Faculty Bibliography

BACKGROUND CONTEXT: Intraoperative CT image-guided navigation (IGN) and robotic assistance have been increasingly implemented during spine surgery to improve accuracy in pedicle screw positioning. However, studies have shown that they increase operative time and ionizing radiation exposure, and it remains controversial whether they improve patient outcomes. PURPOSE: Assess use of IGN and robotic assistance in posterior lumbar surgery and their relationship with patient radiation exposure and perioperative outcomes. STUDY DESIGN/SETTING: Retrospective cohort study at single institution. PATIENT SAMPLE: A total of 165 patients undergoing 1- or 2-level posterior spinal fusion, with or without TLIF. OUTCOME MEASURES: Preoperative CT scan utilization and radiation dose, intraoperative radiation dose (fluoroscopy and/or CT) and total-procedure radiation dose (sum of preoperative CT and intraoperative radiation doses), levels fused, operative time, estimated blood loss (EBL), length of stay (LOS), postoperative complications.
METHOD(S): Patients >=18 years old undergoing 1- or 2-level posterior spinal fusion, with or without TLIF, in a 12-month period included. Chart review performed for pre- and intraoperative data on radiation dose and perioperative outcomes. All radiation doses quantified in milli-Grays (mGy). Univariate analysis and multivariate logistic regression analysis for categorical variables and one-way ANOVA for continuous variables utilized, with significance set at p<0.05.
RESULT(S): A total of 165 patients (51.83% F, 59.13+/-13.18yrs, BMI 29.43+/-6.72, Charlson comorbidity index [CCI] 1.20+/-1.56) were assessed: 12 IGN, 62 robotic, 56 open, 35 MIS without IGN/robotics. Lower proportion of women in open and MIS group (66.67% F IGN, 64.52% robotic, 38.18% open, 45.71% MIS, p=0.010). Younger patients in MIS group (63.42 yrs IGN, 61.74 robotic, 60.63 open, 50.63 MIS, p<0.001). MIS group had lowest mean posterior levels fused (1.42 IGN, 1.27 robotic, 1.32 open, 1.06 MIS, p=0.015). Total-procedure radiation (50.21mGy IGN, 59.84 robotic, 22.56 open, 82.02 MIS), total-procedure radiation/level fused (41.88mGy IGN, 51.18 robotic, 18.56 open, 79.41 MIS) and intraop radiation (44.69mGy IGN, 44.85 robotic, 14.81 open, 80.28 MIS) were lowest in the open group and highest in the MIS group compared to IGN and robotic (all p<0.001). A higher proportion of robotic and lower proportion MIS patients had preop CT (25% IGN, 82.26% robotic, 37.5% open, 8.57% MIS, p<0.001). EBL (441.67mL IGN, 380.24 robotic, 355.36 open, 162.14 MIS, p=0.002) and LOS (4.75 days IGN, 3.89 robotic, 3.89 open, 2.83 MIS, p=0.039) were lowest in the MIS group. Highest operative time for IGN patients (303.5 min vs 264.85 robotic, 229.91 open, 213.43 MIS p<0.001). No differences in BMI, CCI, postoperative complications (p=0.313, 0.051, 0.644, respectively).
CONCLUSION(S): IGN and robotic assistance in posterior lumbar fusion were associated with higher intraoperative and total-procedure radiation exposure than open cases without IGN/robotics, but significantly less than MIS without IGN/robotics, without differences in perioperative outcomes. While MIS procedures reported highest radiation exposure to patient, of more concern is that the proportion of total radiation dose applied to surgeon would also be considerably higher in MIS group. FDA DEVICE/DRUG STATUS: Brainlab Airo Mobile Intraoperative CT scanner (Approved for this indication).
