This Week in Neurosurgery — Oct 28, 2026
Generated Oct 29, 2026 · 10:17
The week's practice-changing Neurosurgery research, summarized for clinicians.
If the audio fails to play, refresh the page to renew the link.
Get next week’s Neurosurgery briefing — free.
In your podcast app, or readable in your inbox with the audio one tap away.
Clinical cutoffs for plasma biomarkers to identify intracranial pathology in traumatic brain injury patients with Glasgow Coma Scale scores 13 to 15: a TRACK-TBI study.
In over 1,700 mild-range head injury patients, day-one GFAP discriminated CT-positive injury well, and prevalence-adjusted cutoffs near 2,900 to 5,200 pg/mL predicted CT pathology with ninety percent confidence.
Journal of Neurosurgery · 2026 · PubMed
This week’s papers
- 01
Clinical cutoffs for plasma biomarkers to identify intracranial pathology in traumatic brain injury patients with Glasgow Coma Scale scores 13 to 15: a TRACK-TBI study.
In over 1,700 mild-range head injury patients, day-one GFAP discriminated CT-positive injury well, and prevalence-adjusted cutoffs near 2,900 to 5,200 pg/mL predicted CT pathology with ninety percent confidence.
Eagle SR et al. · Journal of Neurosurgery · 2026
From the briefingIn the Journal of Neurosurgery, Eagle and colleagues used the TRACK-TBI cohort, drawn from 18 level one trauma centres in the United States, to set blood biomarker thresholds for mild-range injury, meaning Glasgow Coma Scale scores of 13 to 15 [1]. Among more than 1,700 patients, just over a third of patients had acute findings on CT. Day-one glial fibrillary acidic protein, or GFAP, separated CT-positive from CT-negative patients well, while UCH-L1 performed only modestly. The practical contribution is a set of high positive predictive value cutoffs: at the prevalence seen in this cohort, a GFAP level of roughly 2,900 picograms per millilitre corresponded to a ninety percent chance of CT pathology, and the threshold rose to around 5,200 when a lower, literature-based prevalence of ten percent was assumed. Among CT-negative patients who had MRI at two weeks, a quarter of those scanned showed pathology, and GFAP around 1,400 picograms per millilitre flagged MRI-positive injury with similar confidence. Most work on GFAP has focused on ruling injury out; this analysis frames it as a tool for ruling injury in and triaging toward urgent imaging, although the thresholds come from a secondary analysis and depend heavily on local prevalence.
If you only have time for one paper this week, make it the TRACK-TBI biomarker study from Eagle and colleagues [1]. It moves GFAP from a rule-out test toward concrete, prevalence-adjusted thresholds for ruling intracranial injury in, a question that touches nearly every mild head injury pathway.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 02
Interhemispheric hygroma as a predictor of post-traumatic hydrocephalus after decompressive craniectomy: a phenotype-specific systematic review and meta-analysis.
Interhemispheric hygroma after decompressive craniectomy for trauma was linked to roughly six to sevenfold higher odds of clinically significant hydrocephalus, though the primary pooled estimate rests on three studies.
Ganesh S, Joseph J · Journal of Clinical Neuroscience · 2026
From the briefingMoving to the post-surgical phase of trauma care, Ganesh and Joseph, in the Journal of Clinical Neuroscience, pooled studies of hydrocephalus after decompressive craniectomy and argued that subdural hygroma should not be treated as one entity [2]. Focusing specifically on interhemispheric hygroma, they found it was associated with roughly six to seven times the odds of clinically significant hydrocephalus. The signal held in sensitivity analyses, but the primary estimate rests on only three studies with comparable definitions, so this is a well-reasoned imaging marker for closer surveillance rather than a validated trigger for intervention.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 03
Machine Learning Prediction of Treatment Failure After Middle Meningeal Artery Embolization for Subdural Hematoma.
A retrospective gradient boosting model stratified subdural hematomas treated with middle meningeal artery embolization into risk groups with failure rates ranging from under two percent to nearly sixty percent.
