Phase I/IIa trial of fractionated radiotherapy, temozolomide, and autologous formalin-fixed tumor vaccine for newly diagnosed glioblastoma

Clinical article

Full access

Object

Temozolomide (TMZ) may enhance antitumor immunity in patients with glioblastoma multiforme (GBM). In this paper the authors report on a prospective Phase I/IIa clinical trial of fractionated radiotherapy (FRT) concomitant with TMZ therapy, followed by treatment with autologous formalin-fixed tumor vaccine (AFTV) and TMZ maintenance in patients with newly diagnosed GBM.

Methods

Twenty-four patients (age 16–75 years, Karnofsky Performance Scale score ≥ 60% before initiation of FRT) with newly diagnosed GBM received a total dose of 60 Gy of FRT with daily concurrent TMZ. After a 4-week interval, the patients received 3 AFTV injections and the first course of TMZ maintenance chemotherapy for 5 days, followed by multiple courses of TMZ for 5 days in each 28-day cycle.

Results

This treatment regimen was well tolerated by all patients. The percentage of patients with progression-free survival (PFS) ≥ 24 months was 33%. The median PFS, median overall survival (OS), and the actuarial 2- and 3-year survival rates of the 24 patients were 8.2 months, 22.2 months, 47%, and 38%, respectively. The median PFS in patients with a delayed-type hypersensitivity (DTH) response after the third AFTV injection (DTH-2) of 10 mm or larger surpassed the median length of follow-up for progression-free patients (29.5 months), which was significantly greater than the median PFS in patients with a smaller DTH-2 response.

Conclusions

The treatment regimen was well tolerated and resulted in favorable PFS and OS for newly diagnosed GBM patients. Clinical trial registration no.: UMIN000001426 (UMIN clinical trials registry, Japan).

Abbreviations used in this paper:AFTV = autologous formalin-fixed tumor vaccine; CREIL Center = Critical Path Research and Education Integrated Leading; CTCAE = Common Terminology Criteria for Adverse Events; DTH = delayed-type hypersensitivity; FRT = fractionated radiotherapy; GBM = glioblastoma multiforme; IDH1 = isocitrate dehydrogenase-1; JBTRC = Japan Brain Tumor Reference Center; KPS = Karnofsky Performance Scale; MGMT = O6-methylguanine-DNA methyltransferase; MHC = major histocompatibility complex; MRC = Medical Research Council; OS = overall survival; PFS = progression-free survival; RPA = recursive partitioning analysis; TMZ = temozolomide.

Glioblastoma multiforme (GBM) is the most common primary brain tumor, and it has a dismal prognosis despite aggressive treatment using surgery, radiation therapy, and an alkylating agent, temozolomide (TMZ).3,23 Recently, a growing interest in therapeutic modalities based on tumor-specific immune reactions for treatment of patients with GBM has emerged.6,21 We have developed an autologous formalin-fixed tumor vaccine (AFTV) for in vivo induction of killer lymphocytes,5,16 suppression of recurrence of hepatocellular carcinoma after primary resection,12 and treatment of recurrent GBM in a pilot clinical trial.6 Our previous prospective clinical trial for preliminary evaluation of the efficacy of AFTV concomitant with fractionated radiotherapy (FRT), but not with concomitant TMZ, in patients with newly diagnosed GBM (study identification no.: C000000002) demonstrated that this treatment resulted in a median overall survival (OS) duration of 19.8 months and an actuarial 2-year survival rate of 40%. However, at the time of disease progression, TMZ (as approved by the Japanese government on September 15, 2006) at a dose of 150–200 mg/m2 daily for 5 days every 28 days (maintenance TMZ) was administered to 20 out of 22 patients.15 Treatment with AFTV was not accompanied by severe toxicity.12,15

Recent preclinical studies show that vaccination with dendritic cells followed by chemotherapy can significantly increase survival in patients with malignant glioma.13 In animal studies, TMZ shows the potential to enhance antitumor immunity.10,14,19 In our pilot study, AFTV combined with TMZ was used to treat 1 patient with primary GBM and 2 patients with secondary GBM.17 All patients demonstrated pathological improvement associated with the therapy, and the adverse effects related to AFTV plus TMZ were very minor.

These results prompted us to perform the present prospective clinical trial to look for safety and positive efficacy of AFTV in the survival of patients with newly diagnosed GBM treated with standard chemoradiotherapy (that is, primary resection followed by FRT with concomitant TMZ and periodic maintenance TMZ23).

Methods

Study Design

A prospective Phase I/IIa clinical trial of AFTV and maintenance TMZ after FRT concomitant with TMZ for the management of newly diagnosed GBM was conducted by the Association of Cancer Vaccine Therapy in 3 participating hospitals: Tsukuba University Hospital (Ibaraki, Japan), Tokyo Women's Medical University Hospital (Tokyo, Japan), and Oita University Hospital (Oita, Japan). The basic objectives were focused on preliminary evaluation of the therapeutic efficacy and safety of the treatment. The primary end point was OS, and the secondary end point was progression-free survival (PFS). For reference, these outcomes were compared with those in the previous study using AFTV (C000000002).15 The study design and treatment protocol were approved by the ethics committees of all 3 institutions and registered in the University Hospital Medical Information Network (UMIN) clinical trials registry (http://www.umin.ac.jp/ctr/), and the trial's registration number is UMIN000001426. Eligibility and exclusion criteria for patient enrollment are as follows.

The eligibility criteria were 1) age 16–75 years, 2) newly diagnosed GBM with histopathological confirmation of diagnosis of the disease with typical neurological symptoms, 3) maximum possible resection of the tumor (radiologically confirmed maximal removal leaving residual neoplasm within the vital, functionally essential brain areas), 4) availability of at least 1.5 g of neoplastic tissue for AFTV preparation, 5) possibility of in-house AFTV preparation and administration, 6) possibility of complete course of postoperative FRT with a cumulative dose of 60 Gy, 7) Karnofsky Performance Scale (KPS) score ≥ 60% before initiation of FRT, 8) possibility of regular follow-up evaluation, and 9) lymphocyte count ≥ 1000 per mm3 before initiation of FRT.