Copyright

BACKGROUND CONTEXT: Distal junctional kyphosis (DJK) is a relatively new surgical concept often defined in the literature as a change in DJK angle <-10. This cut-off (based on previously established cut-offs for PJK) may carry clinical significance for patient quality of life; however, findings are lacking in the literature. When associated with neurological sequelae or reoperation, DJK may play a more robust role in patient HRQLs. PURPOSE: To establish a clinically relevant definition of symptomatic DJK. STUDY DESIGN/SETTING: Retrospective review of a prospective CD surgery database. PATIENT SAMPLE: One hundred and two patients with cervical deformity. OUTCOME MEASURES: Demographics, HRQL metrics.
METHOD(S): A prospective database of operative CD patients was analyzed. Inclusion criteria were cervical kyphosis >10, cervical scoliosis >10, cSVA >4cm or CBVA >25. DJK angle (DJKA) was defined >10degree change in kyphosis between LIV and LIV-2 in addition to >10degree index angle. Pts with DJKA >10degree & 1) no reop due to DJK(DJF) & 2) no physician-reported neurological sequelae at any time-point up to 1yr were categorized as "mild DJK." Those with a DJKA >10degree in addition to 1) reop due to DJK(DJF) or 2) >1 new-onset neurological sequelae related to DJK (spinal cord deficit, gait disturbance, hyperreflexia, lower-limb spasticity) were categorized as "Symptomatic DJK." ANOVA explored differences in PROMs (NDI, mJOA, EQ5D, EQ5DVAS, NRS) at baseline (BL) up to 1yr follow-up. Multivariate logistic regression analyzed predictors of inferior HRQLs associated with DJK groups, specifically scoring beyond one standard deviation (SD) of the mean value of all DJK pts.
RESULT(S): One hundred and two CD patients were analyzed. Sixty-five pts had no DJK, 25 pts had mild DJK and 12 pts had symptomatic DJK. By 1 year, 6/12 pts had DJF, 6/12 pts had >1 DJK-related neurologic sequelae not present before surgery(20% corticospinal tract deficit, 20% spasticity, 20% gait impairment, 20% hyperreflexia) without DJF, and 2 pts had both. Symptomatic DJK pts had higher mean DJKA than mild DJK pts immediately postop (29.8degree vs 19.1degree, p=0.150). No differences in BL HRQLs were noted between groups, including NDI, mJOA, EQ5D, VAS, and NRS Back (all p>0.05). At 1yr, Symptomatic DJK pts had significantly higher disability (52.3 vs 28.7, p=0.006) and trended lower EQ5D scores (0.75 vs 0.81, p=0.059), higher NRS back pain (5.82 vs 4.06, p=0.119), lower VAS (56.4 vs71.7,p=0.084), and lower mJOA (14.2 vs 15.0,p=0.495) than their mild DJK counterparts. Controlling for age and gender, conditional forward regression analysis revealed symptomatic DJK to a strong predictor of NDI >1 SD of the mean compared to mild DJK pts (OR: 43.4 [2.8 - 668.1], p=0.007); that is, symptoDJK increased the odds of NDI>1SD by at least 2.8x in relation to the mild DJK cohort.
CONCLUSION(S): DJK associated with reoperation and/or neurologic sequelae was associated with inferior quality of life metrics over long-term follow-up. Our results demonstrate that Symptomatic DJK may not necessarily correlate to a greater measured DJK angle per se (ie, >20degree); instead, DJK with concurrent neuro findings or subsequent reop better predicts worsened disability compared to DJK alone. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: The operative treatment of adult spinal deformity (ASD) can be very technically challenging with high complication rates. Despite these challenges it is well established that patients ultimately benefit from such treatment. However the surgical outcomes for patients with severe sagittal deformity have not been reported. PURPOSE: To investigate whether patients with severe sagittal deformity have favorable outcomes following surgical correction. STUDY DESIGN/SETTING: Retrospective review of a prospective, multicenter adult spinal deformity (ASD) database. PATIENT SAMPLE: Inclusion criteria: operative patients with age >=18, SVA >=15cm, PI-LL >=30degree, and/or lumbar kyphosis >=5degree with minimum 2-yr follow-up. OUTCOME MEASURES: Health-related quality of life (HRQOL) scores included: Oswestry Disability Index (ODI), Short Form-36 (SF36), Scoliosis Research Society (SRS22), back/leg pain numerical rating scale (NRS) and minimum clinically important difference (MCID)/substantial clinical benefit (SCB) for pts eligible to meet them. Radiographic values included: max coronal cobb angle, coronal C7 plumb line, pelvic tilt (PT), mismatch between pelvic incidence and lumbar lordosis (PI-LL), thoracic kyphosis (TK), C7 sagittal vertical axis (SVA). Demographic, frailty, surgical, and complications data were also collected.
