Efficacy and safety of anlotinib combined with the STUPP regimen in patients with newly diagnosed glioblastoma: a multicenter, single-arm, phase II trial

Objective: Glioblastomas are highly vascularized malignant tumors. We determined the efficacy and safety of the anti-angiogenic multi-kinase inhibitor, anlotinib, for a newly diagnosed glioblastoma. Methods: This multicenter, single-arm trial (NCT04119674) enrolled 33 treatment-naïve patients with histologically proven glioblastomas between March 2019 and November 2020. Patients underwent treatment with the standard STUPP regimen [fractionated focal irradiation in daily fractions of 1.8-2 Gy given 5 d/w × 6 w (total = 54-60 Gy)] or radiotherapy plus continuous daily temozolomide (TMZ) (75 mg/m2 of body surface area/d, 7 d/w from the first to the last day of radiotherapy), followed by 6 cycles of adjuvant TMZ (150–200 mg/m2 × 5 d during each 28-d cycle) plus anlotinib (8 mg/d on d 1–14 of each 3-w cycle for 2 cycles during concomitant chemoradiotherapy, 8 maximal cycles as adjuvant therapy, followed by maintenance at 8 mg/d. The primary endpoint was progression-free survival (PFS). Secondary endpoints included overall survival (OS) and adverse events (AEs). Results: Thirty-three patients received the planned treatment. The median PFS was 10.9 months (95% CI, 9.9–18.7 months) and the 12-month PFS rate was 48.5%. The median OS was 17.4 months (95% CI, 14.5–21.1 months) and the 12-month OS rate was 81.8%. The most common AEs included hypertriglyceridemia [58% (n = 19)], hypoalbuminemia [46% (n = 15)], and hypercholesterolemia [46% (n = 15)] during concurrent chemoradiotherapy and leukopenia [73% (n = 24)], hypertriglyceridemia [67% (n = 22)], and neutropenia [52% (n = 17)] during adjuvant therapy. Five patients discontinued treatment due to AEs. HEG1 (HR, 5.6; 95% CI, 1.3–23.7; P = 0.021) and RP1L1 alterations (HR, 11.1; 95% CI, 2.2–57.2; P = 0.004) were associated with a significantly shorter PFS. Conclusions: Anlotinib plus the STUPP regimen has promising anti-tumor activity against glioblastoma and manageable toxicity. HEG1 and RP1L1 alterations might be novel predictive biomarkers of the response to anlotinib.


Introduction
Glioblastoma multiforme (GBM) is the most common aggressive intracranial tumor affecting adults, accounting for 48.6% of primary malignant brain tumors 1 .Despite our current understanding of the molecular mechanisms underlying GBM [2][3][4] , using the Stupp protocol as the current standard regimen, which consists of maximal surgical debulking followed by radiotherapy with concurrent and subsequent temozolomide (TMZ) [fractionated focal irradiation in daily fractions of 2 Gy given 5 d/w × 6 w (total = 60 Gy)] or radiotherapy plus continuous daily TMZ (75 mg/m 2 of body surface area/d, 7 d/w from the first to the last day of radiotherapy), followed by 6 cycles of adjuvant TMZ (150-200 mg/m 2 × 5 d during each 28-d cycle), the prognosis of patients with GBM remains dismal, with a median progression-free survival (PFS) of 6.9 months and an overall survival (OS) of 14.6 months 5 .A recent retrospective study showed that GBM patient outcome has not improved over the last decade with the median OS < 16 months, even in the subgroup with an excellent prognosis 2 , highlighting the urgent need for novel effective and safe therapeutic regimens.
GBM is highly invasive and marked for rampant genomic instability.Robust aberrant angiogenesis renders GBM potentially amenable to anti-angiogenic therapy 6,7 .Nevertheless, bevacizumab, an anti-VEGF antibody, failed to significantly extend the OS of newly diagnosed GBM patients when added to the standard of care (SOC) in the first-line setting, and only prolonged the PFS by 3-4 months 8,9 .An important reason for bevacizumab failing to improve the OS is that bevacizumab only targets a single signaling angiogenesis pathway 10 , highlighting the importance of simultaneously targeting multiple pro-angiogenic growth factors.Although attempts have been made to treat GBM with anti-angiogenic multi-kinase inhibitors (MKIs), such as vandetanib and sorafenib, these drugs likely failed to improve prognosis due to inadequate blood-brain barrier (BBB) penetration, limited VEGFR2 and/ or EGFR inhibition, or intolerable toxicities 11,12 .
Anlotinib, a novel MKI, targets onco-angiogenesis and suppresses tumor growth by simultaneously blocking VEGFR, FGFR, PDGFRA, and c-Kit, therefore exerting a broad spectrum of tumor inhibitory effects 13 .Anlotinib has been approved as a 2 nd -line agent in China for advanced soft tissue sarcomas and advanced non-small cell lung cancer (NSCLC).Anlotinib has also been investigated for other advanced solid tumors 14 .Anlotinib has been shown to have activity in patients with brain metastases from NSCLC, delaying the time to brain progression 15 .Our preclinical study showed that radiation renders the BBB more permeable to anlotinib 16 .Together, anlotinib and radiation were shown to be synergistic in suppressing orthotopic glioma growth 16 , suggesting that anlotinib might be active against intracranial tumors.However, no prospective study has involved anlotinib as monotherapy or in combination with TMZ for newly diagnosed GBM.
In this phase II trial we investigated the efficacy and safety of anlotinib added to the current SOC for newly diagnosed GBM.Because no effective predictive biomarkers have been identified that enable biomarker-stratified anti-angiogenic therapies 17 , we further performed an exploratory analysis of genomic profiles in this cohort to identify potential predictors of treatment response using whole exome sequencing (WES).

