OBJECTIVE: To preserve facial nerve function in vestibular schwannoma (VS) microsurgery, some have advocated subtotal resection (STR) if the tumor is densely adherent to a thinned facial nerve. The objective of this study was to determine if residual volume is associated with progression and whether there is a threshold residual volume that should be pursued during STR to prevent recurrence. A secondary objective of this study was to determine whether facial nerve function at last follow-up was associated with extent of resection (EOR). METHODS: Clinical and radiographic data were retrospectively collected from the records of 164 patients with VS who underwent resection. Tumor volumes were measured using Visage, and standard statistical methods were used. The House-Brackmann scale was used to assess changes in facial nerve function before surgery and at last follow-up. RESULTS: Sixty-one patients (37%) received gross-total resection (GTR) and 103 (63%) received STR. The median clinical and radiographic follow-ups were 49 and 48 months, respectively. The median residual volume was 0.5 cm³ after STR. Kaplan-Meier actuarial survival analysis revealed a 96.3% 5-year progression-free survival (PFS) rate after GTR, which was greater than that after STR (84.5%, p = 0.03). Recursive partitioning analysis of patients receiving STR revealed a residual volume of 0.60 cm³ as the optimal threshold for recurrence. Patients with residual volume ≥ 0.60 cm³ had a 76.0% 5-year PFS, regardless of adjuvant SRS, which was lower than that for patients undergoing GTR (96.3%) or STR (95.6%) with residual volumes < 0.60 cm³ (p < 0.01). On Cox regression analysis, residual volume ≥ 0.60 cm³ (HR 14.4, p = 0.01) was independently associated with progression, even when accounting for patient age, adjuvant radiosurgery, and preoperative tumor size. In 112 patients with at least 24 months of follow-up after their last treatment, tumor control was achieved in 111 (99.1%) patients at a median last follow-up of 71 months. Worse facial nerve function at the last follow-up was independently associated with prior treatment for VS (adjusted OR 3.7, p = 0.04), but not residual volume cohort or preoperative tumor volume. CONCLUSIONS: Residual volume > 0.60 cm³ after VS resection was independently associated with tumor progression, even accounting for adjuvant SRS. These data support maximizing the EOR during VS surgery, even if GTR cannot be safely achieved.

Artificial intelligence (AI), deep learning (DL), and machine learning (ML) algorithms are revolutionizing spine surgery. Soon, these technologies may allow the integration of automated devices into clinical practice. The roles of such devices are yet to be imagined and then developed, but one could assume that automated surgical devices can assist spine surgeons in a variety of ways, such as contextual guidance, precise screw placements, or intraoperative monitoring. In the not-too-distant future, such devices may be able to perform entire surgeries autonomously. Current literature suggests that advancements toward autonomous robotic surgery may improve surgical approaches and reduce negative clinical variation in spine surgery outcomes. This review aims to examine the current trends, practices, and advancements in surgical automation and provide an overview of the stages of automation of devices currently employed within spine surgery.

Rapid advances in large language models (LLMs) are transforming the role of students and principal investigators (PIs) in biomedical research. This perspective examines how LLMs can reshape the laboratory model as de facto “Co-PIs” for tasks ranging from literature triage to hypothesis generation. By clarifying both opportunities and risks, we propose a framework for efficient AI collaboration which aims to guide investigators and trainees in harnessing LLMs responsibly.