METHOD(S): Comparisons between 2-yr postop and baseline HRQOL/radiographic data were made. P<0.05 was considered significant.
RESULT(S): A total of 138 pts were included (54.3% Female, ave age 63.3+/-11.5yrs). Ave baseline frailty score was 4.1+/-1.4 indicating the pt to befrail. Ave OR time 386.2+/-136.5min, EBL 1829.8+/-1474.6cc. 71 (51.4%) of the pts had a prior fusion; 89.9% posterior fusion only, 9.4% combined anterior-posterior approach. Mean number posterior levels fused 11.5+/-4.1. 44.9% had a 3-column osteotomy and 68.8% had interbody fusion. All 2-yr postop radiographic parameters were significantly improved compared to baseline (p<0.05 for all) except coronal C7 plumb line (p>0.05). All 2-yr HRQOL measures were significantly improved compared to baseline (p<0.004 for all); 46.6-73.8% of pts met either MCID or SCB for all HRQOL; 74.6% of pts had at least 1 complication, 11.6% had 4 or more complications, 33.3% had at least 1 major complication, and 42 (30.4%) had a postop revision.
CONCLUSION(S): Pts with severe sagittal malalignment benefit from surgical correction at 2-yrs postop both radiographically and clinically despite having a high complication rate. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Analysis of interactions spinal alignment metrics may uncover novel and powerful alignment parameters, similar to pelvic incidence minus lumbar lordosis (PI-LL). PURPOSE: This study utilized a data driven approach to hypothesis generation by testing all possible division interactions between a set of spinal alignment parameters. TPA divided by PI (TPA/PI) was identified and subsequently characterized. STUDY DESIGN/SETTING: Retrospective analysis of a multicenter, prospectively-defined, consecutive cohort of ASD patients. PATIENT SAMPLE: A total of 1,439 patients with baseline ODI were included for hypothesis generation; 666 patients had 2-year postoperative follow-up and were included for validation. OUTCOME MEASURES: The primary outcome measure was Oswestry Disability Index (ODI).
METHOD(S): All possible combinations of division interactions between baseline metrics were assessed with linear regression against baseline ODI. Further analyses also utilized linear regression.
RESULT(S): From 247 raw alignment metrics, 32,398 division interactions were considered in hypothesis generation. Low-cervical and high-thoracic pelvic angles divided by pelvic incidence accounted for 7 of the top 10 most significant interactions. Conceptually, the TPA/PI is a measure of the relative alignment of the line connecting T1 to the femoral head and the line perpendicular to the sacral end plate. Among the validation cohort, the mean TPA/PI was 0.41 (SD 0.24) at baseline and 0.30 at 2 years postoperatively (SD 0.18). Higher TPA/PI was linearly associated with worse ODI at baseline (beta 28.7 for TPA/PI 1 vs 0, p<0.0001), 1 year (beta 16.5, p<0.0001), and 2 years (beta 20.0, p<0.0001). Further, the change in ODI at 2 years from baseline was linearly associated with the change in TPA/PI (beta 9.5, p=0.0172). The optimal statistical cutpoints for TPA/PI discriminating patients based on ODI were 0.21 and 0.41, labeled as 0 (<0.21), + (0.21-0.41), and ++ (>=0.41). At baseline, patients in the 0 group exhibited lower ODI as compared to patients in the + (41.8 vs 32.0, p<0.0001) and ++ (50.7 vs 32.0, p<0.0001) groups. The R-squared for regression of ODI against categorical TPA/PI alone (0.154) was directionally higher than that for the Schwab modifiers (SVA: 0.138, PI-LL 0.111, PT 0.057).