Study population
We conducted this multicenter, single-arm, phase II trial at seven centers in China.Patients 18-70 years of age with histologically proven GBM based on the 2016 World Health Organization (WHO) classification of central nervous system (CNS) tumors were eligible for the study 18 .The patients had adequate organ function and a Karnofsky Performance Status (KPS) score ≥ 60.Patients had to have at least one measurable lesion according to the Response Assessment in Neuro-Oncology (RANO) criteria and received no prior radiotherapy, chemotherapy, immunotherapy, or biologic therapy.Other eligibility criteria are described in the submitted study protocol.
The trial was conducted per the provisions of the

Treatments
Radiotherapy was initiated 4-6 weeks postoperatively at a dose of 1.8-2.0Grays (Gy)/fraction 5 d/w × 6 w for a total dose of 54-60 Gy.TMZ (75 mg/m 2 /d) was administered for a maximum of 49 d.Beginning 4 w after completion of concurrent chemoradiotherapy, patients received adjuvant TMZ for 5 d every 28 d; the first cycle was 150 mg/m 2 /d and subsequent cycles were 200 mg/m 2 /d for a maximum of 6 cycles (Supplementary Figure S1A).The TMZ dose was adjusted upon occurrences of unacceptable toxicities.Chemoradiotherapy was delayed a maximum of 2 w.
Anlotinib (8 mg/d; Chia Tai Tianqing Pharmaceutical Group Co., Ltd., China) was administered orally on d 1-14 of each 3-week cycle × 2 cycles during concomitant therapy for a maximum of 8 cycles during adjuvant chemotherapy.One week after discontinuation of adjuvant chemotherapy, anlotinib (8 mg/d) was given for maintenance (Supplementary Figure S1B).Treatment was continued until progressive disease (PD) per Response Assessment in Neuro-Oncology (RANO).Anlotinib dose reduction was not permitted but a maximum delay of 2 w was allowed for recovery from toxicity.

Patient assessment
The baseline evaluation included full clinical and neurologic evaluations, brain magnetic resonance imaging (MRI), electrocardiography, complete blood count, blood chemical analyses, and urinalysis.Isocitrate dehydrogenase-1 (IDH) mutation and methylguanine-DNA methyltransferase gene (MGMT) promoter methylation status was recorded based on the surgical pathology reports.During concurrent chemoradiation therapy, patients received weekly clinical assessment and laboratory evaluations.Clinical assessment was also performed 1 w before and 3 and 7 d after the end of each adjuvant cycle of TMZ and 1 w before the start of each cycle of anlotinib.Patients had follow-up evaluations every 2 months during the first year and every 3 months thereafter.Radiologic assessments were performed on d 21 and 42 of concurrent chemoradiation therapy, 1 w prior to the start of adjuvant cycle 1, 3, and 5, every 3 cycles of anlotinib, and during follow-up evaluations per RANO criteria.Adverse events (AEs) were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events [NCI CTCAE (version 4.0)].

