Alex Quok An Teo

Study Design: Retrospective cohort study. Objectives: Thoracic ossification of the posterior longitudinal ligament (OPLL) is traditionally considered progressive, raising concern that the ossified mass left after posterior decompression and instrumented fusion (PDF) may enlarge. We characterized long-term CT-based changes in focal OPLL thickness after PDF and assessed whether the residual mass continued to enlarge. Methods: We reviewed all consecutive adult patients who underwent PDF for thoracic OPLL (2009–2023) with ≥2 years of follow-up. Maximum OPLL thickness was measured on axial CT and tracked longitudinally; a predefined 2-mm threshold, prespecified as the minimal clinically important difference, classified change. Thoracic kyphosis (TK) and modified Japanese Orthopaedic Association (mJOA) scores were also assessed. Results: Twenty-six patients (mean age 49.5 ± 12.4 years; 15 men) were followed 5.4 ± 2.7 years (range 2–11). Mean axial OPLL thickness changed from 7.02 ± 1.58 to 6.90 ± 2.01 mm (Δ −0.12 ± 1.03 mm). By the 2-mm threshold, 92.3% were stable, 3.8% decreased, and 3.8% increased; directional decreases occurred in 69.2% and increased with longer follow-up (37.5%, 77.8%, 88.9%; p = 0.023). TK increased (+6.2°, p < 0.001) but was uncorrelated with thickness change (r = 0.250). mJOA improved from 7.3 to 8.7 (p = 0.0001; recovery rate 34.9%; excellent/good 53.8%). The mean change was equivalent to zero within ±2 mm (two one-sided tests p < 0.001). Conclusions: Focal thoracic OPLL thickness remained predominantly stable after PDF, with threshold-level progression infrequent and neurological improvement maintained. Focal radiographic stability, rather than continued enlargement of the residual mass, was the predominant postoperative course after PDF.

Purpose: In a relatively new concept of individualised ACL reconstruction, the restoration or matching of the native ACL footprints at both the tibial and femoral insertions might be sufficient as a surgical objective. This study aimed to determine whether ACL graft size relative to the native ACL tibial footprint influences the rate of graft failure in a Singaporean Asian population. Methods: A retrospective study of all consecutive patients who underwent primary hamstring autograft ACL reconstruction over a 3-year period at a tertiary referral centre was performed. Baseline anthropometric data and native ACL tibial footprint area from pre-operative MRI scans were collected. Graft area was calculated based on the graft tunnel diameter. ACL graft failures were diagnosed clinically and confirmed radiologically. Multivariable analyses were performed using pre-specified variables based on the literature. Results: This cohort comprised 240 patients with a mean age of 24.4 years. There were a total of seven graft failures (2.92%), with two patients undergoing revision surgeries. There was no association between graft-to-footprint ratios and failure rates (71.9 vs. 80.2%, p = 0.386). There were no other significant predictors of graft failure. There was no statistically significant difference in graft failure rate between grafts p = 1.00). Gender, but not the graft-to-footprint ratio, was a significant predictor of graft size. Conclusions: No statistically significant association was identified between graft-to-footprint ratio and early graft failure. However, the low event rate and short-term follow-up limit definitive conclusions.

Purpose: The evidence base for preoperative health-related quality of life (HRQoL) in degenerative lumbar conditions derives predominantly from Western populations. This study characterised HRQoL across four degenerative lumbar conditions in a multi-ethnic Southeast Asian surgical cohort and identified factors associated with physical and mental health at composite and domain levels. Methods: Cross-sectional analysis of 1173 preoperative patients from a Singaporean spine surgery registry (2017–2022). SF-36 norm-based T-scores were computed for eight domains and two component summaries. Diagnosis-stratified stepwise multiple regression identified factors associated with the Physical Component Summary (PCS), Mental Component Summary (MCS), and domain scores. LASSO regression served as a sensitivity analysis. Results: Mean PCS was 35.6 (SD 8.9), 1.4 SD below the US normative mean, whereas MCS was 50.3 (SD 9.7), near normative levels. Physical functioning and role-physical were the most impaired domains. Scores differed across the four diagnoses on six of the ten SF-36 outcomes (one-way ANOVA, all p < 0.05; corroborated by Kruskal–Wallis tests), although effect sizes were small (eta-squared ≤ 0.02). Functional disability showed the strongest association with PCS across all four conditions (β range: − 0.32 to − 0.28; adjusted R ² = 0.48–0.52). MCS models explained limited variance (adjusted R ² = 0.09–0.22). At the domain level, physical functioning had the highest adjusted R ² (0.57–0.67), exceeding even PCS. Indian ethnicity was independently associated with lower MCS in two conditions. LASSO sensitivity analyses corroborated the stepwise findings. Conclusion: Functional disability was most strongly associated with physical HRQoL, channelled through physical functioning limitations. Conventional variables explained limited mental HRQoL variance, suggesting unmeasured psychological and cultural factors play a substantial role. While functional disability’s primacy was consistent with Western evidence, secondary patterns diverged from Western cohorts. These findings underscore the need to incorporate psychological and culturally adapted measures into preoperative assessment.