DeMessie B et al. · Neurosurgery · 2026
From the briefingIn Neurosurgery, DeMessie and colleagues brought machine learning to middle meningeal artery embolization for chronic subdural hematoma [3]. In an international retrospective cohort, about nine percent of treated hematomas in the test set failed, meaning rescue evacuation or retreatment. Their gradient boosting model discriminated well and sorted hematomas into three tiers: failure was under two percent in the low-risk group but close to sixty percent in the small high-risk group, roughly a 37-fold difference. Interestingly, the top predictors were platelet-related measures interacting with age, alongside hematoma thickness. The model is retrospective and has not been prospectively validated, but it suggests embolization failure is concentrated in an identifiable minority of patients.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 04
Diagnostic accuracy of neuropathological assessment across different diagnostic settings for high-grade gliomas: the added value of DNA methylation profiling.
Against DNA methylation profiling as reference, general pathology missed about one in five high-grade gliomas, whereas specialist neuropathology detected all of them with roughly 99 percent overall accuracy.
Christoforou T et al. · Acta Neurochirurgica · 2026
From the briefingIn Acta Neurochirurgica, Christoforou and colleagues compared general pathology, specialist neuropathology, and DNA methylation profiling as the reference standard in 130 patients with suspected high-grade glioma [4]. General pathology never over-called a high-grade glioma but missed about one in five of them, labelling some as low-grade glioma or even non-neoplastic tissue. Specialist neuropathology caught every high-grade case, with about 99 percent overall accuracy, and methylation refined a handful of diagnoses. This is a selected retrospective cohort, but it adds weight to the argument that glioma tissue belongs in specialist hands, with methylation as a complement.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 05
Near real-time intraoperative assessment of pituitary neuroendocrine tumors using stimulated Raman histology during endoscopic transsphenoidal surgery: comparison with conventional histopathology.
Stimulated Raman histology was feasible in all 28 pituitary tumour operations and matched conventional staining well in about 86 percent of samples, distinguishing adenoma from normal gland within minutes.
Millesi M et al. · Journal of Neurosurgery · 2026
From the briefingOn the intraoperative side, Millesi and colleagues, in the Journal of Neurosurgery, tested stimulated Raman histology during endoscopic transsphenoidal surgery for pituitary neuroendocrine tumours [5]. Across 66 samples from 28 patients, imaging was feasible every time, and a blinded neuropathologist rated comparability with conventional staining as good or excellent in about 86 percent of samples, with low comparability in roughly one in seven. Distinguishing adenoma from normal gland, and even subtype features such as Crooke cells, was possible within minutes. The study is small and descriptive and does not link the technique to remission or endocrine outcomes.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 06
Intraoperative dorsal cochlear nucleus action potential monitoring in hearing preservation retrosigmoid surgery for vestibular schwannoma: a predictive analysis of short-term hearing outcomes.
Depth-electrode dorsal cochlear nucleus monitoring recorded stable signals more often than brainstem responses during vestibular schwannoma surgery, and residual amplitude above 0.41 predicted favourable hearing outcomes.
Pang X et al. · Journal of Neurosurgery · 2026
From the briefingAlso in the Journal of Neurosurgery, Pang and colleagues reported continuous monitoring of the dorsal cochlear nucleus action potential with a four-contact depth electrode during retrosigmoid vestibular schwannoma resection in 153 patients [6]. A stable baseline signal was obtained in about eighty percent of patients, compared with about sixty percent for auditory brainstem responses, with no electrode-related complications. Preservation of this signal was the most consistent independent predictor of hearing at one week and six months, and a residual amplitude ratio above 0.41 marked favourable early outcomes. This is a retrospective cohort, and whether acting on the signal changes hearing outcomes remains untested.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 07
Evaluating the impact of intraoperative neuromonitoring on neurological outcomes and extent of resection in pediatric supratentorial brain tumor surgery: a systematic review and meta-analysis.
