Document Type : Original Research


1 Department of Medical Physics, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran

2 Department of Radiation Oncology, Isfahan Milad Hospital, Isfahan, Iran

3 Department of Radiotherapy Oncology, Faculty of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran

4 Department of Genetics and Molecular Biology, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran

5 Applied Physiology Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran

6 Department of Biostatistics, School of Medicine, Shiraz University of Medical Sciences, Shiraz, Iran



Background: Treatment response in High-grade Glioma (HGG) patients changes based on their genetic and biological characteristics. MiRNAs, as important regulators of drug and radiation resistance, and the Apparent Diffusion Coefficients (ADC) value of tumor can be used as a prognostic predictor for glioma.
Objective: This study aimed to identify some of the pre-treatment individual patient features for predicting the treatment response in HGG patients.
Material and Methods: In this prospective study, 18 HGG patients, who were candidated for chemo-radiation treatment, participated after informed consent of the patients. The investigated features were the expression level of miR-222 and miR-205 in plasma, the ADC value of tumor, Body Mass Index (BMI), and age. Treatment response was assessed, and Least Absolute Shrinkage and Selection Operator (LASSO) regression was used to obtain a model to predict the treatment response. Mann-Whitney U test was also applied to select the variables with a significant relationship with patients’ treatment response.
Results: The LASSO coefficients for miR-205, miR-222, tumor’s mean ADC value, BMI, and age were 3.611, -1.683, 2.468, -0.184, and -0.024, respectively. Mann-Whitney U test results showed miR-205 and tumor’s mean ADC significantly related to treatment response (P-value˂0.05). 
Conclusion: The miR-205 expression level of the patient in plasma and tumor’s mean ADC value has the potential for prognostic predictors in HGG.


Maryam Heidari  (Google Scholar)

Parvaneh Shokrani  (Google Scholar)



In adults, gliomas are the most common type of brain tumors divided into four histopathological grades as follows: 1) grades Ι and ΙΙ, known as low-grade glioma, and 2) grades Ⅲ and Ⅳ, recognized as high-grade glioma (HGG) [ 1 , 2 ]. HGG Features such as local invasion, genomic instability, rapid cell proliferation, high angiogenesis, resistance to apoptosis and drug delivery, and the propensity for necrogenesis often lead to poor response to treatment [ 3 , 4 ]. In addition, a treatment protocol is not successful for all HGG patients because of the genetic and cytologic heterogeneity of HGG tumors [ 3 , 5 , 6 ]. Therefore, each patient’s treatment planning must be based on individual biological information. For example, oncologists have used molecular biomarkers to prescribe the most effective chemoradiation therapy regimens [ 3 ], e.g., the expression of microRNAs (miRNAs) associated with glioblastoma progression can be used as a prognostic predictor for glioblastoma [ 4 ]. MiRNAs are non-coding RNA molecules of 18-22 nucleotides for regulating post-transcriptional gene expression and affecting many prominent features of malignant tumors such as cell proliferation, angiogenesis, invasion, and metastasis [ 4 , 7 , 8 ]. MiRNAs are also important regulators of drug and radiation resistance [ 4 ]. Studies have shown that miR-222 is overexpressed in GBM patients. In addition, in GBM patients, expression of miR-205 decreases, resulting in increasing the grade of malignancy. In glioma, miR-205 acts as a tumor suppressor by inhibiting proliferation, reducing invasion, and increasing apoptosis [ 9 , 10 ]. Therefore, information about the expression of dedicated miRNAs before chemo-radiation therapy may help predict the treatment response.

A close relationship between miRNA expression and radiation resistance has also been reported [ 4 ]. Since miRNAs are released into the bloodstream and are quite stable [ 11 ], circulating miRNAs can be used as non-invasive biomarkers in the diagnosis, prognosis, or monitoring of cancer [ 3 , 11 , 12 ]. The first circulating miRNA studied for prognosis was mir-21, in patients with diffuse large B-cell lymphoma [ 13 ]. Since the standard treatment for glioma after surgery is radiation and chemotherapy with temozolomide, and miRNAs effects on developing chemo-radiation therapy resistance, studies have examined the potential role of miRNAs in resistance to these treatments [ 4 , 14 ]. A study showed that miR-222 is overexpressed in GBM patients and had several targets involved in gliomagnesis, such as apoptotic pathways [ 4 ]. In addition, miR-205 is downregulated in glioma patients and could be used as a prognostic biomarker in glioma patients [ 9 ].

