Citations

OBJECTIVE Proximal junctional kyphosis (PJK) affects 5%–61% of patients following thoracolumbar fusion. Many patients are asymptomatic, but a plurality require surgical revision at a cost of $75,000 per case. This analysis sought to analyze the degree to which bone quality and paraspinal muscle sarcopenia influence PJK failure mode. METHODS Patients undergoing thoracolumbar instrumented fusion with an upper instrumented vertebra (UIV) at the thoracolumbar junction (T10–L2) were identified and data were gathered on surgery, bone quality, pre- and postoperative sagittal alignment, and paraspinal muscle cross-sectional area (CSA). PJK was defined as a ≥ 10° increase in proximal junctional angle from the first postoperative radiograph. PJK was classified as discoligamentous failure (type 1), bone failure (type 2), or screw-bone interface failure (type 3) according to the Yagi-Boachie system. Bone quality was assessed by Hounsfield units (HUs) and the vertebral bone quality (VBQ) score at the UIV. RESULTS One hundred fifty patients were identified (median age 67 years, 53.3% female), 46 of whom experienced PJK (22 type 1, 13 type 2, 11 type 3). The median time to onset was most rapid for type 2 events (2.6 months). There were no differences between patients experiencing PJK versus controls regarding bone quality (HUs or VBQ score) or paraspinal muscle CSA on univariate comparison. However, subdivision by PJK type showed patients experiencing bone failure (type 2) PJK had significantly lower HUs at the UIV and UIV+1 relative to those experiencing type 1 PJK or no PJK (all p < 0.05). The VBQ score trended toward being significant, with a higher VBQ score (worse bone quality) in those suffering type 2 PJK, but did not reach statistical significance (p = 0.07). Patients experiencing discoligamentous failure PJK (type 1) had small multifidus CSA (390 mm 2 ) relative to patients experiencing type 2 (516 mm 2 ) or type 3 (440 mm 2 ) PJK and patients who did not experience PJK (481 mm 2 ), although the difference did not reach statistical significance. On time-to-event analysis, low HUs of the UIV/UIV+1 predicted type 2 (hazard ratio [HR] 0.81, 95% CI 0.70–0.93; p = 0.002) and type 2/3 PJK (HR 0.87, 95% CI 0.78–0.96; p = 0.006) but not type 1 PJK. Low UIV multifidus CSA trended toward being a significant predictor of type 1 PJK (HR 0.85, 95% CI 0.69–1.05; p < 0.10). CONCLUSIONS The combination of underlying bone quality and paraspinal musculature CSA at the UIV appeared to influence failure mode among patients who underwent lumbosacral instrumented fusion terminating at the thoracolumbar junction.

OBJECTIVE The present investigation aimed to 1) develop a new CT-based assessment of endplate bone density (endplate Hounsfield unit [EP-HU]) and 2) analyze if EP-HU was a better predictor than vertebral Hounsfield unit (HU) for subsidence after anterior cervical discectomy and fusion (ACDF). METHODS All adult patients who underwent one- to three-level ACDF with a titanium interbody for radiculopathy and/or myelopathy at an academic center between 2018 and 2020 were retrospectively identified. Based on preoperative sagittal CT scans (left, right, and middle cuts), 2-mm superior and inferior endplate regions were circumscribed with the free draw tool to account for endplate surface undulations. The average of the superior and inferior EP-HUs on all CT cuts was used to calculate EP-HU. Cervical vertebral HUs were determined from the average of axial CT cranial, middle, and caudal cuts circumscribing only trabecular bone. The interbody subsidence of the cranial and caudal endplates of each ACDF level was directly measured on the endplate-facing surface of both coronal and sagittal cuts of postoperative CT scans (at 1 year) to determine the maximum subsidence (subsidence defined as ≥ 2 mm). Univariate and stepwise logistic regression analyses were used to compare subsidence based on CT bone metrics. Receiver operating characteristic (ROC) curve analyses were used to determine the probability of subsidence based on EP-HUs and vertebral HUs. RESULTS A total of 35 patients were included. Subsidence occurred at 32 of 67 unique fusion levels. Subsidence was associated with older age (p = 0.008), diabetes diagnosis (p = 0.015), and decreased interbody length (p = 0.019). EP-HUs exhibited moderate correlation with cervical vertebral HUs (Pearson’s ρ = 0.63). Subsidence was significantly associated with decreased total EP-HUs (subsidence: 475 HU, no subsidence: 543 HU; p = 0.019) and decreased lumbar vertebral HUs (subsidence: 296 HU, no subsidence: 341 HU; p = 0.011). ROC curve analysis identified an optimal EP-HU cutoff of 512.30 (area under the curve [AUC] = 0.701) to predict subsidence. The AUC of the vertebral HUs with respect to subsidence was 0.662. EP-HUs < 512.30 predicted subsidence (OR 6.67, p = 0.001) independent of significant demographic and surgical factors. CONCLUSIONS CT cervical EP-HUs rather than vertebral cervical trabecular HUs may be more efficacious in predicting subsidence after ACDF.

