Faculty Bibliography

BACKGROUND CONTEXT: Adult spinal deformity (ASD) patients experience markedly decreased health-related quality of life along many dimensions. PURPOSE: We hypothesized that ASD surgery would be associated with improved work- and school-related productivity, as well as decreased rates of absenteeism. STUDY DESIGN/SETTING: Retrospective cohort study. PATIENT SAMPLE: Only patients eligible for 2-year follow-up were included, and those with a history of previous spinal fusion were excluded. OUTCOME MEASURES: The primary outcome measures in this study were SRS-22r questions 9 ("What is your current level of work/school activity?") and 17 ("In the last 3 months have you taken any days off of work, including household work, or school because of back pain?").
METHOD(S): A repeated measures mixed linear regression was used to analyze responses over time among patients managed operatively (OP) vs nonoperatively (NON-OP). Results were further stratified by baseline employment status, age, SVA, PI-LL, and deformity curve type.
RESULT(S): In total, 1,188 patients were analyzed. Of those, 66.6% (n=792) were managed operatively. The vast majority (78.9%, n=934) were female. Patients were relatively evenly distributed across age groups (27.6% 0-49; 21.1% 50-59; 30.1% 60-69; 21.2% >=70). At baseline, the mean percentage of activity at work/school was 56.4% (SD 35.4%), and the mean days off from work/school over the past 90 days was 1.6 (SD 1.8). Patients undergoing ASD surgery exhibited an 18.1% absolute increase in work/school productivity at 2-year follow-up vs baseline (p<0.0001), while no significant change was observed for the nonoperative cohort (p>0.5). Similarly, the OP cohort experienced 1.1 fewer absent days over the past 90 days at 2 years vs baseline (p<0.0001), while the NON-OP cohort showed no such difference (p>0.3). These differences were largely preserved after stratifying by baseline employment status, age group, SVA, PI-LL, and deformity curve type.
CONCLUSION(S): ASD patients managed operatively exhibited an average increase in work/school productivity of 18.1% and decreased absenteeism of 1.1 per 90 days at 2-year follow-up, while patients managed non-operatively did not exhibit change from baseline. Given the age distribution of patients in this study, these findings should be interpreted as pertaining primarily to obligations at work or within the home. Further study of the direct and indirect economic benefits of ASD surgery to patients is warranted. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Rod failure and pseudarthrosis are common complications following the surgical treatment of adult spinal defomity (ASD). Many strategies have been employed in mitigating these problems, such as interbody fusion, multiple rods, use of more effective biologics, and optimizing spinal alignment. We aimed to study the frequency and type of rod failures in a large ASD population over time. PURPOSE: To study the frequency and type of rod failures in a large ASD population over time. STUDY DESIGN/SETTING: Retrospective review of prospective data. PATIENT SAMPLE: This study included 647 adult spinal deformity patients surgically treated between 2008 and 2018. OUTCOME MEASURES: Rod failures.
METHOD(S): ASD patients with a fusion extended from minimum L1 to pelvis and min 2-year follow-up were included. Radiographs and records were examined to identify characteristics of the rod failures: timing, unilateral vs bilateral, vertebral level, unilateral progressing to bilateral failure, revision and failure rates over time.
RESULT(S): Among the study population, 647/1052 patients met inclusion criteria (age: 64+/-10 yrs., 78% F, BMI: 28.3+/-5.7, Mean follow-up: 37 months +/-13). Surgeries for these patients were performed from 2008 to 2018. The UIV was T7 or above in 306 and T8 or below in 338 patients, most frequent UIV: T4 (121) and T10 (208). A total of 146 patients had a 3COs; 435 had interbody fusion with 187 ALIF, 202 TLIF,135 XLIF; 286 patients had BMP use post only and 203 had BMP use interbody and posteriorly. Overall rod failure rate was 135/647 (21%), of which 9.3% occurred before by 2 years postop. Mean days to failure was 795 day +/- 485 (Median 733). Most frequent failures were in the lower lumbar spine L3-4 (32) 24%, L4-5 (34) 25% and L5-S1 (44) 32%. Of the 97 Unilateral failures, 35 (36.1%) were revised and only 8 progressed to bilateral failure of which 6/8 were revised; 46 Bilateral failure, 24 (52.2%) were revised.Kaplan-Meier analysis shows a survival rate of 89.7% at 2 years with decrease of 5% per year. No significant difference was found between the first 5 years and last 5 years. Comparison of primary vs revision index procedures did not reveal any differences in time to failure nor time to revision. Use of BMP improved survival rate to 91.2% compared to no BMP 83.4% at 2 years.