CONCLUSION(S): This study utilized a data-driven approach for hypothesis generation and identified TPA/PI, the spinopelvic ratio (SPR), as a promising measure of sagittal spinal alignment among ASD patients. This value is strongly associated with the Schwab modifier classification system. Patients with SPR >=0.21 exhibited inferior HRQOL. The SPR classification accounted for a directionally greater proportion of variance in ODI as compared to the three Schwab modifiers. These results highlight the interaction between global sagittal alignment and adequate pelvic structural support. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Patients with severe cervical deformity (CD) often have profound deficits in numerous activities of daily living. Association between health status and postoperative radiographic goals is difficult to quantify. PURPOSE: Investigate the radiographic characteristics of patients who achieved optimal HRQOLs following surgery for CD. STUDY DESIGN/SETTING: Retrospective review of prospective CD database. PATIENT SAMPLE: Forty adult CD patients (C2-C7 Cobb>10degree, CL>10degree, cSVA>4cm, or CBVA>25degree). OUTCOME MEASURES: Health-related-quality of life (HRQL) scores, radiographic alignment.
METHOD(S): Surgical patients with severe CD were isolated based upon a previously presented combination of preop cSVA, T1 slope, maximum focal kyphosis in extension, C2 slope in extension, and number of kyphotic levels in extension. Those with available preop and 1-year postop HRQL data were included. Based on previous study, patients were grouped into 3 distinct patterns of CD: focal deformity (FD), flatneck (FN=large TS-CL and lack of compensation), or cervicothoracic (CT). Postoperative outcomes were defined as "good" if a patient had >2 of the 3 following criteria (1) NDI <20 or meeting MCID, (2) mild myelopathy (mJOA>=13), and (3) NRS-Neck <=5 or improved by >2 points from baseline. Within each distinct deformity group, patients with good outcomes were compared to those with poor outcomes (ie, not meeting the criteria for good) for differences in demographics, HRQL scores, and alignment, via chi-squared or student's t-tests.
RESULT(S): Overall, 83/153 patients met the criteria of severe CD and 40 patients had complete 1-year follow-up of clinical/radiographic data. Patient breakdown by deformity pattern was as follows: CT (N=13), FN (N=17), and FD (N=17), with 7 patients meeting criteria for both FD and FN deformities. Within the FD cohort, maximal focal kyphosis (ie, kyphosis at one level) was better corrected in patients with a "good" outcome (p = 0.03). In the FN cohort, patients who went on to have "good" outcomes presented with worse horizontal gaze (McGregor slope 21degree vs 6degree, p=0.061) and cSVA (72 mm vs 60 mm, p=0.030). "Good" outcome FN patients showed significantly greater postop correction of horizontal gaze (-25degree vs -5degree, p = 0.031). In the CT cohort, patients with "good" outcomes had superior global alignment both pre- (SVA: -17 mm vs 108 mm, p <0.001) and postoperatively (50 mm vs 145 mm, p=0.001). CT patients with "good" outcomes also had better postop cervical alignment (cSVA 35 mm vs 49 mm, p=0.030), and less kyphotic segments during extension (p=0.011). In the FD cohort, there were no differences between "good" and "poor" outcomes patients in preoperative alignment; however, "good" outcome patients showed superior changes in postoperative focal kyphosis (-2degree vs 5degree, p=0.030). Within all three deformity pattern categories, there were no differences between "good" and "poor" outcome patients with respect to demographics or surgical parameters (levels fused, surgical approach, decompression, osteotomy, all p>0.050).
