Survival outcomes of patients with brain metastasis of osteosarcoma can be improved by aggressive multi-disciplinary interventions including chemotherapy

Abstract Background/Objective Brain metastasis in osteosarcoma (BMO) is rare and its clinical characteristics are often buried among studies on brain metastasis of bone and soft tissue sarcomas. The aim of the present study was to summarize the incidence, clinical characteristics, treatment and outcomes of patients with BMO. Methods This retrospective study included 7 patients with BMO who received treatment in our center between 2005 and 2019. The clinical medical records of the 7 patients, together with data of 70 BMO patients published in 33 articles and retrieved by means of PubMed and Medline, were analyzed, retrospectively. Results Data analysis of the 97 BMO patients showed a high correlation between the interval from the primary diagnosis to BMO occurrence and the interval from the primary diagnosis to prior metastases. Multivariate analysis showed that chemotherapy, radiotherapy and surgery were three main factors affecting the overall survival of BMO patients (HR = 0.427; HR = 0.372; HR = 0.296). Surgery combined with chemotherapy or radiotherapy offered a better overall survival than surgery alone. Conclusion Patients with BMO may obtain survival benefits from regular neuroimaging and early aggressive multi-disciplinary interventions including surgical resection, postoperative radiotherapy and chemotherapy. Synopsis This is a retrospective study describing the characteristics of metastasic intervals, locations, clinical features and prognosis in 97 patients with brain metastasis of osteosarcoma (BMO). Multivariate analysis showed that chemotherapy was effective as surgery and radiotherapy for the treatment of BMO. Our findings emphasize the importance of regular neuroimaging and early aggressive multi-disciplinary interventions including surgical resection, postoperative radiotherapy and chemotherapy.


Introduction
Osteosarcoma is one of the most common primary malignant bone tumors.Brain metastasis in osteosarcoma (BMO) is rare, with a reported incidence of 1.8-5.7%, 1,2and is associated with prior local recurrence and pulmonary metastasis.The incidence of BMO is increasing due to the prolongation of survival with systemic disease control but without effective intracranial control.The prognosis of these patients is generally poor. 3Even with treatment, the survival time is usually less than 6 months. 4,5ompared with other bony and soft tissue sarcomas of high heterogeneity, there are some differences in the etiology, histology, clinical features, treatment and prognostic factors in BMO. 5 For example, unlike other histological types of soft tissue and bone sarcomas, osteosarcoma is a radioresistant and chemo-sensitive disease.6][7][8] To the best of our knowledge, there is no independent study to investigate therapeutic options and associated prognosis of BMO.In this study, we reported 7 cases of BMO in our institution together with a summary of case reports and small case series of BMO published in the literature to explore the clinical characteristics, treatments and outcomes of BMO.

Materials and methods
This study was approved by the Institutional Review Board of Jinling Hospital (Nanjing, China).From December 2005 to December 2019, 376 patients with osteosarcoma were retrospectively analyzed, in which 7 (1.9%)progressed to brain metastasis, including 6 male patients and a female patient.BMO was identified by imaging and histological examinations in patients who were scheduled to receive surgery.The clinical data of the 7 BMO patients are listed in Table 1.A literature search was performed in Pubmed and Medline to identify articles by using the following keywords: 'osteosarcoma' AND 'brain' OR 'central nervous system' OR 'CNS' OR 'intracranial'.The exclusion criteria were (1) patients whose primary osteosarcoma had contiguously extended into the intracranial structures; (2) there was a lack of detailed information about the patients; and (3) patients with primary intracerebral osteosarcoma.[6] The characteristics of the retrieved patients are provided in the Supplementary Materials.Survival data were analyzed by Kaplan-Meier survival curve and Cox regression, and correlation analysis was perform by Pearson correlation analysis.p Values are derived from two-tailed tests and statistical significance was established at p < 0.05.All statistics were performed using SPSS version 17.0 and GraphPad Prism version 6.0.

Clinical characteristics and outcomes of the 7 BMO patients in our institution
Of the 376 osteosarcoma patients sought treatment in our center, 7 patients (1.9%) with a median age of 18 years (range: 9-23) developed brain metastasis within a median period of 18 (range 8-43) months after the initial diagnosis of osteosarcoma.The characteristics of the 7 patients are summarized in Table 1.Of them, 6 patients had solitary brain metastatic foci and one patient had multiple brain metastases.Six of the 7 metastases were located in the cerebrum, and the other in the cerebellum.Of the 6 metastases in the cerebellum, 5 occurred in the frontal lobe and one in the parietal lobe.The common presenting symptoms were headache and nausea/vomiting in 4 cases.Two patients presented with hemiplegia-and epilepsy-like motor symptoms.One patient exhibited ocular symptoms such as blurred vision and exophthalmos.Six cases were preceded by lung metastasis and/or local recurrence and metastases at other sites, and simple brain metastasis occurred only in one cases.The median duration from lung metastasis to brain metastasis was 7 (range 2-21) months.
Three patients underwent aggressive therapy, including surgical resection plus postoperative radiotherapy (RT) or chemotherapy (Figure 1).Two patients received RT only, and two patients refused any treatment for personal reasons.Of the 3 patients receiving surgery, 2 were also treated with chemotherapy and apatinib, a broad-spectrum vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR-TKI), which has been approved in China.Five patients who received RT also underwent gamma-knife stereotactic radiosurgery.The treatment journey of each patient is presented in Figure 2. All the 7 patients in our series succumbed to metastatic disease.Of them, the 2 patients who declined treatment after finding brain metastasis died within a month, the 2 patients who received RT died in 3 and 4 months, respectively, and the 3 patients who underwent surgical resection survived 5, 10, 13 months respectively.In addition, both patients who received chemotherapy and apatinib survived for more than 10 months.The median survival time from diagnosis of BMO was 4 (range 1-13) months.