CONCLUSION(S): Rod failures remain a common complication when treating ASD with a rate of 21%. Among those rod failures, 9.3% occurred within 2 years. The most frequent sites of failure were in the lower lumbar spine. Unilateral failures underwent a lower rate of revision than bilateral. BMP increased rod survivability. Kaplan-Meier analysis demonstrated a survival rate of 89.7% at 2 years with a decrease of 5% per year. The failure rate has not improved when comparing the first 5 years. to last 5 years. We must continue to seek solutions to improve rod durability challenges to improve long-term outcomes. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Hounsfield unit values (HU) from computed tomography (CT) have been used to informally assess bone density in patients undergoing spine fusion procedures. HUs are easily obtained from a standard preoperative CT scan, and unlike Dual X-ray Absorptiometry analysis (DEXA) one can focus on specific regions of interest, such as the vertebral body of a planned upper-instrumented vertebrae (UIV). There is still a relative lack of literature on the reliability and utility of HUs to both identify patients with low BMD and guide surgical decision making. PURPOSE: To determine whether there was a significant difference in preoperative HUs, measured at the UIV, in patients that had a bone-density related complication (DRC) within 2 years of their spinal fusion. STUDY DESIGN/SETTING: A retrospective comparative study at a single academic institution. PATIENT SAMPLE: Patients 55 years or older that underwent a spinal fusion procedure in 2017 at a single academic institution. OUTCOME MEASURES: Occurrence of proximal junctional kyphosis, proximal junctional failure, pseudarthrosis, screw loosening or pullout, hardware failure and adjacent segment disease (ASD).
METHOD(S): Baseline preoperative demographic information, smoking history, levels fused, UIV and status as a revision procedure were recorded. All postoperative notes and images for 2 years post-procedure were reviewed for the presence of proximal junctional kyphosis, proximal junctional failure, pseudarthrosis, screw loosening or pullout, hardware failure, and ASD. HUs were measured via regions of interest drawn within the cancellous bone of the mid-vertebral body at the UIV of all patients. Patients were divided into 2 groups for comparison, those who experienced a DRC within 2 years and those who did not. Student's t-test was performed to compare HUs between the groups, chi-square analysis was performed for categorical variables. Dichotomous logistical regression was performed to analyze the relationship between density related complications and HU at the UIV, patient BMI, revision procedure, history of smoking, gender, UIV and number of levels fused. Significance was set at p<0.05.
RESULT(S): A total of 172 consecutive fusion patients with a preoperative CT scan were reviewed. Of these, 49 were revision procedures. 66 had a UIV in the cervical spine, 10 had a UIV in the thoracic spine and 95 had a UIV in the lumbar spine. Ninety-nine were 1 or 2 level fusions, 49 were 3 or 4 level fusions and 23 were long fusions with 4+ levels involved. Forty-eight patients had a DRC. Baseline demographics were similar between the 2 groups, with the exception of more revision procedures in the DRC group (p<0.001). The mean HUs of the UIV in the cohort that had a DRC was 168.92, as compared to 252.66 in the no-DRC group (p<0.001). Regression analysis revealed that low HUs at the UIV and revision procedures were independent risk factors for a DRC. For every 10 unit decrease in HUs, the odds of a DRC rose by 6%. When thoracic and lumbar fusions were analyzed the mean HUs at the UIV in the DRC group were 108.5 vs 152.6 (p<0.001). When cervical fusions were analyzed separately the mean HUs in the DRC group were 308 vs 383.4 (P=0.014).