Diffusion-weighted imaging (DWI) is a form of Magnetic Resonance (MR) imaging in which the random Brownian motion of water molecules within a voxel of tissue or Apparent Diffusion Coefficients (ADC) value is measured and also used to obtain information about tumors environment. Studies have shown that ADC values decrease with increasing degree of malignancy and cellularity of tumor [ 15 , 16 ]. Tumor ADC value can also be used independently as a quantitative predictor of progression-free survival for glioma patients [ 15 ]. Because of the different responses to treatment in HGG patients, depending on the genetic and individual characteristics, identifying some features that can predict the treatment response can aid the physicians in designing individualized treatment protocols. This study aims to identify accessible characteristics of HGG patients for applying as potential prognostic predictors. For this purpose, several features, including absolute levels of miR-205 and miR-222 in plasma, tumor ADC value, age, and Body Mass Index (BMI) were collected from HGG patients before starting the chemo-radiation therapy. This data and the treatment response data were used to design a LASSO regression model for predicting response to the chemoradiation therapy of HGG patients.

Material and Methods


This prospective study protocol was approved by the human ethics board of Isfahan University of Medical Sciences, Isfahan, Iran. From December 2018 to November 2020, patients with pathologically confirmed HGG (grade Ⅲ and Ⅳ), who were candidates for chemo-radiation treatment with MR imaging information of T1, T2, FLAIR, and DWI sequences, were included in this study, after informed consent of the patients or their companion. Patients with other types of brain tumors or brain metastases were excluded, i.e., twenty-one patients were included. Three patients died before starting radiation therapy. Finally, 18 patients: 8 females and 10 males with the median age of 52.44±13.7 years (range 27-72) and median Karnofsky performance scale (KPS) score of 7.38 participated in absolute quantification of two miRNAs (miR-222 and miR-205) levels in plasma.

Before chemo-radiation therapy, blood samples were collected in K2-EDTA, containing tubes, and centrifuged for 12 mins at 1500×g within 2 h after collection. The supernatant was removed to RNase-free tubes and miRNA isolation from fresh plasma was performed.

Mirnas Isolation and Quantitative Reverse-Transcriptase Polymerase Chain Reaction (qRT-PCR)

MiRNAs were extracted from 3 mL plasma from the patients using GeneAll® Hybrid-RTM miRNA isolation kit (Seoul, KOREA 138-859) according to the manufacturer’s protocol.

Purified miRNA samples were stored at -70 °C. MiRNA concentration and purity were then quantified using the Epoch Microplate Spectrophotometer (BioTek- China).

cDNA of miRNAs was synthesized using 10 µL of the total RNA according to the Reverse Transcription System kit (ZistRoyesh, Iran) with miR-specific stem-loop primers [ 17 ]. Real-time Polymerase Chain Reaction (PCR) was performed with the ZistRoyesh kit (Tehran, Iran) using 10 µL of 2x qPCR SYBR green master mix with reactions prepared in duplicate. ABI Applied BiosystemsTM Thermal Cycler (Thermo Scientific, USA) was used with the cycling conditions included an initial 15 min incubation at 95 °C, followed by 40 cycles of 95 °C for 30 s and 56 °C for 1 min. In addition, a dilution series of known template concentrations (101-1012 (molecule/µL)) was prepared for absolute quantification of miRNA levels in plasma, and real-time PCR was performed. Each miRNA and sample determined the average cycle threshold (Ct). The log of each known concentration in the dilution series (x-axis) was plotted against the CT value for that concentration (y-axis). These standards measured the absolute concentration of the miRNAs in the patient’s samples using their Ct values.


According to the standard treatment guidelines and patients’ conditions, surgery followed by chemo-radiation therapy and then adjuvant chemotherapy with temozolomide was prescribed. The dose of temozolomide for concurrent chemoradiation therapy was 75 (mg/m2/day, for no more than 7 weeks) and for adjuvant chemotherapy was 150-200 mg/m2/day on 5 days every 28 days for at least 6 maintenance cycles.