The exclusion criteria were 1) treatment with glucocorticoids or antitumor chemotherapy, 2) intracranial hypertension before initiation of FRT, 3) suppressed hematological function according to the Common Terminology Criteria for Adverse Events (CTCAE) v3.025 or absolute white blood cell count ≤ 2000/mm3, 4) decompensated function of internal organs, 5) malignant tumor other than GBM, 6) planned or existing pregnancy, 7) enrollment in another clinical trial within the 6 months preceding the present study, and 8) ineligibility as judged by the principal investigator of the participating institution.

All criteria except the lymphocyte count were the same as those in the previous clinical trial using AFTV.24 Written informed consent for participation in the study was obtained in each individual case. The 3-year study period started on November 1, 2008, and enrollment of 25 patients was planned, as in the previous study, but only 24 patients were actually enrolled in the present study.

After in-house confirmation of the histopathological diagnosis of GBM following resection of a newly diagnosed parenchymal brain tumor, the eligible patients who agreed to participate in the study were scheduled for FRT and concomitant TMZ, followed by treatment with AFTV and maintenance TMZ according to the standard protocol (Fig. 1A).

Fig. 1.
Fig. 1.

A: Scheme of AFTV clinical studies. The upper panel shows the present study, a prospective clinical trial (UMIN000001426) of FRT concomitant with TMZ, followed by AFTV and TMZ maintenance in patients with newly diagnosed GBM. The lower panel shows, for comparison, the previous prospective clinical trial (C000000002) of AFTV concomitant with FRT in patients with newly diagnosed GBM, in which, at the time of disease progression, maintenance TMZ was administered to 20 of 22 patients.15 B: PFS curve for the 24 patients in the present study. Comparison of PFS in subgroups of patients stratified by recursive partitioning analysis (RPA) Class (Class III, n = 8; Class IV, n = 10; Class V, n = 6) (p = 0.0012, log-rank test); in IDH1R132H-positive (n = 5) and negative (n = 18) cases (p = 0.14, log-rank test); and in DTH-2 positive (n = 9) and negative (n = 14) cases (p = 0.0071, log-rank test). C: OS curve for the 24 patients in the present study. Comparison of OS in subgroups of patients stratified by RPA class (p = 0.0059, log-rank test), IDH1R132H status (p = 0.023, log-rank test), and DTH-2 response cases (p = 0.061, log-rank test). The x-axis indicates months; the y-axis, survival ratio. mOS = median OS; mPFS = median PFS.

The histopathological diagnosis was independently confirmed in the Japan Brain Tumor Reference Center (JBTRC, representative Y.N.) at Gunma University (Maebashi, Japan) according to the current WHO criteria, using paraffin-embedded tissue sections stained with hematoxylin and eosin in each enrolled case. Additional immunohistochemical analysis included evaluation of positive cells using monoclonal antibodies for MIB-1 (Dako), p53 (Dako), major histocompatibility complex (MHC) Class I (AB-46, Hokudo Co.), O6-methylguanine-DNA methyltransferase (MGMT) (clone MT3.1, MAB16200, Merck Millipore), and isocitrate dehydrogenase-1 (IDH1) R132H (D299–3, MBL).

The corresponding staining indices were calculated as the average number of positive cells in the best-stained tumor areas (up to 5) with a total number of cells not less than 1000. The MIB-1, p53, and MGMT indices were expressed as percentages. For category analysis, cases with 10% or more positive cells were rated as positive, and cases with fewer than 10% positive cells were rated as negative for both MGMT and p53.24 The negative MGMT expression was correlated with positive MGMT promoter methylation. For IDH1R132H positivity, cases with 50% or more positive cells were rated as positive, and cases with fewer than 50% positive cells were rated as negative. MHC Class I expression was graded as 0 (absence of staining), + (up to 25% of cells stained), ++ (25%–50% of cells stained), or +++ (more than 50% of cells stained).1,6,15

The baseline clinical investigations at the time of en rollment in the trial included physical examination with evaluation of KPS scores, determination of the Medical Research Council (MRC) neurological functional grade, blood and urine tests, electrocardiogram, chest radiograph, and brain MRI obtained initially within 3 days after surgery and additionally just before the first FRT session. The data were managed by Tsukuba Critical Path Research and Education Integrated Leading (CREIL) Center at the University of Tsukuba and confirmed by the data and safety monitoring committee of this study.

Treatment Protocol

All enrolled patients received the standard Stupp regimen23 with a modified TMZ maintenance cycle. In brief, FRT was started within 2–3 weeks after resection of the neoplasm and included focal irradiation of the tumor cavity or residual lesion including 2 cm of perifocal margin with 2 Gy per fraction up to a total dose of 60 Gy, with daily concurrent TMZ at 75 mg/m2 throughout the course of radiotherapy. Four weeks after the end of FRT, the patients received 150 mg/m2/day of TMZ maintenance chemotherapy for 5 days, followed by 100–200 mg/m2/day of TMZ for 5 days in each 28-day cycle. AFTV was prepared using autologous formalin-fixed GBM tissue according to an established standard operating procedure.6,15 The AFTV treatment consisted of 3 courses of vaccination performed at 1-week intervals. The first course corresponded to the 1st day of TMZ maintenance chemotherapy. AFTV injection was started during the first cycle of TMZ maintenance in 22 patients and during the second cycle of TMZ maintenance in 2 patients with severe lymphopenia. Each course consisted of 5 local intradermal injections of 0.2 ml AFTV per site in the upper arm. Two delayed-type hypersensitivity (DTH) tests were performed 48 hours before the first vaccination (DTH-1) and 2 weeks after the third vaccination (DTH-2). For these tests, fixed autologous tissue fragments (10% [v/v] suspended in 0.1 ml saline without immune adjuvant) were injected intradermally into the forearm, and the response was evaluated by measuring the diameter of the local erythema and induration.

