Patients’ characteristics
A total of 184 patients were enrolled, 56% were followed for lung cancer, 13% for breast cancer, 13% for melanoma, 8.7% for digestive cancer, 4.3% for kidney cancer and 4.9% for other cancers. Table 1 shows the patients and primitive tumor characteristics. One hundred and twenty-two patients (66.3%) received two sessions of SRT, 40 (21.7%) received three sessions, 14 (7.6%) received four sessions, 7 (3.8%) received five sessions, and one (0.5%) received six sessions, totaling 461 administered treatments and 915 treated BM. Twenty-six percent of BMs were irradiated with a monofractionated, 72% of BMs were irradiated with a trifractionated regimen, and the last 2% of BMs were irradiated with other hypofractionated schedules. A mean of two BM was treated per SRT (range: 1–6; 95%CI 1.88–2.12), for a total average of 5 BMs treated during all sessions (range: 2–19; 95%CI 4.52–5.44). All characteristics are shown in Table 2.
Table 1
Patients and primitive tumor characteristics (n = 184)
Characteristics | | Number | Percentage |
Sex | Male | 91 | 49.5% |
| Female | 93 | 50.5% |
Age at diagnosis | | | |
of cancer | Median (range) | 58 (21–87) | |
| ≤ 65 yo | 135 | 73% |
| > 65 yo | 49 | 27% |
of brain metastasis | Median (range) | 61 (24–88) | |
| ≤ 65 yo | 124 | 67% |
| > 65 yo | 60 | 33% |
Primitive cancer | Lung | 103 | 56.0% |
| Adenocarcinoma | 69 | 67% |
| Epidermoid | 19 | 18.4% |
| Small cell | 4 | 3.9% |
| Undifferentiated | 4 | 3.9% |
| Other | 7 | 6.8% |
| Breast | 24 | 13.0% |
| Luminal A | 5 | 20.8% |
| Luminal B | 5 | 20.8% |
| Her2 + | 11 | 45.8% |
| Triple negative | 3 | 12.5% |
| Melanoma | 24 | 13.0% |
| Kidney | 8 | 4.3% |
| Digestive | 16 | 8.7% |
| Other | 9 | 4.9% |
Initial tumor stage | 1 | 23 | 12.5% |
| 2 | 47 | 25.5% |
| 3 | 38 | 20.7% |
| 4 | 30 | 16.3% |
| Unknown | 46 | 25.0% |
Initial node stage | 0 | 37 | 20.1% |
| 1 | 23 | 12.5% |
| 2 | 49 | 26.6% |
| 3 | 26 | 14.1% |
| Unknown | 49 | 26.6% |
Initial metastasis stage | 0 | 81 | 44.0% |
| 1 | 90 | 48.9% |
| Unknown | 13 | 7.1% |
Table 2
Patient characteristics at each session of SRT and before-after comparison between each consecutive session
| | SRT 1 (n = 184) | | SRT 2 (n = 184) | | p | SRT 3 (n = 61) | | p | SRT ≥ 4 (n = 31) | | p |
Age | Median | 61 | | 61.5 | | < 0.001 | 60.6 | | < 0.001 | 61.9 | | < 0.001 |
| Range | 24–88 | | 25–89 | | | 32–79 | | | 47–78 | | |
| 95%CI | 43–77 | | 43.5–77.4 | | | 41-77.8 | | | / | | |
WHO | Median | 1 | | 1 | | 0.39 | 1 | | 0.07 | 1 | | 0.64 |
| 0–1 | 151 | 82% | 149 | 81% | | 49 | 80% | | 25 | 81% | |
| 2–3 | 33 | 18% | 35 | 19% | | 12 | 20% | | 6 | 19% | |
KPS | Median | 90% | | 80% | | 0.02 | 90% | | 0.16 | 90 | | 0.07 |
| < 70% | 8 | 4% | 10 | 5% | | 4 | 7% | | 3 | 10% | |
| ≥ 70 | 127 | 69% | 129 | 70% | | 41 | 68% | | 17 | 55% | |
| Unknown | 49 | 27% | 45 | 24% | | 16 | 26% | | 11 | 35% | |
DS-GPA | Median | 2.5 | | 2.5 | | 0.52 | 2.5 | | 0.89 | 3 | | 0.5 |
| 0–1 | 15 | 8% | 21 | 11% | | 8 | 13% | | 2 | 6% | |
