Glioblastoma management in a lower middle-income country: Nationwide study of compliance with standard care protocols and survival outcomes in Ukraine Free

Abstract

Background

The effective treatment of high-grade gliomas is a complex problem that requires ubiquitous implementation of sophisticated therapy protocols. The present study aimed to perform population-based analysis of glioblastoma management in lower-middle-income countries.

Methods

The National Cancer Registry of Ukraine was screened for the records of adult patients with primary glioblastomas diagnosed in 2015–2019. Survival analysis was performed using Kaplan–Meier method and a multivariable Cox model.

Results

A total of 2973 adult patients with histologically confirmed glioblastoma were included in the study. Mean age of patients was 55.6 ± 11.4 years, males slightly prevailed—1541 (51.8%) cases. The completed clinical protocol including surgery followed by chemoradiotherapy was applied only in 658 (19.0%) patients. The minority of patients 743 (25.0%) were treated at the academic medical centers, where patients were more likely to receive combined treatment 70.1% compared with 57.9% (P  = .0001) at the community hospitals. The overall median survival was 10.6 ± 0.2 months, and the 2-year survival rate was 17%. The number of utilized treatment modalities contributed to better survival rates and was associated with lower hazard ratio: Protocol with 2 modalities — 0.62 (P = .0001), 3 modalities — 0.48 (P = .0001).

Conclusions

The management of glioblastoma in lower-middle-income countries is characterized by insufficient availability of treatment in academic medical centers and low rates of advanced therapy application. Survival analysis showed similar prognostic risk factors and outcomes compared with high-income countries.

Glioblastoma (GBM) remains the most challenging brain tumor, with the lowest survival rates due to insufficient reaction to conventional combined therapy.^1–3 Latest research presented the positive effect of initial supratotal resection of primary GBM facilitated by fluorescence guidance and electrophysiological monitoring.^4–6 Multidisciplinary approach shifted positive quality of life indexes in GBM patients, however slightly affected survival for the last decades all over the world.^7,8

Availability of sophisticated oncological care is a result of the economical and institutional development of each country that leads to different outcomes between populations. Basic socioeconomic factors such as household income, insurance coverage, hospital facilities, and adjuvant therapy accessibility were identified as health-affecting factors in addition to clinically related parameters.^9–11 Available nationwide studies on GBM provided data from high-income countries with lack of reports from developing countries.^{2,3,7,12–14} According to the new World Bank country classification, middle-income countries were divided into lower-middle and upper-middle-income groups, which increased the research field for population-based studies.¹⁵

The present research aimed to reveal characteristics of neurooncological care in lower-middle-income countries and was focused on demographics, utilized treatment options, and survival of patients with GBM.

Methods

Data Selection

The National Cancer Registry of Ukraine (NCR) database was screened for all cases of malignant brain tumors (C71 code according to WHO ICD-10) diagnosed between January 1, 2015 and December 31, 2019. The inclusion criteria were as follows: (1) age ≥18 years, (2) newly diagnosed histologically verified GBM, and (3) follow-up and survival data availability. Patients with concomitant cancers or without histologically verified diagnosis due to failure to undergo surgery were excluded from the study.

The present retrospective study was performed in accordance with recommendations of the Institutional review board and the Declaration of Helsinki.

Statistical Analysis

The demographic and clinical characteristics were assessed using descriptive statistics (mean and standard deviation were used for continuous variables; the proportions and frequency of distribution for categorical variables). The distributions of categorical variables were compared using Chi-square test and Fisher’s exact test.

Patients were divided depending on age at the diagnosis (<45 years, 45–64 years, and > 64 years), number of applied treatment modalities, and hospital type: Community hospital (CH) and academic medical center (AMC), that include university clinics or research institutes.

Survival rates were estimated by the Kaplan–Meier method. Overall survival (OS) was calculated as a function of time for the period from the diagnosis onset until death or last follow-up; disease-specific survival (DS) was calculated from the time of diagnosis until death from progression of GBM. Survival time was censored if the subject was alive or dropped out of study at the time of the last follow-up. Survival curves for the various subgroups were compared using the log-rank Mantel-Cox and Wilcoxon–Breslow tests.

A multivariable Cox proportional hazards ratio (HR) model with 95% confidence intervals was used to examine the adjusted effect of variables: Age, sex, extent of resection (EOR), clinical protocol (1, 2, or 3 modalities), hospital type, and treatment period. The complete-case analysis was performed when missing data were identified. All statistical tests were 2-sided, and P-value less than .05 was considered as statistically significant. Statistical analysis was performed using the Deducer package (Java GUI extensions to statistical programming platform R licensed under the GNU).