CONCLUSION(S): To our knowledge, this is the first study that compares HUs measured at the UIV to the rate of density related complications for single and multilevel fusions in the cervical, thoracic and lumbar spine. This study found that HUs measured at the UIV of a fusion were significantly lower in patients that went on to have a density related complication within 2 years of their index procedure. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Surgeons may preferentially limit fusion levels for younger vs older ASD patients to maintain motion segments and optimize postoperative function. Few data exist comparing the functional impact of upper thoracic (UT) vs thoracolumbar (TL) upper instrumented vertebra (UIV) in younger vs older ASD patients undergoing long fusion to the pelvis. PURPOSE: Evaluate patient reported functional impact of UT vs TL UIV in younger vs older ASD patients undergoing long fusion to the pelvis. STUDY DESIGN/SETTING: Propensity score matched (PSM) analysis of ASD patients prospectively enrolled into a multicenter study. PATIENT SAMPLE: Surgically treated ASD patients prospectively enrolled into multicenter study. OUTCOME MEASURES: Numeric rating scale (NRS) back and leg pain, Scoliosis Research Society-22r questionnaire (SRS-22r), Short Form-36v2 questionnaire (SF-36), Oswestry Disability Index (ODI), estimated blood loss, duration of hospital stay, postoperative complications, revision spine surgery.
METHOD(S): Surgically treated ASD patients prospectively enrolled into a multicenter ASD study were divided into 2 age groups (younger= <65 years, older= > 65 years) and separated according to UIV (TL= L2-T7; UT= T6-T1). Study inclusion criteria; 1) surgery for Lumbar (L), Sagittal (S), or Mixed (M) deformities (as per SRS-Schwab ASD classification), 2) fusion to the pelvis, 3) minimum 5 levels fused, and 4) minimum 2 year postop follow up. Surgery for double major or thoracic scoliosis were excluded. PSM was used to match preop patient demographics, scoliosis, and sagittal spinopelvic parameters including PI-LL, TK, SVS, and TPA. Surgical data evaluated and impact of UIV upon patient reported functional outcomes compared for UT vs TL for younger vs older.
RESULT(S): From 2008-2018, 435 of 717 eligible surgically treated patients were evaluated; younger (n=193; mean age 57.6 years) and older (n=242; mean age 72.3 years), mean levels fused UT=17.4, TL=10.7 (p<0.05). Preop spine deformity, demographics, and performance of osteotomies were similar for matched UT vs TL in younger and older (p>0.05). Surgical blood loss, duration of SICU and hospital stay was greater for UT vs TL in younger and older (p<0.05). UT had more revision surgery than TL due to implant failures in younger (20% vs 3%) and older (16% vs 1%), respectively (p<0.05). Older UT had more major complications than older TL (65% vs 30%). At minimum 2 year postoperative follow up spine deformity correction and all PROMs (including improvements and final values) including SRS-22r activity, SF-36 physical function, SF-36 role physical, SF-36 social function and SF-36 vitality were similar UT vs TL in younger and older (p>0.05).
CONCLUSION(S): Younger ASD patients fused to the pelvis do not report more physical restrictions for UT vs TL UIV compared to older ASD patients, however blood loss, SICU and hospital stay and complications were greater for UT vs TL in younger and older cohorts. When deciding upon UIV for ASD patients, the minimal impact of UT vs TL UIV upon patient reported outcomes must be offset by cognizance of a longer hospital stay and potential for greater postoperative complications for UT fusions especially in older patients. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: As ASD prevalence increases in our ever-aging population there is a hypothetical concomitant increase in poor bone quality, especially if not recognized and not treated. ASD surgery is expensive and carries a high complication profile. It is important to optimize surgical outcomes and reduce complications especially if modifiable preoperative risk factors can be identified, such as osteoporosis. Additional diagnostic modalities such as a DEXA can add cost, delay diagnosis, and can be an additional insurance hurdle. PURPOSE: Our goal was to examine the utility of HU measurement on preoperative CTs for bone health assessment. STUDY DESIGN/SETTING: Retrospective cross-sectional review of a prospective, multicenter ASD cohort. PATIENT SAMPLE: Surgical ASD patients. OUTCOME MEASURES: Hounsfield Units, history of osteoporosis, DEXA results.