| 1.5-2 | 47 | 26% | 44 | 24% | | 10 | 16% | | 3 | 10% | |
| 2.5-3 | 53 | 29% | 54 | 29% | | 17 | 28% | | 11 | 35% | |
| 3.5-4 | 20 | 11% | 20 | 11% | | 10 | 16% | | 4 | 13% | |
| Unknown | 49 | 27% | 45 | 24% | | 16 | 26% | | 11 | 35% | |
RPA | Median | 2 | | 2 | | 0.37 | 2 | | 0.09 | 2 | | 1 |
| 1 | 19 | 10% | 29 | 16% | | 9 | 15% | | 5 | 16% | |
| 2 | 108 | 59% | 100 | 54% | | 32 | 52% | | 12 | 39% | |
| 3 | 8 | 4% | 10 | 5% | | 4 | 6% | | 3 | 10% | |
| Unknown | 49 | 27% | 45 | 24% | | 16 | 26% | | 11 | 35% | |
ECP | Yes | 117 | 64% | 55 | 30% | < 0.001 | 14 | 23% | 1 | 7 | 23% | 1 |
No | 67 | 36% | 129 | 70% | | 47 | 77% | | 24 | 77% | |
Systemic treatment | Yes | 129 | 70% | 126 | 68% | 0.88 | 47 | 77% | 0.37 | 19 | 61% | 0.48 |
No | 55 | 30% | 57 | 31% | | 14 | 23% | | 12 | 39% | |
| Chemotherapy | 76 | 59% | 63 | 50% | | 20 | 43% | | 8 | 42% | |
| Immunotherapy | 14 | 11% | 19 | 15% | | 7 | 15% | | 1 | 5% | |
| Hormonotherapy | 2 | 2% | 1 | 1% | | 0 | 0% | | 0 | 0% | |
| Targeted therapy | 30 | 23% | 28 | 22% | | 15 | 32% | | 9 | 47% | |
| Combined therapy | 3 | 2% | 9 | 7% | | 4 | 9% | | 1 | 5% | |
| Unknown | 4 | 3% | 6 | 5% | | 1 | 2% | | 0 | 0% | |
Number of BM per patient | Median | 1.5 | | 1 | | 0.06 | 1 | | 0.48 | 1 | | 0.8 |
Mean | 2.18 | | 1.9 | | | 1.79 | | | 1.84 | | |
Range | 1–10 | | 1–9 | | | 1–8 | | | 1–8 | | |
| 95%CI | 1.0–6.0 | | 1.0–5.0 | | | 1.0–4.0 | | | | | |
GTV of all BM (mL) | Median | 6.2 | | 0.9 | | < 0.01 | 0.6 | | 0.34 | 1 | | 0.62 |
Mean | 8.5 | | 5.0 | | | 3.1 | | | 5.1 | | |
| Range | 0.1–48.6 | | 0.1–59.7 | | | 0.1–32.2 | | | 0.1–36.5 | | |
| 95%CI | 7.01–9.9 | | 3.3–6.8 | | | 1.3–4.8 | | | 0.1–10.2 | | |
Number of tumor bed metastases per patient | 0 | 114 | 62% | 166 | 90% | < 0.001 | 57 | 93% | 1 | 29 | 94% | 1 |
1 | 66 | 36% | 18 | 10% | | 3 | 5% | | 2 | 6% | |
2 | 4 | 2% | 0 | 0% | | 1 | 2% | | 0 | 0% | |
Number of recurrent BM | 0 | / | | 157 | 85% | / | 51 | 84% | 1 | 23 | 74% | 0.45 |
1 | / | | 26 | 14% | | 9 | 15% | | 8 | 26% | |
2 and more | / | | 1 | 1% | | 1 | 2% | | 0 | 0% | |
Overall condition
The median WHO was 1 ± 0.79 at SRT1, 1 ± 0.75 at SRT2, 1 ± 0.68 at STR3 and 1 ± 0.65 at SRT4 and more (SRT4-5-6), with p = 0.39, p = 0.07 and p = 0.64, respectively, before and after each consecutive session of SRT. The median KPS was 90% ± 16% at SRT1, 80% ± 14% at SRT2, 90% ± 13% at STR3 and 90% ± 16% at SRT4-5-6, with p = 0.02, p = 0.16 and p = 0.07, respectively, between each consecutive session. The KPS at SRT1 was significantly different from the KPS at SRT3 (p = 0.02) but not from the KPS at SRT4 (p = 0.32). The median Ds-GPA was 2.5 ± 0.92 at SRT1, 2.5 ± 1.03 at SRT2, 2.5 ± 1.09 at STR3 and 3 ± 0.65 at SRT4-5-6, with p = 0.52, p = 0.89 and p = 0.5, respectively, between each consecutive session. The median RPA was 2 ± 0.44 at SRT1, 2 ± 0.51 at SRT2, 2 ± 0.53 at STR3 and 2 ± 0.41 at SRT4-5-6, with p < 0.37, p = 0.09 and p = 1, respectively, between each consecutive session of SRT.