CONCLUSION(S): The results of this study highlight intraoperative goals for three distinct sagittal morphotypes of severe CD: cervicothoracic, flat neck, and focal deformity. Distinct deformity specific intraoperative goals include obtaining proper sagittal global/cervical alignment for cervicothoracic patients, correcting maximal focal kyphosis in focal deformity patients, and correcting horizontal gaze for flat neck patients. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Persistent lumbopelvic malalignment following ASD-corrective surgery may impair quality of life and result in persistent pathologic compensation in the lower extremities. Patient-specific age- and BMI-adjusted alignment targets have been proposed to improve alignment outcomes; however, it is unclear whether reaching these postop targets reduces rates of pelvic nonresponse following surgery. PURPOSE: Assess the relationship between pelvic nonresponse to ASD-corrective surgery and persistent lower-extremity compensation. STUDY DESIGN/SETTING: Single center retrospective review. PATIENT SAMPLE: Fifty-eight ASD patients. OUTCOME MEASURES: Sagittal alignment.
METHOD(S): Included: surgical ASD patients with full-spine X-ray imaging at pre- and early postop follow-up (<1Y). Patients were grouped by postop improvement in PT, per SRS-Schwab Classification: those who did not improve (pelvic nonresponders, PNR), and those that did improve (pelvic responders, PR). Groups were propensity score matched for preop PT, and assessed for differences in demographics, surgical factors, and alignment (sagittal spinal and lower extremity) with means comparison tests. Rates of persistent postop lower extremity compensation (defined as no improvement in lower extremity alignment) were compared between groups. Subanalysis assessed the relationship between reaching postop age- and BMI-specific alignment targets and rates of pelvic nonresponse.
RESULT(S): Following propensity score matching, PNR (N=29) and PR (N=29) patients did not differ in age, sex, BMI or preop sagittal spinal alignment (all p>0.05); however, PNR patients presented with less knee flexion (9degree vs 14degree, p=0.043). Groups did not differ in levels fused (10.8 vs 10.8, p=0.974) or osteotomy (93% vs 92%, p=0.902). Postop, PNR patients had inferior lumbopelvic alignment in PT (30degree vs 17degree), PI-LL (17degree vs 3degree), and greater global malalignment for TPA (27degree vs 15degree, all p<0.001). For PNR patients, these changes in alignment were accompanied by greater compensatory anterior hip extension (53mm vs 31mm, p=0.021). PNR patients also showed greater pre- to postop increases in sacrofemoral angle (2degree vs -5degree), and smaller decreases in hip extension (-24mm vs -64mm), pelvic femoral angle (-1.4degree vs -3.8degree), and global sagittal angle (-3.5degree vs -8degree, all p<0.005), indicating persistent lower extremity compensation. PNR patients had higher rates of persistent postop lower extremity compensation for sacrofemoral angle (68% vs 25%), ankle flexion (64% vs 33%), and pelvic shift (28% vs 4%, all p<0.034). PNR and PR groups did not differ in rates of reaching age- and BMI specific ideal postop alignment for PT, SVA, TPA, or PI-LL (all p>0.05). For patients with severe preop SVA deformity, overcorrection relative to ideal postop PT targets was associated with lower rates of pelvic non-response (under: 12%, match: 18%, over: 71%, p<0.001). Lower rates of nonresponse were observed for patients with severe preop PT deformity overcorrected relative to ideal postop PI-LL (under: 0%, match: 30%, over: 70%, p=0.016).
CONCLUSION(S): Pelvic nonresponders following ASD-corrective surgery had higher rates of persistent compensatory action in the lower extremities. Patients with severe preop PT deformity who were surgically overcorrected with respect to ideal PI-LL had lower rates of postop pelvic nonresponse, indicating that for severely malalignmed patients, existing alignment targets may need to be adjusted to optimize alignment outcomes. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Many patients are unable to undergo a major cervical deformity corrective surgery due to deformity severity, age, comorbidities, and overall frailty status. In order to optimize quality of life in patients with cervical deformity, there may be alignment targets to be prioritized. PURPOSE: To prioritize the cervical parameter targets for alignment. STUDY DESIGN/SETTING: Retrospective review of a multicenter prospective cervical deformity database. PATIENT SAMPLE: Seventy-seven patients undergoing cervical deformity (CD) corrective surgery. OUTCOME MEASURES: Cervical regional alignment parameters: cervical sagittal vertical axis(cSVA), cervical lordosis(CL), T1 Slope minus CL(TS-CL), chin brow vertical angle(CBVA), McGregor's slope(MGS), C2-T3 plumb line(C2-T3 SVA), C2-T3 angle, C2 Slope. Health-related quality of life measures: NDI.