Follow-Up

Follow-up investigations were performed 2 weeks (14 ± 2 days) after completion of FRT and every 2 months thereafter. Follow-up examinations included physical examination with evaluation of KPS score, determination of MRC neurological functional grade, a blood test, and brain MRI. Adverse effects of treatment were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) v3.0.25 Axial, coronal, and sagittal views of gadolinium-enhanced T1-weighted MRI sequences (slice thickness ≤ 4 mm) and at least axial images of plain T1- and T2-weighted sequences were acquired every 2 months. The MRI data were evaluated by neuroradiologists in each hospital who were independent of the study, verified by the neuroradiological conference, including members who were blind to this study. All investigated data including evaluation results of the MRI studies were managed by the CREIL Center and confirmed by the data and safety monitoring committee. In this study, disease progression was defined as a 25% or greater increase in the volume of the gadolinium-enhanced lesion or appearance of new brain lesions. Stable disease/incomplete response was defined as a 24% increase to a 99% decrease without appearance of new brain lesions, and complete response was defined as disappearance of all of the enhanced lesion. Pseudoprogression was determined by continued follow-up MRI. If the lesion continued to enlarge, the initial progression was called true disease progression; if the lesion stabilized or shrank, the initial progression was confirmed as pseudoprogression and categorized into stable disease/incomplete response or complete response.

TMZ maintenance chemotherapy for 5 days in each 28-day cycle was only continued until the occurrence of severe adverse effects or tumor progression. At the time of tumor progression, salvage or second-line chemotherapy was administered at the investigators' discretion; most patients received additional chemotherapy with TMZ only or TMZ plus interferon beta and/or bevacizumab.

End Points and Statistical Analysis

The primary end point of the present clinical trial was OS, defined as the time interval from the date of surgery to death from any cause. Secondary end points were PFS and safety of the treatment. The following pre-specified factors were analyzed for their association with OS and PFS: age, sex, tumor size before surgery, preoperative KPS score, recursive partitioning analysis (RPA) class, extent of tumor resection, MIB-1 staining index, p53 staining index, MHC Class I expression grade, and size of DTH-2 response using data from the 24 patients enrolled in the present study for all analyses and the 22 patients enrolled in the previous study for some analyses. Univariate analysis was performed using a log-rank test after construction of Kaplan-Meier survival curves. Continuous variables were dichotomized on the basis of their median values. Factors that showed statistical significance were included in a Cox proportional hazard model for multivariate analysis. For data analysis, the Fisher exact test, Student t-test, or Mann-Whitney U-test was used as appropriate. Differences were considered statistically significant if the 2-tailed p value was < 0.05 in all statistical analyses.

Results

Characteristics of the Enrolled Patients

From November 1, 2008, to December 31, 2011, 26 patients with newly diagnosed GBM were enrolled into this prospective clinical trial (11 in Tsukuba University Hospital, 14 in Tokyo Women's Medical University Hospital, and 1 in Oita University Hospital). Two of the 26 patients were subsequently excluded. The patient in Case 3, after initially agreeing to participate in the study, refused a scheduled AFTV injection after FRT concomitant with TMZ and dropped out of our follow-up program. Another patient (Case 19) was excluded from the final analysis because the investigators noticed that a criterion regarding the KPS score did not meet the eligibility criteria after written informed consent was obtained.

Patient characteristics of the remaining 24 patients are presented in Table 1. These 24 patients included 17 men and 7 women aged 26 to 66 years (mean 48 years). The median preoperative KPS score was 80%; the preoperative KPS score was 90% or 100% in 9 patients, 70% or 80% in 9, 60% in 4, and 50% in 2. All patients had KPS scores of at least 60% by the time of initiation of FRT. Regarding RPA classification,4 8 cases were Class III, 10 cases Class IV, and 6 were Class V. The tumor lesions were mainly located in the frontal lobe in 10 cases, the temporal lobe in 9 cases, the parietal lobe in 3 cases, and the occipital lobe in 2 cases. In 18 cases the lesions were located in a single lobe, and in 6 they were in multiple lobes. In 14 cases, the lesions were on the right side, in 8 they were on the left, and in 2 they were bilateral. The maximum tumor diameter before surgery varied from 14 to 105 mm (mean 54 mm). In 16 cases (67%), resection of the neoplasm was considered total (resection of 98% or more of the contrast-enhancing lesion); and in 8 (33%), resection was partial: 90% in 2 patients, 90%–94% in 3, and 95%–97% in 3.

TABLE 1:

Characteristics of patients in the present study and the previous study*

CharacteristicPresent Study (n = 24)Previous Study (n = 22)p Value
sex>0.99§
 M1715
 F77
age in yrs0.43
 mean4851
 range26–6618–70
preop KPS score (%)0.51**
 median8090
 range50–10030–100
RPA class0.73**
 III87
 IV108
 V67
tumor location>0.99§
 frontal109
 other1413
tumor size, max diameter in mm0.68
 mean5452
 range14–10515–103
extent of tumor removal0.75§
 total1616
 partial86
MIB-1 index (%)0.77**
 median28.729.1
 range8.7–48.07.7–66.8
P53 staining index (%)0.54**
 median3.58.5
 range0–65.50–85.0
MHC Class I staining index (%)0.071**
 median22
 range1–30–3
MGMT (%)
 median15ND
 range0–48ND
IDH1R132H
 positive5ND
 negative18ND
 n.p.1ND
DTH-20.016§
 ≥10 mm917
 <10 mm145
 n.p.10

ND = no data; n.p. = test not performed.

Study identification no.: UMIN000001426.

Study identification no.: C000000002

Fisher exact test.

Student t test.

Mann-Whitney U-test.

In all cases, the histopathological diagnoses independently confirmed in the JBTRC were GBM. The MIB-1 staining index varied from 8.7% to 48.0% (median 28.7%). The p53 staining index varied from 0% to 65.5% (median 3.5%). The grades of MHC Class I expression were evaluated for 22 cases and were + in 4 cases, ++ in 10 cases, and +++ in 8 cases. Positive IDH1R132H staining was seen in 5 of the 23 cases in which it was evaluated, and negative staining was seen in 18 cases. We found no evident differences between the present study and a previous vaccine study (C000000002, n = 22),8 except for a lower positive response (≥10 mm local erythema) ratio in the DTH-2 test (39% vs 77%, p = 0.0156, Fisher direct method). Therefore, we adopted the data from the previous study as a historical control for further comparative analyses.

Patient Survival Benefit

In the present study, the overall length of follow-up varied from 7.3 to 48.7 months (median 19.6 months), and the length of follow-up in the surviving patients varied from 14.5 to 48.7 months (median 29.3 months). At the time of data analysis, 10 patients (42%) were alive and 14 (58%) had died. No patient showed a positive response to the DTH-1 test, whereas response to the DTH-2 test varied in size from 0 to 20 mm (median 5 mm, mean 7 mm).

