Scheduled Brain Tumor Surgery Keeps Patients Home Longer Without Harming Survival

When adults are diagnosed with a high-grade glioma, the most common primary malignant brain tumor in adults, the path to the operating room usually follows one of two routes: emergency admission for immediate resection, or discharge home with a planned operation days or weeks later. A retrospective cohort study now offers one of the clearest pictures yet of who ends up on each track and what those tracks cost in hospital time and safety. Across 166 patients operated on at a single US center between 2021 and 2026, urgent and scheduled surgery produced no detectable difference in complications, extent of resection, timing of adjuvant therapy, or survival. What differed dramatically was where the first month after admission was spent: scheduled patients lived a median of 27 of their first 30 days at home, compared with 15 days for urgent patients, an adjusted difference of 8.9 days.

The study, published in the Journal of Neuro-Oncology, included adults with newly diagnosed WHO grade 3 or 4 gliomas who underwent primary resection at Allegheny General Hospital between January 2021 and July 2026. The treating team’s plan at diagnosis, based on imaging and clinical presentation, determined the pathway classification. Patients resected during their initial hospital admission, or recalled for unplanned surgery before ever establishing care with the neurosurgical oncology service, were labeled urgent. Patients discharged and readmitted for planned surgery, or managed entirely as outpatients, were labeled scheduled. Of the 44 scheduled patients, 40, or 91 percent, proceeded exactly as planned; only four required earlier unplanned operations during the waiting interval and were analyzed in their assigned group.

The two pathways sorted patients into visibly different clinical phenotypes. Urgent patients arrived sicker: their median preoperative Karnofsky Performance Status was 70 versus 80 in the scheduled group, and 17 percent were ASA class IV compared with 4.5 percent. Altered mentation, gait disturbance, visual changes, and headache were all significantly more common at presentation among urgent patients. Seizure told the opposite story, occurring in 39 percent of scheduled patients but only 15 percent of urgent ones, and an incidental, asymptomatic diagnosis was almost exclusively a scheduled phenomenon, 14 percent versus under 1 percent. Nearly all urgent patients, 94 percent, needed corticosteroids at presentation, with 84 percent receiving 16 milligrams per day or more of dexamethasone equivalent, against 48 percent of scheduled patients at lower doses.

Imaging reinforced the same divide. Urgent patients carried larger tumors, a median maximum diameter of 48 millimeters versus 36, and far more frequent midline shift, the displacement of brain structures that signals dangerous intracranial pressure, at 62 percent versus 32 percent. Eloquent-cortex involvement, tumor location, WHO grade 4 histology, IDH mutation status, and MGMT promoter methylation did not differ between the groups, meaning the biological aggressiveness of the tumors was broadly comparable and the separation between pathways lay mainly in presentation severity and mass effect.

The most striking finding concerned healthcare utilization, measured with an endpoint called days alive and at home within 30 days of admission, or DAH-30. The window was anchored at hospital admission rather than surgery specifically so that preoperative inpatient days accrued by urgent patients would count against them. Urgent patients stayed a median of 8 days in their index admission against 3 days for scheduled patients, with the excess driven almost entirely by preoperative days, a median of 4 versus 0. In a negative-binomial model the urgent pathway roughly doubled length of stay, with a rate ratio of 2.30 and an overlap-weighted difference of 4.6 days. Non-home discharge to rehabilitation or other facilities was also more common after urgent surgery, 53 percent versus 27 percent, although the overlap-weighted odds ratio of 1.71 was attenuated.

Safety outcomes showed no detectable penalty for either approach. Any 30-day complication occurred in 36 percent of urgent and 30 percent of scheduled patients, an adjusted odds ratio of 1.15 that crossed unity comfortably, and an overlap-weighted odds ratio of 1.03. Individual events were infrequent on both sides: surgical-site infection, intracranial hemorrhage, cerebrospinal fluid leak, new postoperative seizure, and venous thromboembolism each clustered between roughly 2 and 6 percent with no significant between-group difference. Permanent neurologic deficits persisting to three months affected 12 percent of urgent and 14 percent of scheduled patients, and 30-day mortality was 3.3 percent in the urgent arm and zero in the scheduled arm, a gap that did not reach significance. Extent of resection was essentially identical, with gross-total resection achieved in 39 versus 41 percent of cases.