METHOD(S): Operative ASD patients (scoliosis >20, SVA>5cm, PT>25, or TK>60) were included if they had a preoperative CT. HU were measured by each participating site from axial views within the cancellous body (x3: top, middle, bottom) at both L1 and future UIV. Reliability of the measurement between the 3 acquisitions was performed using instar-class correlation for absolute agreement. Association between HU and patient demographics was assess using Pearson's correlation. Finally, correlation between DEXA measurement and HU was conducted to evaluate relationship between bone quality and HU values.
RESULT(S): There were 694/1493 (46%) patients who had a CT including either L1 or UIV. And 521 patients were identified as having both L1 and UIV measurement. Also, 71.8% were female with a mean age of 63years+/-12.5, 52.6% were revision with mean levels fused of 10.5+/-4.5. The intraclass correlation coefficient (ICC) for UIV and L1 were 0.767 (95CI 0.737-0.796]) and 0.802 (95CI [0.774 0.827]), respectively. Previous instrumentation did not affect L1 HU ICC (r=0.798 vs r=0.809) and showed no significant difference in HU value (p=0.232). Comparison of L1 HU between different sites demonstrated no significant difference (p=0.43). Comparison of L1 and UIV did show a significant difference (L1:151+/-77 vs 160+/-62 p<0.001) although there was a significant correlation (r=0.631 p<0.001). The mean HU value at L1 was consistent with previously published values (p=0.542). There were 116 (22.5%) patients who had a DEXA and 97 (18.6%) patients reported a history of osteoporosis. Comparison of DEXA and HU between patients with and without history of osteoporosis showed a significant difference in HU (155+/-76 vs 134+/-79 p<0.001) and but not in DEXA (p=0.07). A significant but weak association between DEXA and HU measurements (r=0.286 & 0.285 p<0.002). HU did not correlate with baseline demographic parameters such as BMI, CCMI, or frailty but did correlate with age (p<0.009 r=-0.215). Similarly, DEXA did not correlate with baseline demographic parameters except for BMI (p<0.002,r=0.298).
CONCLUSION(S): In this large cohort of surgical ASD patients, bone quality assessment was available for 18% of patients via DEXA or 46% via HU on CT. HU measured from an axial image of L1 and UIV appears to be a reliable assessment of bone quality. Previous instrumentation did not alter the measurements. There was a significant but weak correlation when comparing HU to DEXA. Patients with a reported history of osteoporosis had lower HU. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

OBJECTIVE/UNASSIGNED:Deterioration of global coronal alignment (GCA) may be associated with worse outcomes after adult spinal deformity (ASD) surgery. The impact of fusion length and upper instrumented vertebra (UIV) selection on patients with this complication is unclear. The authors' objective was to compare outcomes between long sacropelvic fusion with upper-thoracic (UT) UIV and those with lower-thoracic (LT) UIV in patients with worsening GCA ≥ 1 cm. METHODS:This was a retrospective analysis of a prospective multicenter database of consecutive ASD patients. Index operations involved instrumented fusion from sacropelvis to thoracic spine. Global coronal deterioration was defined as worsening GCA ≥ 1 cm from preoperation to 2-year follow-up. RESULTS:Of 875 potentially eligible patients, 560 (64%) had complete 2-year follow-up data, of which 144 (25.7%) demonstrated worse GCA at 2-year postoperative follow-up (35.4% of UT patients vs 64.6% of LT patients). At baseline, UT patients were younger (61.6 ± 9.9 vs 64.5 ± 8.6 years, p = 0.008), a greater percentage of UT patients had osteoporosis (35.3% vs 16.1%, p = 0.009), and UT patients had worse scoliosis (51.9° ± 22.5° vs 32.5° ± 16.3°, p < 