Clinical characteristics and outcomes of the 90 BMO patients retried from the published articles
Data collected from 90 patients in 33 published articles lacked consistency.There was less patient information reported in small case series as compared with case reports.Factor analysis could only be performed in patients whose information was available.In following sentences the numbers of total patients were different, because we calculated only those patients with available data.Among the 90 BMO patients, 54 were male (54/90, 60%) and 36 were female, with a median age of 16 (3-60) years, including 50% patients younger than 16 years and 75% younger than 20 years.Twelve patients (12/76, 15.8%) were identified as having extra-skeletal metastases at initial diagnosis, and the primary osteosarcoma lesions were mostly present in the four limbs in the remaining 64 patients (64/76, 84.2%).Before the diagnosis of brain metastasis, 66 patients (66/76, 86.8%) had distant non-brain metastases, including 61 (61/76, 80.3%) with lung metastasis.The median interval from primary disease diagnosis to brain metastasis was 14 (0-124) months.Four-fifths of patients (34/42, 81.0%) developed brain metastasis within 12 months after prior metastasis or recurrence.Correlation analysis showed a high correlation between the interval from primary diagnosis to BMO and interval from primary diagnosis to prior metastasis (r 2 ¼0.7775, p < 0.0001) (Figure 3).
Solitary brain metastasis was detected in 47 patients (47/64, 73.4%), and multiple metastases in 17 patients (17/64, 26.6%) at the time of diagnosis.The cerebrum was the most frequent brain metastatic site, with the frontal lobe predominating (24/58, 41.4%), followed by the parietal lobe (16/58, 27.6%), occipital lobe (16/58, 27.6%) and temporal lobes (6/58, 10.3%).Of the 90 patients, 79 received further treatment, of whom 36 patients (45.6%) received surgical treatment, 24 (30.3%) received RT, and 18 (22.8%)received chemotherapy.Of the 36 patients receiving surgery, 15 underwent comprehensive treatment, including surgery plus RT, surgery plus chemotherapy, or all of them.Unfortunately, the overall survival (OS) of BMO patients was poor, with a median post-brain metastasis survival of 3.5 (0-93) months, and 75% of them had a survival time less than 7.4 months.Considering the bias of case reports that tend to report patients with long survival time, the real prognosis may be even worse.

Survival analysis of post-brain metastasis
We analyzed the prognostic factors in the 7 patients in our series together with 68 patients with survival data available in the literature.Logistic analysis was performed on the following univariate factors: sex, age, lung metastasis, number of brain metastases, interval from first metastasis diagnosis to brain metastasis, chemotherapy, surgery and RT.Univariate analysis revealed that surgery (p ¼ 0.004) and RT (p ¼ 0.000) were significantly associated with a better prognosis (Table 2).Multivariate analysis of survival related factors was statistically analyzed by using Cox proportional hazard model.Knowing that there are studies in the literature reporting that chemotherapy may be a prognostic factor for OS in brain metastasis of bone and soft tissue sarcomas and osteosarcoma is chemo-sensitive, chemotherapy was also included in our multivariate analysis.As shown by multivariate survival analysis in Table 3, chemotherapy can effectively improve the prognosis of patients with BMO (p ¼ 0.019).Patients with a solitary brain lesion who underwent surgery and had a good performance status were chosen for further analysis.Surgery combined with chemotherapy or RT offered better overall survival than surgery alone (Figures 4 and 5), but there was no  significant difference in the clinical outcome between surgery plus chemotherapy and surgery plus RT (p ¼ 0.569).