Mutation calling
To ensure high-quality mutation calls, the following filtering criteria were applied: 1) sequencing depth ≥ 20 × in tumor DNA and ≥ 10 × in germline DNA; 2) variant allele frequency ≥ 0.02 in tumor DNA and < 0.01 in germline DNA; 3) the total number of reads supporting the variant calls was ≥ 4; 4) variant frequency was < 0.01 in ESP6500, 1000 genome, and EXAC databases; and 5) logarithm of odds (LOD) score > 18 (MuTect default was 6.3).We kept the mutations that passed all filtering criteria, except an LOD score < 18, if the identical mutations were present with an LOD score ≥ 18 in other regions within the same tumors.Cancer gene mutations were defined as identical oncogene mutations previously reported, stop gains and frameshift of tumor suppressor genes, and other non-synonymous mutations with a Combined Annotation Dependent Depletion (CADD) score > 20.

Outcomes
The primary endpoint was the PFS, which was calculated from the date of study entry to progressive disease (PD) or death, whichever occurred first.Secondary endpoints included the OS, which was calculated from the date of study entry to death from all causes, treatment failures, and occurrences of treatment-related adverse events (TRAEs).Exploratory outcomes were the association between genetic variants and relapse, PFS, OS, or TRAEs.

Statistical analysis
Assuming a median PFS of 6.9 months based on historical controls, a power of 80%, and a one-sided α error of 10% 5 , a sample size of 31 patients was required to detect a 3.7-month-increase to reach a median PFS of 10.6 months 9 .Assuming a dropout rate of 5%, an accrual size of 33 patients was chosen in this trial.
Statistical analyses were pre-specified and all enrollees were evaluated for efficacy measures and safety.Recursive partitioning analysis (RPA) class was determined using age, KPS at the time of treatment, extent of resection, and mental status 23 .The differences between clinical and mutational frequencies were calculated using a two-sided χ 2 or Fisher's exact test, as indicated.The odds ratio (OR) with a 95% confidence interval (CI) was provided.The Wilcoxon rank-sum test was performed to compare the distribution differences between the two groups.The follow-up duration was calculated using the reverse Kaplan-Meier method.Survival curves were estimated with the Kaplan-Meier method and compared using the log-rank or Tarone-Ware test.Univariate and multivariate Cox regression was used to analyze the influence of genetic mutations on prognosis and clinicopathologic factors that had satisfied the proportional hazards assumption.Spearman correlation was performed to test the association between GBM gene expression in The Cancer Genome Atlas (TCGA) database.
All statistical analyses and graphics used software R (v4.0.2).A P < 0.05 was considered significant for all hypotheses tested.

Patient characteristics
Between March 2019 and November 2020, 36 patients were screened and 33 were eligible and received the allocated study interventions (Supplementary Figure S1C).The median patient age was 52 years (range, 52-69 years), 52% were males, and the median KPS score was 90.Thirty percent of the patients had undergone a gross total tumor resection and 52% had a subtotal tumor resection.Two patients (6%) had an IDH1 mutation (R132H) and 11 had [11/31 (36%)] had MGMT promoter methylation (Table 1).

Treatment compliance
All patients completed scheduled radiotherapy and concomitant chemotherapy.There were no treatment interruptions for anlotinib and no dose reduction for TMZ due to AEs.Twenty-nine patients (88%) completed adjuvant TMZ.Four patients (12%) discontinued adjuvant TMZ due to PD; 2 of the 4 patients completed 5 cycles of anlotinib and the remaining 2 patients received 6 and 7 cycles.Four patients did not receive maintenance therapy with anlotinib due to PD and 1 patient discontinued due to AEs after completion of adjuvant treatment.Twenty-four patients received maintenance therapy with anlotinib.Eighteen patients (55%) discontinued treatment because of PD and 5 (15%) discontinued treatment due to AEs.Ten patients (30%) were still receiving treatment on the data cut-off date (August 31, 2022).The median number of anlotinib cycles was 13 (range, 7-47).
In addition, the log-rank and Tarone-Ware tests showed no center effects for PFS (P = 0.21 and 0.15, respectively) and OS (P = 0.16 and 0.19, respectively).
Furthermore, alterations in the signaling pathways, including RTK, PI3K, RAS, RB/p53, and DNA damage response-associated proteins, did not predict PFS or OS (Supplementary Table S1).

