The standard of care for glioblastoma (GBM) treatment involves maximal resection followed by concomitant radiotherapy and temozolomide. Progression-free survival (PFS) with this treatment is only 6.9 months and relapse is inevitable. At relapse, there is no consensus regarding the optimal therapeutic strategy. The rationale behind the fact that limited chemotherapy agents are available in the treatment of malignant gliomas is related to the blood-brain barrier (BBB), which impedes drug entry to the brain. Intraarterial (IA) chemotherapy allows to circumvent this. Using IA delivery of carboplatin, can produce responses in 70% of patients for a median PFS of 5 months. Median survival from study entry was 11 months, whereas the overall survival (OS) 23 months. How can the OS and PFS be improved? By combining chemotherapeutic agents with different mechanisms of action.

Study design: In this phase II trial, treatment will be offered at relapse. Surgery will be performed for cytoreduction if it is warranted, followed with a combination IA carboplatin + IA Cealyx (liposomal doxorubicin) or IA carboplatin + IA etoposide phosphate. Toxicity will be assessed according to the NCIC common toxicity criteria. Treatment will consist in either IA carboplatin (400 mg/m^2) + IA Cealyx (30 mg/m^2) or IA carboplatin (400 mg/m^2) + IA etoposide phosphate (400 mg/m^2) every 4-6 weeks (1 cycle). Up to twelve cycles will be offered.

Outcome measurements: Tumor response will be evaluated using the RANO criteria by magnetic resonance imaging monthly. Primary outcome will PFS and tumor response. Secondary outcome will include median OS, toxicity, quality of life (QOL), neurocognition (NC).

Putting together these data will allow to correlate clinical and radiological response to QOL and NC.

BACKGROUND - Glioblastoma (GBM) is the most common and aggressive primary brain tumour in adults. The standard first-line treatment for these tumours features maximized surgery followed by radiation with concomitant and adjuvant temozolomide. The overall survival (OS) and progression-free survival (PFS) observed in the clinic with this paradigm are only 14.6 and 6, 9 months respectively. One of the challenges in the treatment of this neoplasm stems from the severe tumour heterogeneity which translates into unpredictable treatment response. As a result, newly diagnosed tumours inevitably relapse after the standard first-line treatment, which is called the Stupp protocol which combines radiation therapy and oral temozolomide. When recurrence occurs, if the patient's functional status is adequate, this will mandate other therapeutic strategies. Interestingly, results obtained in most studies in this setting have been so marginal that there is literally no recognized optimal second and third line of treatment. Admittedly, the access to active therapies is greatly limited by the presence of the blood-brain barrier (BBB), which severely reduces the chemotherapy entry to the CNS.

When one realizes the extensiveness of the vascular network supplying the brain, it appears obvious that a global delivery strategy via this vascular network as a delivery corridor is credible and legitimate. The importance of this vascular system has already been detailed by Bradbury; the author claims that the entire network covers an area of 12 m2/g of cerebral parenchyma. To understand the extensiveness of the cerebral vascularization in a more prosaic way, let us just consider that the brain receives about 20% of the total systemic circulation although it weighs less than 3% of the total body weight.

The access to a patient's cerebral vascular network is technically easy and actually repeatedly performed in the clinic on a regular basis. Via a simple puncture to access the femoral artery, a catheter can be introduced and navigated intraarterially to reach one of the four major cerebral arteries. Once in the target vessel, a therapeutic agent can be administered via the catheter, that is later withdrawn at the conclusion of the procedure. The CIAC allows the construct of a regional chemotherapeutic distribution paradigm within the area irrigated by the targeted vessel.

An increase in the local plasma peak concentration of the drug yields a significantly improved AUC (concentration of drug according to the time) through the first pass effect. This consequently translates in an increased local exposure of the target tissue to the therapeutic agent. Interestingly, as our lab as shown, it is also accompanied by a decreased systemic drug distribution, hence reducing systemic toxicity and potential side effects. Consequently, the therapeutic concentration at the targeted tumour cells is increased by a 3.5 to 5-fold factor. This procedure is performed in the angiographic suite under local anesthesia and typically lasts around 45 minutes.

