Awais K. Hussain

Predicting Surgical Complications in Patients Undergoing Elective Adult Spinal Deformity Procedures Using Machine Learning

STUDY DESIGNnCross-sectional database study.nnnOBJECTIVEnTo train and validate machine learning models to identify risk factors for complications following surgery for adult spinal deformity (ASD).nnnSUMMARY OF BACKGROUND DATAnMachine learning models such as logistic regression (LR) and artificial neural networks (ANNs) are valuable tools for analyzing and interpreting large and complex data sets. ANNs have yet to be used for risk factor analysis in orthopedic surgery.nnnMETHODSnThe American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for patients who underwent surgery for ASD. This query returned 4,073 patients, which data were used to train and evaluate our models. The predictive variables used included sex, age, ethnicity, diabetes, smoking, steroid use, coagulopathy, functional status, American Society of Anesthesiologists (ASA) class >3, body mass index (BMI), pulmonary comorbidities, and cardiac comorbidities. The models were used to predict cardiac complications, wound complications, venous thromboembolism (VTE), and mortality. Using ASA class as a benchmark for prediction, area under receiver operating characteristic curves (AUC) was used to determine the accuracy of our machine learning models.nnnRESULTSnThe mean age of patients was 59.5 years. Forty-one percent of patients were male whereas 59.0% of patients were female. ANN and LR outperformed ASA scoring in predicting every complication (p<.05). The ANN outperformed LR in predicting cardiac complication, wound complication, and mortality (p<.05).nnnCONCLUSIONSnMachine learning algorithms outperform ASA scoring for predicting individual risk prognosis. These algorithms also outperform LR in predicting individual risk for all complications except VTE. With the growing size of medical data, the training of machine learning on these large data sets promises to improve risk prognostication, with the ability of continuously learning making them excellent tools in complex clinical scenarios.nnnLEVEL OF EVIDENCEnLevel III.STUDY DESIGNnCross-sectional database study.nnnOBJECTIVEnTo train and validate machine learning models to identify risk factors for complications following surgery for adult spinal deformity (ASD). Machine learning models such as logistic regression (LR) and artificial neural networks (ANNs) are valuable tools for analyzing and interpreting large and complex data sets. ANNs have yet to be used for risk factor analysis in orthopedic surgery.nnnMETHODSnThe American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for patients who underwent surgery for ASD. This query returned 4,073 patients, which data were used to train and evaluate our models. The predictive variables used included sex, age, ethnicity, diabetes, smoking, steroid use, coagulopathy, functional status, American Society of Anesthesiologists (ASA) class >3, body mass index (BMI), pulmonary comorbidities, and cardiac comorbidities. The models were used to predict cardiac complications, wound complications, venous thromboembolism (VTE), and mortality. Using ASA class as a benchmark for prediction, area under receiver operating characteristic curves (AUC) was used to determine the accuracy of our machine learning models.nnnRESULTSnThe mean age of patients was 59.5 years. Forty-one percent of patients were male whereas 59.0% of patients were female. ANN and LR outperformed ASA scoring in predicting every complication (p<.05). The ANN outperformed LR in predicting cardiac complication, wound complication, and mortality (p<.05).nnnCONCLUSIONSnMachine learning algorithms outperform ASA scoring for predicting individual risk prognosis. These algorithms also outperform LR in predicting individual risk for all complications except VTE. With the growing size of medical data, the training of machine learning on these large data sets promises to improve risk prognostication, with the ability of continuously learning making them excellent tools in complex clinical scenarios.nnnLEVEL OF EVIDENCEnLevel III.

