Abstract
Background and purpose:
Consensus for defining gross tumor volume (GTV) and clinical target volume (CTV) for limited-field radiation therapy (LFRT) of GBM are not well established. We leveraged a department MRI simulator to image patients before and during LFRT to address these questions.
Materials and methods:
Supratentorial GBM patients receiving LFRT (46 Gy + boost to 60 Gy) underwent baseline MRI (MRI1) and interim MRI during RT (MRI2). GTV1 was defined as T1 enhancement + surgical cavity on MRI1 without routine inclusion of T2 abnormality (unless tumor did not enhance). The initial CTV margin was 15 mm from GTV1, and the boost CTV margin was 5–7 mm. The GTV1 characteristics were categorized into three groups: identical T1 and T2 abnormality (Group A), T1 only with larger T2 abnormality not included (Group B), and T2 abnormality when tumor lacked enhancement (Group C). GTV2 was contoured on MRI2 and compared with GTV1 plus 5–15 mm expansions.
Results:
Among 120 patients treated from 2014–2019, 29 patients (24%) underwent replanning based on MRI2. On MRI2, 84% of GTV2 were covered by GTV1+5 mm, 93% by GTV1+7 mm, and 98% by GTV1+15 mm. On MRI1, 43% of GTV1 could be categorized into Group A, 39% Group B, and 18% Group C. Group B’s patterns of failure, local control, or progression-free survival were similar to Group A/C.
Conclusions:
Initial CTV margin of 15 mm followed by a boost CTV margin of 7 mm is a reasonable approach for LFRT of GBM. Omitting routine inclusion of T2 abnormality from GTV delineation may not jeopardize disease control.
Keywords: Glioblastoma, Pattern of Failure, Interim MRI, Limited Field RT
INTRODUCTION
Glioblastoma (GBM) is the most common primary malignant brain tumor in adults and is treated with a combination of surgery, focal radiation therapy (RT), and temozolomide (TMZ) [1]. RT volume definitions gradually emerged based on surgical series demonstrating that 80–90% of recurrences occurred within 2 cm of the enhancing tumor on CT or MRI and that isolated tumor cells were present at the periphery of T2 abnormalities [2–5]. A landmark surgical study demonstrated that the T1 contrast-enhancing abnormality contained high-density GBM tumor cells, while the T2 abnormality represented either brain parenchyma with infiltrating tumor cells or peritumoral edema [3]. There are currently different interpretations of such data for RT treatment planning. In the US, the standard RT guidelines have included an initial field of 46 Gy to the T1 enhancing mass and T2 abnormality plus a clinical target volume (CTV) expansion of 2 cm followed by a sequential boost to the T1 enhancement plus a 2 cm CTV expansion to 60 Gy [6,7]. In contrast, the European practice typically treats the T1 enhancement with a 2 cm CTV expansion to 60 Gy without intentionally using T2 abnormality to generate the expansion, though visually verifying the inclusion of T2 abnormality within the CTV expansion is recommended but not required [8]. Furthermore, some institutions in the US and the Adult Brain Tumor Consortium (ABTC) have adapted limited-field approaches, demonstrating that omission of T2 hyperintensity and reducing margin expansions doesn’t translate into inferior outcomes or alter patterns of failure [9–13]. The impetus to reduce treatment volumes has been motivated by a desire to decrease late effects, including cognitive deficits and RT-induced lymphopenia [14–17].
As target volumes are reduced, the appropriate determination of tumor extent is imperative for treatment planning. The interval between surgery and RT can span 4 to 6 weeks, while the boost phase of RT starts 4 to 5 weeks later, creating a window for tumor volume changes from the postoperative MRI. An interim MRI during RT can assess early changes of tumor growth and evaluate the adequacy of planning margins [18]. Leveraging a departmental MRI simulator, we incorporated a simulation MRI before RT and an interim MRI after two weeks of RT for our limited-field RT (LFRT) approach. We have previously shown that this LFRT approach yielded similar progression-free survival (PFS) and overall survival (OS) while reducing lymphopenia as compared to historical control with standard-field RT at the same institution. Notably, on a matched-pair comparison of the LFRT versus the standard-field approach using propensity-score, there was similar PFS and OS observed [9].This study retrospectively analyzed the radiographical changes of the interim MRIs during the LFRT and their relationship to baseline imaging characteristics and patterns of failure. The purpose is to evaluate if our LFRT approach using smaller margins and the omission of routine inclusion of T2 abnormality from GTV delineation would jeopardize tumor coverage and disease control.
