Introduction Tumor treating fields (TTFields) are low-intensity (1-3 v/m) intermediate
frequency (200 khz for GBM) alternating fields that disrupt cell division. The treatment is
increasingly used as a supplementary modality for patients with glioblastoma (GBM). Recent
randomized clinical studies have demonstrated tolerability of TTFields (Optune®, Novocure®)
for both newly diagnosed and recurrent GBM and a significant and sustained survival benefit
in the former population. Specifically, the addition of TTFields to maintenance temozolomide
(TMZ) therapy for newly diagnosed GBM patients, who had completed initial concurrent
radio-chemotherapy, resulted in significantly prolonged median progression-free survival (PFS
= 6.7 vs.#46; 4.0 months, p<0.001) and median overall survival (OS = 20.9 vs.#46; 16.0 months,
p<0.001) compared to maintenance TMZ therapy alone (ef-14 trial) (4). Consequently, TTFields
were recently approved by the US FDA and included as a category 1 recommendation in the most
recent guidelines from the National Comprehensive Cancer Network (NCCN) for the treatment of
newly diagnosed GBM. For recurrent GBM, TTFields monotherapy has been shown to have
equivalent efficacy and a superior toxicity profile compared to best practice medical
oncological therapy. This led to US FDA approval for this indication in 2011. In addition, a
post-hoc analysis of the EF-14 dataset has indicated that the addition of TTFields to 2nd
line medical oncological treatment at first disease recurrence leads to an overall survival
benefit. Finally, the pride dataset showed that the addition of TTFields to current practice
of medical oncological treatment was beneficial.
Recently, the investigators proposed a novel approach to calculate the electrical field
distribution induced by TTFields therapy of GBM. The approach utilizes finite element
analysis and incorporates personalized MRI data to provide a realistic and individualized
estimate of the field intensity distribution, which is associated with the antimitotic
efficacy of the treatment. Using this approach, the investigators showed that targeted
skull-remodeling surgery, e.g. craniectomy or multiple burr-holes, can provide a substantial
and highly focused enhancement (~100%) of the median field intensity in supratentorial
tumors, while leaving the field distribution in healthy tissues unaffected.
These results support the theoretical rationale that small and clinically feasible
craniectomies may improve the clinical efficacy of TTFields for cerebral hemispheric tumors
without compromising patient safety. Based on these findings, the investigators recently
concluded an open-label, single arm, prospective phase 1 trial (NCT02893137) to investigate
safety and feasibility of SR-surgery with TTFields and best practice neuro-oncological
therapy for recurrent GBM. The trial was active from December 2016 to May 2019. The planned
total number of patients was fifteen. Eligibility criteria included age > 18 years, first
recurrence focal supratentorial GBM (RANO), and KPS ≥ 70. Patients were right-censored for OS
and PFS at the time of analysis and excluded upon progression, death, SUSARs, or unacceptable
AEs. The primary endpoint was toxicity (CTCAEv4.0). The secondary endpoints were OS, PFS, PFS
rate at six months (PFS6), objective response rate (ORR) based on the iRANO criteria (14),
quality of life (QoL), Karnofsky performance score (KPS), and steroid dose.20 patients were
screened, 2 declined, and 3 had KPS < 70. Of the 15 enrolled patients, 4 were excluded prior
to TTFields due to radionecrosis/non-recurrence, post-op infection, neurodeficit and
withdrawal of consent, respectively. All included patients (11 male and 2 female) had GBM
IDH-wt tumors (4 MGMT-methylated). Median KPS was 90 (range 70-100) and median age 57 years
(range 39 to 67). All patients received maximum safe resection at recurrence (4 had no
residual tumor (RANO), 5 non-measurable disease, and 2 measurable disease). 9 patients
received adjuvant bevacizumab monotherapy and 2 temozolomide rechallenge. The mean
skull-defect area was 10.6 cm2 (range 7 to 37 cm2) and the median field enhancement 43 %
(range 25 to 59%). The median follow-up duration was 10 months and the median TTFields
compliance 90% (range 48 to 98%). We observed no SUSARs, no grade 4/5 SAEs, 12 grade 3 SAEs
(6 generalized seizures in patients with known epilepsy, 1 post-op infection, 1 diarrhea,
1TIA, 1 fatigue, 1 headache and 1 DVT). The most common grade AE 1-2 was headache 60% CI95%=
[32; 84], fatigue 53%, CI95%= [27; 79], skin rash 47%, CI95%= [21; 73], and nausea 40%,
CI95%= [16; 68].