The duration of PFS of the 24 patients varied from 3.4 months to 46.4 months (median 8.2 months) (Table 2 and Fig. 1B). The length of follow-up in the progression-free patients varied from 18.9 to 48.7 months (median 29.5 months). In particular, we note that many patients had a long PFS in the present study. The percentage of patients with a PFS of 24 months or more was 33%; PFS showed a statistically significant association with patient age, KPS score, RPA class, extent of tumor removal, IDH1 positivity, and DTH-2 response in the 24 patients in the present study (Table 2). The duration of PFS in patients with a DTH-2 response of 10 mm or larger was significantly longer than in patients with a smaller DTH-2 response. The median PFS values in cases with RPA Class III, IV, and V were over 29.5 months (PFS curve not reaching the 50% survival point), 8.2 months, and 4.9 months, respectively (Fig. 1B). When analysis was performed with data from the 18 patients with preoperative KPS scores of 70% or greater, the median PFS was 16.4 months.

TABLE 2:

Progression-free survival, overall survival, and results of subgroup analyses using the log-rank test*

VariablePFSOS
YesNop ValueYesNop Value
patient sex (M)14.4 (17)5.2 (7)0.11n.r.14.90.091
patient age (<50 yrs)35.0 (12)6.0 (12)0.031n.r.14.90.066
KPS score (≥80%)33.6 (14)5.4 (10)0.0056n.r.15.70.014
RPA class (III + IV)16.4 (18)4.9 (6)0.0034n.r.13.50.0043
tumor location (frontal)16.4 (10)6.1 (14)0.28n.r.19.40.34
tumor size (<50 mm)10.0 (10)7.7 (14)0.4022.220.30.55
tumor removal (total)16.4 (16)4.9 (8)0.0079n.r.14.30.042
MIB-1 index (<30%)9.0 (13)5.7 (11)0.2835.020.30.53
P53 index (<10%)10.9 (14)6.4 (9)0.9235.016.50.52
MHC Class I index (≥2)8.2 (18)n.r. (4)0.1322.2n.r.0.28
MGMT index (<10%)5.7 (5)13.6 (18)0.3712.035.00.14
IDH1R132H status (positive)n.r. (5)6.1 (18)0.023n.r.20.30.14
DTH-2 (≥10 mm)n.r. (9)5.7 (14)0.0071n.r.18.40.061

Unless otherwise stated, the values in this table represent median duration of OS or PFS (as indicated) in months, with the number of patients in parentheses. The median PFS for the whole group (24 patients) was 8.2 months; the median OS was 22.2 months. n.r. = not reached (OS or PFS curve did not reach the 50% survival point).

Test was not performed in 1 patient.

Test was not performed in 2 patients.

As shown in Table 2 and Fig. 1C, the median OS of the 24 patients in the present study was 22.2 months (95% CI 2.7–41.7 months). The actuarial 2- and 3-year survival rates were 47% and 38%, respectively. The duration of OS showed a statistically significant association with preoperative KPS score, RPA class (III/IV vs V), and extent of tumor removal with the log-rank test. The median OS values in cases with RPA Class III, IV, and V were over 29.3 (OS curve not reaching the 50% survival point), 20.3, and 12.0 months, respectively (Fig. 1C). When analysis was performed with the 18 patients with preoperative KPS scores of 70% or greater, the median OS value did not reach the 50% survival point, and the actuarial 2-year survival rate was 60.6%. In patients with a DTH-2 response size of 10 mm or larger, the OS curve did not reach 50% survival point. The OS tended to be longer than in patients who had a DTH-2 response size smaller than 10 mm, although the difference was not statistically significant with the log-rank test (Table 2). The duration of PFS showed a statistically significant association with IDH1R132H positivity (Fig. 1B), but no association with p53, MIB-1 indexes, or MHC Class I positivity. We found a statistically significant relationship between IDH1R132H positivity and the MIB-1 index (< 30%) (p = 0.045, Fisher direct method), but we found no relationship between IDH1R132H positivity and any other patient factors, including DTH-2 (p > 0.1, Fisher direct method).

When analysis was performed with all 46 patients from the 2 studies (the present study and the previous study, C000000002), the median PFS and OS were 7.5 months (95% CI 4.8–9.8 months) and 21.4 months (95% CI 14.3–28.5 months), respectively. Univariate analysis using combined patient data from the 46 patients in the 2 studies showed that the OS differed significantly between the subgroups for patient age (≥ 50 vs < 50 years, p = 0.0076), preoperative KPS score (≥ 80% vs < 80%, p = 0.0077), and RPA class (III/IV vs V, p = 0.0003). Progression-free survival was significantly different between subgroups in the study type (present study vs previous study, p = 0.038), patient age (p = 0.022), preoperative KPS score (p = 0.030), RPA class (p = 0.0084), extent of tumor removal (p = 0.0033), and DTH-2 response (p = 0.013), and tended to be different between subgroups stratified by sex (men vs women, p = 0.095). In particular, as shown in Fig. 2A, among these 7 factors, study type and DTH-2 response were independently associated with a favorable PFS outcome on multivariate analysis using the Cox proportional hazard model. Figure 2B–E shows that we observed clear differences between the PFS curve in the subgroup of patients categorized by study type and DTH-2 response, whereas we found only a tendency in OS. The DTH-2–positive subgroup showed a median OS of 26.9 months and a median PFS of 13.8 months.

Fig. 2.
Fig. 2.

A: Hazard ratios of various factors related to a favorable PFS outcome in the 46 patients enrolled in the 2 studies in a Cox proportional hazard model for multivariate analysis. Two factors (study type and DTH-2 response) were statistically significant in this model. B and C: Comparison of PFS (B, p = 0.038; C, p = 0.013, log-rank test) in subgroups of patients in the present study (n = 24, violet line) and the previous study (n = 22, blue line), and in subgroups of patients with a DTH-2 response of 10 mm or more (n = 26, violet line) and less than 10 mm (n = 19, blue line). D and E: Comparison of OS in subgroups divided by study group (D, p = 0.45) and DTH-2 response (E, p = 0.14). The x-axis indicates months; the y-axis, survival ratio.

