The Surveillance Window That Glioblastoma Does Not Forgive
Glioblastoma (GBM) is the most aggressive primary brain tumor in adults. Real-world data shows median overall survival around 12 months, based on a 2023 analysis of 467 consecutive GBM patients, and most tumors recur within 6 to 9 months of initial resection. This short timeline means every scheduled post-operative MRI matters. A missed scan is serious. It is a lost opportunity in a disease where time is limited.
Neuro-oncology clinics already know this. Most lack a systematic workflow to prevent this gap.
What the Compliance Data Actually Shows
In 2024, the INTERVAL-GB multi-centre cohort study examined MRI surveillance timing across centres in Great Britain and Ireland. The findings showed serious gaps. Only 52.8 percent of patients received MRI surveillance that matched NICE follow-up imaging recommendations. Compliance with the tighter EANO recommendations fell to 24.9 percent, according to published findings in Neuro-Oncology. Roughly three in four patients did not receive imaging at the frequency their guidelines recommend.
This is not a skill problem. Clinicians know what to do. The problem is coordination and tracking. Neuroimaging after GBM resection involves multiple handoffs: the neurosurgeon's discharge plan, the neuro-oncologist's treatment schedule, the radiology department's booking capacity, and the patient's ability to attend. Each handoff is a place where timing slips.
Recommended Intervals and Why They Drift
Post-operative MRI should occur within 48 hours of resection to establish a clean extent-of-resection assessment, before post-surgical swelling obscures imaging margins. A second MRI typically follows 2 to 8 weeks after completion of chemoradiation. After that, NICE recommends imaging every 3 to 6 months and EANO recommends every 3 months for the first two years.
Research published in Neuro-Oncology Advances proposed an evidence-based schedule derived from modeling of 277 high-grade glioma patients. The optimal schedule: imaging roughly every 7.4 weeks for the first 120 weeks after treatment completion, followed by every 27.6 weeks thereafter. The key finding is that many clinics determine radiological surveillance plans arbitrarily in daily practice rather than using evidence-based protocol.
The gap between recommended and actual intervals widens for several practical reasons.
- Chemoradiation schedules are managed by radiation oncology, while follow-up MRIs are typically ordered by neuro-oncology. When the two teams use separate scheduling systems, post-treatment imaging can fall through.
- Pseudoprogression - a treatment-related imaging change that mimics recurrence - prompts short-interval repeat scans at many centres, but these extra appointments are often handled one-off rather than built into a protocol track.
- Most scheduling tools are calendar-based rather than protocol-driven. They record an appointment once it is booked. They do not alert staff when a required appointment was never booked at all.
That last point matters most. A calendar that is silent about a missing MRI looks identical to one where the MRI is already ordered. The gap is invisible until it is too late.
Early Post-Operative Imaging and Its Prognostic Role
Early post-operative MRI does more than confirm resection extent. A study of 187 consecutive GBM patients receiving standard therapy, published in Scientific Reports, found that radiological findings on early post-operative MRI had independent prognostic relevance for overall survival. For clinic operations, the key point is direct: the quality of your surveillance workflow shapes the data available for clinical decisions from the first week after surgery.
Extent of resection is one of the strongest modifiable prognostic factors in GBM. Confirming that extent within 48 hours - before post-surgical swelling alters the imaging picture - is guideline standard. Delays at this early stage cascade through the entire treatment and surveillance chain.
Where the Workflow Breaks Down
Three failure modes appear consistently across neuro-oncology programmes.
Handoff from surgical to oncology scheduling. Neurosurgeons often discharge patients with an instruction to arrange MRI within 48 hours alongside an outpatient neuro-oncology referral. If the oncology clinic's first available appointment is two weeks out, the 48-hour window closes without a formal scan order. Most scheduling platforms have no automatic trigger to catch this.
Treatment-phase tracking lag. During the Stupp protocol, patients attend radiation oncology daily for roughly six weeks. The post-chemoradiation MRI is the next key imaging touchpoint for neuro-oncology. If that appointment is not proactively booked during the radiation phase - often by a different administrative team - it drifts. Median time to recurrence is 6 to 9 months, so a scan that slips by even four weeks represents a meaningful share of the expected progression-free window.
Recurrence detection delay. When surveillance is intermittent rather than protocol-driven, progression may not be detected until the patient becomes symptomatic. The clinical question is not only whether recurrence occurred, but whether the surveillance programme detected it at a point where intervention remains viable. For a related look at how post-surgical surveillance gaps play out in another cancer type, see our analysis of post-surgery surveillance scheduling in colorectal cancer.
What a Protocol-Driven Scheduling System Changes
A protocol-driven approach replaces the silent calendar with an active alert layer. Rather than recording only booked appointments, the system tracks expected appointments relative to the patient's treatment phase and flags any that have not been ordered or completed by a defined threshold date.
In the Rucja platform, a clinic can define GBM post-resection milestones: 48-hour post-operative MRI, post-chemoradiation MRI, and quarterly surveillance imaging. The platform shows which patients are within their expected imaging window and which have passed it without a scan on record. Staff do not need to audit each chart individually. The gap appears in the queue view before it becomes a missed appointment.
AI-powered report reading sits alongside this scheduling layer. When a radiology report arrives - whether as a structured HL7 message or a PDF upload - the AI extraction layer parses the impression section, flags terms consistent with progression, and routes the case for clinical review. A neuro-oncologist reviewing their morning list sees a trend across recent scans for each patient, not just an isolated report from one date. That context matters when distinguishing pseudoprogression from true recurrence. For a broader look at how AI report extraction changes the daily clinical routine, see our post on moving from PDFs to patient insights.
Scheduling tools built around oncology treatment protocols - rather than adapted from general outpatient calendars - also support the multi-team visibility that GBM care requires. Radiation oncology, neuro-oncology, and neurosurgery can each view the patient's imaging timeline and know which team holds responsibility for the next scan order. Our post on why oncology clinics need treatment timelines, not just calendars covers this broader point in more depth.
Operational Considerations for Clinic Leads
For hospital administrators and clinic operations leads evaluating workflow platforms, GBM surveillance is a useful test. If a platform cannot answer the question - which of our post-resection GBM patients is currently overdue for an MRI - from a single queue view, it is functioning as a general appointment scheduler rather than a surveillance management system.
Useful questions to bring to a platform evaluation:
- Can the system track expected imaging events independently of whether they have been booked?
- Does the platform support multi-specialty visibility so that radiation, neuro-oncology, and neurosurgery teams share a single patient timeline?
- When a radiology report arrives, does it surface in the relevant clinician's workflow within the same session - rather than sitting in a separate document repository?
- Does the alert logic distinguish between a scan that is not yet due, a scan that is due and booked, and a scan that is due but not yet booked?
These distinctions separate a scheduling tool from a surveillance management system. For a disease like GBM, where the recurrence window is measured in weeks, that distinction carries direct clinical and operational weight.
Demos take 30 minutes. We will walk you through this article's surveillance workflow on your live clinic data. Book a demo.