Survival also failed to distinguish the pathways. With a median follow-up of 26.2 months and 94 deaths, median overall survival was 15.4 months for urgent patients and 19.0 months for scheduled patients, with a log-rank p-value of 0.758. Median progression-free survival was 6.6 versus 7.0 months. The crude hazard ratio of 1.08 shifted to 0.64 after adjustment for age, performance status, extent of resection, and MGMT methylation, a move the authors attribute to confounding by indication, since the same features that route a patient to urgent surgery also worsen prognosis. Notably, IDH-mutant tumors made up 4.9 percent of the urgent arm versus 11 percent of the scheduled arm, and none of the 11 patients with IDH-mutant tumors died during follow-up. On an absolute scale, the adjusted difference in restricted mean survival time at 24 months was 0.34 months, with confidence intervals spanning roughly minus 2.5 to plus 3.2 months. Subgroup analysis of the 152 patients with IDH-wildtype grade 4 glioblastoma reproduced the pattern: fewer days at home for urgent patients, no detectable difference in complications or survival.

Scheduling also unlocked planning-intensive surgical techniques. Awake craniotomy, which requires preoperative neuropsychological screening and specialized anesthesia arrangements incompatible with emergency timelines, was used in 18 percent of scheduled cases but only 6.6 percent of urgent ones. By contrast, 5-aminolevulinic acid, a tumor-fluorescence agent taken orally two to four hours before surgery, was deployable in both pathways at statistically similar rates. Crucially, deferring surgery did not delay adjuvant treatment: the intervals from surgery to radiotherapy, 36 versus 39 days, and to chemotherapy, 36.5 versus 39.5 days, did not differ and both fell within the windows recommended by the American Society for Radiation Oncology. The authors also examined the seven patients whose diagnosis-to-surgery interval exceeded 90 days, all monitored with serial imaging, and found that excluding them did not materially change any estimate.

The four scheduled patients who crossed over to earlier unplanned surgery offer a cautionary footnote. All four presented with a Karnofsky score of 70, the median of the urgent arm, three of the four had midline shift, aphasia, or tumors exceeding 50 millimeters, and none presented with seizure. The subgroup is too small for statistical inference, but its members shared features characteristic of the urgent phenotype, suggesting that the same clinical markers separating the pathways may also flag scheduled patients at risk of needing earlier intervention. Corticosteroid dependence may be another signal: three of the four crossover patients went home on 16 milligrams per day of dexamethasone equivalent while awaiting surgery.

The study’s authors argue that a substantial share of patients currently funneled into emergency resection could instead be managed on an expedited outpatient basis, noting that 15.6 percent of urgent patients had a Karnofsky score of at least 80, tumors smaller than 5 centimeters, and no midline shift. In a disease whose median survival in this cohort was 15.7 months, nearly nine additional days at home in the first month, well beyond the 3-day threshold considered minimally clinically important for this measure, represents time for patients to consult family, weigh decisions about extent of resection and clinical trial enrollment, and prepare for multimodal therapy from their own beds rather than hospital wards. The authors caution that their data reflect institutional selection practices rather than formalized criteria, that confounding by unmeasured variables cannot be excluded, and that molecular characterization did not include TERT promoter or CDKN2A/B status. Whether patients now treated urgently could be safely scheduled is a question they say demands prospective, multi-institutional evaluation, supported by multidisciplinary clinics of nurse navigators, neuro-oncologists, radiation oncologists, and neurosurgeons. For now, the single-center evidence establishes that when carefully selected patients wait for planned surgery, they spend more of their limited time at home, and no measurable harm has been observed.

Subject of Research: Urgent versus scheduled resection pathways, patient selection, and healthcare utilization in newly diagnosed high-grade glioma

Article Title: Urgent versus scheduled resection in high-grade glioma: patient selection and healthcare utilization in a single-center cohort

Article References: Mahmoudi, D. A., Kite, T., Singh, E., McKenna, A., Shaikh, S., Vyas, P. S., Herbst, J., Wegner, R. E., Karlovits, S., Yu, A., Whiting, A. C., & Shepard, M. J. (2026). Urgent versus scheduled resection in high-grade glioma: patient selection and healthcare utilization in a single-center cohort. Journal of Neuro-Oncology, 180(2), Article 29. https://doi.org/10.1007/s11060-026-05841-8

Image Credits: AI Generated

DOI: 10.1007/s11060-026-05841-8

Keywords: high-grade glioma, glioblastoma, neurosurgery, scheduled resection, urgent surgery, days alive at home, healthcare utilization, Karnofsky performance status, patient selection, retrospective cohort, brain tumor, perioperative outcomes