), max Cobb angle concavity opposite the CVA (OR 2.10; p = 0.033), and operative time (OR 1.16; p = 0.045). Postoperative coronal malalignment patients were more likely to sustain a major complication (31% vs. 14%; p = 0.01), yet no differences were seen in readmissions (p = 0.72) or reoperations (p = 0.98). No significant differences were seen in 2-year PROs (p > 0.05). CONCLUSIONS: Postoperative coronal malalignment occurred in 18% of ASD patients and was most associated with preoperative CVA/SVA, pelvic obliquity, Qiu B/C curves, LSF curve concavity to the same side as the CVA, and maximum Cobb angle concavity opposite side of the CVA. Postoperative coronal malalignment was significantly associated with increased complications but not readmission, reoperation, or 2-year PROs.

Rod fractures after multi-rod constructs in adult spinal deformity patients fused to the sacrum/pelvis: where do they occur and why?

PURPOSE: (1) To describe the use of multi-rod constructs (MRCs) in adult spinal deformity (ASD) surgery, (2) to report rod fractures occurring at MRC sites, and (3) to evaluate risk factors for rod fractures. METHODS: A single-center, retrospective cohort study was conducted of patients undergoing ASD surgery with these inclusion criteria: minimum 2-year follow-up, MRCs used, ≥ 10-level fusion, and fused to sacrum/pelvis. The primary outcome was rod fracture. Univariate/multivariate logistic regression was performed controlling for age, kickstand rod usage, number of rods across the lumbosacral junction (LSJ), and the amount of coronal/sagittal Cobb correction. RESULTS: Among 57 patients undergoing ASD surgery with MRCs, mean age was 60 ± 11 years. With respect to MRCs, 32 (56%) patients had 3 rods, 18 (32%) had 4, and 7 (12%) had 5. Rods crossing the LSJ were most often three (63%), followed by four (25%) and five (5%) rods. Nine (16%) patients experienced rod fractures with eight (89%) patients having no more than three rods crossing the LSJ. A coronal correction > 30 mm was more often seen in patients with rod fracture (p = 0.030), while an SVA correction > 50 mm was not significantly different (p = 0.608). Multivariate logistic regression revealed that the amount of coronal correction was significantly associated with rod fracture (OR 1.03, 95% CI 1.01-1.07, p = 0.044), as was achieving a coronal correction > 30 mm (OR 7.72, 95% CI 1.17-51.10, p = 0.034), with no association between the amount of sagittal correction obtained and rod fracture. CONCLUSION: This study found that greater coronal correction was associated with an increased odds of rod fracture. We suggest adding at least four rods across the LSJ cephalad to the interbody fusions to avoid rod fractures in these high demand areas. LEVEL OF EVIDENCE: III.

Incidence, mechanism, and protective strategies for 2-year pelvic fixation failure after adult spinal deformity surgery with a minimum six-level fusion.

OBJECTIVE: The purpose of this study was to determine the incidence, mechanism, and potential protective strategies for pelvic fixation failure (PFF) within 2 years after adult spinal deformity (ASD) surgery. METHODS: Data for ASD patients (age ≥ 18 years, minimum of six instrumented levels) with pelvic fixation (S2-alar-iliac [S2AI] and/or iliac screws) with a minimum 2-year follow-up were consecutively collected (2015-2019). Patients with prior pelvic fixation were excluded. PFF was defined as any revision to pelvic screws, which may include broken rods across the lumbosacral junction requiring revision to pelvic screws, pseudarthrosis across the lumbosacral junction requiring revision to pelvic screws, a broken or loose pelvic screw, or sacral/iliac fracture. Patient information including demographic data and health history (age, sex, BMI, smoking status, American Society of Anesthesiologists score, osteoporosis), operative (total instrumented levels [TIL], three-column osteotomy [3CO], interbody fusion), screw (iliac, S2AI, length, diameter), rod (diameter, kickstand), rod pattern (number crossing lumbopelvic junction, lowest instrumented vertebra [LIV] of accessory rod[s], lateral connectors, dual-headed screws), and pre- and postradiographic (lumbar lordosis, pelvic incidence, pelvic tilt, major Cobb angle, lumbosacral fractional curve, C7 coronal vertical axis [CVA], T1 pelvic angle, C7 sagittal vertical axis) parameters was collected. All rods across the lumbosacral junction were cobalt-chrome. All iliac and S2AI screws were closed-headed tulips. Both univariate and multivariate analyses were performed to determine risk factors for PFF. RESULTS: Of 253 patients (mean age 58.9 years, mean TIL 13.6, 3CO 15.8%, L5-S1 interbody 74.7%, mean pelvic screw diameter/length 8.6/87 mm), the 2-year failure rate was 4.3% (n = 11). The mechanisms of failure included broken rods across the lumbosacral junction (n = 4), pseudarthrosis across the lumbosacral junction requiring revision to pelvic screws (n = 3), broken pelvic screw (n = 1), loose pelvic screw (n = 1), sacral/iliac fracture (n = 1), and painful/prominent pelvic screw (n = 1). A higher number of rods crossing the lumbopelvic junction (mean 3.8 no failure vs 2.9 failure, p = 0.009) and accessory rod LIV to S2/ilium (no failure 54.2% vs failure 18.2%, p = 0.003) were protective for failure. Multivariate analysis demonstrated that accessory rod LIV to S2/ilium versus S1 (OR 0.2, p = 0.004) and number of rods crossing the lumbar to pelvis (OR 0.15, p = 0.002) were protective, while worse postoperative CVA (OR 1.5, p = 0.028) was an independent risk factor for failure. CONCLUSIONS: The 2-year PFF rate was low relative to what is reported in the literature, despite patients undergoing long fusion constructs for ASD. The number of rods crossing the lumbopelvic junction and accessory rod LIV to S2/ilium relative to S1 alone likely increase construct stiffness. Residual postoperative coronal malalignment should be avoided to reduce PFF.