Pooled pediatric series showed higher gross total resection and fewer deficits with intraoperative neuromonitoring, but studies lacked direct within-cohort comparisons, so causal benefit remains unproven.
Vadali S et al. · Journal of Clinical Neuroscience · 2026
From the briefingA broader but weaker picture of monitoring comes from Vadali and colleagues in the Journal of Clinical Neuroscience, who pooled 41 studies of nearly 1,900 children undergoing supratentorial tumour resection [7]. Gross total resection was about 77 percent in monitored series versus about 53 percent in unmonitored series, with fewer postoperative deficits, particularly near eloquent cortex. The authors are explicit that these are descriptive pooled estimates from studies that did not compare monitored and unmonitored children directly, so they cannot be read as causal evidence.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 08
Urgent versus scheduled resection in high-grade glioma: patient selection and healthcare utilization in a single-center cohort.
Patients with presumed high-grade glioma, preserved performance status and no midline shift underwent scheduled resection with more days at home and no detectable difference in complications or survival.
Mahmoudi DA et al. · Journal of Neuro-Oncology · 2026
From the briefingFinally in this theme, Mahmoudi and colleagues, in the Journal of Neuro-Oncology, asked who with presumed high-grade glioma actually needs surgery during the index admission [8]. In 166 patients at a single centre, those taken to scheduled resection had better performance status and less midline shift, and they spent about nine more days at home in the first month. Complications and overall survival did not differ detectably. Because the groups were selected by clinical status, this shows that a distinct, recognisable population can follow a scheduled pathway without an observed harm signal, not that scheduling itself improves outcomes.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 09
American Epilepsy Society/Congress of Neurological Surgeons Epilepsy Surgery Clinical Practice Guideline.
A GRADE-based joint guideline updates 2003 epilepsy surgery parameters, covering temporal, neocortical, hamartoma and callosotomy surgery while weighting memory, quality of life and safety alongside seizure control.
Gonzalez-Martinez J et al. · Neurosurgery · 2026
From the briefingIn Neurosurgery, Gonzalez-Martinez and colleagues present the joint American Epilepsy Society and Congress of Neurological Surgeons clinical practice guideline on epilepsy surgery, updating practice parameters from 2003 [9]. Using GRADE methodology, the panel covered resection for mesial temporal and neocortical epilepsy, resection or disconnection of hypothalamic hamartoma, and callosotomy in developmental and epileptic encephalopathies, deliberately leaving neurostimulation aside. Notably, it treated quality of life, mortality, adverse events and verbal memory loss as critical outcomes alongside seizure control. The abstract does not state the specific recommendations, so the full document is needed for the graded statements.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
- 10
Is There a Cervical Alignment Threshold Warranting Cervical Inclusion in the Construct?
About fourteen percent of adults developed cervical decompensation after thoracolumbar deformity fusion, predicted by baseline alignment and upper thoracic instrumentation, with a proposed C2-pelvic angle threshold awaiting validation.
Passias PG et al. · Neurosurgery · 2026
From the briefingAlso in Neurosurgery, Passias and colleagues examined cervical decompensation after thoracolumbar fusion to the sacrum in 1,190 adults with spinal deformity [10]. About fourteen percent of patients decompensated within two years, though only 24 patients needed cervical extension. Upper thoracic instrumentation roughly tripled the hazard, and higher baseline C2-pelvic angle predicted earlier decompensation, with a proposed threshold of 26.2 degrees. Decompensation was linked to worse patient-reported outcomes, but the authors themselves state the threshold requires external validation before it informs construct planning.
Written by AI from the published abstract, for healthcare professionals. Not medical advice; check the paper before acting on it.
The full briefing
This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.
Welcome to This Week in Neurosurgery. This week we're covering 10 notable papers spanning neurotrauma and subdural disease, intraoperative diagnosis and monitoring in tumour surgery, and new guidance in epilepsy and spinal deformity surgery. Let's dive in.