Age Stratification of 30-Day Postoperative Outcomes Following Excisional Laminectomy for Extradural Cervical and Thoracic Tumors:

Study Design: Retrospective cohort study. Objectives: To evaluate age as an independent predictive factor for perioperative morbidity and mortality in patients undergoing surgical decompression for metastatic cervical and thoracic spinal tumors using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2011 to 2014. Methods: We identified 1673 adult patients undergoing excisional laminectomy of cervical and thoracic extradural tumors. Patients were stratified into quartiles based on age, with Q1 including patients aged 18 to 49 years, Q2 including patients aged 50 to 60 years, Q3 including patients aged 61 to 69 years, and Q4 including patients ≥70 years. Univariate and multivariate regression analyses were performed to examine the association between age and 30-day perioperative morbidity and mortality. Results: Age was an independent risk factor for 30-day venous thromboembolism (VTE) and reoperation. Patients in Q3 for age had nearly a 4 times increased risk of VTE than patients in Q1 (odds ratio [OR] 3.97; 95% CI 1.91-8.25; P < .001). However, there was no significant difference in VTE between patients in Q4 and Q1 (P = .069). Patients in Q2 (OR 1.99; 95% CI 1.06-3.74; P = .032) and Q4 (OR 2.18; 95% CI 1.06-4.52; P = .036) for age had a 2 times increased risk of reoperation compared with patients in Q1. Conclusions: Age was an independent predictive factor for perioperative VTE and reoperation, but there was no clear age-dependent relationship between increasing age and the risk of these perioperative complications.

Impact of Obesity on Surgical Outcomes Following Laminectomy for Spinal Metastases

Study Design: Retrospective cohort study. Objectives: To determine the effect of obesity (body mass index >30 kg/m2) on perioperative morbidity and mortality after surgical decompression of spinal metastases. Methods: The American College of Surgeons National Surgical Quality Improvement Program database is a large multicenter clinical registry that collects preoperative risk factors, intraoperative variables, and 30-day postoperative morbidity and mortality outcomes from hospitals nationwide. Current Procedural Terminology codes were used to query the database for adults who underwent decompression with laminectomy for treatment of metastatic spinal lesions between 2010 and 2014. Patients were separated into 2 cohorts based on the presence of absence of obesity. Univariate analysis and multivariate logistic regression analysis were used to analyze the effect of obesity on perioperative morbidity and mortality. Results: There was a significantly higher rate of venous thromboembolism (VTE; obese 6.6% vs nonobese 4.2%; P = .01) and pulmonary complications (obese 2.6% vs nonobese 2.2%; P = .046) in the obese group compared with the nonobese group. The nonobese group had prolonged hospitalization (obese 62.0% vs nonobese 69.0%; P = .001) and a higher incidence of blood transfusions (obese 26.8% vs nonobese 34.2%; P < .001). On multivariate analysis, obesity was found to be an independent risk factor for VTE (odds ratio = 1.75, confidence interval = 1.17-2.63, P = .007). Conclusions: Obese patients were predisposed to an elevated risk of VTE following laminectomy for spinal metastases. Early postoperative mobilization and a low threshold to evaluate for perioperative VTE are important in these patients in order to appropriately diagnose and treat these complications and minimize morbidity.

Hypoalbuminemia as an Independent Risk Factor for Perioperative Complications Following Surgical Decompression of Spinal Metastases

Study Design: Retrospective cohort study. Objective: Malnutrition has been shown to be a risk factor for poor perioperative outcomes in multiple surgical subspecialties, but few studies have specifically investigated the effect of hypoalbuminemia in patients undergoing operative treatment of metastatic spinal tumors. The aim of this study was to assess the role of hypoalbuminemia as an independent risk factor for 30-day perioperative mortality and morbidity after surgical decompression of metastatic spinal tumors using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2011 to 2014. Methods: We identified 1498 adult patients in the ACS-NSQIP database who underwent laminectomy and excision of metastatic extradural spinal tumors. Patients were categorized into normoalbuminemic and hypoalbuminemic (ie, albumin level <3.5 g/dL) groups. Univariate and multivariate regression analyses were performed to examine the association between preoperative hypoalbuminemia and 30-day perioperative mortality and morbidity. Subgroup analysis was performed in the hypoalbuminemic group to assess the dose-dependent effect of albumin depletion. Results: Hypoalbuminemia was associated with increased risk of perioperative mortality, any complication, sepsis, intra- or postoperative transfusion, prolonged hospitalization, and non-home discharge. However, albumin depletion was also associated with decreased risk of readmission. There was an albumin level–dependent effect of increasing mortality and complication rates with worsening albumin depletion. Conclusions: Hypoalbuminemia is an independent risk factor for perioperative mortality and morbidity following surgical decompression of metastatic spinal tumors with a dose-dependent effect on mortality and complication rates. Therefore, it is important to address malnutrition and optimize nutritional status prior to surgery.

