Gamma Knife vs LINAC-based radiosurgery for vestibular schwannoma (retrospective review)
ForVestibular schwannoma, mostly Koos II solid, intracanalicular or mixed
TL;DRLocal control 98.8%; no difference in new vertigo, gait disturbance or facial palsy with GKRS (n=36) vs LINAC (n=49) SRS for vestibular schwannoma.
The 98.8% local control is freedom from further VS-directed intervention; volumetric enlargement still occurred in 6/36 GKRS and 9/49 LINAC pts. The LINAC arm pools margin-free 12 Gy to the 80-99% IDL with post-May 2017 frameless 12 Gy plus a 5 mm PTV margin, era split unreported, so it cannot say which LINAC approach matches GK.
In small, mostly Koos II vestibular schwannoma, this supports GK or LINAC single-fraction SRS chosen by local availability when weighing new vertigo, gait disturbance or facial palsy; it does not inform hearing preservation, which the series did not report.
The 98.8% local control is freedom from further VS-directed intervention; volumetric enlargement still occurred in 6/36 GKRS and 9/49 LINAC pts. The LINAC arm pools margin-free 12 Gy to the 80-99% IDL with post-May 2017 frameless 12 Gy plus a 5 mm PTV margin, era split unreported, so it cannot say which LINAC approach matches GK.
9 details
Retrospective chart review across multiple sites of a single institution, VS treated 2008-2017; N=85 (36 GKRS, 49 LINAC). Platform set by provider preference, as GKRS and LINAC sit at different centers. Median follow-up 27 mo.
Vestibular schwannoma, 95.3% solid, intracanalicular (45.9%) or intra- plus extracanalicular (43.5%); Koos II majority with no Koos difference between cohorts (p=0.65). Pre-SRS symptoms were common: tinnitus 62.4%, gait disturbance 48.2%, vertigo 30.6%.
All single-fraction SRS. GKRS: 12.5 Gy to the 50% IDL, stereotactic head frame. LINAC: 12 Gy to the 80-99% IDL with no CTV/PTV margin, then from May 2017 frameless, 12 Gy to the 100% IDL with a 5 mm PTV margin by VMAT.
No primary endpoint stated in source. Outcomes were new tinnitus, vertigo, gait disturbance and facial palsy (HB grade 3 or higher) in pts free of each at baseline, plus volumetric MRI tumor control. Hearing not reported; trigeminal dysfunction excluded for inconsistent documentation.
Local control 98.8%, with 1/85 pts needing salvage (surgery, LINAC arm). No difference in direction of tumor size change (p=0.199).
| New symptom | GKRS | LINAC | p |
|---|---|---|---|
| Tinnitus | 1/13 (7.7%) | 1/19 (5.5%) | 0.0007 (MVA) |
| Vertigo | 1/28 (3.6%) | 2/31 (6.5%) | 0.785 |
| Gait disturbance | 4/18 (22.3%) | 4/26 (15.4%) | 1.000 |
| Facial palsy (HB ≥3) | 0/34 (0.0%) | 2/48 (4.2%) | 0.509 |
| Tumor outcome | GKRS (n=36) | LINAC (n=49) |
|---|---|---|
| Decreased or stable | 30/36 | 40/49 (81.6%) |
| Increased size | 6/36 (16.7%) | 9/49 (18.4%) |
| Salvage required | None | 1/49 (2.0%) |
Both new LINAC facial palsies resolved (2/2); none occurred after GKRS. New gait disturbance resolved in 3/4 GKRS and 2/4 LINAC pts, and the single new GKRS tinnitus persisted.
Authors position this against prior series reporting favorable local control with GKRS or LINAC individually, where direct symptom and QOL comparisons were lacking. The source names no randomized GK vs LINAC comparison; the authors call for prospective data.
The 'QOL' outcomes are chart-abstracted symptoms, with HB grade assigned by the reviewing author when not documented. Mean follow-up differs (24 vs 36 mo), and the tinnitus p=0.0007 on multivariable analysis rests on 1 event per arm. Methods and results disagree on pre-2017 LINAC immobilization.
Median follow-up of 27 mo sits inside the 3-year window in which the authors cite 20-30% pseudoprogression, so the enlargements in both arms are not yet resolved as benign. The data show no signal of platform-specific vestibular or facial harm; they do not show parity.
Retrospective, provider-preference allocation, N=85; symptom comparisons use difference tests on single-digit event counts, so the 'interchangeable' conclusion is not something this design can establish.
- Does hearing preservation differ between GK and LINAC SRS?
- Does a 5 mm PTV margin on frameless LINAC add cranial nerve toxicity?
- Tumor control beyond the 3-year pseudoprogression window
📚 Sources · 📄 1 paper
The longer read
The narrow question this series can answer is whether the delivery platform changes early cranial nerve and vestibular symptoms after single-fraction SRS for vestibular schwannoma, and on that question it finds no signal. New vertigo, gait disturbance and facial palsy did not differ between 36 GKRS and 49 LINAC pts, and 98.8% of the cohort needed no further VS-directed intervention. For a department that owns only one platform, that is reassurance worth having. The authors go further and call the approaches interchangeable, and that claim sits beyond what the design can carry.
The distance between those two readings is statistical. New-symptom counts per arm top out at 4 (gait disturbance, 4/18 GKRS and 4/26 LINAC), the comparisons are difference tests rather than an equivalence design, and at these counts a null result mainly reports wide uncertainty. The single positive finding shows how fragile the modelling is: a multivariable p=0.0007 for tinnitus, built on one new case in each arm, is not a result to carry forward, and the abstract itself describes the tinnitus difference as not clinically significant.
On the radiation side the comparison is less clean than two labels suggest. GKRS prescribed 12.5 Gy to the 50% isodose line, which leaves the tumor interior well above the marginal dose. LINAC prescribed 12 Gy to the 80-99% isodose line with no margin, then from May 2017 moved to a frameless mask, 12 Gy to the 100% isodose line and a 5 mm PTV margin by VMAT. That is three dose distributions, not two, and the source does not report how many LINAC pts fell in each era. A 5 mm expansion around a tumor in the internal auditory canal widens the treated volume along the facial and cochlear nerves; whether the two LINAC facial palsies, both of which resolved, came from the margin era is the question the pooled analysis cannot answer. Allocation followed provider and campus, and the GKRS cohort carried more extracanalicular-only tumors (25.0% vs 0.0%) and brainstem mass effect (19.4% vs 4.1%), so the arms were not interchangeable at baseline either.
The local control figure needs its definition attached. The 98.8% is freedom from salvage, while volumetric enlargement was recorded in 6/36 GKRS and 9/49 LINAC pts. The authors cite pseudoprogression in 20-30% of pts within three years of radiosurgery, and with a median follow-up of 27 mo much of this cohort has not cleared that window, so those enlargements are unresolved rather than settled as benign. Hearing, central to counseling any VS pt about SRS, is not reported, although cochlear dose was collected.
What the study supports is modest: for a small, mostly Koos II tumor, there is no signal here that the choice between GK and LINAC moves vestibular or facial outcomes over a median 27 mo. What would make the interchangeability claim wrong is a difference in hearing or late control that this design cannot see. The authors call for prospective data; the useful version would carry hearing endpoints and report frameless, margin-expanded LINAC separately from the margin-free technique.