In Milad hospital, Isfahan, for each patient, a 3-dimensional conformal RT treatment plan was performed using a Prowess Panther DAQ treatment planning system, commissioned for the 15-MV photon beam of a Siemens Artiste linear accelerator (Siemens Medical Solutions, Muenchen, Germany).

Total doses of 60-70 Gy were delivered to each patient during 5-6 weeks over 5 days per week (dose/fraction=1.8-2 Gy). In this study protocol, two phases of radiation therapy were performed where Planning Target Volume (PTV1) and PTV2 received approximately 45 and 16 Gy, respectively. In the initial plan, PTV1 included clinical target volume (CTV1) and 0.3-0.5 cm margin while CTV1 was defined as postoperative peritumoral edema and 2 cm margin, followed by a boost field with enhanced tumor and 2.3-2.5 cm margin as PTV2 (according to RTOG 0525 and RTOG 0825) [ 18 ].

The clinical assessment of the patients, Karnofsky performance scale (KPS), was scored at three points (before chemo-radiation therapy, one and three months after receiving radiation therapy). Based on the general condition of patients, KPS was scored from 0 to 100, i.e., 100 and 0 are considered perfect performance status and death, respectively.

Three months after the chemo-radiation therapy course, treatment was evaluated by an experienced radiation oncologist and an experienced radiologist using KPS and the standard measure of changes in the volume of enhanced and non-enhanced anatomic lesions in MR images according to Response Assessment in Neuro-Oncology (RANO) criteria [ 19 ]. In RANO criteria, treatment response was divided into four types, including complete response, partial response, stable diseases, and progression. Despite a few studied patients, they were divided into two groups: the patients who responded to treatment and those who did not. Patients that indicated partial response were included in the response group and those with disease progression were included in the non-response group. Moreover, patients with stable disease were included in the response group, providing their KPS score improved after treatment. However, reducing the KPS score post-treatment led to the involvement in the non-response group.

MR Imaging

All patients underwent MRI using T1W (pre-and post-contrast), T2W, Fluid Attenuated Inversion Recovery (FLAIR), and DWI sequences in three-time points, before treatment planning, one month, and three months after radiation therapy. Imaging was done using 1.5 T MRI units (Siemens MAGNETOM Aera or Siemens MAGNETOM Avanto or PHILIPS, Ingenia).

A free open-source toolkit, ITK-SNAP, version 3.6.0 ( was used to extract the mean Apparent Diffusion Coefficients (ADC) values of the tumor region in the ADC map. After importing the ADC map of each patient into the toolkit, the tumor was delineated and the mean of ADC values was obtained. Figure 1 shows an example of tumor contouring in the ITK-SNAP toolkit based on T1 images.

Figure 1. Tumor delineation in Apparent Diffusion Coefficients (ADC) map. In order to calculate the tumor mean ADC value in ADC map of each patient, tumor was delineated using ITK-SNAP toolkit (left images), based on T1 images (right images) and the mean ADC value was obtained.

Statistical Analysis

Different statistical methods could be used depending on the sample size and the type of independent and dependent variables to investigate the predictive effect of several patient characteristics. Since the sample size was very small compared to the number of independent variables, LASSO method was used, i.e. the number of variables is controlled by defining the penalty function on the sum of the absolute values of the model coefficients. Then coding was conducted using RStudio software (1.4.1103) with the Glmnet package [ 20 ] and LASSO model coefficients were obtained for the related variables; deviance of the model was also obtained. Finally, the Mann–Whitney U test using SPSS software (IBM Statistics 26) was applied to find the variables with a significant relationship with the response to treatment.


Treatment was assessed, and the patients were divided into response and non-response groups. Ten patients were included in the response group, and eight did not respond to the treatment. Further, a complete response was not observed among the patients. The absolute expression levels of miR-222 and miR-205 in the plasma of patients were obtained based on the number of molecules per microliter of plasma, using real-time PCR and the standard curves, and the results are shown in Figure 2. The optimal lambda value and the coefficients of the LASSO regression are shown in Table 1. Mann-Whitney U non-parametric test was also applied to evaluate the significance of the relationship between dependent and independent variables as seen in Table 2.