Results

A total of 10 221 individual records of patients with brain tumors (WHO International Classification of Diseases Code C71) were identified in the NCR database for the 2015–2019 period. There were 5737 patients with glial brain tumors, including 5385 adult patients (over 17 years of age). The high-grade gliomas were presented with GBM in 3090 (57.4%) cases, and the second large group was almost 3.5 times less and consisted of anaplastic astrocytomas—872 (16.2%) cases.

The histologically confirmed GBM was diagnosed in 2973 adult patients. GBM selection group included pathology types listed as glioblastoma multiforme (9440/3) − 2857 (96.1%) records, giant-cell glioblastoma (9441/3) − 41 (1.4%), and gliosarcoma (9442/3) − 75 (2.5%).

The mean age of GBM patients was 55.6 ± 11.4 years, and the median was 57 years; a total of 1032 (34.7%) patients were within the fifth decade of life (50–59 years). Only 491 (16.5%) patients were younger than 45 years, 1839 (61.9%) were in 45–64 years age group, other 643 (21.6%) were 65 years and older (P = .003). The male population was predominant with a total of 1541 (51.8 %), P = .003.

Median period from diagnosis to treatment was 0.5 months, with significant decrease during the study period from 0.6 to 0.2 months (P = .003). Most of the patients received surgical treatment combined with radiotherapy and/or chemotherapy—2114 (61.0%) cases, but only 658 (19.0%) were operated on with subsequent chemoradiotherapy (Stupp protocol, modified Stupp protocol, or CCNU). Surgery was the only treatment modality in 1352 (39.0%) cases, surgery combined with radiotherapy registered in 1364 (39.5%) patients, and with chemotherapy only in 88 (2.5%) patients (Table 1).

The difference in number of used treatment modalities was not significant between time periods (P = .35), but was significant for age groups (P < .001). Surgery as the only modality was most dominant in patients 
over 64 years of age (51.5%), and surgery with chemotherapy was used in only 0.6% of patients in this age group (Table 2).

Data for the EOR were available only for 2102 (70.7%) patients. The rate of biopsy was only 2.1% (62 cases), and there were no differences in biopsy rates between age groups, treatment period, and hospital type (P = .076). The majority of patients were registered for treatment at the CH - 2230 (75.0%), the rest 743 (25.0%) were registered at the AMC. The number of patients treated at the AMC did not increase significantly during the study period (P = .771). Patients treated at the AMC were more likely to receive combined treatment at 70.1% compared with 57.9% (P = .0001) treated at the CH.

The follow-up data were available for all included patients with GBM between 2015 and 2019. Most of the patients 2154 (72.5%) died due to GBM progression, and 113 (3.8%) died due to other causes, and 22 (0.7%) patients changed their place of residence, last of 684 (23.0%) patients were under supervision.

The Cox proportional hazards modeling revealed better prognosis for patients who received treatment at the AMC (HR 0.85, 95% CI 0.77–0.94; P = .002). Younger patients (18–44 years) were likely to live longer, than patients within the 45–64 years group (HR 1.57, 95% CI 1.39–1.79; P < 0.001) and older (HR 1.88, 95% CI 1.61–2.19; P = .0001). The number of utilized treatment modalities contributed to better survival rates and was associated with lower HR: Protocol with 2 modalities—0.62 (95% CI 0.56–0.68; P = .0001), three modalities—0.48 (95% CI 0.42–0.54; P = .0001) (Table 3).

The overall median survival was 10.6 ± 0.2 (95% CI 10.2–11.0) months. The disease-specific survival was 11.0 ± 0.2 (CI 95% 10.5–11.5) months. The Kaplan–Mayer survival analysis did not show differences between groups of patients according to sex (P = .29). The highest survival rates were observed in young adults (18–45 years)—15.3 ± 0.6 (95% CI 14.3–16.9) months (P < .001) (Table 4).

The median overall survival was the highest in the combined clinical protocol group—16.0 ± 0.4 months (95% CI: 15.3–16.8), and the worst prognosis was in surgery alone with median OS 4.9 ± 0.2 (95% CI 4.5–5.3) months at the CH, and 6.9 ± 0.9 (95% CI 5.2–8.6) months at the AMC (P < .001). For patients with survival of more than a 1-month (period of surgical complications), the median OS raised up to 6.1 ± 0.2 (95% CI 5.6–6.6) months at the CH, and to 7.2 ± 0.9 (95% CI 5.4–9.0) at the AMC (Table 5).