Importance Multiparametric magnetic resonance imaging (MRI), with or without prostate biopsy, has become the standard of care for diagnosing clinically significant prostate cancer. Resource capacity limits widespread adoption. Biparametric MRI, which omits the gadolinium contrast sequence, is a shorter and cheaper alternative offering time-saving capacity gains for health systems globally. Objective To assess whether biparametric MRI is noninferior to multiparametric MRI for diagnosis of clinically significant prostate cancer. Design, Setting, and Participants A prospective, multicenter, within-patient, noninferiority trial of biopsy-naive men from 22 centers (12 countries) with clinical suspicion of prostate cancer (elevated prostate-specific antigen [PSA] level and/or abnormal digital rectal examination findings) from April 2022 to September 2023, with the last follow-up conducted on December 3, 2024. Interventions Participants underwent multiparametric MRI, comprising T2-weighted, diffusion-weighted, and dynamic contrast–enhanced (DCE) sequences. Radiologists reported abbreviated biparametric MRI first (T2-weighted and diffusion-weighted), blinded to the DCE sequence. After unblinding, radiologists reported the full multiparametric MRI. Patients underwent a targeted biopsy with or without systematic biopsy if either biparametric MRI or multiparametric MRI was suggestive of clinically significant prostate cancer. Main outcomes and measures The primary outcome was the proportion of men with clinically significant prostate cancer. Secondary outcomes included the proportion of men with clinically insignificant cancer. The noninferiority margin was 5%. Results Of 555 men recruited, 490 were included for primary outcome analysis. Median age was 65 (IQR, 59-70) years and median PSA level was 5.6 (IQR, 4.4-8.0) ng/mL. The proportion of patients with abnormal digital rectal examination findings was 12.7%. Biparametric MRI was noninferior to multiparametric MRI, detecting clinically significant prostate cancer in 143 of 490 men (29.2%), compared with 145 of 490 men (29.6%) (difference, −0.4 [95% CI, −1.2 to 0.4] percentage points; P = .50). Biparametric MRI detected clinically insignificant cancer in 45 of 490 men (9.2%), compared with 47 of 490 men (9.6%) with the use of multiparametric MRI (difference, −0.4 [95% CI, −1.2 to 0.4] percentage points). Central quality control demonstrated that 99% of scans were of adequate diagnostic quality. Conclusion and relevance In men with suspected prostate cancer, provided image quality is adequate, an abbreviated biparametric MRI scan, with or without targeted biopsy, could become the new standard of care for prostate cancer diagnosis. With approximately 4 million prostate MRIs performed globally annually, adopting biparametric MRI could substantially increase scanner throughput and reduce costs worldwide. Trial registration ClinicalTrials.gov Identifier: NCT04571840