Figure 2. Standard curves of A) miR-205, B) miR-222. The log of each known concentration in the dilution series (x-axis) was plotted against the Ct value for that concentration (y-axis). From these standards, the absolute concentration of the miRNAs in the patient’s samples was measured using their Ct values.

Variables Lasso coefficients
miR-205 3.611
ADC 2.468
miR-222 -1.683
BMI -0.184
age -0.024
Intercept 4.656
Deviance (model) 11.988
lambda 0.039
ADC: Apparent Diffusion Coefficients, BMI: Body Mass Index
Table 1.Results of Least Absolute Shrinkage and Selection Operator (LASSO) regression modeling for variables associated with treatment response
Variables Response N Mean St. Error Mean P-value
miR-205 Non-response 8 0.030 0.026 0.009
response 10 0.200 0.097
ADC Non-response 8 0.387 0.130 0.043
response 10 0.733 0.084
miR-222 Non-response 8 0.291 0.136 0.068
response 10 0.076 0.074
BMI Non-response 8 27.350 1.812 0.083
response 10 23.920 0.968
Age Non-response 8 58.000 4.101 0.203
response 10 48.000 4.705
ADC: Apparent Diffusion Coefficients, BMI: Body Mass Index
Table 2.Results of Mann-Whitney non-parametric test


In this study, the relationship between the treatment response and several characteristics of patients before the treatment was investigated using LASSO regression. The characteristics studied for each patient include the expression level of miR-205 and miR-222 in plasma, the mean ADC value of tumor, age, and BMI. The LASSO coefficient for miR-205 with a positive value was larger than those of other variables, i.e., the expression level of this miRNA is directly related to the response to treatment. Vascular Endothelial Growth Factor (VEGF) is a direct target of miR-205, resulting in miR-205 acting as a glioma tumor suppressor [ 10 ]. Studies have shown that downregulation of miR-205 in tissue specimens, glioma cell line, and serum is associated with poor prognosis in patients with glioma [ 9 , 21 - 23 ]. On the other hand, the LASSO coefficient for miR-222 was negative, indicating that the survival of glioma patients decreases with increasing the expression of this miRNA. Some studies have acknowledged the role of miR-222 in increasing resistance to chemotherapy and radiation therapy [ 4 , 24 , 25 ]. This miRNA is known as an MGMT-regulating miRNA in glioblastoma that one of the miR-222 targets is the P53 Upregulated Modulator of Apoptosis (PUMA) gene, regulating cell death. Under normal circumstances, PUMA binds to BCL-2 and CVL-x, leading to apoptosis; however, increased miR-222 expression and downregulation of PUMA improve cell survival [ 26 , 27 ]. According to the results, the second feature associated with treatment outcome was the tumor ADC value that the LASSO coefficient of this feature was also positive, i.e., patients with a higher level of tumor ADC value are more likely to respond to the treatment. The amount of ADC decreases with increasing malignancy and the grade of disease [ 15 ] consistent with the results obtained in this study. The LASSO coefficients for BMI and age were -0.184 and -0.024, respectively, indicating that both are inversely related to the response to treatment, although age had a very weak relationship. LASSO regression modeling does not provide a P-value for the model and only provides model deviance. Here, the model deviance was 11.988 and the Null deviance was 24.73. Mann-Whitney test results were used to confirm the LASSO regression coefficients. For this purpose, P-value obtained for each variable in the Mann-Whitney test were compared with LASSO regression coefficients of the variables in terms of their correlation degree with treatment response. As shown in Tables 1 and 2, these two results are completely consistent. For example, the P-value for the miR-205 expression variable was the lowest and its LASSO coefficient was the largest. A limitation of this study is its small sample size, while regression models with larger sample sizes have more stable coefficients and a much higher validity. So, more study with a larger number of data is needed before clinical application.