Overall median survival was 2.1 months longer in patients at the AMC (12.2 ± 0.5, 95% CI 11.3–13.1) compared with the CH (10.1 ± 0.2, 95% CI 9.6–10.5). The treatment at the AMC provided better outcomes for any type of clinical protocol and any age group.

The higher EOR was a predictor of better survival in each age group. The median OS for patients after biopsy for GBM was 5.9 ± 0.6 (95% CI 4.7–7.0) months, for patients who underwent tumor resection it was 11.0 ± 0.2 (95% CI 10.5–11.4) months.

The median follow-up period was 33.3 ± 1.5 (95% CI 30.9–38.8) months. The addition of each treatment modality increased the 2-year survival rate from 11% to 28% and the median follow-up period from 31.8 ± 3.6 (95% CI 24.7–38.8) to 38.0 ± 2.2 (95% CI 33.7–42.3) months (Table 6).

Discussion

The study by Griswold et al. showed that most of the randomized trials in neurosurgery (73.3%) were performed in high-income countries with a prevalence of industrial and institutional funding.¹⁶ In accordance with this, the recent published nationwide research in GBM outcomes presented the data for high-income countries without comparison to low-income countries and could not reflect the global trends.^{7,9,10,12,13,17}

Furthermore, according to the latest World Bank classification of countries by income, the middle-income group was divided into lower-middle-income and upper-middle-income countries.¹⁵ Additionally, the population-based studies on GBM outcomes in developing countries are not available. The present research aimed to fill this gap, discovering the patterns of care and survival outcomes in patients with newly diagnosed GBM in lower-middle-income countries.

The present research reveals the GBM as a most common glial tumor in adults (57.4%), which is in line with Ostrom et al. data (57.7%) described the brain tumors in the United States population for 2013–2017 period.² Proportions of GBM subtypes were with huge prevalence of multiforme GBM (96.1%), which corresponds with Fabbro-Peray et al. revealed 96.8% of them in France.¹² However, further demographic analysis showed significant differences compared to high-income countries. It was established that the median age of patients at the time of diagnosis was 57 (56–58) years. In other studies, it ranged from 61.3 in the Netherlands to 66 years in the US and Denmark.^{2,7,12,13,17,18} Stratification by age groups also showed that the majority of the patients (78.4%) were at active working age—18–64 years old. And the prevailed population of GBM patients (65.9%) were in 50–69 years group, which differs from Efremov et al. work for Germany where persons in 50–69 years group reached 50.6%.⁷

The previous studies documented increasing in GBM incidence in elderly populations of high-income countries due to higher life length, health awareness, and diagnostic tools accessibility.^2,11,17 This is despite the fact, that the patients of older age groups for the economic, social, and medical reasons are less able to be applied to opportune health care.^19,20 And obviously in the developing counties these effects could have higher presentation but not only in persons of old age.

Bower et al.¹⁰ demonstrated the heterogeneity of treatment options and outcomes in GBM patients between different economic communities of one population, although there were the same median age both in low- and high-income groups. Considering the data from Nilsson et al.²¹ defining the effect of household income on brain tumors incidence, the results of our study could reflect the presence of general economic factors affecting population health in lower-middle-income countries. The further research and meta-analysis in countries of different economical levels are needed.

Another prominent discrepancy we found was that males slightly dominated: male to female ratio was 1,07 with 51.8% of males. In contrast to the high-income counties, where GBM were found in males 1.6 times more often with 57%–62%.^2,7,12,13,17 The recent explanations of these sex differences were found on molecular basis with focus on the variance in gen expressions, metabolic and immune processes.²² In this context, our results could complement to the general epidemiological landscape of GBM and present new reasons for further studies of gliomagenesis risk factors.

The compliance with the combined clinical protocols is the essential precondition for better prognosis in patients with GBM. The technical and financial availability of individualized therapy regimens for each patient was described as a prominent improvement that provided superior outcomes in high-income countries.^10,11,23,24 Efremov et al. reported the rising of chemotherapy application rate for GBM treatment almost on twice in Germany, from 28.3% to 60.0% between 1999 and 2014, at the same time Hansen et al. declared that 76% of the patients received radiotherapy and 59% chemotherapy in Denmark.^3,7 We observed low coverage of entire combined therapy protocols in patients with GBM: only 19% of them received chemoradiotherapy after surgery, when surgery followed by radiotherapy was applied in 39.5% of patients. Hereby, the levels of surgery as a single treatment option remained high during the research period (34.8%–41.1%), particularly, in aged patients (51.5%). That is almost two times higher in contrast to the study by Fabbro-Peray et al., where the surgery without any adjuvant therapy was performed only in 20% of patients in France.¹² These notable differences in levels of received treatment indicate limited compliance with approved oncological protocols for GBM in a lower middle-income country. Inadequate financing of health care system, especially in developing countries, seems to be a reasonable reason for the low distribution of adjuvant therapy, considering the results by Bower et al.¹⁰