Analysis of Lymphocyte Count and DTH-2 Response

In the present study, the mean lymphocyte counts just before chemoradiotherapy and just before AFTV injection were 1459 cells/mm3 (range 749–2319 cells/mm3) and 1036 cells/mm3 (range 178–4636 cells/mm3), respectively, with a statistically significant difference (p = 0.025, Student t-test). We found no correlation between any 2 sets of data among these mean lymphocyte counts and the size of the DTH-2 response. The mean value for the minimum lymphocyte count during chemoradiotherapy was 457 cells/mm3 (range 71–1028 cells/mm3). As shown in Fig. 3A–C, the minimum lymphocyte count was moderately correlated with the lymphocyte count just before chemoradiotherapy and was weakly correlated with the size of the DTH-2 response. The minimum lymphocyte count was the only factor associated with the size of the DTH-2 response. When OS was compared in subgroups stratified by lymphopenia grade (Grade 0–2, 3, and 4), patients with Grade 3 lymphopenia (median OS surpassed the follow-up period) probably had better survival than others (median OS 16.5 months), including those with Grade 0–2 (median OS 22.2 months) or Grade 4 (median OS 12.9 months) (Fig. 3D), although the difference in DTH-2 positivity between the 2 groups was not statistically significant (p > 0.99, Fisher direct method).

Fig. 3.
Fig. 3.

A: Correlation between the lymphocyte count just before chemoradiotherapy and the minimum (min) lymphocyte count during chemoradiotherapy (|R| = 0.59; p = 0.0026, single regression analysis). B: Correlation between the DTH-2 response size and the minimum lymphocyte count during chemoradiotherapy (dotted line, |R| = 0.43, p = 0.043, single regression analysis; solid line, |R| = 0.65; p = 0.0090, multiple regression analysis). C: Correlation between the DTH-2 response size and the lymphocyte count just before chemoradiotherapy (|R| = 0.045; p > 0.1, single regression analysis). D: Comparison of OS in subgroups divided by lymphopenia (CTCAE Grade 0–2, 3, and 4) (Grade 3 vs Grade 0–2 and 4, log-rank, p = 0.024). The x-axis indicates months; the y-axis, survival ratio.

Treatment Safety

An AFTV treatment with TMZ maintenance after FRT concomitant with TMZ treatment was well tolerated by all patients (Table 3). We observed 28 events probably related to AFTV injection and 160 events mainly related to TMZ, FRT, and antiepileptic drugs. The most frequent AFTV-related adverse effect was skin irritation consisting of local erythema, induration, and swelling at the injection sites, and included 19 events corresponding to Grade 1 toxicity and 1 event corresponding to Grade 2. Other AFTV-related adverse effects were fever in 3 cases, appetite loss in 1 case, seizures in 2 cases, and headache in 2 cases, all of which corresponded to Grade 1 or 2 toxicities. We found no evidence of any autoimmune phenomena as AFTV-related adverse effects.

TABLE 3:

Adverse events in the 24 patients in the present study*

Type of Adverse EventBefore AFTVDuring/After AFTV
Gr 1Gr 2Gr 3Gr 4Gr 1Gr 2Gr 3Gr 4
leukopenia77213200
neutropenia12225000
lymphopenia27861100
decreased platelet count41202110
decreased hemoglobin level21002100
liver dysfunction32301120
other abnormal lab findings63011000
skin reactions including local erythema, induration, & swelling at injection1170020 (19)1 (1)00
sites
hair loss120000000
drug-induced skin reactions23100100
fever01106 (3)100
nausea, appetite loss87004 (1)200
constipation50101000
stomatitis12000000
seizures020007 (2)20
vertigo, dizziness10002000
headache, other pains31002 (2)000
other adverse events26001200

Values in parentheses indicate adverse events that were probably related to AFTV. Gr = CTCAE grade.

Discussion

One of the basic objectives of this study was to test the safety of AFTV treatment concomitant with TMZ after FRT and TMZ treatment. Indeed, AFTV treatment with TMZ resulted in only minor adverse events categorized as CTCAE Grade 1 or Grade 2, similar to those in the previous study using EGFRvIII (epidermal growth factor receptor variant III) peptide vaccine.18 Aggressive treatments using radiotherapy followed by adjuvant TMZ with chemotherapeutic agents previously showed the best OS (28 months).9 However, in that study, the enrollment was stopped due to an unacceptable frequency of toxicity related to the neoadjuvant chemotherapeutic agents. Clinical use of 6 weeks of continuous TMZ often causes Grade 3 or 4 lymphopenia,7,11 and this was seen in the present study. Some animal studies have indicated that TMZ may enhance antitumor immunity.10,14,19 In a murine model of an established intracerebral tumor, vaccination-induced immunity in the setting of myeloablative TMZ, but not non-myeloablative TMZ, leads to significantly prolonged survival.19 These data are consistent with our interesting result that the median OS in patients with Grade 3 lymphopenia was superior to that in patients with Grade 0–2 lymphopenia. A similar result was reported in melanoma patients in whom the development of Grade 2 lymphopenia was positively correlated with the clinical outcome. The data demonstrated that specific CD8+ T-cell responses against selected tumor-associated antigens are enhanced by the administration of TMZ.8

On the other hand, severe lymphopenia (Grade 4) resulted in worse OS in the patients in the present study, probably due to minimal immunoreactivity. Indeed, our studies indicated that the DTH-2-positive ratio in patients with severe lymphopenia was lower than in those with milder grades of lymphopenia. This lower immunoreactivity may be due to improper timing of AFTV injection after myeloablative TMZ in the present study. If DTH-2 positivity is established even in patients with severe lymphopenia after using myeloablative TMZ, we believe that a median OS over 24 months can be expected.

These results were not inferior to those in other studies using TMZ2,18,20,23 in which patients with a preoperative KPS score of 70% (or 80%) or more were enrolled. However, to be exact, the results are comparable only to previously published data from our group, because our study differs from the previous other studies in inclusion criteria and neuroradiological assessments. In the combined 46 patients in the present study and our previous study using AFTV, the DTH-2 response and study type, but not patients' KPS score or extent of tumor removal, were independent factors associated with a favorable PFS outcome on multivariate and univariate analyses. The DTH-2-positive subgroup showed a median OS of 26.9 months and a median PFS of 13.8 months. Thus, we speculate that AFTV therapy combined with chemoradiotherapy led to the favorable outcomes.