Objective The objective of this study was to determine the predictive value of amyloid‐positron emission tomography (PET) versus the plasma ratio of phosphorylated tau at threonine 217 (p‐tau217) to non‐phosphorylated tau217 (%p‐tau217) for tau‐PET transitions (T− to T+). The added value of combining plasma amyloid‐β 42 and amyloid‐β 40 (Aβ42/40) and %p‐tau217 into an amyloid probability score (APS2) was also assessed. Methods Mayo Clinic Study of Aging (MCSA) participants had plasma markers measured at via mass spectrometry (MS), an amyloid‐PET scan, and a tau‐PET (meta‐temporal region of interest [ROI]) negative scan (standardized uptake value ratio [SUVR] <1.29) at the index (baseline) date, along with one or more follow‐up tau‐PET scans. The BioFINDER‐2 cohort was used for validation. Cox proportional hazards models adjusted for age, sex, and apolipoprotein ( APOE ) ε4 were used to assess predictors, with scaling to the interquartile range (IQR) for comparability of hazard ratios (HR). Results Among 255 tau‐PET negative MCSA participants (median age: 71.9 years), 37 converted to tau‐PET positive (median follow‐up time: 3.81 years). Higher %p‐tau217 (HR: 1.52 [95% CI: 1.28–1.80]), amyloid‐PET centiloid (HR: 1.47 [95% CI: 1.20–1.79]), and APS2 (HR: 1.62 [95% CI: 1.22–2.16]) predicted tau‐PET conversion. However, Aβ42/40 (HR: 0.94 [95% CI: 0.54–1.66]) was not associated with tau‐PET conversion. In the BioFINDER‐2 cohort (605 tau‐negative, median age: 70.2), 33 converted to tau‐positive (median follow‐up time: 2 years), with higher %p‐tau217 (HR: 1.80 [95% CI: 1.50–2.17]), amyloid‐PET centiloid (HR: 2.29 [95% CI: 1.77–2.97]), and lower Aβ42/40 (HR: 2.38 [95% CI: 1.17–4.83]) predicting conversion. Interpretation In two cohorts, %p‐tau217 was associated with tau‐PET conversion, comparable to amyloid‐PET. APS2 also predicted conversion in the MCSA cohort, whereas Aβ42/40 predicted conversion in the BioFINDER‐2 cohort, which had more individuals with cognitive impairment. ANN NEUROL 2025;98:1249–1260

ABSTRACTBackgroundDifferentiating metastatic melanoma from histologic mimics such as malignant gastrointestinal neuroectodermal tumor (GNET), malignant melanotic nerve sheath tumor (MMNST), and epithelioid malignant peripheral nerve sheath tumor (EMPNST) poses significant diagnostic challenges due to overlapping morphology and immunophenotypes. PRAME is a novel immunohistochemical marker increasingly used to distinguish melanoma from its mimics, but remains underexplored in these rare tumor types.MethodsPRAME immunohistochemistry was performed on four GNETs, seven MMNSTs, 10 EMPNSTs, 16 metastatic melanomas (including eight undifferentiated melanomas), and two MITF‐rearranged melanocytic tumors. PRAME expression was scored from 0 to 4+ based on the percentage of tumor nuclei showing moderate to strong staining.ResultsAll GNETs, MMNSTs, and MITF‐rearranged tumors were PRAME‐negative. EMPNSTs showed variable expression: six were negative (0–1), one equivocal (2+), and three positive (3–4+). Fifteen of 16 melanomas were PRAME‐positive. PRAME scores differed significantly among tumor types (p = 1.99 × 10−5). PRAME demonstrated high sensitivity and specificity for distinguishing metastatic melanoma from primary mimickers including GNET and MMNST, but low specificity in EMPNST (71.4%).ConclusionsPRAME reliably distinguishes metastatic melanoma from GNET and MMNST, supporting its use in this differential. However, its reduced specificity in EMPNST limits its standalone diagnostic value in this context, emphasizing the need for a multimodal approach.

Abstract Objectives Large language models (LLMs) offer potential in natural language processing tasks in healthcare. Due to the need for high accuracy, understanding their limitations is essential. The purpose of this study was to evaluate the performance of LLMs in classifying radiology reports for the presence of pulmonary embolism (PE) under various conditions, including different prompt designs and data perturbations. Materials and Methods In this retrospective, institutional review board approved study, we evaluated 3 Google’s LLMs including Gemini-1.5-Pro, Gemini-1.5-Flash-001, and Gemini-1.5-Flash-002, in classifying 11 999 pulmonary CT angiography radiology reports for PE. Ground truth labels were determined by concordance between a computer vision-based PE detection (CVPED) algorithm and multiple LLM runs under various configurations. Discrepancies between algorithms’ classifications were aggregated and manually reviewed. We evaluated the effects of prompt design, data perturbations, and repeated analyses across geographic cloud regions. Performance metrics were calculated. Results Of 11 999 reports, 1296 (10.8%) were PE-positive. Accuracy across LLMs ranged between 0.953 and 0.996. The highest recall rate for a prompt modified after a review of the misclassified cases (up to 0.997). Few-shot prompting improved recall (up to 0.99), while chain-of-thought generally degraded performance. Gemini-1.5-Flash-002 demonstrated the highest robustness against data perturbations. Geographic cloud region variability was minimal for Gemini-1.5+-Pro, while the Flash models showed stable performance. Discussion and Conclusion LLMs demonstrated high performance in classifying radiology reports, though results varied with prompt design and data quality. These findings underscore the need for systematic evaluation and validation of LLMs for clinical applications, particularly in high-stakes scenarios.