Regarding survival results were PFS6 was 64%, CI95%= [35; 85], PFS= 8.8 months, CI95%= [6.2;
13.2], OS= 15.0 months, CI95%= [9.6;16.2], and OS12= 64%, CI95%= [35;85]. One patient had
complete response during TTFields. Based on these results, the investigators concluded that
skull-remodelling surgery incl. craniectomy is not associated with additional toxicity in
combination with TTFields and best practice medical oncological treatment and holds promising
potential for improving TTFields outcome by focally enhancing the field intensity in the
tumor. The proposed phase 2 trial aims to validate this hypothesis in a randomized
comparative setting.
METHODS AND ANALYSIS Trial design The study is designed as an investigator-initiated,
prospective, multi-center, multi-national, randomized, minimax two-stage, comparative, phase
2 trial, investigating superior efficacy of the intervention. Patients randomized to the
control arm will receive TTFields plus best physician's choice medical oncological treatment
(BPC treatment) and patients in the interventional arm will receive SR-surgery in addition to
TTFields and BPC treatment. The trial will enroll an expected sample size of 70 patients from
4-6 scandinavian countries with Aarhus University Hospital, Denmark, as the Sponsor and
coordinating site. The primary outcome will be overall survival at 12 months (OS12) and the
trial is designed to detect a 20% increase in OS12 in the interventional arm compared to
control (from 40% in the control arm to 60% in the interventional arm). Secondary outcomes
will include progression-free survival (PFS), quality of life (QoL), clinical performance,
and objective response rate (ORR). Interim futility analysis will be conducted after 12
months follow-up of the first 52 patients. The trial will be stopped at interim analysis if
the OS12 months is lower in the control arm compared to the interventional arm. The trial
design and data reporting is in accordance with the consort 2010 statement, the gnosis
standards for neuro-oncology trials, recent RANO recommendations for phase 2 trial design in
neuro-oncology, and general recommendations for phase 2/3 trial design (16-19). The study
will be performed in the period march 2020 to march 2023. The inclusion period is expected to
be the first 24 months of the trial.
Participants Seventy (70) patients with progressive GBM according to RANO criteria will be
enrolled. Potential trial participants will be identified at an institutional
multidisciplinary neuro-oncological tumor boards and subsequently referred for eligibility
screening. All inclusion scans are assessed by a trained neuroradiologist. Patients who are
immediately ineligible, e.g. due to poor performance status, multifocal disease, significant
comorbidity, evidence of extracranial primary tumor, or other excluding circumstances, will
not be referred for screening.
Experimental intervention Patients will be randomised 1:1 to one of two study arms to receive
either, 1) SR-surgery, TTFields and best practice medical oncological therapy (interventional
arm), or 2) TTFields and best practice medical oncological therapy alone (control arm). Best
practice treatment will typically also include maximum safe resection. Since the intervention
(SR-surgery) is only effective during TTFields therapy, TTFields will continue until 1)
patient exclusion from the trial, 2) occurrence of a suspected unexpected serious adverse
reaction (SUSAR) related to the intervention, or 3) ethical or medical safety
contraindication for further TTFields therapy determined by the investigators. This means
that treatment may continue beyond progression, as long as active treatment is indicated, and
in some cases as compassionate use.
Sample size and statistical considerations The sample size calculations were based on a
randomized, comparative, minimax two-stage design and the binomial primary outcome of OS12
for the patient population treated per protocol, i.e. for whom TTFields therapy was
initiated. This means that patients excluded prior to TTFields therapy initiation are not
included in the primary outcome analysis. Therefore, additional patients may be enrolled to
ensure that the planned sample receives active TTFields treatment. The aim of the trial is to
determine whether the intervention is superior to control and thus worthy of further phase 3
investigation. The expected OS12 in the control arm is set to 40% based on the EORTC 26101
trial, in which patients with first recurrence of GBM were treated with lomustine monotherapy
or lomustine/bevacizumab combination therapy. The trial showed an OS12 of approximately 30%
and we therefore set the expected level of OS12 to be slightly above this threshold given the
fact that TTFields is added to the treatment which otherwise represents recommended practice.
There are no current studies providing accurate OS12 estimates for a population comparable to
the study control arm and we therefore based the estimate on an expected benefit of 10% from
TTFields alone compared to EORTC 26101. Setting the probabilities of false-positive and
false-negative trial results to α = 0.15 and 1-β = 0.80, respectively, and defining the
expected target level of OS12 to be OS12 = 0.6, i.e. a 20% absolute increase (and 50%
relative increase) compared to the control arm, we calculated a maximum sample size of n = 42
patients in each arm (total n = 84) and an expected total sample size (en = 69.8). The
expected sample size en is calculated as en = n1 × pet0 + n × (1
- - pet0), where n1 = 52 is
the collective sample size of both groups at interim (stage 1) analysis, pet0 = 0.44 is the
probability of early futility termination.