Previously described prognostic factors associated with better survival in GBM were also valid for our target population. Combined treatment for GBM provided at high-volume hospitals with academic facilities leads to better outcomes in patients.^3,14 The AMC has higher amount of financing, modern equipment and experienced specialists and engaged about 40% of patients.^14,25,26 In our study, specialized oncological care at the AMC was provided only for a quarter of patients and was stable over the research period. Nevertheless, the time from diagnosis to treatment start has decreased. This positive shift may be linked to higher public awareness, and better communication between neurosurgical clinics and diagnostic centers. The access to care and distance to the treatment location were also previously identified as factors influencing outcomes in GBM patients.^10,11,24,25

Clinical studies of GBM treatment outcomes noticed an increase in survival rates after widespread application of new approaches to adjuvant and concomitant chemotherapy after 2005 when population-based studies still show fever outcomes improvement. Efremov et al. assumed the accompanying development of neurosurgical techniques and tools also could affect survival rates positively, even for GBM without chemotherapy sensitiveness.⁷

The setting of contraindications to aggressive treatment is more likely for aged patients that usually underwent biopsy followed chemoradiotherapy. Also, the initial low index of performance status in aged patients is associated with higher levels of complications and early mortality. On another side, the adjuvant therapy toxicity often defines its spontaneous withdrawal in disabled aged patients before the end of the course was achieved. The incomplete clinical protocol in patients older than 65 years leads to decrease in median OS, Fatehi et al. found 7.1 months, in our study, it was 4.4 months.⁸ Wherein adolescent and younger adults do not achieve so significant benefit from chemotherapy in comparison to an older patient, but receive it concomitant and commonly after radiotherapy. Comparing survival outcomes in series with different levels of adjuvant therapy application can demonstrate the value of treatment options.

According to the population-based European research median OS for GBM patients range from 10.0 in Germany to 11.7 in Finland, that compatible with our data (10.6 months) and SEER research in the United States (11 months).^{7,12,13,17,23,26} Also 1-year and 2-year survival rates (45% and 17%) are also compatible with SEER data (45.7% and 19.0%) and other European studies.^{2,12,13,17,23}

This notable equality of survival rates between our research data and studies from high-income countries calls to look for a reason that younger median age at the diagnosis compensates for lower levels of adjuvant therapy. Otherwise, it could be assumed that in high-income countries the levels of chemotherapy use can overlap the real GBM sensitivity, and expansion of adjuvant therapy routine prescription would not improve survival. Anyway, the issue of an individualized approach in adjuvant therapy remains unresolved and further population studies could identify real explanations for comparable survival rates between countries with different levels of income.

Limitations of the Study

Presented research has restrictions typical for retrospective studies with incompleteness of case data registered at the cancer register. The analysis of prognostic survival factors was limited by the lack of data on the of performance status of patients, a correct estimate of EOR, as well as the molecular profile of the tumor (MGMT gene promoter methylation and IDH status) and the type of applied chemotherapy.

Conclusions

The presented research is the first national wide study of glioblastoma treatment and outcomes in lower-middle-income countries. Our results showed that compared to high-income countries the glioblastomas debuted in younger persons and were less likely treated at AMC. Also, the application of standard combined treatment protocols has relatively low coverage of radiochemotherapy after surgical resection. Survival analysis revealed the same prognostic risk factors between high and lower-middle-income countries, and comparable survival outcomes. Awareness of these aspects requires further complementary multicenter or population-based studies in developing countries.

Funding

No funding was received for conducting this study.

Conflicts of Interest

The authors have no relevant financial or nonfinancial interests to disclose.

Authorship

Conception and design: AR, VR. Administrative support: VR, ZF. Provision of study materials or patients: AR, VK, AD, ZF. Collection and assembly of data: AR, ZF. Data analysis and interpretation: AR, VK, AD. Manuscript writing: AR, VK, AD, VR, ZF. Final approval of manuscript: AR, VK, AD, VR, ZF.

References