METHOD(S): Included:CD patients (C2-C7 Cobb>10degree, CL>10degree, cSVA>4cm, or CBVA>25degree) with full baseline (BL) and 1-year (1Y) radiographic parameters and NDI scores; patients with cervical or cervicothoracic Primary Driver Ames type. Patients with BL Ames classified as low CD for both parameters of cSVA(<4cm) and TS-CL(<15degree) were excluded. Patients assessed: Meeting MCID for NDI(<-15 DELTANDI). Ratios of correction were found for regional parameters (cSVA, CL, T1 Slope, TS-CL, CBVA, MGS, C2-T3 SVA, C2-T3 angle, C2 Slope) categorized by Primary Ames Driver (cervical[C] or cervicothoracic[CT]). Decision tree analysis assessed cut-offs for differences associated with meeting NDI MCID at 1Y.
RESULT(S): Seventy-seven CD patients included (62.1yrs, 64%F, 28.8kg/m2). Average CCI: 0.94, 7% current smokers. By approach, anterior: 19.4%, posterior: 41.6%, combined approach: 39%. Mean anterior levels fused: 3.5, posterior: 8.3, total: 7.5. Average op time: 553.1min; mean EBL: 1128.1ccs. 41.6% met MCID for NDI. A backwards linear regression model including radiographic differences as predictors from BL to 1 year for meeting MCID for NDI demonstrated an R2 of 0.820 (p=0.032) included TS-CL, cSVA, MGS, C2SS, C2-T3 angle, C2-T3 SVA, CL. By primary Ames driver, 67.5% of patients were categorized as C, and 32.5% as CT. Ratios of change in predictors for MCID NDI patients (BL -1Y) for C driver patients: 260.8% MGS, 140.3% CL, 121.2% C2-T3 angle, 49.6% C2 slope, 41.1% cSVA, 20.5% TS-CL, 3.1%C2-T3 SVA. While correction in CT driver patients included: 168.7% CL, 93% MGS, 70.8% C2-T3 angle, 31.1% cSVA, 27.5% C2 slope, 24.9% TS-CL, 13.7% C2-T3 SVA. The ratios were not significant between the two groups(p>0.050). Decision tree analysis determined cut-offs for radiographic change, prioritizing in the following order (based upon ordinal regression values): a correction <=42.5degreeC2-T3 angle (OR: 5.667[1.074-29.891], p=0.041), <35.4degreeCL (OR:4.636[0.857-25.071], p=0.075), >-31.76degreeC2 slope (OR: 3.2 [0.852-12.026], p=0.085), >-11.57mm cSVA (OR: 3.185[1.137-8.917], p=0.027), >-2.16degree MGS (OR: 2.724[0.971-7.636], p=0.057).
CONCLUSION(S): Certain ratios of correction of cervical parameters contribute to improving neck disability. Specific cut-offs of radiographic differences from baseline to 1 year were found prioritizing C2-T3 angle, followed by cervical lordosis, C2 slope, C2-C7 plumb line, and McGregor's slope, all strongly associated with meeting the minimal clinically important difference for the neck disability index score. Prioritizing these radiographic alignment parameters will optimize patient-reported outcomes for patients undergoing cervical deformity surgery. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Adult spinal deformity (ASD) patients exhibit complex and highly variable pathology. The decision to manage patients operatively is subjective and varies based on training and experience. Machine learning algorithms have shown promise in supporting clinical decision-making. PURPOSE: We sought to develop models capable of accurately discriminating between patients receiving operative vs nonoperative treatment based only on baseline radiographic and clinical data at enrollment. STUDY DESIGN/SETTING: Retrospective analysis of a multicenter, prospectively-defined, consecutive cohort of ASD patients. PATIENT SAMPLE: A total of 1,503 patients were included in this study, divided in a 70:30 split for training and testing. OUTCOME MEASURES: Patients receiving operative treatment were defined as those undergoing surgery up to 1 year after their baseline visit.