The IA procedure is a very safe procedure. Indeed, this procedure has been used at our institution for over 15 years using various chemotherapeutic agents and thus have precise statistics on the risks and complications. Indeed, 722 different patients have been treated adding up to 3600 procedures and have compiled the following events. During the MRI that followed the IA infusion, 66 complications were identified (1.84%), 27 of which were associated to symptoms (0.75%). During the infusion, 39 episodes of seizures occurred (1.08%), all of which were successfully controlled with anti-seizure medication. Moreover, a significant reduction in white, red or platelet blood cell count occurred in 52 patients during the treatment phase (7.2%). This study will investigate the efficacy of using combined chemotherapeutic agents described above. Our team currently uses intraarterial (IA) infusion to alleviate the effects of the BBB. This delivery strategy was shown to be well tolerated, triggered very few discomforts and side effects, and significantly improved survival. So much so, that it is nowadays considered a standard of care for relapsing tumours in our institution.

Like cisplatin, carboplatin is a molecule made of a platinum atom surrounded in a plane by two ammonia groups and two other ligands in the cis position. Unlike the chloride atoms found in cisplatin, the ligands in carboplatin are esther functional groups that form a ring structure. As such, carboplatin is more stable, causes less vomiting and is less neurotoxic, less ototoxic and less nephrotoxic. Carboplatin's exact mechanism of action remains unclear. However, it is well known that carboplatin is activated inside the cell into reactive platinum species. These reactive complexes react with DNA bases to create inter- and intrastrand crosslinks which prevent cell division by hindering DNA synthesis.

At our institution, carboplatin is the primary chemotherapeutic agent for IA infusions and yields positive tumour responses in 70% of patients for a median PFS of 5 months. Although interesting, there is obviously room for improvement in the care of these patients. Hence the current proposal.

For patients in which carboplatin fails, other chemotherapeutics are chosen arbitrarily from a list of agents available for IA infusion. As such, our team has successfully treated relapsing GBM patients with IA delivery of methotrexate, melphalan, etoposide phosphate or Caelyx (liposomal doxorubicin). At the heart of the present study, carboplatin, which will be combined with either one of two agents found to be ideally suited in this setting: Caelyx (liposomal doxorubicin) or etoposide phosphate.

Doxorubicin is an anthracycline, an antineoplastic antibiotic developed from Streptomyces peucetius subsp. Cassius. It is a very potent antitumour agent and is considered one of the most active antineoplastic drugs developed to date. Its effect is produced via different mechanisms: DNA binding and cross-linking, interference with DNA strand separation, inhibition of RNA polymerase, inhibition of topoisomerase II, formation of free radicals and membrane peroxidation have all been suggested.

In vitro studies in malignant glial cell lines have demonstrated that doxorubicin induces a halt in cell growth within 24 hours, and results in apoptosis within 48 hours. It has been identified as one of the most potent chemotherapeutic drugs against malignant glioma cell lines in vitro. However, in vivo, the use of doxorubicin is limited by its inability to cross the BBB.

Doxorubicin is rapidly distributed in the body tissues, and binds to plasma protein and cell membranes. The clinical application of this agent is unfortunately limited by its dose-related side effects such as cardiotoxicity and myleotoxicity.

Caelyx is a chlorhydrate of doxorubicin encapsulated within a pegylated liposome. The liposomal formulation of doxorubicin (Caelyx) exhibits an altered pharmacokinetic profile favouring the use of this drug formulation in brain tumour treatment. It has a longer terminal half-life than free doxorubicin, and reaches greater concentration in the tumour. Because of a decreased uptake by the reticuloendothelial system, the drug remains in circulation much longer. This seems to be especially true in glioblastoma, where it tends to accumulate in significant concentration due to the increase in neovascularisation. This has been shown in experimental settings, as well as in the clinic. Interestingly, because of its altered pharmacokinetic properties, it also presents a reduced toxicity profile. The liposomal formulation of doxorubicin causes less myelosuppression, nausea, vomiting and alopecia than standard doxorubicin. The cardiotoxicity is also reduced.

However, even with the greater accumulation of the drug in the tumour cells, its rate of BBB penetration when administered via IV infusion remains a limiting factor. Indeed, it is too low to yield a significant concentration accumulation within the tumour site to produce a therapeutic benefit.

