Could upfront temozolomide chemotherapy postpone the intervene of radiotherapy in young patients with high-risk low-grade gliomas?

TMT time Part B groups, p value], At the same time, we compared cognitive results of patients receiving chemotherapy alone with that of 11 patients receiving postoperative radiotherapy with or without adjuvant chemotherapy for patients’ choice or other pathologies. The results showed an obvious improvement in MMSE score, AVLT/HVLT-R total and TMT time Part B tests


Background
Gliomas can be classi ed into 3 categories according to treatment strategy and prognosis. [1] The rst category is tumors with early recurrence and poor prognosis after operation such as glioblastoma, and early radiotherapy and chemotherapy are needed after operation. The second category is tumors with good prognosis, 10-year survival rate, such as pilocytic astrocytoma and pleomorphic xanthoastrocytoma. Only regular follow-up is needed after operation. The last category is advanced recurrent tumors such as low-grade gliomas. National Comprehensive Cancer Network (NCCN) points out that for LGGs patients in high-risk group patients (below 40 years old and subtotal resection), standard strategies are maximum safe resection with radiotherapy and adjuvant chemotherapy, but didn't address the question of whether patients need adjuvant therapy immediately after diagnosis because of RTOG 9802 [2].
EORTC 22845 has shown that compared with delayed radiotherapy, early radiotherapy prolongs PFS, but does not prolong OS signi cantly [3]. Worse still, Radiotherapy brings signi cant treatment-related side effects as cognitive dysfunction in the meantime of its controlling tumor recurrence according to a 12-year follow-up for patients below 40 years old [4]. A 10-year follow-up suggested that greater dose of radiotherapy was associated with a greater cognitive decline [5].
Temozolomide (TMZ) is one of the rst line regimens for LGGs chemotherapy [6], but whether temozolomide has neurotoxicity still remains unknown. The results of the EORTC 22033-26033 has shown that compared with temozolomide chemotherapy, radiotherapy didn't prolong PFS in LGGs patients signi cantly, and OS outcomes still remains unknown [7]. EORTC 22033-26033 also has not gured out the intervention time of radiotherapy and the sequence of intervention for radiotherapy and chemotherapy. For residual tumors that can't be totally removed, if upfront chemotherapy can decrease the burden of the tumor, it will contribute to improving curative effect of radiotherapy theoretically.
Because young patients need to avoid cognitive impairment, the intervention of postoperative radiotherapy immediately is questionable. We wonder that chemotherapy can be used to postpone the intervention of radiotherapy to avoid cognitive impairment at early stage. And IDH mutation is considered to be a prognostic indicator identifying a subgroup of gliomas with an improved survival, but whether a predictive factor related to a better temozolomide response is questionable [8,9].
Our study focused on LGGs patients under 40 years old with non-totally resected tumors in the eloquent areas, and tested temozolomide response. We focused on three questions: 1) which groups/biomarkers favor LGGs patients to higher temozolomide chemotherapy sensitivity and longer survival; 2) how does neurocognitive function change with temozolomide chemotherapy; 3) does chemotherapy better protects cognitive function than radiotherapy and can we use temozolomide to postpone the intervene of radiation to prolong survival and delay cognitive impairments.
We therefore initiated an interventional study (NCT02209428) to validate whether young patients with non-totally resected LGGs in eloquent areas could be controlled by upfront temozolomide, and whether radiotherapy could be postponed until progression to avoid potential cognitive impairments. We anticipate our results can enhance and further facilitate the decision-making process in creating an individualized therapeutic approach for these young LGGs patients to ensure the highquality survival and the healthy social life.