Objective: To examine sociodemographic, clinical, and healthcare-related factors associated with preoperative health-related quality of life (HRQoL) among patients undergoing surgery for degenerative lumbar spine conditions in a multi-ethnic Asian population. Methods: This cross-sectional study used baseline data from the Spine PROM Surgery Registry, including 1194 patients scheduled for surgery within a Singapore healthcare cluster between 2017 and 2022. HRQoL was measured using the EQ-5D-3L, with utility scores crosswalked to the EQ-5D-5L index using the van Hout crosswalk. Hierarchical linear regression assessed factors associated with HRQoL across three blocks: sociodemographic, clinical, and healthcare/lifestyle. Multivariable logistic regression identified factors associated with reporting problems within each EQ-5D dimension. Results: Mean age was 58.1 years (SD 16.1); 51.5% were male. Mean EQ-5D-5L crosswalk index was 0.43 (SD 0.38). Pain/discomfort (93.6%) and usual activities problems (84.3%) were most commonly reported. Lower EQ-5D scores were independently associated with non-outpatient presentation (β = −0.37), non-Chinese ethnicity, primary education or below relative to higher levels, and history of accident or trauma (β = −0.11). Dimension-level analyses showed primary education or below was associated with higher odds of problems across mobility, self-care, usual activities, and anxiety/depression compared with higher education levels, with a clear monotonic gradient. Female sex was associated with higher odds of pain/discomfort and anxiety/depression problems compared with male sex. Non-outpatient presentation was associated with markedly higher odds of self-care problems (OR = 2.98). Conclusions: Patients awaiting lumbar spine surgery appear to have impaired preoperative HRQoL. Although the modest explained variance limits robust risk prediction, preoperative profiles may still help inform clinical discussions and shared decision-making. Non-outpatient presentation may help identify patients who could benefit from enhanced preoperative support, although this requires prospective validation. Differences by ethnicity and education suggest opportunities for culturally tailored counselling. EQ-5D dimension profiles may indicate targets for prehabilitation and provide Singapore-based benchmark data for a lumbar spine surgery cohort for patient-centred care, service benchmarking, and health technology assessment.

Background: Degenerative cervical spondylosis (DCS) is a common and potentially debilitating condition, with surgery indicated in selected patients. Deep learning models (DLMs) can improve consistency in grading DCS neural stenosis on magnetic resonance imaging (MRI), though existing models focus on axial images, and comparisons are mostly limited to radiologists. Methods: We developed an enhanced transformer-based DLM that trains on sagittal images and optimizes axial and foraminal classification using a maximized dataset. DLM training utilized 648 scans, with internal testing on 75 scans and external testing on an independent 75-scan dataset. Performance of the DLM, spine surgeons, and radiologists of varying subspecialities/seniority were compared against a consensus reference standard. Results: On internal testing, the DLM achieved high agreement for all-class classification: axial spinal canal κ = 0.80 (95%CI: 0.72–0.82), sagittal spinal canal κ = 0.83 (95%CI: 0.81–0.85), and neural foramina κ = 0.81 (95%CI: 0.77–0.84). In comparison, human readers demonstrated lower levels of agreement (κ = 0.60–0.80). External testing showed modestly degraded model performance (κ = 0.68–0.77). Conclusions: These results demonstrate the utility of transformer-based DLMs in multiplanar MRI interpretation, surpassing spine surgeons and radiologists on internal testing and highlighting its potential for real-world clinical adoption.

Purpose To compare deep learning models of different architecture for automated lumbar spinal stenosis classification on MRI and benchmark their performance against radiologists and orthopedists. Methods Lumbar spine MRI studies from Sep-2015 to Sep-2019 were retrospectively obtained. Exclusion criteria included previous spinal instrumentation, suboptimal image quality, post-gadolinium studies, and severe scoliosis. Axial T2-weighted and sagittal T1-weighted images were used. Studies were split into training/validation and test sets. An external test set of 100 studies was used. Training data were labelled by 4 radiologists using predefined gradings. Two models, CNN-based and transformer-based, were developed. Consensus labelling by two expert spine radiologists served as the reference standard. Test sets were labelled by 8 participants (2 general radiologists, 2 radiologists-in-training, 2 orthopedists, 2 orthopedists-in-training). Detection recall (%), interrater agreement (Gwet κ), sensitivity, and specificity were evaluated. Results 564 MRI lumbar spines were included (mean age = 52 ± 19[SD]; 302 women), with 464(82%) and 100(18%) for training/validation and internal testing, respectively. Both models showed high recall for all regions of interest (> 94%), similar to participants. Dichotomous classification (normal/mild vs. moderate/severe) by the CNN model, transformer model, and participants showed respective kappas for central canal 0.99/0.99/0.97–0.98, lateral recesses 0.98/0.94/0.81–0.94, and neural foramina 0.98/0.95/0.91–0.95 on internal testing (p < 0.001); for central canal 0.99/0.97/0.92–0.97, lateral recess 0.97/0.90/0.61–0.91, and neural foramina 0.99/0.94/0.87–0.93 on external testing (p < 0.001). Conclusion The CNN model showed superior performance, and the transformer model showed similar to superior performance compared to clinicians for classifying lumbar spinal stenosis. These models could assist clinicians in report generation, surgical planning and education.