METHODS
Patient Selection
Adult patients with newly diagnosed, supratentorial, histologic or molecular GBM treated with a planned six-week course of LFRT with concurrent TMZ between November 2014 to September 2019 who had a baseline simulation and interim MRI after two weeks of RT were retrospectively reviewed. Our department’s MRI simulator allowed for pretreatment MRI simulation in the treatment position without additional financial costs or visits to diagnostic radiology scanners. In November 2014, we adopted a standard LFRT approach for treating GBM and a uniform method for RT simulation that included a baseline MRI simulation concurrently with CT simulation approximately two weeks before RT and an interim MRI evaluation approximately two weeks after the start of RT. The interim MRI was conducted to ensure that no additional treatment modifications are needed before the smaller RT boost and to provide enough time for the treating team to replan if needed. Since this approach was our standard clinical workflow for all GBM patients, it was exempted from Institutional Review Board (IRB) review. Patients were provided informed consent regarding the simulation and imaging process before treatment. We conducted a retrospective review to analyze the clinical outcomes from this approach and received IRB approval for this analysis (IRB number: 201312136). Molecular GBM was defined as histological grade II-III astrocytoma without isocitrate dehydrogenase (IDH) mutation but with at least one of the following molecular alterations: telomerase reverse transcriptase (TERT) promoter mutation, epidermal growth factor receptor (EGFR) amplification, or a combination of whole chromosome 7 gain and whole chromosome 10 loss in concordance with published recommended diagnostic criteria [19]. LFRT was adopted as our institutional standard to treat GBM patients in November 2014 and was previously described [9]. Patients with a prior course of cranial RT or who didn’t receive 60 Gy dose were excluded.
Treatment
As previously described, LFRT consisted of 46 Gy in 2-Gy fractions delivered to the initial planning target volume (PTV4600), followed by a sequential boost of 14 Gy in 2-Gy fractions delivered to the boost PTV6000. PTV4600 was generated as gross tumor volume (GTV) plus a CTV margin of 15 mm and a PTV margin of 3–5 mm; PTV6000 was generated as GTV plus a CTV margin of 5–7 mm and a PTV margin of 3–5 mm. The slight variation of the boost CTV margins and the PTV margins are due to different clinical trial requirements or physician preference. All patients received concurrent daily TMZ at 75 mg/m2 during RT. Adjuvant TMZ was administered at a dose of 150–200 mg/m2 on days 1–5 in a 28-day cycle.
Simulation and Interim MRI
All patients underwent a simulation MRI (MRI1) for RT planning 1–2 weeks before RT using a departmental MRI scanner (Philips 1.5 Tesla Ingenia MRI). The MRI sequences included axial fluid attenuation inversion recovery (FLAIR) sequence at 3 mm thickness and magnetization-prepared rapid gradient echo (MPRAGE) sequences with and without gadolinium contrast at 1.2 mm thickness. GTV1 was defined as T1 enhancement plus the surgical cavity on the MRI1 without routine inclusion of T2 abnormality (unless tumor did not enhance well). GTV1s were categorized into three groups based on MRI1 (the pretreatment MRI simulation) imaging characteristics and contouring patterns: identical T1-enhancing volume (including surgical cavity) and T2/FLAIR abnormality for contouring GTV1 (Group A), only T1 used for contouring GTV1 with the larger T2/FLAIR abnormality omitted from GTV1 contour (Group B), and T2/FLAIR abnormality used for contouring GTV1 for patients with predominantly non-enhancing tumors (Group C). Although the post-operative simulation MRIs were used to define the GTV1, the pre-operative MRIs were also reviewed to guide the contouring pattern, especially for Group C. Figure 1 depicts examples of each GTV1 group.
FIGURE 1.
Representative MRI Examples of the three GTV1 Groups from the Baseline MRI. T1-weighed sequence with contrast (top) and T2/FLAIR sequence (bottom), GTV contour outlined in red. (A) Group A with identical T1 and T2 abnormality. (B) Group B with T1 only with larger T2 abnormality not included. (C) Group C with T2 abnormality when tumor lacked enhancement.
All patients subsequently underwent an interim MRI (MRI2) approximately two weeks after initiating RT using the same simulation MRI scanner and was fused to MRI1 for comparison. As mentioned above, the clinical workflow incorporating both a simulation MRI1 and an interim MRI2 was utilized for all clinically appropriate patients. If the enhancing tumor was suspected of having progressed beyond the boost CTV based on the imaging characteristics on the MRI2, the treating physician revised the boost plan to ensure the new enhancing abnormality was adequately covered.
As a component of our retrospective analysis, GTV2 was contoured on MRI2 and compared to GTV1. In addition, volumetric expansions of GTV1 were generated with different hypothetical CTV margins (5–15 mm) to evaluate their coverage of GTV2. Appropriate coverage was defined as no component of GTV2 extending beyond the hypothetical CTV expansions.
Follow-up and Patterns of Failure
Patients underwent serial MRI imaging 4–6 weeks after RT completion and every 8 weeks thereafter. As previously described, local and distant failures were determined based on the multi-disciplinary assessment of histopathologic, clinical, and radiographic data [20]. The pattern of failure was determined by comparing the MRI at time of progression with the delivered RT treatment plan using isodose lines (IDLs). The pattern of failure category was dependent on whether the progressing lesion(s) were located within certain IDLs: central (within 57 Gy IDL [95% of 60 Gy]), in-field (outside 57Gy IDL but within 44 Gy IDL [95% of 46Gy]), marginal (outside 44Gy IDL but within 12 Gy IDL [20% of 60 Gy]), distant (outside of 12 Gy IDL). The progressing lesion was typically defined as the entire enhancing abnormality that had been enlarging compared to the prior MRIs and had been deemed a recurrent tumor by multi-disciplinary assessment. In cases with multiple discrete lesions on follow-up MRI, each lesion was assessed with respect to the IDLs mentioned above.