0.001). Index operations were comparable, except UT patients had longer fusions (16.4 ± 0.9 vs 9.7 ± 1.2 levels, p < 0.001) and operative duration (8.6 ± 3.2 vs 7.6 ± 3.0 hours, p = 0.023). At 2-year follow-up, global coronal deterioration averaged 2.7 ± 1.4 cm (1.9 to 4.6 cm, p < 0.001), scoliosis improved (39.3° ± 20.8° to 18.0° ± 14.8°, p < 0.001), and sagittal spinopelvic alignment improved significantly in all patients. UT patients maintained smaller positive C7 sagittal vertical axis (2.7 ± 5.7 vs 4.7 ± 5.7 cm, p = 0.014). Postoperative 2-year health-related quality of life (HRQL) significantly improved from baseline for all patients. HRQL comparisons demonstrated that UT patients had worse Scoliosis Research Society-22r (SRS-22r) Activity (3.2 ± 1.0 vs 3.6 ± 0.8, p = 0.040) and SRS-22r Satisfaction (3.9 ± 1.1 vs 4.3 ± 0.8, p = 0.021) scores. Also, fewer UT patients improved by ≥ 1 minimal clinically important difference in numerical rating scale scores for leg pain (41.3% vs 62.7%, p = 0.020). Comparable percentages of UT and LT patients had complications (208 total, including 53 reoperations, 77 major complications, and 78 minor complications), but the percentage of reoperated patients was higher among UT patients (35.3% vs 18.3%, p = 0.023). UT patients had higher reoperation rates of rod fracture (13.7% vs 2.2%, p = 0.006) and pseudarthrosis (7.8% vs 1.1%, p = 0.006) but not proximal junctional kyphosis (9.8% vs 8.6%, p = 0.810). CONCLUSIONS:In ASD patients with worse 2-year GCA after long sacropelvic fusion, UT UIV was associated with worse 2-year HRQL compared with LT UIV. This may suggest that residual global coronal malalignment is clinically less tolerated in ASD patients with longer fusion to the proximal thoracic spine. These results may inform operative planning and improve patient counseling.

OBJECTIVE:Few studies have compared fractional curve correction after long fusion between transforaminal lumbar interbody fusion (TLIF) and anterior lumbar interbody fusion (ALIF) for adult symptomatic thoracolumbar/lumbar scoliosis (ASLS). The objective of this study was to compare fractional correction, health-related quality of life (HRQL), and complications associated with L4-S1 TLIF versus those of ALIF as an operative treatment of ASLS. METHODS:The authors retrospectively analyzed a prospective multicenter adult spinal deformity database. Inclusion required a fractional curve ≥ 10°, a thoracolumbar/lumbar curve ≥ 30°, index TLIF or ALIF performed at L4-5 and/or L5-S1, and a minimum 2-year follow-up. TLIF and ALIF patients were propensity matched according to the number and type of interbody fusion at L4-S1. RESULTS:Of 135 potentially eligible consecutive patients, 106 (78.5%) achieved the minimum 2-year follow-up (mean ± SD age 60.6 ± 9.3 years, 85% women, 44.3% underwent TLIF, and 55.7% underwent ALIF). Index operations had mean ± SD 12.2 ± 3.6 posterior levels, 86.6% of patients underwent iliac fixation, 67.0% underwent TLIF/ALIF at L4-5, and 84.0% underwent TLIF/ALIF at L5-S1. Compared with TLIF patients, ALIF patients had greater cage height (10.9 ± 2.1 mm for TLIF patients vs 14.5 ± 3.0 mm for ALIF patients, p = 0.001) and lordosis (6.3° ± 1.6° for TLIF patients vs 17.0° ± 9.9° for ALIF patients, p = 0.001) and longer operative duration (6.7 ± 1.5 hours for TLIF patients vs 8.9 ± 2.5 hours for ALIF patients, p < 0.001). In all patients, final alignment improved significantly in terms of the fractional curve (20.2° ± 7.0° to 6.9° ± 5.2°), maximum coronal Cobb angle (55.0° ± 14.8° to 23.9° ± 14.3°), C7 sagittal vertical axis (5.1 ± 6.2 cm to 2.3 ± 5.4 cm), pelvic tilt (24.6° ± 8.1° to 22.7° ± 9.5°), and lumbar lordosis (32.3° ± 18.8° to 51.4° ± 14.1°) (all p < 0.05). Matched analysis demonstrated comparable fractional correction (-13.6° ± 