Safety
Thirty-three patients received anlotinib plus SOC and were included in the safety analysis.The 3 most frequent treatment-emergent AEs (TEAEs) during concurrent chemoradiotherapy plus anlotinib were hypertriglyceridemia (58%), hypercholesterolemia (46%), and hypoalbuminemia (46%), while the 3 most common TEAEs during adjuvant chemotherapy plus anlotinib were leukopenia (73%), hypertriglyceridemia (67%), and neutropenia (52%).Two patients receiving concurrent chemoradiotherapy plus anlotinib developed grade 3 hypertriglyceridemia.One patient with grade 4 and 2 patients with grade 3 thrombocytopenia, 1 patient with grade 3 leukopenia, and 1 patient with grade 3 hypertriglyceridemia were reported during adjuvant therapy (Table 3).Six patients (18%) required hospitalization due to AEs, including 1 with grade 1 seizures and 5 with grade 2 cerebral ischemia during maintenance therapy.Two patients with cerebral ischemia resumed treatment after appropriate management and 3 discontinued treatment.Interestingly, PIK3CA mutations were significantly associated with cerebral ischemia (P = 0.032; OR = 16.4);specifically, 3 patients with grade 2 cerebral ischemia who underwent WES all harbored PIK3CA mutations (Supplementary Table S2).Meanwhile, duration of anlotinib treatment was not related to cerebral ischemia (P = 0.63).Grade 3 body weight reduction occurred in 1 patient at the end of cycle 8 of maintenance therapy and one patient developed grade 4 thrombocytopenia during adjuvant therapy and discontinued treatment.No death was reported.

Discussion
We hypothesized that a MKI simultaneously targeting multiple pro-angiogenic factors with broad tumor inhibitory activities could improve the outcome of newly diagnosed GBM patients receiving SOC.In this multicenter, single-arm, phase  II trial, newly diagnosed GBM patients attained a median PFS of 10.9 months (95% CI, 9.9-18.7 months), meeting the primary endpoint of the study, with a 12-month PFS of 48.5%.Furthermore, the median OS reached 17.4 months and the probabilities of survival at 12 months were 81.8%.The safety profile for anlotinib is acceptable and consistent with that observed in other tumors 14 .
This study was the first trial to evaluate the efficacy and safety of anlotinib added to the SOC for first-line treatment of GBM, attaining a median PFS of 10.9 months, which is higher than that of historical controls (6.9 months) 5 .The PFS in our cohort is comparable to bevacizumab added to SOC for newly diagnosed GBM in RTOG-0825 (10.7 months) and AVAglio (10.6 months).PFS improvement with bevacizumab was significant versus placebo in AVAglio but did not reach the pre-specified threshold for significance in RTOG-0825 (P < 0.004) 8,9 .It is worth noting that progression assessment using the RANO criteria in our study was more accurate The median OS was 17.4 months in this trial, 16.8 months in AVAglio, and 15.7 months in RTOG-0825.The OS in all 3 trials was longer than the Stupp trial (14.6 months) 5 .However, bevacizumab failed to improve the OS rates versus placebo in AVAglio and RTOG-0825.A retrospective analysis of the AVAglio data suggested that subgroups with IDH1 wild-type pro-neural glioblastoma derived an OS benefit from bevacizumab, suggesting that the high tumor heterogeneity of GBM leads to different treatment responses.In our study MGMT methylation status was significantly associated with PFS, which is consistent with RTOG-0825.In contrast, the OS of patients with methylated MGMT promoter was not longer than patients with unmethylated MGMT promoter in AVAglio.MGMT methylation status is a predictor of TMZ resistance and a strong prognostic biomarker of newly diagnosed GBM regardless of treatment 24 .Our study did not identify subgroups of newly diagnosed GBM patients who achieved an OS or PFS benefit by the addition of anlotinib, including MGMT status, Ki-67, and co-morbidities.Future investigations involving predictive imaging markers and biomarkers that could stratify patients for anti-angiogenic therapy added to the SOC in newly diagnosed GBM patients are warranted.
A variety of methods have been used to study the biomarkers related to treatment response or drug resistance and to develop individualized precision treatment strategies 3,4 .WES based on next-generation sequencing (NGS) has been proven remarkably powerful in genomic analysis and is appropriate for both screening previously discovered and novel significant alterations 25 .In the secondary analysis both HEG1 and RP1L1 variants were independent biomarkers of PFS, and the latter variant was significantly associated with OS.Although none of the prognostic biomarkers of anti-angiogenic therapy have been previously proven, newly diagnosed GBM with a pro-neural gene expression signature was identified as a subgroup with better OS compared to placebo 26 .HEG1 has a critical role in angiogenesis, which is related to tumor progression.An HEG1 defect may result in deficiency of vessel formation, leading to a worse survival outcome of patients with less vascularized GBMs who receive anlotinib therapy 27 .RP1L1 is a component of photoreceptor cilium, which is involved in cancer by association with hedgehog, Wnt, and PDGF signaling, and is pivotal in regulating the fate of glioma stem cells 28,29 .The effect and potential predictive values of HEG1 and RP1L1 alterations in response to anti-angiogenesis therapy of GBM should be the focus of corollary studies.
Overall, the study regimen was tolerable.There were no increased or unanticipated toxicities.Palmar-plantar erythrodysesthesia syndrome, a frequent AE of anlotinib 30 , occurred in 4 patients (12%) during adjuvant treatment, but none were grade 3 or higher.Hypothyroidism, a common AE of MKI 31 , occurred in 30% of the patients receiving adjuvant TMZ plus anlotinib, but none were grade 3 or higher.Hypertension, another common AE associated with angiogenesis inhibitors 32 , occurred in 4 patients (12%) during concurrent chemoradiotherapy and 5 (15%) during adjuvant therapy.Bleeding, a major concern with anti-angiogenic therapy, was not reported in this study.However, grade 2 cerebral ischemia occurred in 5 patients during maintenance therapy, 2 of whom resumed treatment after appropriate management.Cerebrovascular ischemia occurred in 9% of newly diagnosed GBM patients receiving bevacizumab added to concurrent chemoradiotherapy and adjuvant TMZ in a phase II trial, which may arise from radiation-induced occlusive arteriopathy potentiated by bevacizumab 33 .Lipid abnormalities are another possible contributor to cerebrovascular ischemia.In the current study lipid abnormalities were observed during both concurrent chemoradiotherapy plus anlotinib [hypercholesterolemia (46%) and hypertriglyceridemia (58%)] and adjuvant chemotherapy plus anlotinib [hypertriglyceridemia (67%)] and should be carefully monitored because of the association with arterial thromboembolic events.Interestingly, PIK3CA mutations were significantly associated with cerebrovascular ischemia in the secondary analysis but ischemia did not increase with prolongation of treatment.The overall acceptable toxicity profile of the study regimen may be partially due to the low dose (8 mg) of anlotinib, which is typically given at 10 or 12 mg in other tumors.
The current study had several limitations.This was a single arm study without a control group and the sample size was limited.Based on the encouraging results of our preclinical study and the current trial 16 , a randomized controlled study (NCT04959500) is ongoing that includes a larger sample size.In addition, the current study did not evaluate changes in neurocognitive function and quality of life among the patients in the cohort.In the current study a decline in KPS scores was only noted in patients during concurrent radiochemotherapy.No KPS decline due to AEs was observed.