Statistical Analysis
Chi-square (or Fisher’s exact test when appropriate) and Mann-Whitney U were used to compare categorical and continuous variables, respectively. Local control (LC), PFS, and OS were estimated using the Kaplan-Meier method and compared using log-rank test. All time-to-event endpoints were calculated from the start of RT. Univariable (UVA) and multivariable (MVA) were performed using the Cox proportional hazard model to identify factors associated with endpoints. All UVA variables with p < 0.20 were selected for MVA. Log of negative log plotting was used to assess the proportional hazards assumption. All tests were 2-sided, and significance was defined as p < 0.05. Statistical analyses were performed using SPSS Statistics (version 23.0, IBM, Armonk, NY) and R Studio (version 3.6.2).
RESULTS
One hundred twenty patients met inclusion criteria. All patients received 60 Gy in 30 once-daily fractions. Two patients were excluded because they did not receive the standard RT dose: one received 66 Gy due to significant treatment break, and the other only completed 42 Gy. Table 1 outlines baseline characteristics. Median follow up was 13 months (range: 1.1–57 months), median age at diagnosis was 59 years (range: 21–79 years), and median Karnofsky performance status (KPS) was 80 (range: 50–100). Fifty percent were female, and 36% achieved gross total resection (GTR). Five percent were molecular GBM, 39% had O6-methylguanine-DNA-methyltransferase (MGMT) promoter methylation, and 7% had IDH mutation. Regarding GTV1 categorization, 52 patients (43%) were Group A, 47 patients (39%) were Group B, and 21 patients (18%) were Group C. No significant differences in baseline characteristics were identified between Groups (Table 1).
TABLE 1.
Baseline Characteristics
| Characteristic | All (n = 122) | Group A (n=52) | Group B (n=49) | P (A vs B) | Group C (n=21) | P (A vs C) |
|---|---|---|---|---|---|---|
| Age - year | 0.93 | 0.91 | ||||
| Median (range) | 59 (21–79) | 59 (23–77) | 59 (27–79) | 62 (21–69) | ||
| Sex - n (%) | 0.14 | 0.25 | ||||
| Male | 60 (50) | 30 (58) | 21 (45) | 9 (43) | ||
| Female | 60 (50) | 22 (42) | 26 (55) | 12 (57) | ||
| KPS - n (%) | 0.69 | 0.81 | ||||
| Median (range) | 80 (50–100) | 80 (50–100) | 80 (50–100) | 80 (50–100) | ||
| KPS, n (%) | 0.12 | 0.54 | ||||
| ≥70 | 106 (88) | 43 (83) | 44 (94) | 19 (90) | ||
| <70 | 14 (12) | 9 (17) | 3 (6) | 2 (10) | ||
| Race, n (%) | 0.62 | 0.53 | ||||
| White | 117 (96) | 49 (94) | 45 (96) | 21 (100) | ||
| Other | 5 (4) | 3 (6) | 2 (4) | 0 | ||
| Extent of Resection, n (%) | 0.48 | 0.07 | ||||
| Biopsy | 23 (19) | 6 (12) | 8 (17) | 7 (33) | ||
| STR | 55 (45) | 25 (48) | 21 (45) | 9 (43) | ||
| GTR | 44 (36) | 21 (40) | 18 (38) | 5 (24) | ||
| GBM type, (n (%) | 0.67 | 0.57 | ||||
| Histologic GBM | 116 (95) | 50 (96) | 44 (94) | 19 (91) | ||
| Molecular GBM | 6 (5) | 2 (4) | 3 (6) | 2 (10) | ||
| MGMT methylation, n (%) | 0.62 | 0.07 | ||||
| Yes | 49 (40) | 19 (37) | 19 (42) | 9 (43) | ||
| No | 67 (55) | 32 (62) | 26 (58) | 9 (43) | ||
| Not Performed | 6 (5) | 1 (2) | 2 (4) | 3 (14) | ||
| IDH Mutation, n (%) | 0.55 | 0.21 | ||||
| Yes | 8 (7) | 2 (4) | 3 (6) | 3 (14) | ||
| No | 112 (92) | 49 (94) | 44 (94) | 17 (81) | ||
| Not Performed | 2 (2) | 1 (2) | 0 | 1 (5) | ||
PFS = progression free survival, OS = overall survival, KPS = Karnofsky performance status, STR = subtotal resection, GTR = gross total resection, GBM = glioblastoma, MGMT = O6-methylguanine-DNA-methyltransferase, IDH = isocitrate dehydrogenase
In addition to concurrent TMZ, some patients received experimental agents (or placebo): disulfram (n=12, 10%), nivolumab/placebo (n=4, 3%), and veliparib/placebo (n=3, 2%). The majority (92.5%) of patients received adjuvant TMZ.