Methods
Study design and participants NCT02209428 was initiated in 2014, as a prospective, one arm, open-label study in a single tertiary specialized center (Glioma Surgery Division, Neurological Surgery Department of Huashan Hospital, Fudan University) in Shanghai, China. The selection criteria were: 1) 18 years < age ≤ 40 years, both genders. 2) No neurologic cognitive de cits (MMSE ≥ 27), no psychiatric abnormalities before surgery, pre-operative KPS ≥ 80. 3) Tumor located in eloquent areas or deeply located nuclei, rendering radiological complete resection inapplicable. [10] 4) Post-operative histological pathology con rms LGGs (astrocytomas, oligodendrogliomas, or oligoastrocytomas, 2007 WHO classi cation Grade II). 5) No contraindications to temozolomide chemotherapy. 6) Informed consent to temozolomide chemotherapy. They were informed and provided proof of their written agreement to participate to the present study. The study was approved by the Huashan Hospital's institutional review board.

Chemotherapy
Chemotherapy was started within 2 weeks to 3 months after surgery and continued for 6 cycles. Metronomic temozolomide regimen [11] was administrated from day 1 to 21 at a dose of 75 mg/m2/day, repeated every 28 days. (Additional le 1) Toxicity and adverse effects were graded according to the National Cancer Institute-Common Toxicity Criteria (NCI-CTC v4.0).

Follow up
All patients were evaluated by an independent neurosurgeon and a neuropsychologist, before and after operation, every 2 months during the course of chemotherapy, every 3 months during the half year after chemotherapy and later on every 6 months until progression.
The baseline evaluation and follow-up evaluation included 2 parts: 1) MRI for tumor volume.. Real-time calculation of dynamic tumor volume (Osirix) on T2/FLAIR sequence was performed and objective response was assessed according to RANO standard [12]. 2) Neuropsychological evaluation. Neuropsychological evaluation was performed using comprehensive neurocognitive battery in Huashan hospital, including Learning Test-Revised (HVLT-R), Trail Making Test time A and B (TMT time A and B), Mini-Mental State Examination (MMSE), Multilingual aphasia examination controlled oral word association (COWAT).

Pathological and molecular diagnosis
Pathological diagnosis of tumor tissue from surgery con rms LGGs. In August 2017, we centrally reviewed all enrolled tumor pathology information and renewal them as 2016 WHO classi cation of tumors of the central nervous system according to the test of IDH1/2 [15,16], 1p/19q co-deletion, TP53 and ATRX [17].

Endpoints
The primary endpoint is objective response rate (ORR), which is de ned as response and no response. According to RANO standard [12], response includes complete response (CR), partial response (PR) and minor response (MR), while no response includes stable disease (SD) and disease progression (PD). Secondary endpoints are as follows: 1) Intensity of response (IOR) is de ned as the ratio of maximum volume reduction to residual volume after operation. 2) Duration of response (DOR) is used instead of progression-free survival and de ned as the time between the date of operation and the date of treatment escape or relapse. [18] 3) Malignant progression-free survival (MPFS), cognitive function results and the safety of chemotherapy.

Statistical analysis
We enrolled 11 patients as the pilot study, and the response rate of IDH mutant group and IDH wildtype group is 61% and 17% respectively. IDH mutant accounts for 80% of low-grade gliomas, and we estimated 54 patients were su cient to provide 80% power and a significance level of 5% to detect difference in ORR between IDH mutant group and IDH wild-type group.
Dynamic assessment of sample size was su cient to re ect the trial e cacy, and thus adjusted the sample size. In our study, no patients were lost for a specially-assigned person reminding them of regular follow-ups in our hospital. Because patients who met the admission criteria and were willing to receive temozolomide chemotherapy alone were rare, especially for IDH wild-type LGGs patients, and in order to better explore the e ciency of chemotherapy, we nally enrolled 54 patients and had analysis of 65 patients including 11 patients in pilot study (Fig 1). Comparisons for ORR were performed using the Fisher's exact test, which was de ned as response and no response. Unpaired nonparametric test (Wilcoxon-Mann-Whitney U test) was performed for comparison of IOR between groups, and logistic regression for correlation between tumors residual volume and ORR. Kaplan-Meier method are used for DOR. Repeated mixed model was used for comparison among cognitive test scores, which could handle missing values. And unpaired t test with Welch's correction was used for comparing our cognitive test data with that in historical literature. For all analyses, a P value of less than 0.05 was regarded as signi cant.
Statistical analysis was done with SPSS version 19.0 (SPSS Inc. at IBM Company) and charted with GraphPad Prism version This study is registered at ClinicalTrials.gov, number NCT02209428.