6.7° for TLIF patients vs -13.6° ± 8.1° for ALIF patients, p = 0.982). In all patients, final HRQL improved significantly in terms of Oswestry Disability Index (ODI) score (42.4 ± 16.3 to 24.2 ± 19.9), physical component summary (PCS) score of the 36-item Short-Form Health Survey (32.6 ± 9.3 to 41.3 ± 11.7), and Scoliosis Research Society-22r score (2.9 ± 0.6 to 3.7 ± 0.7) (all p < 0.05). Matched analysis demonstrated worse ODI (30.9 ± 21.1 for TLIF patients vs 17.9 ± 17.1 for ALIF patients, p = 0.017) and PCS (38.3 ± 12.0 for TLIF patients vs 45.3 ± 10.1 for ALIF patients, p = 0.020) scores for TLIF patients at the last follow-up (despite no differences in these parameters at baseline). The rates of total complications were similar (76.6% for TLIF patients vs 71.2% for ALIF patients, p = 0.530), but significantly more TLIF patients had rod fracture (28.6% of TLIF patients vs 7.1% of ALIF patients, p = 0.036). Multiple regression analysis demonstrated that a 1-mm increase in L4-5 TLIF cage height led to a 2.2° reduction in L4 coronal tilt (p = 0.011), and a 1° increase in L5-S1 ALIF cage lordosis led to a 0.4° increase in L5-S1 segmental lordosis (p = 0.045). CONCLUSIONS:Operative treatment of ASLS with L4-S1 TLIF versus ALIF demonstrated comparable mean fractional curve correction (66.7% vs 64.8%), despite use of significantly larger, more lordotic ALIF cages. TLIF cage height had a significant impact on leveling L4 coronal tilt, whereas ALIF cage lordosis had a significant impact on restoration of lumbosacral lordosis. The advantages of TLIF may include reduced operative duration and hospitalization; however, associated HRQL was inferior and more rod fractures were detected in the TLIF patients included in this study.

BACKGROUND:Operative treatment of adult spinal deformity (ASD) can be very challenging with high complication rates. It is well established that patients benefit from such treatment; however, the surgical outcomes for patients with severe sagittal deformity have not been reported. OBJECTIVE:To report the outcomes of patients undergoing surgical correction for severe sagittal deformity. METHODS:Retrospective review of a prospective, multicenter ASD database. Inclusion criteria: operative patients age ≥18, sagittal vertical axis (SVA) ≥15 cm, mismatch between pelvic incidence and lumbar lordosis (PI-LL) ≥30°, and/or lumbar kyphosis ≥5° with minimum 2 yr follow-up. Health-related quality of life (HRQOL) scores including minimal clinically important difference (MCID)/substantial clinical benefit (SCB), sagittal and coronal radiographic values, demographic, frailty, surgical, and complication data were collected. Comparisons between 2 yr postoperative and baseline HRQOL/radiographic data were made. P < .05 was significant. RESULTS:A total of 138 patients were included from 502 operative patients (54.3% Female, Average (Avg) age 63.3 ± 11.5 yr). Avg operating room (OR) time 386.2 ± 136.5 min, estimated blood loss (EBL) 1829.8 ± 1474.6 cc. A total of 71(51.4%) had prior fusion. A total of 89.9% were posterior fusion only. Mean posterior levels fused 11.5 ± 4.1. A total of 44.9% had a 3-column osteotomy. All 2 yr postoperative radiographic parameters were significantly improved compared to baseline (P < .001 for all). All 2yr HRQOL measures were significantly improved compared to baseline (P < .004 for all). A total of 46.6% to 73.8% of patients met either MCID/SCB for all HRQOL. A total of 74.6% of patients had at least 1 complication, 11.6% had 4 or more complications, 33.3% had minimum 1 major complication, and 42(30.4%) had a postop revision. CONCLUSION/CONCLUSIONS:Patients with severe sagittal malalignment benefit from surgical correction at 2 yr postoperative both radiographically and clinically despite having a high complication rate.