Background/Objectives: Minimally invasive spine surgery (MISS) is complex and requires proficiency with a variety of technological and robotic modalities. Acquiring these skills is a long and involved process, often with a steep learning curve. This paper seeks to characterize the state of MISS training in neurosurgical and orthopedic residency programs, focusing on their effectiveness at minimizing substantial learning curves in the field, as well as highlighting potential areas for future growth. Methods: We conducted a scoping review of the PubMed, Scopus, and Embase databases utilizing the PRISMA extension for scoping reviews. Results: Of the 100 studies initially identified, 16 were included in our final analysis. MISS training types could be broadly grouped into four categories: virtual simulation (including AR and VR), physical models, hybrid didactic and simulation, and mentored training. Training with these modalities led to improvements in resident performance across multiple different MISS techniques, including percutaneous pedicle screw fixation, MIS dural repair, MIS-TLIF, MIS-LLIF, MIS-ULBD, microscopic discectomy/disk herniation repair, percutaneous needle placement, and surgical navigation. Specific improvements included reduced error rate, operation time, and fluoroscopy exposure, as well as increased procedural knowledge, accuracy, and confidence. Conclusions: The incorporation of MISS training modalities in spine surgery residency leads to increases in simulated performance and could serve as a means of overcoming significant learning curves in the field.

Stem cells are essential for tissue maintenance, repair, and regeneration, yet their dysregulation gives rise to cancer stem cells (CSCs), which drive tumor progression, metastasis, and therapy resistance. Despite extensive research on stemness and oncogenesis, a critical gap remains in our understanding of how the transcriptomic landscapes of normal somatic stem cells (SSCs) diverge from those of CSCs to enable malignancy. This review synthesizes current knowledge of the key signaling pathways (Wnt, Notch, Hedgehog, TGF-β), transcription factors (Oct4, Sox2, Nanog, c-Myc, YAP/TAZ), and epigenetic mechanisms (chromatin remodeling, DNA methylation, microRNA regulation) that govern stemness in SSCs and are hijacked or dysregulated in CSCs. We highlight how context-specific modulation of these pathways distinguishes physiological regeneration from tumorigenesis. Importantly, we discuss the role of epithelial–mesenchymal transition (EMT), cellular plasticity, and microenvironmental cues in reprogramming and maintaining CSC phenotypes. By integrating transcriptomic and epigenetic insights across cancer biology and regenerative medicine, this review provides a framework for identifying vulnerabilities specific to CSCs while still preserving normal stem cell function. Understanding these distinctions is essential for the development of targeted therapies that minimize damage to healthy tissues and advance precision oncology.

OBJECTIVE: Accurate vertebral segmentation is an important step in imaging analysis pipelines for diagnosis and subsequent treatment of spinal metastases. Segmenting these metastases is especially challenging given their radiological heterogeneity. Conventional approaches for segmenting vertebrae have included manual review or deep learning; however, manual review is time-intensive with interrater reliability issues, while deep learning requires large datasets to build. The rise of generative AI, notably tools such as Meta’s Segment Anything Model 2 (SAM 2), holds promise in its ability to rapidly generate segmentations of any image without pretraining (zero-shot). The authors of this study aimed to assess the ability of SAM 2 to segment vertebrae with metastases. METHODS: A publicly available set of spinal CT scans from The Cancer Imaging Archive was used, which included patient sex, BMI, vertebral locations, types of metastatic lesion (lytic, blastic, or mixed), and primary cancer type. Ground-truth segmentations for each vertebra, derived by neuroradiologists, were further extracted from the dataset. SAM 2 then produced segmentations for each vertebral slice without any training data, all of which were compared to gold standard segmentations using the Dice similarity coefficient (DSC). Relative performance differences were assessed across clinical subgroups using standard statistical techniques. RESULTS: Imaging data were extracted for 55 patients and 779 unique thoracolumbar vertebrae, 167 of which had metastatic tumor involvement. Across these vertebrae, SAM 2 had a mean volumetric DSC of 0.833 ± 0.053. SAM 2 performed significantly worse on thoracic vertebrae relative to lumbar vertebrae, female patients relative to male patients, and obese patients relative to non-obese patients. CONCLUSIONS: These results demonstrate that general-purpose segmentation models like SAM 2 can provide reasonable vertebral segmentation accuracy with no pretraining, with efficacy comparable to previously published trained models. Future research should include optimizations of spine segmentation models for vertebral location and patient body habitus, as well as for variations in imaging quality approaches.