Repeat radiation with bevacizumab and minocycline in bevacizumab-refractory high grade gliomas: a prospective phase 1 trial

An open-label, single arm, Phase 1, dose-escalation and dose-expansion clinical trial was conducted with patients enrolled at a single institution. The trial included adults with radiologically proven recurrent, intracranial glioma who had been previously treated with radiation, temozolomide and bevacizumab. Progression must have been documented by Response Assessment in Neuro-Oncology (RANO) criteria [27], and patients must have been at least 6 months from prior radiation. Subjects were required to have adequate hematologic, hepatic, and renal function and a Karnofsky Performance Status (KPS) of at least 50. The protocol did not limit the size of the tumor as long as the radiation oncologist determined reirradiation was feasible. Exclusion criteria included contraindications to bevacizumab, including uncontrolled hypertension, intracranial hemorrhage, recent surgery, or non-healing wounds, and contraindications to minocycline, such as lupus or connective tissue disease. Patients were considered evaluable if they had a dose limiting toxicity (DLT) during the DLT window or if they completed at least 28 days of minocycline and two infusions of bevacizumab without a DLT. The DLT window included the radiation and 28 days afterward. During dose escalation, unevaluable patients were replaced.

Bevacizumab was given at the standard, FDA-approved dose of 10 mg/kg IV every two weeks. Radiation was individualized based on the best judgment of the radiation oncologist. Most subjects received 3750 centigray (cGy) in 15 fractions or 4000 cGy in 20 fractions. Details of radiation planning are given in Table 2. The gross tumor volume (GTV) was defined mostly as the T1 enhancing tumor, but given the effects on MRI imaging of anti-angiogenic agents the T2/FLAIR abnormality could be included in the GTV at the judgment of the treating radiation oncologist. The median clinical target volume (CTV) margin was 0.5 cm (Range 0–2 cm). The size of the CTV and the decision of whether to include T2/FLAIR in the GTV were based on the judgment of the treating radiation oncologist taking into consideration MRI changes over time, prior radiation, and anatomic location. The median planning treatment volume (PTV) margin was 0.3 cm (Range 0–0.7 cm). The median PTV size was 192.3 cm³ (Range 0.3–487 cm³). The median RT dose was 3750 cGy (Range 2000–5400 cGy) in 15 fractions (Range 5–30). Intensity modulated radiation therapy (IMRT) was used in all patients but one, who received stereotactic radiosurgery (SRS). Minocycline was taken orally twice a day starting on the day prior to radiation and continuing until progression or intolerance. Three dose levels of minocycline were tested: 100 mg twice a day, 200 mg twice a day, and 400 mg twice a day.

The primary endpoint was the rate of adverse events during and up to 28 days after radiation. All patients who received a dose of minocycline were included in the safety analysis. DLTs were defined as grade 3 or intolerable grade 2 toxicities (other than anemia, lymphopenia, high cholesterol, or weight gain) related to the minocycline occurring during or within 28 days of the end of radiation.

The secondary endpoints were progression-free survival (PFS) at three (PFS3) and 6 (PFS6) months from the beginning of study treatment, the overall response rate, and changes over time in symptom burden and cognitive function. PFS was defined from the date of start of minocycline until the day of documented disease progression or death. PFS was only assessed in patients on the maximal tolerated dose (MTD) of minocycline. Patients who declined follow-up or did not complete radiation were censored at the time of study withdrawal. Patient symptom burden and symptom interference were measured using the MDASI-BT [28]. Cognitive function was measured using COGSTATE brief battery (detection test (DET), identification test (IDN), one card learning test (OCLT), and Groton Maze Learning Test (GMLT)) [29] Symptom burden and cognitive function were measured at baseline, week 4, week 12, and week 26 before patients were seen by the neuro-oncologist or received any MRI results.

The trial was approved by the University of Utah Institutional Review Board (IRB). It was registered on clinicaltrials.gov, NCT01580969, prior to beginning enrollment.

Twenty-three patients were screened between July, 2012 and November, 2017, of which twenty-two were eligible and enrolled on the trial. Clinical characteristics are summarized in Table 1. Six people enrolled at the 100 mg twice a day dose level, three at the 200 mg twice a day dose level and thirteen at the 400 mg twice a day dose level. All patients have completed follow up and have died. Three people enrolled at the lowest dose did not complete radiation due to patient preference despite not having any DLT, so they were replaced (Table 2).

