Recurrence pattern of glioblastoma treated with intensity-modulated radiation therapy versus three-dimensional conformal radiation therapy

1. Patient selection

We retrospectively examined 91 patients with glioblastomas treated with radiation therapy between January 2013 and December 2019 at Seoul St. Mary’s Hospital. The inclusion criteria were histologically confirmed glioblastoma classified according to the 2016 World Health Organization Classification of Tumors of the Central Nervous System [11], following surgery or biopsy studies, and received adjuvant radiation therapy. The exclusion criteria were previous brain radiation therapy, radiation therapy discontinued midway, no follow-up diagnostic images after radiation therapy, and diagnosed with another primary cancer within 5 years before glioblastoma diagnosis (exceptions were made for primary cancers not affecting the treatment and follow-up of glioblastomas).

2. Treatment protocol

Brain magnetic resonance imaging (MRI) and computed tomography (CT) were performed to determine tumor location. All patients underwent surgery or biopsy studies for histological confirmation. All received adjuvant radiation therapy either with or without concurrent temozolomide. Concurrent chemoradiation therapy involved the administration of temozolomide (TMZ) at a dose of 60-80 mg/m²/day throughout the radiation therapy period. Adjuvant TMZ was administered at 150–200 mg/m² for 5 days in 28-day cycles, for up to six cycles. The detailed chemotherapy regimen and dosage are summarized in Table 1. CT simulation was conducted with the head secured using thermoplasty. The gross target volume (GTV) was defined as the surgical resection cavity plus any residual tumor identified in post-contrast T1-weighted MRI scans. The clinical target volume (CTV) was defined as the GTV with an added 2 cm margin. The planning target volume was defined as the CTV with an additional 3–5 mm margin, in accordance with the ESTRO-ACROP guidelines [12]. Both 3D-CRT and IMRT were performed using CT-based treatment planning. The radiation therapy dose for all patients was 59.4–60 Gy, administered in 30–33 fractions (1.8–2.0 Gy daily fraction).

Table 1.

Patient characteristics

3. Dosimetric analysis of recurrence

To analyze recurrent tumor dosimetry, the initial radiotherapy-planning CT scan was fused with the follow-up brain MRI scan using MIM software (version 7.1.0; MIM Software, Cleveland, OH, USA). The radiation oncologist then defined and contoured the recurrent tumor volume based on the T1-weighted enhancing mass on brain MRI.

Recurrence patterns were classified as follows [13-16] (Fig. 1). In-field recurrence: 80%–100% of the recurrent gross tumor volume (rGTV) intersects with the prescribed 95% isodose line of the radiation target field volume. Marginal recurrence: 20%–80% of the rGTV intersects with the prescribed 95% isodose line of the radiation target field volume. Out-of-field recurrence: <20% of the rGTV intersects with the prescribed 95% isodose line of the radiation target field volume.

Fig. 1.

Illustration of recurrence patterns post-radiation therapy in glioblastoma. The yellow circle indicates the recurred gross tumor volume, and the red line delineates the prescribed radiation field. (A) In-field recurrence, (B) marginal recurrence, and (C) out-field recurrence.

4. Statistical analysis

Statistical analyses were performed using R version 4.3 (https://www.r-project.org) and Statistical Package for Social Sciences version 18 (SPSS Inc., Chicago, IL, USA). The primary endpoints were overall survival (OS) and progression-free survival (PFS). Survival rates were estimated using the Kaplan–Meier method, and differences between groups were tested using the log-rank test. Cox proportional hazards model was used to perform univariate analysis. Categorical variables were compared using the chi-square or Fisher exact test. Statistical significance was set at p < 0.05.

1. Patient outcomes

The median follow-up time was 18.9 months (range, 4.9 to 121.7). The median age was 59 years, with a balanced distribution between the two treatment groups. IMRT (n = 60) and 3D-CRT (n = 31) were the treatment modalities. Patient characteristics are summarized in Table 1. The median OS was 18.9 months. The median PFS was 9.4 months. For patients treated with 3D-CRT, the median OS and PFS were 19.3 and 10.8 months, respectively. For patients treated with IMRT, the median OS and PFS were 18.4 and 8.9 months, respectively. The p-values for OS and PFS between the two treatment groups were 0.363 and 0.731, respectively (Fig. 2). Patients who underwent gross total resection (GTR) had a median OS and PFS of 25.8 and 12 months, respectively. Patients who did not undergo GTR had a median OS and PFS of 17.3 and 8.3 months, respectively. The p-values for OS and PFS between GTR and non-GTR groups were 0.003 and 0.001, respectively. In patients with MGMT (O⁶-methylguanine-DNA-methyltransferase) methylation, the median OS and PFS were 25.1 and 15.2 months, respectively. The p-values for OS and PFS between MGMT methylated and non-methylated groups were 0.094 and 0.075, respectively (Fig. 3).

Fig. 2.

Kaplan-Meier survival curves of outcomes between two treatment groups: (A) overall survival (OS) of all patients, (B) progression-free survival (PFS) of all patients, (C) OS of 3D-CRT versus IMRT, and (D) PFS of 3D-CRT versus IMRT. 3D-CRT, three-dimensional conformal radiation therapy; IMRT, intensity-modulated radiation therapy.