As a limitation of the present study, some characteristics of our patient group, including age, were not typical for GBM patients in general (patient age in this study was slightly younger than those in other studies).2,18,20 Indeed, about 20% of GBM patients in each hospital were enrolled in this study, mainly owing to patients' participation in other clinical studies and not meeting the inclusion criteria (for example, some patients had smaller tumors or could not be treated with maximal resection because of tumor depth). These selection biases might affect outcome. In the present study, we found that 5 of the patients had IDH1R132H-positive GBM, which is usually associated with a better clinical outcome. However, even in patients with IDH1R132H-negative GBM, the median OS was 20.3 months. IDH1R132H and DTH-2 were not significantly associated with each other. This study does not prove efficacy for the combination of FRT, TMZ, and AFTV, which was just compared with a previous similar study consisting of FRT and AFTV injection with different timing. To address this, we are now beginning a new prospective randomized study of FRT, TMZ, and AFTV versus FRT, TMZ, and placebo. In addition, the present study may contain a bias due to a high total resection rate (67%), which was higher than in previous studies from other institutions. Stummer et al. reported that the median OS surpassed the follow-up period (24 months) in patients with no residual tumor and was 16.9 months in patients with residual tumor diameters of no more than 1.5 cm.22 However, even in the partial resection subgroup or the patients with lower KPS scores (< 80%) in our study, the median OS was over 14 months.

Conclusions

The regimen in this clinical trial consisting of FRT concomitant with TMZ followed by AFTV and TMZ maintenance in patients with newly diagnosed GBM was well tolerated and resulted in a median OS of 22.2 months and a median PFS of 8.2 months. A further randomized trial using AFTV will be promising.

Acknowledgment

We wish to thank the members of the CREIL Center at the University of Tsukuba for their critical advice in conducting the study and data management during the study period.

Disclosure

This study was supported by a project for promoting practical applications of advanced medical technologies in Tsukuba University Hospital. A portion of the materials for the tumor vaccine was provided by Cell-Medicine, Inc. (CMI) free of charge. CMI is a venture company for research and development of immunotherapy born from RIKEN (The Institute of Physical and Chemical Research) and University of Tsukuba in Japan. K.T, Y.U., T.I., are members/employees of CMI; T.O. is president and chief executive officer; and K.T, T.O and E.I. are stockholders.

Author contributions CMI to the study and manuscript preparation include the following. Conception and design: Ishikawa, Muragaki, Maruyama. Acquisition of data: Muragaki, Maruyama, Ikuta, Hashimoto, Matsuda, Abe. Analysis and interpretation of data: Ishikawa. Drafting the article: Ishikawa. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Ishikawa. Statistical analysis: Ishikawa. Administrative/technical/material support: Tsuboi, Uemae, Ishihara, Matsutani, Karasawa, Nakazato, Ohno. Study supervision: Matsumura.

This article contains some figures that are displayed in color online but in black-and-white in the print edition.

References

  • 1

    Al-Batran SERafiyan MRAtmaca ANeumann AKarbach JBender A: Intratumoral T-cell infiltrates and MHC class I expression in patients with stage IV melanoma. Cancer Res 65:393739412005

  • 2

    Clarke JLIwamoto FMSul JPanageas KLassman ABDeAngelis LM: Randomized phase II trial of chemoradiotherapy followed by either dose-dense or metronomic temozolomide for newly diagnosed glioblastoma. J Clin Oncol 27:386138672009

  • 3

    Committee of Brain Tumor Registry of Japan: Report of Brain Tumor Registry of Japan (1969–1996). Neurol Med Chir (Tokyo) 43:Supplivii11112003

  • 4

    Curran WJ JrScott CBHorton JNelson JSWeinstein ASFischbach AJ: Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials. J Natl Cancer Inst 85:7047101993

  • 5

    Ishikawa ETsuboi KTakano SUchimura ENose TOhno T: Intratumoral injection of IL-2-activated NK cells enhances the antitumor effect of intradermally injected paraformaldehy-defixed tumor vaccine in a rat intracranial brain tumor model. Cancer Sci 95:981032004

  • 6

    Ishikawa ETsuboi KYamamoto TMuroi ATakano SEnomoto T: Clinical trial of autologous formalin-fixed tumor vaccine for glioblastoma multiforme patients. Cancer Sci 98:122612332007

  • 7

    Ishikawa EYamamoto TSakamoto NNakai KAkutsu HTsuboi K: Low peripheral lymphocyte count before focal radiotherapy plus concomitant temozolomide predicts severe lymphopenia during malignant glioma treatment. Neurol Med Chir (Tokyo) 50:6386442010

  • 8

    Iversen TZBrimnes MKNikolajsen KAndersen RSHadrup SRAndersen MH: Depletion of T lymphocytes is correlated with response to temozolomide in melanoma patients. OncoImmunology 2:e232882013

  • 9

    Kim IHPark CKHeo DSKim CYRhee CHNam DH: Radiotherapy followed by adjuvant temozolomide with or without neoadjuvant ACNU-CDDP chemotherapy in newly diagnosed glioblastomas: a prospective randomized controlled multicenter phase III trial. J Neurooncol 103:5956022011

  • 10

    Kim TGKim CHPark JSPark SDKim CKChung DS: Immunological factors relating to the antitumor effect of temozolomide chemoimmunotherapy in a murine glioma model. Clin Vaccine Immunol 17:1431532010

  • 11

    Kocher MFrommolt PBorberg SKRühl USteingräber MNiewald M: Randomized study of postoperative radiotherapy and simultaneous temozolomide without adjuvant chemotherapy for glioblastoma. Strahlenther Onkol 184:5725792008

  • 12

    Kuang MPeng BGLu MDLiang LJHuang JFHe Q: Phase II randomized trial of autologous formalin-fixed tumor vaccine for postsurgical recurrence of hepatocellular carcinoma. Clin Cancer Res 10:157415792004

  • 13

    Liu GBlack KLYu JS: Sensitization of malignant glioma to chemotherapy through dendritic cell vaccination. Expert Rev Vaccines 5:2332472006