Etoposide phosphate (Etopophos; Bristol-Myers Squibb Company, Princeton, NJ) is a water-soluble prodrug of etoposide that is rapidly and completely converted to the parent compound after intravenous dosing. The pharmacokinetic profile of either etoposide or etoposide phosphate is identical. Toxicity and clinical activity are also the same. Since etoposide phosphate is water soluble, solutions of up to 20 mg/mL can be prepared. However, in high doses, it can only be given as a 5-minute bolus, in small volumes and as a continuous infusion. Furthermore, it is not formulated with polyethylene glycol, polysorbate 80 (Tween; ICI Americas, Wilmington, DE), and ethanol, and does not cause acidosis when given at high doses. The easier-to-use etoposide phosphate represents an improved formulation of etoposide.

Classically, etoposide must be diluted prior to use with sodium chloride (0.9% w/v) or glucose (5% w/v) solutions to concentration of 0.2 mg/mL (i.e., 1 ml of concentrate in 100 ml of vehicle) up to 0.4 mg/mL (i.e., 2 ml of concentrate in 100 ml of vehicle). Evidently, this cannot be something considered in a setting of IA administration, as the volume administered would be excessive. Hence, the use of etoposide phosphate, for which 100-fold increased concentration can be prepared in a volume accessible for an IA administration: 200 cc.

STUDY DESIGN - This clinical trial will be an open label randomized phase II study in which intraarterial administration of carboplatin (400 mg/m2) combined with Caelyx (30 mg/m2) will be compared with intraarterial administration of carboplatin (400 mg/m2) combined with etoposide phosphate (400 mg/m2). Patients that have failed the standard first line of treatment (Stupp protocol) and that are diagnosed with recurrent GBM will be randomly distributed to one of the two second-line treatment paradigms using the block randomization method. Each recruited patient will undergo maximal resection before beginning treatments. Treatment cycles will be administered on a monthly basis until a progression is identified on the magnetic resonance imaging (MRI) scan or until a total of 12 cycles have been completed. The cohort will count 120 patients that will be divided into two groups of 60 patients receiving one of the two chemotherapeutic combinations. As to which of the two combinations will be best remains to be determined. For that reason, data from our latest published clinical trial and patients treated with intraarterial carboplatin at our institution will be used as benchmarks for baseline comparisons (OS of 11 months and PFS of 5 months from study entry).

AIM - By using carboplatin in combination with Caelyx or etoposide phosphate in the setting of an IA infusion, our intention is to optimally deliver carboplatin-based chemotherapy combinations to the brain beyond the BBB, and more specifically to the tumour cells.

HYPOTHESES - In patients treated with either combination, our prediction is that this will lead to an improved tumour response and control rate, with minimal impact on the quality of life. Our preliminary clinical data seems to support this hypothesis.

Inclusion Criteria:

1. Histological diagnosis of glioblastoma multiforme.

2. Radiological progression on an MRI scan, according to the RANO criteria, in the context of a known glioblastoma multiforme, already treated with the Stupp protocol of combined radiotherapy-Temozolomide. This implies a measurable disease on MRI.

3. Prior radiotherapy and temozolomide, as per the Stupp protocol, no sooner than 4 weeks, is permitted.

4. Eighteen or more years of age.

5. Performance status: Karnofsky ranging from 60 to 100%.

6. Haematopoietic parameters at recruitment:

   • Platelet counts > 100,000/mm3.
   • Hemoglobin > 8 g/dL.
   • Absolute neutrophil count > 1,500/mm3.

7. No impaired bone marrow function.

8. Hepatic parameters at recruitment:

   • Bilirubin ≤ 2 times normal value.
   • AST and ALT ≤ 2 times upper limit of normal (ULN).
   • Alkaline phosphatase ≤ 2 times ULN (unless attributed to the tumour).
   • No impaired hepatic function.

9. Renal parameters at recruitment:

   • No impaired renal function.
   • Creatinine no greater than 1.5 fold of the normal value.
   • Creatinine clearance > 30 ml/min.

10. Normal ECG.

11. Written informed consent obtained.

    • Patients should be either sterile or else use a contraceptive strategy (for at least 2 months prior to study accruals).

Exclusion Criteria:

1. Presence of a severe psychiatric or medical condition that would interfere with treatment administration or study recruitment.
2. Presence of an active autoimmune disease.
3. No prior cardiac disease within the past 5 years OR LVEF of at least 50% at baseline ultrasound.
4. Occurrence of another malignancy within the past 5 years except curatively treated basal cell or squamous cell skin cancer or in situ cervical carcinoma.
5. Pregnancy (as confirmed by a positive b-HCG) or actively nursing.
6. Presence of an uncontrolled systemic infection.