Median time between initiation of RT and MRI2 was 16 days (range, 9–33 days). Overall, 37 patients (31%) had GTV2 larger than GTV1 on MRI2, and 29 patients underwent replanning based on the MRI2. Eight patients with GTV2 larger than GTV1 did not undergo replanning due to only minimal enlargement on the interim MRI, and therefore their GTV2s were still covered by the boost CTV: 5 patients covered by GTV1+5 mm and 3 patients covered by GTV1+7 mm. When comparing GTV2 with the hypothetical CTV expansions of GTV1, 84% would be covered by GTV1+5 mm, 93% covered by GTV1+7 mm, 95% covered by GTV1+10 mm, and 98% covered by GTV1+15 mm (Table 2). Two patients with GTV2 beyond the 15 mm margin had multi-focal progression during RT distant from the initial field. Table 2 displays approximately 95% of Group A and C would have adequate coverage of GTV2 by GTV1+7 mm, while only 89% of Group B would be covered by GTV1+7 mm. In contrast, approximately 90% of Group A and C would be covered by GTV1+5 mm, while only 77% of Group B would be covered by GTV1+5 mm. Similar patterns were observed for the subset of patients with and without MGMT methylation (supplemental Table S1 and S2).
TABLE 2.
Cumulative Coverage of GTV2 by GTV1 Margins Based on Interim MRI.
| GTV1 + 5 mm | GTV1 + 7 mm | GTV1 + 10 mm | GTV1 + 15 mm | Total | |
|---|---|---|---|---|---|
| Group A* - n (%) | 46 (88) | 49 (94) | 50 (96) | 50 (96) | 52 |
| Group B* - n (%) | 36 (77) | 42 (89) | 43 (91) | 47 (100) | 47 |
| Group C* - n (%) | 19 (91) | 20 (95) | 21 (100) | 21 (100) | 21 |
| Overall - n (%) | 101 (84) | 111 (93) | 114 (95) | 118 (98) | 120 |
GTV1 = gross tumor volume defined as T1 enhancement plus surgical cavity on the MRI1 without routine inclusion of T2 abnormality (unless tumor did not enhance well).
GTV2 = gross tumor volume retrospectively contoured on MRI2 and compared to GTV1
GTV1s were categorized into 3 groups based on MRI1 imaging characteristics and contouring patterns:
Group A: identical T1-enhancing volume (including surgical cavity) + T2/FLAIR abnormality for GTV1
Group B: only T1 used for contouring GTV1 with the larger T2/FLAIR abnormality omitted
Group C: T2/FLAIR abnormality used for contouring GTV1 for patients with non-enhancing tumors
Overall, 74 patients (62%) developed progression, including 69 patients with local recurrence and 7 with distant failure (including 5 patients with both local and distant recurrences). Sixty-six patients with local recurrences were included in the pattern of failure analysis. Three patients lacked available MRI images for analysis and were not evaluable. As seen in Table 3, the vast majority of patients (97%) had only central or in-field recurrences, while only 3% had a marginal recurrence but in the presence of concomitant central recurrence. For GTV1 Groups A, B, and C, the central/in-field recurrences were similar: 97%, 100%, and 92%, respectively. Notably, Group B, whose GTV didn’t include the larger T2/FLAIR abnormality, didn’t demonstrate different pattern of failure compared to Groups A and C (Table 3).
TABLE 3.
Patterns of Local Failure
| Central1 only | In-field2 only | Marginal3 only | Central + In-field | Central + Marginal | Central + In-field + Marginal | Total | |
|---|---|---|---|---|---|---|---|
| Group A* - n (%) | 20 (67) | 1 (3) | 0 | 8 (27) | 1 (3) | 0 | 30 |
| Group B* - n (%) | 19 (83) | 0 | 0 | 4 (17) | 0 | 0 | 23 |
| Group C* - n (%) | 10 (77) | 0 | 0 | 2 (15) | 0 | 1 (8) | 13 |
| Overall - n (%) | 50 (74) | 1 (2) | 0 | 14 (21) | 1 (2) | 1 (2) | 66 |
Central recurrence: progressing lesion(s) located within 57 Gy isodose line [95% of 60 Gy]
In-field recurrence: progressing lesion(s) located outside 57Gy isodose line but within 44 Gy isodose line [95% of 46Gy]
Marginal recurrence: progression lesion(s) located outside 44Gy isodose line but within 12 Gy isodose line [20% of 60 Gy])
GTV1s were categorized into 3 groups based on MRI1 imaging characteristics and contouring patterns:
Group A: identical T1-enhancing volume (including surgical cavity) + T2/FLAIR abnormality for GTV1
Group B: only T1 used for contouring GTV1 with the larger T2/FLAIR abnormality omitted
Group C: T2/FLAIR abnormality used for contouring GTV1 for patients with non-enhancing tumors
Overall, median PFS was 10 months (95% CI 7.6–12.4), and median OS was 15.2 months (95% CI 12.5–17.9). LC and PFS were not significantly different between the 3 GTV1 groups (all p > 0.05; Figure 2A and 2B). Group B had a non-significantly lower OS than Group A (44% vs 68% at 1 year, respectively, p=0.07).
FIGURE 2.
Kaplan Meier curves of local control (A), progression free survival (B), and overall survival (C) comparing GTV1 Groups A, B, and C.
On MVA, incomplete resection (HR 1.88, 95% CI 1.14–3.09, p= 0.01), and unmethylated MGMT (HR 2.29, 95% CI 1.37–3.83, p= 0.002) were significantly associated with poor PFS, while only incomplete resection (HR 2.31, 95% CI 1.42–3.76, p= 0.001) was associated with poor OS. The omission of peritumoral T2 abnormality from GTV as in Group B was not associated with worse PFS or OS (Table 4).