METHOD(S): Potential predictors included available demographics, past medical history, HRQoL, and premeasured radiographic parameters from AP and lateral films. Variables with >10% missing data were discarded, and the remainder underwent median imputation. In total, 321 potential predictors were included. Random forest, elastic net regression, and support vector machines (SVMs) with radial and linear kernels were trained. Model performance was evaluated by area under the receiver operating characteristic curve (AUC).
RESULT(S): Of the patients, 69.0% (n=727) and 69.1% (n=311) in the training and testing sets received operative management, respectively. Upon evaluation with the testing dataset, performance for SVM linear (AUC=0.910), elastic net (0.913), and SVM radial (0.914) models was excellent, and the random forest model performed very well (0.830). In our SVM radial model, HRQoL metrics were particularly important for making predictions; the top 3 most important variables were SRS appearance, SRS total, and ODI.
CONCLUSION(S): This study developed models exhibiting excellent discrimination (AUC>0.9) between patients receiving operative vs nonoperative management, based solely on baseline enrollment values. HRQoL metrics were strongest in making these predictions. Future investigations may evaluate the implementation of such models for decision support in the clinical setting. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright

BACKGROUND CONTEXT: Lateral interbody fusion (LLIF) is often performed with biologic adjuvants to promote fusion. Commercially available bone allograft containing allogeneic stem cells (ASC) and bone morphogenetic protein-2 (BMP) are designed to promote fusion while avoiding the morbidity of iliac crest autograft; however, no study to date has directly compared the two in LLIF. The ASC studied is Osteocel Pro (NuVasive, Inc). PURPOSE: This non-industry funded study compares fusion rate, complications, and costs between LLIF with BMP and ASC. STUDY DESIGN/SETTING: Single center retrospective comparative study. PATIENT SAMPLE: Patients with 1-3 lumbar levels treated with LLIF. OUTCOME MEASURES: Outcomes measures are fusion at 1 year postoperative, complication rates, length of stay, and costs.
METHOD(S): A retrospective chart review was conducted to identify patients treated with LLIF and ASC or BMP from February 2012 through September 2017. Patients were included who had from 1-3 lumbar levels treated with LLIF and at least 1 year of radiographic follow up. Interbody fusion was assessed on lumbar X-ray images using a validated scale.
RESULT(S): A total of 94 patients were included representing 162 levels fused. Of these, 74 patients and 133 levels were treated with BMP; 20 patients and 29 levels were treated with ASC. Comparing patients treated with BMP or ASC, there were no differences in age [61.6 vs 60.4, p=0.7], BMI [29.8 vs 28.3, p=0.3], gender [60.8% vs 55.0% female], smoking status [12.2% vs 10.0%, p=1], diabetes [28.4% vs 15.0%, p=0.2], Charleston Comorbidity Index [4.3 vs 3.5, p=0.2], revision status [47.3% vs 45.0%, p=0.9], intraoperative complications [4.1% vs 5.0%, p=1], postoperative complications [37.8% vs 30.0%, p=0.5], or blood loss [881 vs 528ml, p=0.2]. More levels were fused in the BMP group (1.8 vs 1.45, p=0.04) and the BMP group tended toward a longer length of stay [4.8 vs 3.8 days, p=0.06]. There was a nonsignificant trend toward a higher fusion rate with BMP vs ASC[98.5% vs 93.1%, p=0.1]. The average amount of rhBMP used per level was 2.0 cc compared to 5.9 cc of ASC. There was no difference in the cost of the BMP per level compared with ASC [4.45% vs 4.80%, p=0.33], but the BMP group tended toward a higher cost of total care [103.5% vs 87.6%, p=0.1].