The prognostic predictors for chemo-radiation treatment response of HGG patients were studied using the LASSO model and Mann-Whitney test. Based on the results, patients’ miR-205 expression level in plasma and tumor’s mean ADC value significantly correlated with treatment response.


The authors would like to acknowledge Dr. Reza Moghareabed, Dr. Narges Motamedi, and Fariba Davanian for their contributions to this study.

Authors’ Contribution

M. Heidari, under the supervision of P. Shokrani, made a contribution in collecting patient information, performing experimental works, measuring data, and preparing the manuscript. Admission of patients to the study, their treatment, and their treatment evaluation were done by AR. Amouheidari and S. Hemati. AR. Amouheidari also collaborated in the tumor contouring of the patients. H. Khanahmad and I. Rahimmanesh provided guidance and participation in the genetic part of the study. P. Jafari had cooperation in statistical data analysis and coding for Lasso modeling. The final version of the manuscript has been approved by all authors.

Ethical Approval

This study protocol was approved by the human ethics board of Isfahan University of Medical Sciences, Isfahan, Iran (Ethical code: 1397.198).

Informed consent

Patients were included in the study after informed consent of them or their companion.


This study was financially supported by Isfahan University of Medical Sciences, grants No. 397552.

Conflict of Interest



  1. Schwartzbaum JA, Fisher JL, Aldape KD, Wrensch M. Epidemiology and molecular pathology of glioma. Nat Clin Pract Neurol. 2006; 2(9):494-503. DOI | PubMed
  2. Kleihues P, Burger PC, Scheithauer BW. The new WHO classification of brain tumours. Brain Pathol. 1993; 3(3):255-68. DOI | PubMed
  3. Fathi Kazerooni A, Bakas S, Saligheh Rad H, Davatzikos C. Imaging signatures of glioblastoma molecular characteristics: A radiogenomics review. J Magn Reson Imaging. 2020; 52(1):54-69. Publisher Full Text | DOI | PubMed
  4. Shea A, Harish V, Afzal Z, Chijioke J, Kedir H, Dusmatova S, et al. MicroRNAs in glioblastoma multiforme pathogenesis and therapeutics. Cancer Med. 2016; 5(8):1917-46. Publisher Full Text | DOI | PubMed
  5. Huse JT, Phillips HS, Brennan CW. Molecular subclassification of diffuse gliomas: seeing order in the chaos. Glia. 2011; 59(8):1190-9. DOI | PubMed
  6. Riemenschneider MJ, Jeuken JW, Wesseling P, Reifenberger G. Molecular diagnostics of gliomas: state of the art. Acta Neuropathol. 2010; 120(5):567-84. Publisher Full Text | DOI | PubMed
  7. Jonas S, Izaurralde E. Towards a molecular understanding of microRNA-mediate­d gene silencing. Nat Rev Genet. 2015; 16(7):421-33. DOI | PubMed
  8. Santangelo A, Tamanini A, Cabrini G, Dechecchi MC. Circulating microRNAs as emerging non-invasive biomarkers for gliomas. Ann Transl Med. 2017; 5(13):277. Publisher Full Text | DOI | PubMed
  9. Yue X, Lan F, Hu M, Pan Q, Wang Q, Wang J. Downregulation of serum microRNA-205 as a potential diagnostic and prognostic biomarker for human glioma. J Neurosurg. 2016; 124(1):122-8. DOI | PubMed
  10. Yue X, Wang P, Xu J, Zhu Y, Sun G, Pang Q, Tao R. MicroRNA-205 functions as a tumor suppressor in human glioblastoma cells by targeting VEGF-A. Oncol Rep. 2012; 27(4):1200-6. Publisher Full Text | DOI | PubMed
  11. Xiao Y, Zhang L, Song Z, Guo C, Zhu J, Li Z, Zhu S. Potential Diagnostic and Prognostic Value of Plasma Circulating MicroRNA-182 in Human Glioma. Med Sci Monit. 2016; 22:855-62. Publisher Full Text | DOI | PubMed