TABLE 4.
Univariable and Multivariable Analyses for Progression-free Survival (PFS) and Overall Survival (OS)
| Characteristic | PFS UVA HR (95% CI) | P | PFS MVA HR (95% CI) | P | OS UVA HR (95% CI) | P | OS MVA HR (95% CI) | P |
|---|---|---|---|---|---|---|---|---|
| Age - year | 1.05 (1.03–1.07) | 0.79 | - | - | 1.03 (0.99–1.05) | 0.14 | NS | NS |
| Sex | ||||||||
| Male | Reference | - | - | Reference | - | - | ||
| Female | 0.88 (0.57–1.38) | 0.57 | 1.01 (0.67–1.57) | 0.95 | ||||
| KPS Score | 0.99 (0.98–1.01) | 0.27 | - | - | 0.98 (0.97–0.99) | 0.01 | NS | NS |
| Race | ||||||||
| White | Reference | - | - | Reference | - | - | ||
| Other | 0.80 (0.11–5.78) | 0.83 | 1.56 (0.22–11.26) | 0.66 | ||||
| Extent of Resection | ||||||||
| GTR | Reference | Reference | Reference | Reference | ||||
| Biopsy/STR | 1.73 (1.07–2.79) | 0.03 | 1.88 (1.14–3.09) | 0.01 | 2.22 (1.38–3.58) | 0.001 | 2.31 (1.42–3.76) | 0.001 |
| MGMT methylation | ||||||||
| Yes | Reference | Reference | Reference | NS | NS | |||
| No | 2.38 (1.43–3.97) | 0.001 | 2.29 (1.37–3.83) | 0.002 | 1.50 (0.94–2.39) | 0.09 | ||
| IDH Mutation | ||||||||
| Yes | Reference | - | - | Reference | NS | NS | ||
| No | 1.98 (0.62–6.30) | 0.21 | 3.20 (0.79–13.02) | 0.11 | ||||
| GTV1 Group | ||||||||
| Group A/C | Reference | - | - | Reference | NS | NS | ||
| Group B | 1.20 (0.72–1.99) | 0.38 | 1.59 (1.03–2.47) | 0.04 | ||||
KPS = Karnofsky performance status, STR = subtotal resection, GTR = gross total resection, GBM = glioblastoma, MGMT = O6-methylguanine-DNA-methyltransferase, IDH = isocitrate dehydrogenase
DISCUSSION
Using a cohort of newly diagnosed GBM patients treated with standardized LFRT with paired MRI scans before and during RT, this study demonstrates that an initial CTV margin of 15 mm followed by a boost CTV margin of 7 mm may be a reasonable approach to reduce treatment volumes for most GBM cases. For the classic GBM cases where the tumor is avidly enhancing on the T1-weighted sequence and is surrounded by a larger T2/FLAIR abnormality (Group B), omission of T2 abnormality from GTV delineation doesn’t appear to jeopardize tumor control.
However, a significant proportion of the Group B patients may develop larger enhancing abnormality beyond the 5–7 mm CTV expansion approximately two weeks after starting RT, which may represent either pseudoprogression or active disease. This study also observed that almost 20% of GBM (Group C) might not enhance on the T1-weighted sequence, so GTV delineation would typically rely on T2/FLAIR abnormality. Advanced imaging techniques able to more accurately delineate high-risk tumor regions would significantly aid in treatment planning of these cases and warrant future investigation.
The reduced CTV margins used in our LFRT approach yielded comparable PFS and OS rates as historical controls [1,13,21–23]. Given only 2% of cases developed progression beyond the 15 mm CTV margin on interim MRI, the initial 15 mm margin appears reasonable. As for the boost CTV margin, the 5 mm expansion appears to only cover 84% of the tumors on the interim MRIs, suggesting this boost margin may not be sufficient in the absence of an interim MRI evaluation. In contrast, a boost CTV margin of 7 mm appears to cover 93% of the tumors on the interim MRI, including approximately 95% of Group A and C cases. The 7 mm boost CTV expansions may only cover approximately 89% of Group B whose GTVs didn’t intentionally include peritumoral T2/FLAIR abnormality, so a more generous boost margin may be considered depending on other clinical factors, such as target volume, surrounding critical structures, and availability of interim MRI. However, for a sequential boost technique using photon therapy, an additional 14 Gy boost to a smaller PTV6000 (with a 7 mm CTV expansion from the GTV) would typically ensure that the initial PTV4600 (with a 15 mm CTV expansion) would receive at least 54 Gy due to the entrance and exit dose of photon beams, which could have compensated for the minority of patients with potentially progressing lesions. Thus, reduced CTV margins may become more relevant for proton therapy given the lack of exit dose due to the Bragg peak’s sharp dose gradient [24]. Notably, the ongoing NRG BN001, a randomized phase II study comparing standard-dose photon versus dose-escalated proton therapy for newly diagnosed GBM, uses a 5 mm CTV expansion of the T1 enhancement to generate the boost PTV (NCT02179086).