Fig. 3.

Kaplan-Meier survival curves of patients who underwent gross total resection (GTR): (A) OS of GTR versus PTR or biopsy only, (B) PFS of GTR versus PTR or biopsy only, (C) OS of MGMT status, and (D) PFS of MGMT status. OS, overall survival; PFS, progression-free survival; GTR, gross total resection; PTR, partial tumor resection; MGMT, O⁶-methylguanine-DNA-methyltransferase promoter methylations.

2. Recurrence pattern

Seventy-eight patients experienced recurrence. Twelve patients did not experience recurrence. The recurrent cases were as follows: 67 (73.6%), in-field; 5 (5.5%), marginal; and 19 (20.9%), out-of-field (Table 2). Multiple recurrence types were observed in 12 cases. Among 3D-CRT-treated cases, 24 (70.6%) were in-field; 1 (2.9%), marginal; and 9 (26.5%) out-of-field recurrence. Among IMRT-treated cases, 43 (75.5%) were in-field, 4 (7.0%) had recurrence, and 10 (17.5%) had out-of-field recurrence. The IMRT group had a higher marginal recurrence rate (7.0%, four cases) than the 3D-CRT group (2.9%, one case).

Table 2.

Comparison of recurrence patterns between 3D-CRT and IMRT treatments

Notably, out-of-field recurrence was more frequent in IMRT-treated patients who underwent GTR than 3D-CRT-treated patients (20.0% vs. 8.3%). However, the odds ratio (OR) was 0.37 (range, 0.01 to 4.49) without statistical significance (p = 0.626). Among patients who did not underwent GTR, out-of-field recurrence was observed in 16.2% of IMRT-treated and 36.3% of 3D-CRT-treated patients with OR of 2.95 (range, 0.86 to 10.12) with marginal significance (p = 0.079) (Tables 2, 3).

Table 3.

Comparative OR and p-values for recurrence patterns glioblastoma patients treated with 3D-CRT versus IMRT

1. Analysis of recurrence patterns

In-field recurrence remains the predominant challenge, consistent with previous studies, suggesting that even with dose escalation, local control of glioblastoma is not substantially improved [14]. This is indicative of the aggressive biological behavior of glioblastoma, which often extends beyond radiographically defined tumor margins and remains elusive even with the most targeted radiation techniques [13].

Marginal recurrence, although less frequent, is a concern, particularly in the context of IMRT. Our data showed a non-significant trend toward higher marginal recurrence rates with IMRT (6.7%) than with 3D-CRT (3.2%). This marginal difference, although not significant (OR = 0.41; p = 0.647), invites a re-evaluation of the IMRT planning processes to ensure that treatment margins are adequate (Fig. 4, Table 3). This trend emphasizes the need for a nuanced understanding of glioblastoma infiltration patterns and perhaps supports a more aggressive margin delineation strategy in line with the ESTRO-ACROP guidelines [12].

Fig. 4.

Comparative odds ratios (OR) and p-values for recurrence patterns of glioblastoma patients treated with 3D-CRT versus IMRT: (A) all patients, (B) patients who received GTR, and (C) patients who did not received GTR. Error bars represent the 95% confidence intervals for the estimated ORs. 3D-CRT, three-dimensional conformal radiation therapy; IMRT, intensity-modulated radiation therapy; GTR, gross total resection.

2. Impact of surgical extent on recurrence and survival

Patients who underwent GTR showed better survival outcomes, echoing the principle that the extent of surgical resection plays a critical role in glioblastoma management. This is substantiated by significant differences in OS and PFS in favor of GTR (p = 0.003 and p = 0.001, respectively) (Fig. 3) and by existing evidence suggesting that maximal tumor resection is associated with improved survival [1,18]. When comparing the impact of GTR between 3D-CRT and IMRT, the benefit of GTR remains evident. For the 3D-CRT group, GTR had a hazard ratio (HR) of 1.732 for OS (p = 0.176) and 1.796 for PFS (p = 0.141), while for the IMRT group, GTR had an HR of 2.294 for OS (p = 0.004) and 2.413 for PFS (p = 0.004). These results suggest that the advantage of GTR is consistent across different radiotherapy modalities, but it is more statistically significant in the IMRT group (Table 4).

Table 4.