Results
Up to December 30th 2019, 65 patients receiving adjuvant temozolomide only were recruited with a median follow-up of 39.6 months (Fig 1), and 50 patients had long-term evaluation with extensive serial neuropsychological batteries. The patients' basic characteristics were shown on Table 1. Malignant progression rate for IDH-mutant was 6/34(17.65%), and that for diffuse astrocytoma, IDH-wildtype was 44.44%.
(Additional le 2) 50 of 65 patients who only received temozolomide chemotherapy had long-term evaluation with extensive serial neuropsychological batteries. (Fig 3) We analyzed the cognitive data at 5 time points: before chemotherapy, after chemotherapy, 1 year after surgery, 2 years after surgery and 3 years after surgery, and results are as follows. (Table 2) Mean HVLT-R score for each group is 14.5, 17.6, 20.5, 22.4, 23.9, and scores of 2 and 3 years after surgery were statistically higher than those of 1 year after surgery and before. Although the mean HVLT-R score of 3 years after surgery is higher than that of 2 years after surgery, there was no statistically difference. There was no statistical difference in HVLT-R percentage of retention among every group, which means there is no learning effect affects validity. COWAT showed during follow-up, the number of naming had statistically increased as compared to that before chemotherapy. Mean MMSE score for each group is 26.8, 28.2, 28.9, 28.7, 29.1, and scores of 1, 2 and 3 years after surgery were statistically higher than those of before and after chemotherapy. Although the mean MMSE score of 1, 2 and 3 years after operation increased gradually, there was no statistically difference among these 3 groups. And we compared our patients with patients undergoing early radiotherapy (NCCTG protocol 86-72-51) [ Table 3).
The temozolomide chemotherapy was well tolerated by all recruited patients, and all adverse effects could be relieved by expectant treatment. Adverse effects were shown on Table 1. Leucogen, peanut coat or recombinant granulocyte colony stimulating factor were used for hematologic adverse reactions, and Bicyclol or other hepatic protectants were used for elevated live enzymes. And no discontinuation of chemotherapy occurred among all patients except one patient for Grade 3 thrombocytopenia.