What Radiographic and Clinical Factors Ultimately Necessitate a C2-Sacrum Instrumented Posterior Spinal Fusion?

OBJECTIVE: /Hypothesis: Patients undergoing C2-sacrum PSF have unique medical histories and multiple prior operations over an extended period. DESIGN: Single center, retrospective cohort. METHODS: Consecutive C2-sacrum PSF patients operated on by 4 surgeons at a single-center from 2015-2020 were reviewed. Demographics, comorbidities, indications, surgical history, and radiographic parameters were collected. RESULTS: 23 patients underwent C2-sacrum PSF. 13 (57%) were male, and 21 (91.3%) were adults. Mean age at time of first spine surgery was 44 years (range 5-71) and 53 years (range 14-72) at the time of C2-sacrum PSF. Six patients (26%) had osteoporosis, and 6 patients (26%) had neurologic comorbidities-including Parkinson's disease (4), cerebral palsy (1), and Brown Sequard syndrome (1). Four (17%) had connective tissue disease. Two patients underwent C2-sacrum PSF as an index procedure: (1) 67M with myelomatous fractures and 124° of cervicothoracic kyphosis; (2) 28F with severe Marfan syndrome with 140° thoracic scoliosis and 130° thoracic kyphosis. The remaining 21 (91%) underwent C2-sacrum PSF as a revision following prior spinal surgeries on average, 4 previous surgeries (range 1-13) over 10.5 years (range .3-37.4). Indications for the remaining 21 C2-sacrum PSF revision procedures included 17 (81%) for kyphosis (5 of whom also had significant coronal deformity), 1 (5%) for only coronal malalignment, 2 (10%) for instrumentation failure, and 1 (5%) for myelopathy. CONCLUSIONS: 91% (21/23) of patients requiring C2-sacrum PSF were treated as revisions of prior fusions, with a mean of 4 prior surgeries over 10 years. Over 80% of these patients underwent C2-sacrum PSF to address kyphosis. 26% had neurologic conditions, and 26% had osteoporosis.

Operative Time Less Than 1.5 Hours, Male Sex, Dependent Functional Status, Presence of Dyspnea, and Reoperations Within 30 days Are Independent Risk Factors for Readmission After ACLR.

Purpose: The purposes of this study are to use a large, patient-centered database to describe the 30-day readmission rate and to identify predictive risk factors for readmission after elective isolated ACLR. Methods: The National Surgical Quality Improvement Program Database was retrospectively queried for isolated ACLR procedures between 2011 and 2017. Current Procedural Terminology (CPT) codes were used to identify isolated ACLR patients. Those undergoing additional procedures such as meniscectomy or multi-ligamentous reconstruction were excluded. Readmissions were analyzed against demographic variables with bivariate analysis. Multivariate logistic regression was used to find independent risk factors for 30-day readmissions after ACLR. Results: A total of 11,060 patients (37.2% female) were included with an average age of 32.2 ± 10.6 years and mean body mass index (BMI) of 27.9 ± 6.5 kg/m2 (29.2% were >30). The overall readmission rate was 0.59%. The most reported reason for readmission was infection 0.22 (24 out of 11,060). The following variables were associated with significantly higher readmission rates: male sex (P = .001), history of severe chronic obstructive pulmonary disease (COPD) (P = .025), cardiac comorbidity (P = .034), operative time >1.5 hours (P <.001), partially dependent functional health status (P = .002), high preoperative creatinine (P = .009), normal preoperative albumin (P = .020), hypertension (P = .034), and reoperations (P < .001). Operative time >1.5 hours, male sex, dependent functional status, the presence of dyspnea, and undergoing a reoperation were identified as independent risk factors for 30-day readmissions (P < .05 for all). Conclusions: Isolated ACLR is associated with low 30-day readmission rates. Operative time >1.5 hours, male sex, dependent functional status, the presence of dyspnea, and 30-day reoperations are independent risk factors for readmission that should be considered in patient selection and addressed with preoperative counseling. Level of Evidence: Level III, retrospective cohort study.