The omission of T2/FLAIR abnormalities from the delineation of GTV didn’t alter patterns of failure nor worsen clinical outcomes, as seen when comparing Group B to Groups A and C. Although Group B demonstrated non-significantly worse OS, a similar trend was not observed for LC nor PFS (Figure 2A–C). Group B was not associated with a significantly worse OS on MVA after adjusting for other confounding factors (Table 4), suggesting its association with worse OS on UVA may be due to underlying confounders or biological differences. These results are consistent with several other studies intentionally omitting T2/FLAIR abnormality from their treatment volumes [1,11,25]. Notably, Kumar et al. performed a small prospective phase II study randomizing patients to the RTOG approach of 2 cm expansion on T2 abnormality for the 46 Gy volume and 2.5 cm on T1 enhancement for the 14 Gy boost versus the MD Anderson approach of 2.5 cm expansion on the GTV without intentional inclusion of T2 abnormalities, and this under-powered study showed improved OS, PFS, and quality of life with the MD Anderson LFRT approach [26]. However, given that more Group B patients in our study had larger enhancing abnormality beyond the 5–7 mm boost margins on the interim MRI than Groups A and C, the possibility remained that at least part of the initial T2/FLAIR abnormality in Group B harbored gross disease. Prior studies with amino acid PET have shown that glioma boundaries are under-estimated by contrast-enhancing MRI. Radiolabeled amino acids, which are actively transported by system L in glioma cells, can more accurately delineate contrast-enhancing and non-enhancing tumors [27–29]. Other advances in MRI techniques such as whole-brain spectroscopic MRI or diffusion basis spectrum imaging have also demonstrated promising potential to more accurately predict regions of high tumor cellularity for GBM [30,31]. Incorporation of these imaging techniques may help to improve GTV delineation for LFRT in the future. In the meantime, when LFRT is employed without using T2/FLAIR abnormality to delineate GTV, visual inspection to ensure its inclusion by the initial CTV expansion may be reasonable as per the European approach [8].
Our study showed that the pattern of failure for GBM patients treated with LFRT is predominantly central or in-field, which is consistent with other patterns of failure studies [10,12,13,23,32]. Studies that used CTV margins as small as 5 mm have also reported central or in-field recurrences as the predominant mode of progression with relatively rare marginal relapses. Given that the standard 60 Gy with concurrent TMZ is inadequate to control the gross tumor and prevent central recurrences, elective treatment of large fields of relatively low-risk regions may not be high yield and may subject patients to unnecessary treatment-related toxicities and poorer quality of life. For example, previous studies have shown larger RT volumes correlating with the magnitude of lymphopenia, negatively impacting survival [9,16,33]. LFRT may also maximize the tolerability of novel systemic agents when administered concurrently with RT for GBM.
Given this is a single-institutional retrospective study, its findings should be considered hypothesis generating and should be validated with prospective data. The sample size was not prospectively designed and may be insufficient to detect small differences. The absence of a control group further limits definitive conclusions regarding the accuracy and efficacy of the proposed delineation strategies for different subsets of GBM (ie Group A-C). Although our previous propensity-matched comparison of the LFRT approach with historical standard-field RT at the same institution has demonstrated comparable PFS and OS [9], larger prospective studies are warranted to confirm that such approach does not reduce disease control and survival. Another limitation was the ambiguity in differentiating local recurrence from pseudoprogression or RT necrosis. However, local recurrence was determined by a multi-disciplinary neurooncology team with a thorough review of clinical, histopathologic, and radiographic evidence. Furthermore, early radiographic changes on interim MRI may also represent pseudoprogression rather than true disease progression. Future studies should examine the incorporation of multiparametric advanced neuro-imaging to further refine RT planning for GBM and to improve the therapeutic ratio [34].
Supplementary Material
Highlights.