CONCLUSION(S): ASC and BMP are both acceptable adjuvants in LLIF that demonstrate comparable fusion rates at 1 year with comparable cost in the setting of similar groups of patients. The radiographic fusion rate seen in our study compares to previous reports in the literature using ASC. Cost considerations are becoming ever more cogent in spine surgery; the results of this study can inform decision making regarding which biologic adjuvant to use in lumbar interbody fusion. FDA DEVICE/DRUG STATUS: Osteocel (Approved for this indication), rhBMP (Infuse) (Not approved for this indication)
Copyright

BACKGROUND CONTEXT: Lumbar herniated nucleus pulposus (HNP) is a common spinal pathology often treated by microscopic lumbar discectomy (MLD), though prior reports have not demonstrated which preoperative MRI factors may contribute to significant clinical improvement after MLD. PURPOSE: To analyze the MRI characteristics in patients with HNP that predict meaningful clinical improvement in Health Related Quality of Life scores (HRQL) after MLD. STUDY DESIGN/SETTING: Retrospective clinical and radiological study of patients undergoing MLD for HNP at a single institution over a two year period of time. PATIENT SAMPLE: Eighty-eight patients receiving MLD treatment for HNP. OUTCOME MEASURES: Cephalocaudal canal migration; canal & HNP anterior-posterior (AP) lengths and ratio; canal & HNP axial areas and ratio; hemi-canal & hemi-HNP axial areas and ratio; disc appearance (black, grey or mixed), baseline (BL) and 3-month (3M) postoperative HRQL scores.
METHOD(S): Patients >18 years old who received MLD for HNP with BL and 3M HRQL scores of PROMIS (Physical Function, Pain Interference, and Pain Intensity), ODI, VAS Back, and VAS Leg scores were included. HNP and spinal canal measurements of cephalocaudal migration, AP length, area, hemi-area, and disc appearance were performed using T2 axial and sagittal MRI. HNP measurements were divided by corresponding canal measurements to calculate AP, Area, and Hemi-Area ratios. Using known minimal clinically importance differences (MCID) for each DELTAHRQoL score, patients were separated into two groups based on whether they reached MCID (+) or did not reach MCID (-). The MCID for Pain Intensity was calculated using a decision tree. A linear regression illustrated correlations between PROMIS vs ODI and VAS Back/Leg scores. Independent t-tests and chi-square tests were utilized to investigate significant differences in HNP measurements between the (+) and (-) MCID groups.
RESULT(S): Eighty-eight MLD patients were included (age=44.6+/-14.9, 38.6% female). Pain Interference and pain intensity were strongly correlated with ODI and VAS Back/Leg (R>=.505), and physical function was significantly correlated with ODI and VAS Back/Leg (R=-.349) (all p<.01). The strongest MRI predictors of meeting HRQL MCID were grey disc appearance, HNP area (>116.6 mm²), hemi-HNP Area (>84.6 mm²), and Hemi-Area Ratio (>51.8%); (+) patients were 2.7 times more likely to have a grey HNP than (-) patients in 5 out of 6 HRQL score comparisons (p<.025). Also, (+) patients had larger HNP areas than (-) patients had in 5 out of 6 HRQoL score comparisons (116.6 mm² +/- 46.4 vs 90.0 mm² +/- 43.2, p<.04), and had larger hemi-HNP areas than (-) patients had in 4 out of 6 HRQL score comparisons (84.6 mm² +/- 38.8 vs 66.3 mm² +/- 29.7, p<.04). (+) patients had a greater hemi-area ratio than (-) patients had in 4 out of 6 HRQL score comparisons (51.8% +/- 14.7 vs 43.9% +/- 14.9, p<.05).
CONCLUSION(S): Patients who met MCID after MLD had larger HNP areas by 26.6 mm² and larger hemi-HNP areas by 18.3 mm² than those who did not meet MCID. These patients were also 2.7x more likely to have a grey HNP compared to patients who did not meet MCID. When accounting for HNP area relative to canal area, patients who met MCID had a 7.9% greater Hemi-HNP canal occupation than patients who did not meet MCID. The results of this study suggest that preoperative MRI parameters can be useful in predicting patient reported improvement after MLD. FDA DEVICE/DRUG STATUS: This abstract does not discuss or include any applicable devices or drugs.
Copyright