Oncological status
Sixty-four percent of patients were diagnosed with ECP at SRT1, and 37.5% of patients were diagnosed with brain metastatic disease at initial diagnosis. Twenty-eight percent of patients have ECP at SRT1 during follow-up of cancer. Synchronous BM diagnosis with primitive tumors was observed in 53.4% of lung cancer cases, 37.5% of kidney cancer cases, 20.8% of melanoma cases and 18.8% of digestive cancer cases but not in any breast cancer cases. Figure 1 represents the evolution of oncological status at each SRT. There was no statistically significant difference between patients with synchronous or metachronous brain metastatic at SRT1 and ECP patients at SRT1 who received more than 3 SRT sessions (p = 0.49 and p = 0.8, respectively) or more than 4 SRT sessions (p = 0.41 and p = 0.64, respectively). There was no significant difference in ECP between SRT1 and noninitial brain metastatic patients who received more than 3 SRT sessions (p = 0.31) or more than 4 SRTs (p = 0.29). ECP was present for 64% of patients at SRT1 compared to 30% at SRT2 (p < 0.001), 23% at SRT3 (p = 1) and 23% at SRT4-5-6 (p = 1) in before-after analysis between each succeeding session. By comparing patients with no extracerebral synchronous BM with others at SRT1, there was no significant difference between ECP between SRT1 and SRT2 (p = 0.73). Patients who were extracranial metastasis (ECM)-free at cancer diagnosis and ECP-free at SRT1 and SRT2 were not significantly more likely to receive three or more SRT sessions than those with ECM at diagnosis (p = 0.58) and/or ECP at SRT1 and SRT2 (p = 0.8 and p = 0.27, respectively).
Systemic therapy
Seventy percent, 68%, 77%, and 61% of patients were on systemic therapy at SRT1, SRT2, SRT3 and SRT4-5-6, respectively. Patients who had uncontrolled primary tumors at diagnosis of brain metastases and ECM were more likely to receive systemic therapy at SRT1 (p = 0.04 and p < 0.01, respectively). Patients who had ECM at SRT2 and/or SRT3 were more likely to receive systemic therapies than patients with no ECM (p < 0.001 and < 0.001, respectively). Patients who had ECP at SRT2 and/or SRT3 were more likely to receive systemic therapies than patients who were ECP-free (p < 0.001 and p = 0.03, respectively). Systemic therapies were divided into 5 classes: chemotherapy (CT), targeted therapy (TT), immunotherapy (IT), hormonotherapy (HT), and combination therapy. In the before-after comparison between each consecutive session, there was no significant difference in the use of the different classes of treatments between SRT1 and SRT2 or between SRT2 and SRT3 (p = 0.8 and p = 0.8, respectively). Fifty-nine percent, 50%, 43% and 26% of patients received CT at SRT1, SRT2, SRT3 and SRT4 and above, respectively, and there was no significant difference between each consecutive session in the before-after comparison (p = 0.11, p = 1, and p = 0.48, respectively). The use of CT was statistically associated with progression of the primary tumor site (p < 0.001 at SRT1 and p < 0.001 at SRT2). A total of 11%, 15%, 15% and 3% of patients received IT at SRT1, SRT2, SRT3 and SRT4 and above, respectively, and there was no significant difference between each session (p = 0.45, p = 0.48 and p = 1, respectively). Twenty-three percent, 22%, 32% and 29% of patients received TT at SRT1, SRT2, SRT3 and SRT4 and above, respectively, and there was no significant difference between each session (p = 0.72, p = 0.68 and p = 1, respectively).
Patients treated with TT at SRT1 and/or SRT2 tended to be more likely to receive more than three SRT sessions than other patients (p = 0.16 and p = 0.15, respectively). TT was statistically associated with primitive tumoral control at SRT1 (p < 0.001) and SRT2 (p = 0.01) and ECP at SRT1 (p = 0.05). TT tended to be statistically associated with primitive tumoral control at SRT3 (p = 0.11). TT was not statistically associated with primitive tumoral control at SRT4 (p = 0.59) or ECP at SRT2 (p = 0.50). The following associations were not calculable due to the small number of patients and events.
Interval time between consecutive SRT
The median delay between each consecutive SRT session was 9.2 months (range 0.9–69.9; 95%CI 7.9–10.6) between SRT1 and SRT2, 5.3 months (range 1.1–61.5; 95%CI 6.1–11.5) between SRT2 and SRT3, 7.8 months (range 2.7–54.8; 95%CI 2.9–19.1) between SRT3 and SRT4 and 7.2 months (range 3.1–52.5; 95%CI 5.3–15.1) between SRT4 and the following SRT sessions. There was no significant difference in the before-after analysis between each consecutive time interval (p = 0.57, p = 0.66 and p = 0.11, respectively).