Discussion
BMO is rare, accounting for approximately 4.7-31.3% of brain metastases from sarcomas. 23,26Published data of BMO are derived mainly from small retrospective case series of brain metastases in sarcomas and case reports.To the best of our knowledge, the present study represents the most detailed and comprehensive pooled analysis about BMO.Four of the 7 patients in our series were diagnosed in the past three years, and all of them had a history of administering the targeted drug apatinib to control the lung metastasis.The incidence of BMO in patients with pulmonary metastases is reported to be 10-15%. 36t was reported that the existence of metastatic disease at initial diagnosis or recurrence within a year was related to the increased risk of BMO. 26 Consistent with previous reports, we found that most cases (67%) of BMO occurred in children and adolescents under 18 years of age.There are controversies over whether periodical neuroimaging should be performed in metastatic patients, especially those with pulmonary metastases.Yonemoto et al. 25 and Marina et al. 4 both suggested that routine brain imaging should be performed in patients with active lung metastasis or relapse regularly, while others maintained that early detection of brain metastasis did not increase the survival and therefore suggested that brain imaging should be performed only when neurological symptoms appeared. 2As some patients were asymptomatic at the diagnosis of brain metastasis, early detection before symptoms appear may help timely treatment.However, no study mentioned the optimal time of performing imaging investigations in high-risk populations of BMO.In a retrospective study of 112 cases of brain metastasis from sarcomas including osteosarcoma, the median interval from primary sarcoma diagnosis to brain metastasis was 2.2 years, and that from prior systemic metastasis to brain metastasis was 1.1 years. 37In our study, including 7 cases in our institution, the median interval from primary disease diagnosis to brain metastasis was 14.5 months, and 76% (72/92) of cases were diagnosed with BMO within 3 years.Our findings are consistent with those reported in the literature.Correlation analysis showed that BMO was more likely to occur around 12 months after prior metastases (Figure 3).Therefore, we suggest performing MRI to detect BMO every 12 months for 3 years after relapse or metastasis.
The treatment of BMO patients is heterogeneous in the literature. 2,3,36In most cases, the therapeutic strategies rely on the clinicians' limited experience about the disease.In the few reports of several cases from a single center, all were usually treated in the same manner.According to our statistics, about 25.6% patients (22/86) gave up further treatment and all died within 3 months.Since most chemotherapy drugs cannot pass the blood-brain barrier, surgery and radiotherapy have become the mainstay of treatment for BMO patients.Previous reports have indicated that complete resection of brain metastases is beneficial to selected patients. 6,25Surgery can immediately reduce the burden of the tumor, but the brain metastasis treatment without other additional treatments may lead to local failure in most cases.Combination of surgery with adjuvant radiotherapy or chemotherapy can reduce such risks.Our study has demonstrated that aggressive multi-disciplinary treatment strategies, such as surgery plus chemotherapy or radiotherapy, offer better clinical outcomes than surgery alone, but we had too few cases to detect differences in the overall effect between surgery plus chemotherapy and surgery plus radiotherapy.
The role of chemotherapy in the treatment of brain metastasis is controversial. 5Our results demonstrated that chemotherapy might be an independent factor for survival aside from surgery and radiotherapy in the multivariate analysis.The reasons may be due to chemo-sensitivity of osteosarcoma and partial disruption of the blood-brain barrier in brain metastasis.A French multi-center retrospective study of 246 patients with sarcoma brain metastases in 2018 also found that chemotherapy became a prognostic factor for OS, although chemotherapy for brain metastases may not be able to achieve the desired results. 7hemotherapy sensitivity can be assessed empirically.A longer free-event interval between initial diagnosis and progress may suggest higher sensitivity to previous chemotherapy drugs, which means that the patients can continue to have the pre-chemotherapy drugs in the treatment plan.Similarly, chemotherapy could be co-administered in combination with surgery or radiotherapy.Three of the 7 BOM patients in our series received aggressive treatment, and their OS after brain metastasis was generally higher than that of patients without the treatment.Local treatment of brain lesions may relieve symptoms and improve the prognosis of BMO patients.But unfortunately, all 7 patients with brain metastases died within 13 months after the onset of brain metastases, demonstrating that local treatment has limited effect in improving the survival.
At present, there are no reports or studies on targeted drugs for brain metastasis in sarcomas.Apatinib is one of the TKIs that have effects on advanced osteosarcoma. 38Of the 7 patients reported here, the two receiving apatinib therapy had the longest survival.But this may also have relevance to their comprehensive treatment and better systemic conditions.
Our research has several limitations.First, the sample size is relatively small because of rarity of the disease.Second, this is a retrospective study and the data about BMO were retrieved from small retrospective case series of brain metastases in sarcomas and case reports.Unfortunately, information was incomplete in retrospective studies of brain metastases in sarcomas and there was publication bias of case reports engaged in long-term survival cases.
We conclude that radiotherapy, chemotherapy and surgery are independent favorable prognostic factors.BMO patients may obtain survival benefits from regular neuroimaging and early aggressive multi-disciplinary interventions including surgical resection, postoperative RT, and chemotherapy.

Figure 2 .
Figure 2. The treatment journey of seven BMO patients.

Figure 3 .
Figure 3. Correlation analysis of interval from primary diagnosis to BMO and interval from primary diagnosis to prior metastases.

Table 1 .
Patients characteristics for 7 patients with BMO in our institution.

Table 2 .
Univariate analysis of factors associated with Post-Brain metastasis survival.
MST: median survival time; IFB: interval from first metastasis diagnosis to brain metastasis;.Ã : The results are statistically significant (p＜0.05).

Table 3 .
Multivariate analysis of prognostic factors on overall survival.