Age Is a Risk Factor for Postoperative Complications Following Excisional Laminectomy for Intradural Extramedullary Spinal Tumors

Study Design: Retrospective analysis. Objective: The incidence of intradural extramedullary (IDEM) spinal tumors is increasing. Excisional laminectomy for removal and decompression is the standard of care, but complications associated with patient age are unreported in the literature. Our objective is to identify if age is a risk factor for postoperative complications after excisional laminectomy of IDEM spinal tumors. Methods: A retrospective analysis was performed on the 2011 to 2014 ACS-NSQIP (American College of Surgeons National Surgical Quality Improvement Program) database for patients undergoing excisional laminectomy of IDEM spinal tumors. Age groups were determined by interquartile analysis. Chi-squared tests, t tests, and multivariate logistic regression models were employed to identify independent risk factors. Institutional review board approval was not needed. Results: A total of 1368 patients met the inclusion criteria for the study. Group 1 (age ≤ 44) contained 372 patients, group 2 (age 45-54) contained 314 patients, group 3 (age 55-66) contained 364 patients, and group 4 (age > 66) contained 318 patients. The univariate analysis showed that mortality and unplanned readmission were highest among patients in group 4 (1.26%, P = .011, and 10.00%, P = .039, respectively). Postoperative wound complications were highest among patients in group 1 (2.15%, P = .009), and postoperative venous thromboembolism and cardiac complications were highest among patients in group 3 (4.4%, P = .007, and 1.10%, P = .032, respectively). Multivariate logistic regression revealed that elderly age was an independent risk factor for postoperative venous thromboembolism (group 3 vs group 1; odds ratio = 6.739, confidence interval = 1.522-29.831, P = .012). Conclusions: This analysis revealed that increased age is an independent risk factor for postoperative venous thromboembolism in patients undergoing excisional laminectomy for IDEM spinal tumors.

Relationship between sagittal balance and adjacent segment disease in surgical treatment of degenerative lumbar spine disease: meta-analysis and implications for choice of fusion technique

AbstractStudy designMeta-analysis.ObjectiveTo conduct a meta-analysis investigating the relationship between spinopelvic alignment parameters and development of adjacent level disease (ALD) following lumbar fusion for degenerative disease.nSummary of background dataALD is a degenerative pathology that develops at mobile segments above or below fused spinal segments. Patient outcomes are worse, and the likelihood of requiring revision surgery is higher in ALD compared to patients without ALD. Spinopelvic sagittal alignment has been found to have a significant effect on outcomes post-fusion; however, studies investigating the relationship between spinopelvic sagittal alignment parameters and ALD in degenerative lumbar disease are limited.nMethodsSix e-databases were searched. Predefined endpoints were extracted and meta-analyzed from the identified studies.ResultsThere was a significantly larger pre-operative PT in the ALD cohort versus control (WMD 3.99, CI 1.97–6.00, pu2009=u20090.0001), a smaller pre-operative SS (WMD −u20092.74; CI −u20095.14 to 0.34, pu2009=u20090.03), and a smaller pre-operative LL (WMD −u20094.76; CI −u20097.66 to 1.86, pu2009=u20090.001). There was a significantly larger pre-operative PI-LL in the ALD cohort (WMD 8.74; CI 3.12–14.37, pu2009=u20090.002). There was a significantly larger postoperative PI in the ALD cohort (WMD 2.08; CI 0.26–3.90, pu2009=u20090.03) and a larger postoperative PT (WMD 5.23; CI 3.18–7.27, pu2009<u20090.00001).ConclusionThe sagittal parameters: PT, SS, PI-LL, and LL may predict development of ALD in patients’ post-lumbar fusion for degenerative disease. Decision-making aimed at correcting these parameters may decrease risk of developing ALD in this cohort.Graphical abstractThese slides can be retrieved under Electronic Supplementary Material.