Univariate Cox proportional hazards analysis for OS and PFS

Interestingly, the study found a trend toward higher out-of-field recurrence with IMRT in patients undergoing GTR, although the difference was not significant (OR = 0.37; p = 0.626). This suggests that 3D-CRT is more effective than IMRT in controlling the disease spread outside the primary treatment zone in this subgroup. However, the lack of significance warrants a cautious interpretation. After GTR, the surgical cavity and surrounding tissues may harbor residual microscopic glioma cells not visible on imaging [19]. The less conformal nature of 3D-CRT may cover these regions more effectively, despite the potential for slightly increased toxicity [8,9]. Conversely, the study found a trend toward lower out-of-field recurrence with IMRT in patients who did not receive GTR (OR = 2.95; p = 0.079), although this did not reach conventional levels of significance (p = 0.079). This implies that without complete surgical removal of the tumor, 3D-CRT may not be as effective in managing potential tumor spread. The near-significant p-value suggests that there is an approximately 195% increase in the odds of out-of-field recurrence with 3D-CRT compared with IMRT, which may be clinically relevant, indicating that IMRT could be a better option in cases where GTR is not achieved. Buglione et al. [20] analyzed the pattern of recurrence in patients treated with the Stupp protocol. They found only one case of out-field recurrence in those treated with 3D-CRT, and no out-field recurrence was observed in the IMRT group. Moreover, Morganti et al. [21] specifically analyzed IMRT in a cohort where 12 patients underwent GTR and seven patients did not receive GTR. The out-field recurrence rate was 13.3%, which is lower than our study. The results indicated that IMRT was more effective in reducing out-field recurrences compared to 3D-CRT, highlighting its advantage in non-GTR cases.

3. Comparison of recurrence patterns across studies

Our study's comparison of recurrence patterns (Table 5) is essential for evaluating the effectiveness of radiotherapy in glioblastoma treatment. In our study, the in-field recurrence rate for 3D-CRT was 70.6%, which aligns with historical data from Wallner et al. [22] and Minniti et al. [23], who reported rates of 78% and 81%, respectively. For IMRT, our study showed an in-field recurrence rate of 75.5%, similar to the findings of Langhans et al. [16] and Zheng et al. [24], who reported rates of 75.4% and 83.6%, respectively. Faustino et al. [18] reported an even higher in-field recurrence rate of 83% for their mixed technique group (IMRT of 57%, 3D-CRT of 43%).

Table 5.

Studies of recurrence patterns in glioblastoma

In our study, the marginal recurrence rates were 2.9% for 3D-CRT and 7.0% for IMRT. These rates are consistent with those reported by Oppitz et al. [25] for 3D-CRT (3%). However, the marginal recurrence rates reported by Minniti et al. [23] for 3D-CRT (5.7%) and Faustino et al. [18] for their mixed group (9.7%) are broader than our study's definitions, which could explain the higher rate of marginal recurrence. However, although 3D-CRT appears to have a lower tendency for marginal recurrence compared to IMRT, it is difficult to interpret the statistical correlation due to the varying definitions of marginal recurrence in each study.

The out-field recurrence rates in our study were different between 3D-CRT (26.5%) and IMRT (17.5%), even though lacking statistical significance. The out-field recurrence rate is consistent with Morganti et al. [21], who reported a 13.3% out-field recurrence rate for their IMRT group. Faustino et al. [18] reported a lower out-field recurrence rate of 7.3%, which could be attributed to differences in patient selection and treatment protocols.

4. Novelty and limitations of the study

Although previous studies have explored various aspects of glioblastoma recurrence patterns [13,16,18,22,23,25-27], our study is the first to specifically compare the recurrence patterns of glioblastoma treated using IMRT and 3D-CRT.

While our findings are promising, they are tempered by certain inherent limitations. The retrospective design, while expedient, is susceptible to biases inherent to post hoc analyses and may not provide the control and randomization offered by prospective studies. The sample size, although adequate for initial observations, requires expansion in future studies to confirm these findings across broader demographic and clinical settings. The single-institution study design may have affected the generalizability of the results, underscoring the need for multicenter studies to validate and extend our findings.

5. Future directions

Future research should be directed toward enhancing the precision of targeted radiation therapy for glioblastoma. This involves not only improving imaging technologies for tumor delineation but also potentially integrating biomarker-driven strategies to personalize treatment. Such advancements could lead to a reduction in both marginal and out-of-field recurrence, paving the way for improved glioblastoma management and outcomes.

6. Conclusion

The data presented in this study provide a nuanced perspective on the recurrence patterns of glioblastoma treated with radiotherapy. Despite technological advancements in radiotherapy, such as IMRT, our results did not reveal a significant advantage in terms of OS or PFS compared with 3D-CRT. This suggests that factors beyond the scope of radiation technique is instrumental in influencing patient outcomes.

The observed recurrence patterns reinforce the notion that the aggressive biological behavior of glioblastoma may diminish the potential benefits of more precise radiation modalities. In-field recurrence rate remains a principal concern, highlighting the necessity for therapeutic strategies that extend beyond dose escalation. Our study reported slightly higher marginal recurrence rates with IMRT, indicating that even with advanced targeting techniques, achieving complete margin coverage is challenging.

Furthermore, patients who underwent GTR had better survival outcomes, echoing the principle that the extent of surgical resection plays a critical role in glioblastoma management. Conversely, for patients who did not receive GTR, our findings indicated a trend toward lower out-of-field recurrence with IMRT, albeit not significant.

In conclusion, our study provides valuable insights into the collective understanding that personalized and flexible treatment planning are crucial for glioblastoma management. Future research should aim to harness new developments in imaging, radiation planning, and targeted therapies to improve glioblastoma treatment. Our ongoing pursuit is to better manage this challenging cancer and improve patients' quality of life.