We start with trauma, where three papers try to make risk prediction more precise. In the Journal of Neurosurgery, Eagle and colleagues used the TRACK-TBI cohort, drawn from 18 level one trauma centres in the United States, to set blood biomarker thresholds for mild-range injury, meaning Glasgow Coma Scale scores of 13 to 15 [1]. Among more than 1,700 patients, just over a third of patients had acute findings on CT. Day-one glial fibrillary acidic protein, or GFAP, separated CT-positive from CT-negative patients well, while UCH-L1 performed only modestly. The practical contribution is a set of high positive predictive value cutoffs: at the prevalence seen in this cohort, a GFAP level of roughly 2,900 picograms per millilitre corresponded to a ninety percent chance of CT pathology, and the threshold rose to around 5,200 when a lower, literature-based prevalence of ten percent was assumed. Among CT-negative patients who had MRI at two weeks, a quarter of those scanned showed pathology, and GFAP around 1,400 picograms per millilitre flagged MRI-positive injury with similar confidence. Most work on GFAP has focused on ruling injury out; this analysis frames it as a tool for ruling injury in and triaging toward urgent imaging, although the thresholds come from a secondary analysis and depend heavily on local prevalence. Moving to the post-surgical phase of trauma care, Ganesh and Joseph, in the Journal of Clinical Neuroscience, pooled studies of hydrocephalus after decompressive craniectomy and argued that subdural hygroma should not be treated as one entity [2]. Focusing specifically on interhemispheric hygroma, they found it was associated with roughly six to seven times the odds of clinically significant hydrocephalus. The signal held in sensitivity analyses, but the primary estimate rests on only three studies with comparable definitions, so this is a well-reasoned imaging marker for closer surveillance rather than a validated trigger for intervention. In Neurosurgery, DeMessie and colleagues brought machine learning to middle meningeal artery embolization for chronic subdural hematoma [3]. In an international retrospective cohort, about nine percent of treated hematomas in the test set failed, meaning rescue evacuation or retreatment. Their gradient boosting model discriminated well and sorted hematomas into three tiers: failure was under two percent in the low-risk group but close to sixty percent in the small high-risk group, roughly a 37-fold difference. Interestingly, the top predictors were platelet-related measures interacting with age, alongside hematoma thickness. The model is retrospective and has not been prospectively validated, but it suggests embolization failure is concentrated in an identifiable minority of patients.
Our second theme is how we know what we are removing, and what we are preserving, during tumour surgery. In Acta Neurochirurgica, Christoforou and colleagues compared general pathology, specialist neuropathology, and DNA methylation profiling as the reference standard in 130 patients with suspected high-grade glioma [4]. General pathology never over-called a high-grade glioma but missed about one in five of them, labelling some as low-grade glioma or even non-neoplastic tissue. Specialist neuropathology caught every high-grade case, with about 99 percent overall accuracy, and methylation refined a handful of diagnoses. This is a selected retrospective cohort, but it adds weight to the argument that glioma tissue belongs in specialist hands, with methylation as a complement. On the intraoperative side, Millesi and colleagues, in the Journal of Neurosurgery, tested stimulated Raman histology during endoscopic transsphenoidal surgery for pituitary neuroendocrine tumours [5]. Across 66 samples from 28 patients, imaging was feasible every time, and a blinded neuropathologist rated comparability with conventional staining as good or excellent in about 86 percent of samples, with low comparability in roughly one in seven. Distinguishing adenoma from normal gland, and even subtype features such as Crooke cells, was possible within minutes. The study is small and descriptive and does not link the technique to remission or endocrine outcomes. Also in the Journal of Neurosurgery, Pang and colleagues reported continuous monitoring of the dorsal cochlear nucleus action potential with a four-contact depth electrode during retrosigmoid vestibular schwannoma resection in 153 patients [6]. A stable baseline signal was obtained in about eighty percent of patients, compared with about sixty percent for auditory brainstem responses, with no electrode-related complications. Preservation of this signal was the most consistent independent predictor of hearing at one week and six months, and a residual amplitude ratio above 0.41 