Optimal limited-field radiation therapy approaches for GBM are not well defined
Early radiological changes during RT provide insights on adequacy of limited-field
Strategies to reduced CTV margins appears not to jeopardize outcomes
Omission of T2 abnormality in selected cases is a reasonable approach
Funding:
Supported by Clinical and Translational Science Award (CTSA) Grant [UL1 TR000448] and Siteman Comprehensive Cancer Center and NCI Cancer Center Support Grant [P30 CA091842] for author JC
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- [1].Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJB, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 2005;352:987–96. 10.1056/NEJMoa043330. [DOI] [PubMed] [Google Scholar]
- [2].Hochberg FH, Pruitt A. Assumptions in the radiotherapy of glioblastoma. Neurology 1980;30:907. 10.1212/WNL.30.9.907. [DOI] [PubMed] [Google Scholar]
- [3].Kelly PJ, Daumas-Duport C, Kispert DB, Kall BA, Scheithauer BW, Illig JJ. Imaging-based stereotaxic serial biopsies in untreated intracranial glial neoplasms. J Neurosurg 1987;66:865–74. 10.3171/jns.1987.66.6.0865. [DOI] [PubMed] [Google Scholar]
- [4].Halperin EC, Bentel G, Heinz ER, Burger PC. Radiation therapy treatment planning in supratentorial glioblastoma multiforme: An analysis based on post mortem topographic anatomy with ct correlations. Int J Radiat Oncol 1989;17:1347–50. 10.1016/0360-3016(89)90548-8. [DOI] [PubMed] [Google Scholar]
- [5].Burger PC, Dubois PJ, Schold SC, Smith KR, Odom GL, Crafts DC, et al. Computerized tomographic and pathologic studies of the untreated, quiescent, and recurrent glioblastoma multiforme. J Neurosurg 1983;58:159–69. 10.3171/jns.1983.58.2.0159. [DOI] [PubMed] [Google Scholar]
- [6].Cabrera AR, Kirkpatrick JP, Fiveash JB, Shih HA, Koay EJ, Lutz S, et al. Radiation therapy for glioblastoma: Executive summary of an American Society for Radiation Oncology Evidence-Based Clinical Practice Guideline. Pract Radiat Oncol 2016;6:217–25. 10.1016/j.prro.2016.03.007. [DOI] [PubMed] [Google Scholar]
- [7].Kruser TJ, Bosch WR, Badiyan SN, Bovi JA, Ghia AJ, Kim MM, et al. NRG brain tumor specialists consensus guidelines for glioblastoma contouring. J Neurooncol 2019;143:157–66. 10.1007/s11060-019-03152-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Niyazi M, Brada M, Chalmers AJ, Combs SE, Erridge SC, Fiorentino A, et al. ESTRO-ACROP guideline “target delineation of glioblastomas.” Radiother Oncol J Eur Soc Ther Radiol Oncol 2016;118:35–42. 10.1016/j.radonc.2015.12.003. [DOI] [PubMed] [Google Scholar]
- [9].Rudra S, Hui C, Rao YJ, Samson P, Lin AJ, Chang X, et al. Effect of Radiation Treatment Volume Reduction on Lymphopenia in Patients Receiving Chemoradiotherapy for Glioblastoma. Int J Radiat Oncol 2018;101:217–25. 10.1016/j.ijrobp.2018.01.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Paulsson AK, McMullen KP, Peiffer AM, Hinson WH, Kearns WT, Johnson AJ, et al. Limited Margins Using Modern Radiotherapy Techniques Does Not Increase Marginal Failure Rate of Glioblastoma. Am J Clin Oncol 2014;37:177–81. 10.1097/COC.0b013e318271ae03. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Chang EL, Akyurek S, Avalos T, Rebueno N, Spicer C, Garcia J, et al. Evaluation of Peritumoral Edema in the Delineation of Radiotherapy Clinical Target Volumes for Glioblastoma. Int J Radiat Oncol 2007;68:144–50. 10.1016/j.ijrobp.2006.12.009. [DOI] [PubMed] [Google Scholar]
- [12].Gebhardt BJ, Dobelbower MC, Ennis WH, Bag AK, Markert JM, Fiveash JB. Patterns of failure for glioblastoma multiforme following limited-margin radiation and concurrent temozolomide. Radiat Oncol Lond Engl 2014;9:130. 10.1186/1748-717X-9-130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].McDonald MW, Shu H-KG, Curran WJ, Crocker IR. Pattern of Failure After Limited Margin Radiotherapy and Temozolomide for Glioblastoma. Int J Radiat Oncol 2011;79:130–6. 10.1016/j.ijrobp.2009.10.048. [DOI] [PubMed] [Google Scholar]
- [14].Scheibel RS, Meyers CA, Levin VA. Cognitive dysfunction following surgery for intracerebral glioma: influence of histopathology, lesion location, and treatment. J Neurooncol 1996;30:61–9. 10.1007/BF00177444. [DOI] [PubMed] [Google Scholar]
- [15].Corn BW, Wang M, Fox S, Michalski J, Purdy J, Simpson J, et al. Health related quality of life and cognitive status in patients with glioblastoma multiforme receiving escalating doses of conformal three dimensional radiation on RTOG 98–03. J Neurooncol 2009;95:247–57. 10.1007/s11060-009-9923-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Yovino S, Grossman SA. Severity, etiology and possible consequences of treatment-related lymphopenia in patients with newly diagnosed high-grade gliomas. CNS Oncol 2012;1:149–54. 10.2217/cns.12.14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Grossman SA, Ellsworth S, Campian J, Wild AT, Herman JM, Laheru D, et al. Survival in Patients With Severe Lymphopenia Following Treatment With Radiation and Chemotherapy for Newly Diagnosed Solid Tumors. J Natl Compr Cancer Netw JNCCN 2015;13:1225–31. 10.6004/jnccn.2015.0151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Manon R, Hui S, Chinnaiyan P, Suh J, Chang E, Timmerman R, et al. The Impact of Mid-Treatment MRI on Defining Boost Volumes in the Radiation Treatment of Glioblastoma Multiforme. Technol Cancer Res Treat 2004;3:303–7. 10.1177/153303460400300308. [DOI] [PubMed] [Google Scholar]