Conclusions
In conclusion, anlotinib added to the current SOC for newly diagnosed GBM exhibited encouraging anti-tumor activities and had an acceptable toxicity profile.HEG1 and RP1L1 alterations could be novel predictive biomarkers in patients with newly diagnosed GBM treated with anlotinib plus SOC.Future randomized controlled trials and application of biomarkers are warranted for further clinical development of anlotinib for precise treatment of GBM in different line settings and combination regimens, such as immunotherapy.

1 Figure 1
Figure 1 Survival outcomes of the study patients.The Kaplan-Meier curves of progression-free survival (PFS) (A) and overall survival (OS) (B).Forest plot of PFS (C) and OS (D) per patient subgroup.

Figure 2
Figure 2 Continued

Figure 2
Figure 2 Distribution of genomic alterations with regard to response to treatment and survival analysis based on Kaplan-Meier curve in GBM patients in secondary analysis.(A) Mutational landscape of patients with glioblastoma multiforme (GBM) detected by whole exome sequencing (WES).(B) The frequency of gene variants between relapse and non-relapse patients.PFS in patients harboring RP1L1 (C) and HEG1 (D) alterations and wild-type.OS in patients with harboring RP1L1 (E) and HEG1 (F) alterations and wild-type.
Declaration of Helsinki and the International Conference on Harmonisation guidelines for Good Clinical Practice.The trial protocol was approved by the Cancer Hospital Institutional Review Board (IRB) of the University of Chinese Academy of Sciences [Zhejiang Cancer Hospital (IRB-2018-238)].All patients provided written informed consent before enrollment.The trial is registered with CLINICALTRIALS.GOV (NCT04119674).The study protocol adhered to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) statement and the study report adhered to the Consolidated Standards of Reporting Trials (CONSORT) statement.

Table 1
Demographic and baseline characteristics of the patients and those detected by WES