  • 14

    Mitchell DACui XSchmittling RJSanchez-Perez LSnyder DJCongdon KL: Monoclonal antibody blockade of IL-2 receptor α during lymphopenia selectively depletes regulatory T cells in mice and humans. Blood 118:300330122011

  • 15

    Muragaki YMaruyama TIseki HTanaka MShinohara CTakakura K: Phase I/IIa trial of autologous formalin-fixed tumor vaccine concomitant with fractionated radiotherapy for newly diagnosed glioblastoma. Clinical article. J Neurosurg 115:2482552011. (Erratum in J Neurosurg 118:

  • 16

    Ohno T: Autologous formalin-fixed tumor vaccine. Curr Pharm Des 11:118111882005

  • 17

    Sakamoto NIshikawa EYamamoto TSatomi KNakai KSato M: Pathological changes after autologous formalin-fixed tumor vaccine therapy combined with temozolomide for glioblastoma. Three case reports. Neurol Med Chir (Tokyo) 51:3193252011

  • 18

    Sampson JHHeimberger ABArcher GEAldape KDFriedman AHFriedman HS: Immunologic escape after prolonged progression-free survival with epidermal growth factor receptor variant III peptide vaccination in patients with newly diagnosed glioblastoma. J Clin Oncol 28:472247292010

  • 19

    Sanchez-Perez LAChoi BDArcher GECui XFlores CJohnson LA: Myeloablative temozolomide enhances CD8+ T-cell responses to vaccine and is required for efficacy against brain tumors in mice. PLoS ONE 8:e590822013

  • 20

    Smith KAAshby LSGonzalez LFBrachman DGThomas TCoons SW: Prospective trial of gross-total resection with Gliadel wafers followed by early postoperative Gamma Knife radiosurgery and conformal fractionated radiotherapy as the initial treatment for patients with radiographically suspected, newly diagnosed glioblastoma multiforme. J Neurosurg 109:Suppl1061172008. (Errata in J Neurosurg 110:J Neurosurg 111:

  • 21

    Steiner HHBonsanto MMBeckhove PBrysch MGeletneky KAhmadi R: Antitumor vaccination of patients with glioblastoma multiforme: a pilot study to assess feasibility, safety, and clinical benefit. J Clin Oncol 22:427242812004

  • 22

    Stummer WMeinel TEwelt CMartus PJakobs OFelsberg J: Prospective cohort study of radiotherapy with concomitant and adjuvant temozolomide chemotherapy for glioblastoma patients with no or minimal residual enhancing tumor load after surgery. J Neurooncol 108:89972012

  • 23

    Stupp RHegi MEMason WPvan den Bent MJTaphoorn MJJanzer RC: Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol 10:4594662009

  • 24

    Takano SKato YYamamoto TKaneko MKIshikawa ETsujimoto Y: Immunohistochemical detection of IDH1 mutation, p53, and internexin as prognostic factors of glial tumors. J Neurooncol 108:3613732012

  • 25

    Trotti AColevas ADSetser ARusch VJaques DBudach V: CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 13:1761812003

Article Information

Address correspondence to: Eiichi Ishikawa, M.D., Ph.D., Department of Neurosurgery, Graduate School of Comprehensive Human Sciences, Faculty of Medicine, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8575, Japan. email: e-ishikawa@md.tsukuba.ac.jp.

Please include this information when citing this paper: published online July 4, 2014; DOI: 10.3171/2014.5.JNS132392.

© AANS, except where prohibited by US copyright law."

Headings

Figures

  • View in gallery

    A: Scheme of AFTV clinical studies. The upper panel shows the present study, a prospective clinical trial (UMIN000001426) of FRT concomitant with TMZ, followed by AFTV and TMZ maintenance in patients with newly diagnosed GBM. The lower panel shows, for comparison, the previous prospective clinical trial (C000000002) of AFTV concomitant with FRT in patients with newly diagnosed GBM, in which, at the time of disease progression, maintenance TMZ was administered to 20 of 22 patients.15 B: PFS curve for the 24 patients in the present study. Comparison of PFS in subgroups of patients stratified by recursive partitioning analysis (RPA) Class (Class III, n = 8; Class IV, n = 10; Class V, n = 6) (p = 0.0012, log-rank test); in IDH1R132H-positive (n = 5) and negative (n = 18) cases (p = 0.14, log-rank test); and in DTH-2 positive (n = 9) and negative (n = 14) cases (p = 0.0071, log-rank test). C: OS curve for the 24 patients in the present study. Comparison of OS in subgroups of patients stratified by RPA class (p = 0.0059, log-rank test), IDH1R132H status (p = 0.023, log-rank test), and DTH-2 response cases (p = 0.061, log-rank test). The x-axis indicates months; the y-axis, survival ratio. mOS = median OS; mPFS = median PFS.

  • View in gallery

    A: Hazard ratios of various factors related to a favorable PFS outcome in the 46 patients enrolled in the 2 studies in a Cox proportional hazard model for multivariate analysis. Two factors (study type and DTH-2 response) were statistically significant in this model. B and C: Comparison of PFS (B, p = 0.038; C, p = 0.013, log-rank test) in subgroups of patients in the present study (n = 24, violet line) and the previous study (n = 22, blue line), and in subgroups of patients with a DTH-2 response of 10 mm or more (n = 26, violet line) and less than 10 mm (n = 19, blue line). D and E: Comparison of OS in subgroups divided by study group (D, p = 0.45) and DTH-2 response (E, p = 0.14). The x-axis indicates months; the y-axis, survival ratio.

  • View in gallery

    A: Correlation between the lymphocyte count just before chemoradiotherapy and the minimum (min) lymphocyte count during chemoradiotherapy (|R| = 0.59; p = 0.0026, single regression analysis). B: Correlation between the DTH-2 response size and the minimum lymphocyte count during chemoradiotherapy (dotted line, |R| = 0.43, p = 0.043, single regression analysis; solid line, |R| = 0.65; p = 0.0090, multiple regression analysis). C: Correlation between the DTH-2 response size and the lymphocyte count just before chemoradiotherapy (|R| = 0.045; p > 0.1, single regression analysis). D: Comparison of OS in subgroups divided by lymphopenia (CTCAE Grade 0–2, 3, and 4) (Grade 3 vs Grade 0–2 and 4, log-rank, p = 0.024). The x-axis indicates months; the y-axis, survival ratio.