Discussion
For LGGs, IDH mutation and 1p/19q codeletion are tightly associated with characteristics of gliomas at better biological behavior and clinical prognosis [20]. In our study, IDH mutant group had a better ORR than IDH wild-type group. It suggested that IDH mutation was a predictive factor for better temozolomide response, and 64.29% of LGG patients with IDH mutation can achieve partial or minor response after temozolomide chemotherapy.
IDH mutant group had a longer DOR than IDH wild-type group, and 1p/19q codeletion group also had a longer DOR than 1p/19q retain group. This suggested that both IDH mutation and 1p/19q codeletion were meaningful prognostic markers, which helped us to decide whether the patient needed to receive chemotherapy alone at early stage to delay radiotherapy. The median DOR for the diffuse astrocytoma, IDH-mutant subgroup was 44.5 months, and that for oligodendroglioma, IDHmutant and 1p/19q-codeleted was 52.4 months. This suggests that for the majority of IDH mutant patients, temozolomide chemotherapy alone can control tumor progression for 4 years or longer. We recommend that providers should engage close follow-up for patients on upfront temozolomide chemotherapy after surgery, and cautiously decide whether the radiotherapy should intervene based on their response. As for IDH wild-type group, IOR and DOR showed poor response and prognosis.
This group of tumors biologically and clinically resembles the behavior of glioblastomas, and these patients might need to receive concurrent chemo and radiotherapy at early stage.
Compared with EORTC 22033-26033 [7], our study focused on young patients below 40 years old, who needed to protect their cognitive function. The age of temozolomide alone group in EORTC 22033-26033 ranged from 37-53, and 64% patients were above 40. EORTC 22033-26033 didn't gure out whether patients need to receive chemotherapy immediately after operation, and some of them received chemotherapy 30 months after diagnosis. As for extension of resection, the EORTC 22033-26033 contained biopsy, partial resection and total resection, while all patients in our study receive subtotal resection.
Temozolomide treatment of LGGs might induce driver mutations in the RB and AKT-mTOR pathways, which may drive malignant progression to secondary GBM [21]. Some chemotherapy-induced metabolic stress in IDH1 mutant gliomas might lead to poor chemotherapy response [22]. In our study, malignant progression occurred in 17.65% of diffuse astrocytoma, IDHmutant and in 9.09% of oligodendroglioma, IDH-mutant and 1p/19q-codeleted. Temozolomide treatment was safe for most of gliomas with IDH mutation, especially for oligodendroglioma, IDH-mutant and 1p/19q-codeleted. Our next step is to explore the mechanism in IDH mutated tumors with poor temozolomide response and prognosis.
EORTC 22033-26033 [23] showed that the effect of temozolomide chemotherapy or radiotherapy on health-related quality of life (HRQOL) or MMSE scores did not differ in patients with LGG, and it wasn't consistent with a prior report of 12-year followup [4]. In our study, TMT time A and B and COWAT showed that executive function, visual spatial perception ability, language pro ciency and plasticity of patients improved at the end of chemotherapy. MMSE showed cognitive function was still statistically improved even 1 year after surgery, and HVLT-R showed memory and attentional function was still statistically improved even 2 years after surgery. This suggests that temozolomide chemotherapy might has no obvious neurotoxicity, which leads to cognitive impairment. Cognitive function can be improved compared to that before chemotherapy. Meanwhile, the comparison between our results to the prospective study with the similar group of patients undergoing radiotherapy (NCCTG protocol 86-72-51) [19] suggested that chemotherapy might protect neurocognitive function better than radiotherapy. This answers our previous questions ---it is worth considering upfront adjuvant temozolomide chemotherapy in biologically favorable groups early after tumor resection to control tumor progression and delay radiotherapy intervention to prevent early cognition decline in young high-functioning patients. This is an ongoing study with continuous long-term follow-up for cognitive function and survival data. Limitations of the study lie in the inter-study comparison of neurocognitive evaluation results between chemo and radiation therapy. The demographics, diagnostic and treatment protocols might not be completely comparable between the studies. This could affect the validity of the comparison as well as the generalization of the results. Future prospective phase III studies should involve randomization between the two adjuvant therapies in the same demographic group of patients.

Conclusion
Temozolomide is the rst line for LGGs chemotherapy, and generally well tolerated. We infer that 1) IDH mutation is a predictive factor for better temozolomide response, and a prognostic factor for longer survival. Upfront temozolomide therapy might be more appropriate for IDH mutant LGGs to postpone radiation; 2) Compared with early radiotherapy, patients undergoing adjuvant chemotherapy would have a better cognitive rehabilitation at an early stage. Therefore, IDH mutant young patients could choose upfront chemotherapy to postpone the intervene of radiation and delay potential cognitive impairments.

Abbreviations
LGG Low-grade gliomas ORR objective response rate IOR intensity of response DOR duration of response MPFS malignant progression free survival Declarations Ethics approval and consent to participate. This study was approved by the Huashan Hospital's institutional review board.
Informed consent obtained from all individual participants included in the study was written.
Consent for publication. All subjects participating in the image acquisition signed the consent form.
Availability of data and materials. The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Competing interests. None declared.

Funding
This work was supported by Chinese Society of Clinical Oncology (CSNO-2015-MSD06), and Shiwen Yuan's effort was supported by National Institute of Mental Health (R25MH101076). Sponsors had no in uence on study design, in the collection, analysis, and interpretation of data, in the writing of the report, and in the decision to submit the paper for publication. The corresponding Author had full access to all the data in the study and had full responsibility for decision to submit for publication.    Our study RT. 11 patients received postoperative radiotherapy with or without adjuvant chemotherapy for patients' choice or other pathologies, but we still followed their cognitive function. Figure 1 Trail pro le.  SupplementaryFigures.zip