  12. Zhao H, Shen J, Hodges TR, Song R, Fuller GN, Heimberger AB. Serum microRNA profiling in patients with glioblastoma: a survival analysis. Mol Cancer. 2017; 16(1):59. Publisher Full Text | DOI | PubMed
  13. Lawrie CH, Gal S, Dunlop HM, Pushkaran B, Liggins AP, Pulford K, et al. Detection of elevated levels of tumour-associated microRNAs in serum of patients with diffuse large B-cell lymphoma. Br J Haematol. 2008; 141(5):672-5. DOI | PubMed
  14. Ujifuku K, Mitsutake N, Takakura S, Matsuse M, Saenko V, Suzuki K, et al. miR-195, miR-455-3p and miR-10a(*) are implicated in acquired temozolomide resistance in glioblastoma multiforme cells. Cancer Lett. 2010; 296(2):241-8. DOI | PubMed
  15. Durand-Muñoz C, Flores-Alvarez E, Moreno-Jimenez S, Roldan-Valadez E. Pre-operative apparent diffusion coefficient values and tumour region volumes as prognostic biomarkers in glioblastoma: correlation and progression-free survival analyses. Insights Imaging. 2019; 10(1):36. Publisher Full Text | DOI | PubMed
  16. Sugahara T, Korogi Y, Kochi M, Ikushima I, Shigematu Y, Hirai T, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging. 1999; 9(1):53-60. DOI | PubMed
  17. Mohammadi-Yeganeh S, Paryan M, Arefian E, Vasei M, Ghanbarian H, Mahdian R, et al. MicroRNA-340 inhibits the migration, invasion, and metastasis of breast cancer cells by targeting Wnt pathway. Tumour Biol. 2016; 37(7):8993-9000. DOI | PubMed
  18. Zhao F, Li M, Kong L, Zhang G, Yu J. Delineation of radiation therapy target volumes for patients with postoperative glioblastoma: a review. Onco Targets Ther. 2016; 9:3197-204. Publisher Full Text | DOI | PubMed
  19. Wen PY, Macdonald DR, Reardon DA, Cloughesy TF, Sorensen AG, Galanis E, et al. Updated response assessment criteria for high-grade gliomas: response assessment in neuro-oncology working group. J Clin Oncol. 2010; 28(11):1963-72. DOI | PubMed
  20. Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010; 33(1):1-22. Publisher Full Text | PubMed
  21. Yang C, Wang C, Chen X, Chen S, Zhang Y, Zhi F, et al. Identification of seven serum microRNAs from a genome-wide serum microRNA expression profile as potential noninvasive biomarkers for malignant astrocytomas. Int J Cancer. 2013; 132(1):116-27. DOI | PubMed
  22. Chen W, Kong KK, Xu XK, Chen C, Li H, Wang FY, et al. Downregulation of miR 205 is associated with glioblastoma cell migration, invasion, and the epithelial-mesenchymal transition, by targeting ZEB1 via the Akt/mTOR signaling pathway. Int J Oncol. 2018; 52(2):485-95. DOI | PubMed
  23. Li FF, Xing C, Wu LL, Xue F. MiR-205 enhances cisplatin sensitivity of glioma cells by targeting E2F1. Eur Rev Med Pharmacol Sci. 2018; 22(2):299-306. DOI | PubMed
  24. Li W, Guo F, Wang P, Hong S, Zhang C. miR-221/222 confers radioresistance in glioblastoma cells through activating Akt independent of PTEN status. Curr Mol Med. 2014; 14(1):185-95. DOI | PubMed
  25. Zhang C, Zhang J, Hao J, Shi Z, Wang Y, Han L, et al. High level of miR-221/222 confers increased cell invasion and poor prognosis in glioma. J Transl Med. 2012; 10:119. Publisher Full Text | DOI | PubMed
  26. Zhang CZ, Zhang JX, Zhang AL, Shi ZD, Han L, Jia ZF, et al. MiR-221 and miR-222 target PUMA to induce cell survival in glioblastoma. Mol Cancer. 2010; 9:229. Publisher Full Text | DOI | PubMed
  27. Chen L, Zhang J, Han L, Zhang A, Zhang C, Zheng Y, et al. Downregulation of miR-221/222 sensitizes glioma cells to temozolomide by regulating apoptosis independently of p53 status. Oncol Rep. 2012; 27(3):854-60. DOI | PubMed