Role of Repeat Resection in Recurrent Glioblastoma

Patients with recurrent Glioblastoma (GBM) are commonly presented to surgeons, along with the question of whether or not to re-resect the recurrence. There is no Level 1 evidence to support a role for repeat surgery in this context, but a multitude of observational research suggests that repeat surgery may improve quality survival. Unfortunately, these studies all suffer from selection bias. The goal of this study is to provide a care trial context to help neurosurgeons manage patients presenting with recurrent GBM, with no additional risks, tests, or interventions than what they would normally encounter in routine care. Secondary goals include a test of the hypothesis that repeat resection can improve median overall survival, and that it can increase the number of days of survival outside of a hospital/nursing/palliative care facility.

Glioblastoma (GBM) is the most common malignant brain tumor (48.3% of malignant tumors).1 It has a predilection for Caucasian men with a mean age of 65.1, 2 There are approximately 13,000 new diagnoses of GBM made every year in the United States.1 Patients have an average survival of 12-15 months, with a 5-year survival rate of 6.8% despite various treatments.1 If the prognosis at the time of initial diagnosis is grim, it is even worse at the time of recurrence. Recurrence after the initial resection, which can be symptomatic or discovered on surveillance MRI imaging, occurs in nearly all patients, usually within the first year, even when the initial management is aggressive.3-6 There is no standard way to care for recurrent GBM patients and treatment may include one or more of the following: repeat surgery, radiation therapy, or second-line chemotherapy.7 Repeat surgical management carries a greater risk of wound infection and cerebrospinal fluid leak than the initial surgery, especially in patients who received radiation.8 When maximal resection was attempted at the time of the initial procedure, the second surgery is more at risk of neurological injury. Repeat surgical management often also entails a delay in the initiation of further chemotherapy, as these agents are not administered peri-operatively because of potentially deleterious effects on wound healing. Nonetheless, despite, or because of the desperate setting, patients and surgeons are often pressed to consider repeat surgery, which is performed in at least 10-30% of patients.9 Repeat surgical treatment may be more favorable with younger age (\<60 years), and preoperative Karnofsky performance status (KPS) of at least 70.9 Median survival following repeat resection varies between 13-54 weeks,10, 11 but all reports suffer from the limitations of retrospective studies, the most obvious being selection bias in terms of patient age, functional status, tumor location and size.11, 12 Yet, despite favorable biases some studies report no improvement in terms of survival or quality of life.13, 14 Two recent systematic meta-analyses of case series offer conflicting conclusions about the merit of re-operation.15, 16 If the intent of repeat surgery is to improve quality and quantity of life, there is no level 1 evidence it is effective.3, 16, 17 Thus repeat surgery should be considered the experimental arm of a randomized trial designed in the best interest of the patient, being only offered as a 50% chance, always balanced by a 50% chance of being allocated non-surgical care, or a chance to escape what could turn out to be unnecessary or harmful surgery.18 For patients may be better served by being free to pursue other life priorities in the company of loved ones rather than being scheduled to endure repeat craniotomy and multiple clinical or research follow-up visits. There is sufficient community equipoise to support the conduct of such a randomized trial.19 Tria Design: The trial is a simple, all-inclusive, prospective, multicenter, randomized care trial18 that allocates 1:1 re-operation (or not) for patients with recurrent Glioblastoma. The primary outcome is overall survival. Secondary outcomes include standard peri-operative safety outcomes and (a notion of) 'quality survival', or survival at home, measured by counting days of survival minus days in hospital/nursing home/palliative care setting. Blinding is not feasible; treatment allocation will not be masked. The trial allows for pre-randomization.22 The burden of the proof of benefit is on surgery. While some patients allocated non-surgical management may still want surgery, and some patients allocated surgery may prefer non-surgical management, we feel participating patients will be better informed by the discussion around trial participation. Hypothesis: Patients with recurrent Glioblastoma, at the time they are considered for repeat resection, who undergo repeat resection for the GBM, will experience an increase in median overall survival from 6 to 9 months.