References

1

Al-Batran SERafiyan MRAtmaca ANeumann AKarbach JBender A: Intratumoral T-cell infiltrates and MHC class I expression in patients with stage IV melanoma. Cancer Res 65:393739412005

2

Clarke JLIwamoto FMSul JPanageas KLassman ABDeAngelis LM: Randomized phase II trial of chemoradiotherapy followed by either dose-dense or metronomic temozolomide for newly diagnosed glioblastoma. J Clin Oncol 27:386138672009

3

Committee of Brain Tumor Registry of Japan: Report of Brain Tumor Registry of Japan (1969–1996). Neurol Med Chir (Tokyo) 43:Supplivii11112003

4

Curran WJ JrScott CBHorton JNelson JSWeinstein ASFischbach AJ: Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials. J Natl Cancer Inst 85:7047101993

5

Ishikawa ETsuboi KTakano SUchimura ENose TOhno T: Intratumoral injection of IL-2-activated NK cells enhances the antitumor effect of intradermally injected paraformaldehy-defixed tumor vaccine in a rat intracranial brain tumor model. Cancer Sci 95:981032004

6

Ishikawa ETsuboi KYamamoto TMuroi ATakano SEnomoto T: Clinical trial of autologous formalin-fixed tumor vaccine for glioblastoma multiforme patients. Cancer Sci 98:122612332007

7

Ishikawa EYamamoto TSakamoto NNakai KAkutsu HTsuboi K: Low peripheral lymphocyte count before focal radiotherapy plus concomitant temozolomide predicts severe lymphopenia during malignant glioma treatment. Neurol Med Chir (Tokyo) 50:6386442010

8

Iversen TZBrimnes MKNikolajsen KAndersen RSHadrup SRAndersen MH: Depletion of T lymphocytes is correlated with response to temozolomide in melanoma patients. OncoImmunology 2:e232882013

9

Kim IHPark CKHeo DSKim CYRhee CHNam DH: Radiotherapy followed by adjuvant temozolomide with or without neoadjuvant ACNU-CDDP chemotherapy in newly diagnosed glioblastomas: a prospective randomized controlled multicenter phase III trial. J Neurooncol 103:5956022011

10

Kim TGKim CHPark JSPark SDKim CKChung DS: Immunological factors relating to the antitumor effect of temozolomide chemoimmunotherapy in a murine glioma model. Clin Vaccine Immunol 17:1431532010

11

Kocher MFrommolt PBorberg SKRühl USteingräber MNiewald M: Randomized study of postoperative radiotherapy and simultaneous temozolomide without adjuvant chemotherapy for glioblastoma. Strahlenther Onkol 184:5725792008

12

Kuang MPeng BGLu MDLiang LJHuang JFHe Q: Phase II randomized trial of autologous formalin-fixed tumor vaccine for postsurgical recurrence of hepatocellular carcinoma. Clin Cancer Res 10:157415792004

13

Liu GBlack KLYu JS: Sensitization of malignant glioma to chemotherapy through dendritic cell vaccination. Expert Rev Vaccines 5:2332472006

14

Mitchell DACui XSchmittling RJSanchez-Perez LSnyder DJCongdon KL: Monoclonal antibody blockade of IL-2 receptor α during lymphopenia selectively depletes regulatory T cells in mice and humans. Blood 118:300330122011

15

Muragaki YMaruyama TIseki HTanaka MShinohara CTakakura K: Phase I/IIa trial of autologous formalin-fixed tumor vaccine concomitant with fractionated radiotherapy for newly diagnosed glioblastoma. Clinical article. J Neurosurg 115:2482552011. (Erratum in J Neurosurg 118:

16

Ohno T: Autologous formalin-fixed tumor vaccine. Curr Pharm Des 11:118111882005

17

Sakamoto NIshikawa EYamamoto TSatomi KNakai KSato M: Pathological changes after autologous formalin-fixed tumor vaccine therapy combined with temozolomide for glioblastoma. Three case reports. Neurol Med Chir (Tokyo) 51:3193252011

18

Sampson JHHeimberger ABArcher GEAldape KDFriedman AHFriedman HS: Immunologic escape after prolonged progression-free survival with epidermal growth factor receptor variant III peptide vaccination in patients with newly diagnosed glioblastoma. J Clin Oncol 28:472247292010

19

Sanchez-Perez LAChoi BDArcher GECui XFlores CJohnson LA: Myeloablative temozolomide enhances CD8+ T-cell responses to vaccine and is required for efficacy against brain tumors in mice. PLoS ONE 8:e590822013

20

Smith KAAshby LSGonzalez LFBrachman DGThomas TCoons SW: Prospective trial of gross-total resection with Gliadel wafers followed by early postoperative Gamma Knife radiosurgery and conformal fractionated radiotherapy as the initial treatment for patients with radiographically suspected, newly diagnosed glioblastoma multiforme. J Neurosurg 109:Suppl1061172008. (Errata in J Neurosurg 110:J Neurosurg 111:

21

Steiner HHBonsanto MMBeckhove PBrysch MGeletneky KAhmadi R: Antitumor vaccination of patients with glioblastoma multiforme: a pilot study to assess feasibility, safety, and clinical benefit. J Clin Oncol 22:427242812004

22

Stummer WMeinel TEwelt CMartus PJakobs OFelsberg J: Prospective cohort study of radiotherapy with concomitant and adjuvant temozolomide chemotherapy for glioblastoma patients with no or minimal residual enhancing tumor load after surgery. J Neurooncol 108:89972012

23

Stupp RHegi MEMason WPvan den Bent MJTaphoorn MJJanzer RC: Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol 10:4594662009

24

Takano SKato YYamamoto TKaneko MKIshikawa ETsujimoto Y: Immunohistochemical detection of IDH1 mutation, p53, and internexin as prognostic factors of glial tumors. J Neurooncol 108:3613732012

25

Trotti AColevas ADSetser ARusch VJaques DBudach V: CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 13:1761812003

TrendMD

Cited By

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 82 82 44
PDF Downloads 55 55 34
EPUB Downloads 0 0 0

PubMed

Google Scholar