PURPOSE/OBJECTIVE:AI algorithms have shown promise in medical image analysis. Previous studies of ASD clusters have analyzed alignment metrics-this study sought to complement these efforts by analyzing images of sagittal anatomical spinopelvic landmarks. We hypothesized that an AI algorithm would cluster preoperative lateral radiographs into groups with distinct morphology. METHODS:This was a retrospective review of a multicenter, prospectively collected database of adult spinal deformity. A total of 915 patients with adult spinal deformity and preoperative lateral radiographs were included. A 2 × 3, self-organizing map-a form of artificial neural network frequently employed in unsupervised classification tasks-was developed. The mean spine shape was plotted for each of the six clusters. Alignment, surgical characteristics, and outcomes were compared. RESULTS:Qualitatively, clusters C and D exhibited only mild sagittal plane deformity. Clusters B, E, and F, however, exhibited marked positive sagittal balance and loss of lumbar lordosis. Cluster A had mixed characteristics, likely representing compensated deformity. Patients in clusters B, E, and F disproportionately underwent 3-CO. PJK and PJF were particularly prevalent among clusters A and E. Among clusters B and F, patients who experienced PJK had significantly greater positive sagittal balance than those who did not. CONCLUSIONS:This study clustered preoperative lateral radiographs of ASD patients into groups with highly distinct overall spinal morphology and association with sagittal alignment parameters, baseline HRQOL, and surgical characteristics. The relationship between SVA and PJK differed by cluster. This study represents significant progress toward incorporation of computer vision into clinically relevant classification systems in adult spinal deformity. LEVEL OF EVIDENCE IV/UNASSIGNED:Diagnostic: individual cross-sectional studies with the consistently applied reference standard and blinding.

STUDY DESIGN/METHODS:Retrospective review. OBJECTIVES/OBJECTIVE:The aim of this study was to define a surgical invasiveness threshold that predicts major complications after adult spinal deformity (ASD) surgery; use this threshold to categorize patients into quartiles by invasiveness; and determine the odds of major complications by quartile. SUMMARY OF BACKGROUND DATA/BACKGROUND:Understanding the relationship between surgical invasiveness and major complications is important for estimating the likelihood of major complications after ASD surgery. METHODS:Using a multicenter database, we identified 574 ASD patients (more than 5 levels fused; mean age, 60 ± 15 years) with minimum 2-year follow-up. Invasiveness was calculated as the ASD Surgical and Radiographic (ASD-SR) score. Youden index was used to identify the invasiveness score cut-off associated with optimal sensitivity and specificity for predicting major complications. Resulting high- and low-invasiveness groups were divided in half to create quartiles. Odds of developing a major complication were analyzed for each quartile using logistic regression (alpha = 0.05). RESULTS:The ASD-SR cutoff score that maximally predicted major complications was 90 points. ASD-SR quartiles were 0 to 65 (Q1), 66 to 89 (Q2), 90 to 119 (Q3), and ≥120 (Q4). Risk of a major complication was 17% in Q1, 21% in Q2, 35% in Q3, and 33% in Q4 (P < 0.001). Comparisons of adjacent quartiles showed an increase in the odds of a major complication from Q2 to Q3 (odds ratio [OR] 1.8; 95% confidence interval [CI]: 1.0-3.0), but not from Q1 to Q2 or from Q3 to Q4. Patients with ASD-SR scores ≥90 were 1.9 times as likely to have a major complication than patients with scores <90 (OR 1.9, 95% CI 1.3-2.9). Mean ASD-SR scores above and below 90 points were 121 ± 25 and 63 ± 17, respectively. CONCLUSION/CONCLUSIONS:The odds of major complications after ASD surgery are significantly greater when the procedure has an ASD-SR score ≥90. ASD-SR score can be used to counsel patients regarding these increased odds.Level of Evidence: 3.