The numbers are based on the given statistical
parameters and a two-stage minimax design as given in sh jung 2018. Interim futility analysis
will be conducted after endpoint assessment of 52 patients and the trial will be terminated
if the experimental arm performs worse than the control arm. In this case the futility
criterion will be defined as OS12experimental < OS12control and this conclusion will
determine futility at marginal power and significance levels of 1-β* = 0.80 and of α* = 0.20,
respectively. If the trial is not terminated at interim stage 1 analysis, the trial will
proceed to enroll a total of 84 patients and terminate when the final OS12 endpoint data have
been obtained for all patients.
The secondary outcome measures of hazard rate ratios of PFS and OS will be tested using the
log-rank test at the 0.05 alpha level. Time-to-event endpoints will be estimated from the
time of randomization. Toxicity and adverse events will be reported using absolute numbers
and appropriate risk estimates. Final analysis will be conducted when all patients have been
excluded or censored, e.g. due to end-of-trial, or loss-to-follow-up. Subgroup analysis based
on prognostic factors will be performed, incl. analysis of patient characteristics in each
arm and identification of characteristics of potential responders to the intervention. This
will also include a correlation analysis between the calculated field distribution and
outcome estimates. Analyses will be based on both the intent-to-treat population and the
population proceeding to active therapy (per protocol), i.e. At least four weeks of active
TTFields therapy with ≥75% average compliance.
Randomisation and blinding Patients will be randomised immediately upon enrollment in the
trial, i.e. prior to initial surgery. Randomisation will be performed using randomization.com
and patients will be randomised in a 1:1 ratio to receive TTFields and best practice medical
oncological treatment w/o skull remodeling surgery. The study endpoint assessment committee
will be blinded to the intervention. Local principal investigators, co-investigators and
sub-investigators will not be blinded, as optimal array planning will require knowledge about
skull-remodeling surgery. The operating surgeon performing the skull-remodeling surgery will
not be blinded. Patients will not be blinded, as it will be possible for the patient to feel
the skull configuration and thereby determine whether SR-surgery was performed.
Risks and safety precautions General safety considerations concerning Optune® therapy are
outlined in the related technical and safety material provided by Novocure®. In general,
TTFields therapy is regarded a safe and recognized treatment for GBM. The most common adverse
effect is local skin rash in the area underlying the transducer arrays. The risk of skin rash
will be reduced by slightly moving the arrays during regular array changes. Skin rash may be
treated with topical corticosteroids, more frequent electrode replacement and short pauses in
TTFields therapy, if needed. The risk of brain injury due to the skull remodeling aspect of
surgery is minimal, as the entire operation is performed extra-durally. Minor craniectomies
do not pose a significant risk of mechanical brain injury. For larger craniectomies, external
protective headgear, such as a cap with mechanical reinforcement over the unprotected brain
area, may be used to reduce the risk of traumatic brain injury after craniectomy. In the case
of post-operative infection or impaired wound healing, antibiotic therapy will be
administered. If needed, surgical revision/debridement of the wound and potentially removal
of the bone flap may be performed. In such cases, the patient may proceed to receive or
continue TTFields therapy at an appropriate time, if it is in the patient's best interest.
TTFields therapy following removal of the infected bone flap may be accepted even if the
craniectomy size exceeds 30 cm2. In all cases, the decision whether to continue with TTFields
therapy after surgery will be based on careful consideration of the expected safety and
benefit for the patient. All participants will be advised of situations and adverse events,
in which medical attention is required. Investigators or properly trained medical personnel
will be available for immediate attention or medical advice specific to the treatment or
protocol. Patients will receive contact information to the device support specialist, whom
they may contact in case of technical problems or considerations.
Monitoring and quality control The project will comply with international guidelines good
clinical practice (ICH-GCP and ds/en iso 14155:2012. Clinical investigation of medical
devices for human subjects
- - good clinical practice) and be monitored by the local contract
research organization (CRO).
Source data, verification of consent, and handling of adverse
events and near-miss incidents will be subject to 20% sampled monitoring. Full access to all
source data, incl. patient records, will be granted in relation to trial inspection by the
relevant national competent authorities. Experimental "waivers" from the protocol will be
registered in the project trial master file and reported to the relevant authorities. Changes
to the protocol will be submitted for approval by the relevant authorities in accordance with
GCP guidelines and regulations.
Dissemination All results will be published in peer-reviewed international scientific
journals and presented at international scientific conferences, regardless of academic
conclusions. Positive, negative, and inconclusive results will be publicly available.