The median time interval between SRT1 and SRT2 was 8.1 months (95%CI 7.7–11.0) for patients with no systemic treatment at SRT2, compared with 7.0 months (95%CI 8.4–16.4) for patients with maintenance treatment and 5.3 months (95%CI 5.3–8.8) for patients with active treatment (p < 0.001). There was also a trend towards a statistical association for the time interval between SRT2 and SRT3 and maintenance treatment at SRT3 (p = 0.09). There was no statistical association between the time interval between consecutive SRT and ECP at SRT2, SRT3 and SRT4 or higher (p = 0.24, p = 0.33 and p = 0.65, respectively).
Cerebral status
The median number of BMs was 1.5 (95%CI 1.0–6.0) at SRT1, 1 (95%CI 1.0–5.0) at SRT2, 1 (95%CI 1.0–4.0) at SRT3 and 1 (95%CI 1.4–2.2) at SRT4-5-6. Among the 915 BMs, 96 were operated on before SRT (10%). Seventy patients (38%) underwent surgery for one or two BMs at SRT1, 18 (10%) at SRT2, 4 (7%) at SRT3 and 2 at SRT4-5-6. The number of BMs between each SRT was not significantly different between each SRT, but there was a tendency for more metastases in SRT1 (p = 0.06). The number of BMs at SRT1 tended to be associated with control of the primitive tumoral site (p = 0.14), ECM at cancer diagnosis (p = 0.1) and ECP at SRT1 (p = 0.24) but was not associated with the use of a third SRT session (p = 0.95). Breast cancer, lung cancer and melanoma provided more BMs at SRT1 (p < 0.001) and more total BMs (p = 0.05) than other primitive cancers.
The median GTVs of each BM at SRT1, SRT2, SRT3 and SRT4-5-6 were 0.4 mL (95%CI 1.84–2.94), 0.4 mL (95%CI 2.67–4.77), 0.25 mL (95%CI 1.42–4.43), and 0.35 mL (95%CI 1.16–4.16), respectively. Total GTV at SRT1 was statistically higher than GTV at SRT2 (p < 0.001), but this difference was not found between SRT2 and SRT3 (p = 0.34) or between SRT3 and SRT4-5-6 (p = 0.62). A high GTV at SRT1 was not statistically associated with synchronous brain metastases at diagnosis (p = 0.43).
Twenty-seven patients (15%) were reirradiated for one or more local relapses at SRT2, 10 (17%) at SRT3 and 8 (26%) at SRT4 and above. Among the 93 BMs with local relapse, 63.4% recurred after the first session, 26.8% after the second session, 4.3% after the third session, 4.3% after the fourth session, and 1% after the fifth session (p < 0.001). The median interval time between SRT and local relapse was 9.2 months (95%CI 10.6–18.0) for BMs treated at SRT1 and 7.5 months (95%CI 7.1–13.6) for BMs treated at SRT2 and more (p = 0.28). Fifty-one BMs were reirradiated after local recurrence, 56.8% were reirradiated at SRT2, 27.5% were reirradiated at SRT3, and 15.7% were reirradiated at SRT4 and above (p < 0.001).
BMV grade and patient outcome
At SRT2, 8.1%, 33.7% and 58.2% of patients were classified as having high, intermediate, and low BMV, respectively; at SRT3, 7.6%, 36.4% and 56%, respectively; and at the last SRT session, 6.5%, 37.5% and 56%, respectively. There was almost perfect concordance between the BMV grade calculated at the last SRT session and that calculated at SRT2 (r = 0.89; p < 0.001). Salvage WBRT was used in 18.5% of patients after a median time of 9.8 months (95%CI 8.6–15.9) after SRT1 and 3.9 months (95%CI 3.5–5.2) after the last SRT. This WBRT was used after SRT2, SRT3 and SRT4-5-6 in 24.4%, 7.7% and 4.5% of patients, respectively. Among patients with high BMV grade, 60% received 2 SRT sessions without WBRT, 20% received 3 SRT sessions or more without WBRT, and 20% ultimately received WBRT. Among patients with intermediate BMV grade, 54.8% received only 2 SRT sessions without WBRT, 16.1% received 3 SRT sessions or more without WBRT, and 29% ultimately received WBRT. Among patients with low BMV grade, 46.7% received only 2 SRT sessions without WBRT, 41.1% received 3 SRT sessions or more without WBRT, and 12.1% ultimately received WBRT (p = 0.02). The median interval time between SRT1 and WBRT was statistically longer for low-BMV patients than for other patients (p < 0.001). Patient outcomes according to BMV grade at SRT2 are summarized in Fig. 1.