marked favourable early outcomes. This is a retrospective cohort, and whether acting on the signal changes hearing outcomes remains untested. A broader but weaker picture of monitoring comes from Vadali and colleagues in the Journal of Clinical Neuroscience, who pooled 41 studies of nearly 1,900 children undergoing supratentorial tumour resection [7]. Gross total resection was about 77 percent in monitored series versus about 53 percent in unmonitored series, with fewer postoperative deficits, particularly near eloquent cortex. The authors are explicit that these are descriptive pooled estimates from studies that did not compare monitored and unmonitored children directly, so they cannot be read as causal evidence. Finally in this theme, Mahmoudi and colleagues, in the Journal of Neuro-Oncology, asked who with presumed high-grade glioma actually needs surgery during the index admission [8]. In 166 patients at a single centre, those taken to scheduled resection had better performance status and less midline shift, and they spent about nine more days at home in the first month. Complications and overall survival did not differ detectably. Because the groups were selected by clinical status, this shows that a distinct, recognisable population can follow a scheduled pathway without an observed harm signal, not that scheduling itself improves outcomes.
Our last theme covers guidance and thresholds in functional and spine surgery. In Neurosurgery, Gonzalez-Martinez and colleagues present the joint American Epilepsy Society and Congress of Neurological Surgeons clinical practice guideline on epilepsy surgery, updating practice parameters from 2003 [9]. Using GRADE methodology, the panel covered resection for mesial temporal and neocortical epilepsy, resection or disconnection of hypothalamic hamartoma, and callosotomy in developmental and epileptic encephalopathies, deliberately leaving neurostimulation aside. Notably, it treated quality of life, mortality, adverse events and verbal memory loss as critical outcomes alongside seizure control. The abstract does not state the specific recommendations, so the full document is needed for the graded statements. Also in Neurosurgery, Passias and colleagues examined cervical decompensation after thoracolumbar fusion to the sacrum in 1,190 adults with spinal deformity [10]. About fourteen percent of patients decompensated within two years, though only 24 patients needed cervical extension. Upper thoracic instrumentation roughly tripled the hazard, and higher baseline C2-pelvic angle predicted earlier decompensation, with a proposed threshold of 26.2 degrees. Decompensation was linked to worse patient-reported outcomes, but the authors themselves state the threshold requires external validation before it informs construct planning.
If you only have time for one paper this week, make it the TRACK-TBI biomarker study from Eagle and colleagues [1]. It moves GFAP from a rule-out test toward concrete, prevalence-adjusted thresholds for ruling intracranial injury in, a question that touches nearly every mild head injury pathway.
Here is what this week's evidence adds up to in Neurosurgery. First, blood GFAP now has proposed high-confidence cutoffs for identifying CT and MRI injury in mild-range trauma, but they come from one large observational cohort and shift with local prevalence. Second, risk prediction in subdural and post-craniectomy care is sharpening, with a machine learning model separating very low from very high embolization failure risk and interhemispheric hygroma emerging as a hydrocephalus marker, both still lacking prospective validation. Third, diagnostic precision in tumour surgery is improving, with specialist neuropathology markedly outperforming general pathology for high-grade glioma and Raman histology and brainstem-level hearing monitoring showing promise in small or retrospective series. Fourth, pediatric neuromonitoring and scheduled glioma resection both look favourable in observational data, but neither has comparative evidence showing causal benefit. And finally, the new epilepsy surgery guideline puts memory and quality of life on par with seizure freedom, while the cervical alignment threshold in deformity surgery awaits external validation.
That's your roundup for This Week in Neurosurgery. The full transcript and references are available on the episode page. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.
If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if it saves you time, a quick rating in your podcast app helps other physicians find the show.
This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.
Spot something worth flagging?
Get this every week in your podcast app — free.
New neurosurgery episodes land in your feed automatically — listen on your commute.