- [19].Brat DJ, Aldape K, Colman H, Holland EC, Louis DN, Jenkins RB, et al. cIMPACT-NOW Update 3: Recommended diagnostic criteria for “Diffuse astrocytic glioma, IDH-wildtype, with molecular features of glioblastoma, WHO grade IV.” Acta Neuropathol (Berl) 2018;136:805–10. 10.1007/s00401-018-1913-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Hassanzadeh C, Rao YJ, Chundury A, Rowe J, Ponisio MR, Sharma A, et al. Multiparametric MRI and [18F]Fluorodeoxyglucose Positron Emission Tomography Imaging Is a Potential Prognostic Imaging Biomarker in Recurrent Glioblastoma. Front Oncol 2017;7. 10.3389/fonc.2017.00178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Stupp R, Taillibert S, Kanner AA, Kesari S, Steinberg DM, Toms SA, et al. Maintenance Therapy With Tumor-Treating Fields Plus Temozolomide vs Temozolomide Alone for Glioblastoma: A Randomized Clinical Trial. JAMA 2015;314:2535–43. 10.1001/jama.2015.16669. [DOI] [PubMed] [Google Scholar]
- [22].Gorlia T, Bent MJ van den, Hegi ME, Mirimanoff RO, Weller M, Cairncross JG, et al. Nomograms for predicting survival of patients with newly diagnosed glioblastoma: prognostic factor analysis of EORTC and NCIC trial 26981–22981/CE.3. Lancet Oncol 2008;9:29–38. 10.1016/S1470-2045(07)70384-4. [DOI] [PubMed] [Google Scholar]
- [23].Minniti G, Amelio D, Amichetti M, Salvati M, Muni R, Bozzao A, et al. Patterns of failure and comparison of different target volume delineations in patients with glioblastoma treated with conformal radiotherapy plus concomitant and adjuvant temozolomide. Radiother Oncol J Eur Soc Ther Radiol Oncol 2010;97:377–81. 10.1016/j.radonc.2010.08.020. [DOI] [PubMed] [Google Scholar]
- [24].Lomax AJ. Charged particle therapy: the physics of interaction. Cancer J Sudbury Mass 2009;15:285–91. 10.1097/PPO.0b013e3181af5cc7. [DOI] [PubMed] [Google Scholar]
- [25].Mason WP, Maestro RD, Eisenstat D, Forsyth P, Fulton D, Laperrière N, et al. Canadian recommendations for the treatment of glioblastoma multiforme. Curr Oncol Tor Ont 2007;14:110–7. 10.3747/co.2007.119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Kumar N, Kumar R, Sharma SC, Mukherjee A, Khandelwal N, Tripathi M, et al. Impact of volume of irradiation on survival and quality of life in glioblastoma: a prospective, phase 2, randomized comparison of RTOG and MDACC protocols. Neuro-Oncol Pract 2020;7:86–93. 10.1093/nop/npz024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Kracht LW, Miletic H, Busch S, Jacobs AH, Voges J, Hoevels M, et al. Delineation of Brain Tumor Extent with [11C]l-Methionine Positron Emission Tomography: Local Comparison with Stereotactic Histopathology. Clin Cancer Res 2004;10:7163–70. 10.1158/1078-0432.CCR-04-0262. [DOI] [PubMed] [Google Scholar]
- [28].Pauleit D, Floeth F, Hamacher K, Riemenschneider MJ, Reifenberger G, Müller H-W, et al. O-(2-[18F]fluoroethyl)-L-tyrosine PET combined with MRI improves the diagnostic assessment of cerebral gliomas. Brain J Neurol 2005;128:678–87. 10.1093/brain/awh399. [DOI] [PubMed] [Google Scholar]
- [29].Pafundi DH, Laack NN, Youland RS, Parney IF, Lowe VJ, Giannini C, et al. Biopsy validation of 18F-DOPA PET and biodistribution in gliomas for neurosurgical planning and radiotherapy target delineation: results of a prospective pilot study. Neuro-Oncol 2013;15:1058–67. 10.1093/neuonc/not002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Cordova JS, Shu H-KG, Liang Z, Gurbani SS, Cooper LAD, Holder CA, et al. Whole-brain spectroscopic MRI biomarkers identify infiltrating margins in glioblastoma patients. Neuro-Oncol 2016;18:1180–9. 10.1093/neuonc/now036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Ye Z, Price RL, Liu X, Lin J, Yang Q, Sun P, et al. Diffusion Histology Imaging Combining Diffusion Basis Spectrum Imaging (DBSI) and Machine Learning Improves Detection and Classification of Glioblastoma Pathology. Clin Cancer Res Off J Am Assoc Cancer Res 2020. 10.1158/1078-0432.CCR-20-0736. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Sherriff J, Tamangani J, Senthil L, Cruickshank G, Spooner D, Jones B, et al. Patterns of relapse in glioblastoma multiforme following concomitant chemoradiotherapy with temozolomide. Br J Radiol 2013;86:20120414. 10.1259/bjr.20120414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Huang J, DeWees TA, Badiyan SN, Speirs CK, Mullen DF, Fergus S, et al. Clinical and Dosimetric Predictors of Acute Severe Lymphopenia During Radiation Therapy and Concurrent Temozolomide for High-Grade Glioma. Int J Radiat Oncol 2015;92:1000–7. 10.1016/j.ijrobp.2015.04.005. [DOI] [PubMed] [Google Scholar]
- [34].Whitfield GA, Kennedy SR, Djoukhadar IK, Jackson A. Imaging and Target Volume Delineation in Glioma. Clin Oncol 2014;26:364–76. 10.1016/j.clon.2014.04.026. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.