Most patients had GBM or multiply recurrent lower grade gliomas without approved systemic treatment options. All patients treated at the 400 mg twice a day dose had either grade III or IV gliomas at initial diagnosis. Of the 22 patients, 13 had unknown IDH status, although one of these was a 1p/19q codeleted anaplastic oligodendroglioma that was presumably IDH mutated and two were GBMs with EGFRvIII mutation that were presumably wild-type. Of the rest, 6 were IDH wild-type, including 4 GBMs, 1 grade IV glioneuronal tumor, and 1 grade II astrocytoma., and three were IDH mutated, including one GBM and two anaplastic astrocytomas. The tumor that was called anaplastic oligodendroglioma histologically was 1p/19q codeleted, but the tumor that was called grade II oligodendroglioma was not 1p/19q codeleted so by current criteria would have been called an astrocytoma.

The maximum tolerated dose of minocycline in this population was 400 mg twice a day. Of the patients who started at 400 mg twice a day, nine (70%) were able to complete radiation without stopping or reducing the minocycline. Adverse events were consistent with the known side effect profiles of minocycline, bevacizumab, and radiation. Grade 3 or higher adverse events and adverse events that occurred in more than one person are summarized in Table 3. Grade 3 or higher adverse events occurred in 24% of subjects and all resolved with supportive care or holding the minocycline. No patient had symptomatic radiation necrosis or radiation necrosis requiring treatment. There were no DLTs in the dose escalation and one DLT in the dose expansion (grade 3 nausea, dehydration, and confusion possibly related to the minocycline).

PFS and OS were assessed in all patients who started treatment at the 400 mg twice a day dose. The PFS3 in this group was 64.6% (95% confidence interval (CI) 42.8–99.8%), and the PFS6 was 27.7% (95% CI 10.6–72.3%), which compares favorably with PFS3 and PFS6 seen in other trials of patients who progressed on bevacizumab. The median PFS was 3.8 months (95% CI 2.5 months–infinity), and the median OS was 6.4 months (95% CI 4.2 months–12.1 months). Radiographic responses were rare, with one person with a gemistocytic astrocytoma treated at the 100 mg twice a day dose level having a complete response and one person with a GBM treated at the 200 mg twice a day level having a partial response. For all 22 people who enrolled on the trial, the median PFS was 3.8 months (95% CI 2.5–8.1 months). The PFS3 was 65.3% (95% CI 47.5–89.8%), and the PFS6 was 35.2%, (95% CI 19.4–63.8%).

Compliance with the MDASI-BT was high with 100% compliance at baseline (22 out of 22 eligible patients), week 12 (nine out of nine eligible patients), and week 26 (four out of four eligible patients) and 93% compliance at week 4 (13 out of 14 eligible patients). At baseline, the average MDASI-BT score was in the moderate range (4–6) for fatigue and memory problems. At week 4, the average MDAS-BT score was in the moderate range for fatigue and memory, and the average interference score was in the moderate range for general activity, walking and enjoyment of life, while the interference with work score had gone into the severe range (> 6). By week 12, however, the interference scores were all below moderate, except for interference with work, which was still in the moderate range. Average scores for the two subscales of the MDASI-BT and for fatigue, interference with general activity, interference with work, interference with walking, and interference with employment are shown in Fig. 1.

Compliance with cognitive testing was 73% at baseline, 69% at week 4, 55% at week 12, and 75% at week 26. Missing test data was almost exclusively due to inability to complete the testing due to slow cognition or difficulty with attention. Thus, the cognitive function of the patients without data on cognitive testing is likely worse than those with testing. Summary scores are shown in Table 4. Of those that completed testing, no significant cognitive declines were seen from baseline, except for the IDN, which declined from an average of 2.79 s to 2.86 s (p = 0.02) from baseline to 4 weeks. The IDN test measures processing speed. In particular, the DET, which is a measurement of attention, OCLT, which is a measurement of working memory, and GMLT, which is a measurement of visual learning and memory, did not worsen for people who did not progress.