Brain metastases (BMs) in CRC are uncommon, with an incidence of only 0.3–9%[3, 6, 13]. The somewhat lower incidence in our cohort (0.26%) can be explained by the fact that our study was limited to synchronous brain metastases, which account for 3.4–43% of brain metastases in CRC [4, 13]. The prognosis of CRC patients with brain metastases is very poor, as the median survival is only 3 months (range: 2–9.6 months) [3, 6, 14–19]. BMs have a significant impact on the quality of life of CRC patients once related symptoms appear . Understanding the prognostic variables for BMs is critical for determining survival and treatment options. However, there have been very few studies and a consensus on treatment for BMs in CRC is lacking. From 2010 to 2016, the SEER database included 678 CRC patients with synchronous BMs. According to multivariate Cox regression analysis performed in this study, age, CEA level, and Primary tumor resection were independent risk factors for synchronous BMs in CRC.
A growing amount of research suggests that patients who have their primary tumor removed have a better chance of surviving. In epithelial ovarian and renal cancers, improved survival results linked with surgical debulking have a wellestablished data foundation[20, 21]. However, in advanced CRC, the survival benefit of the PTR is debatable, and current NCCN guidelines support PTR only when symptoms are present. According to some clinicians, the improvement in OS is unclear, such that the morbidity and mortality resulting from tumor resection should be avoided as it will delay the start of chemotherapy, which may in turn reduce survival [12, 23]. However, some studies have shown that primary tumor resection considerably improves OS [24, 25], as also demonstrated in our cohort of CRC patients with synchronous brain metastases. In the primary tumor resection group of this study, the 1-year survival rate was 38.8% compared to 20.9% in patients with unresected primary tumors (p < 0.0001). In order to balance the selection bias between the PTR and non-PTR groups, we further analyzed the corresponding subgroups. Our results showed that PTR had a survival benefit regardless of the site of the primary tumor and with or without extracranial metastases. The higher survival rate following PTR can be related to the decrease of cancer stem cells that are resistant to chemotherapy as well as the reduction of primary tumor load. PTR significantly lowered the risk of CRC-related complications such acute bleeding, perforation, and obstruction, which might result in increased surgical mortality and morbidity. Our forest plot of subgroup analysis revealed that PTR improved survival in CEA-positive CRC BMs patients, but it had no effect on CEA-negative patients. This suggests that we should be more cautious about whether to perform surgery for the primary tumor in patients with CRC BMs complicated with CEA negative.
Our results also identified age as a significant prognostic factor for brain metastases in CRC. Patients in our study were divided into two age groups: <60 and ≥ 60 years. Patients < 60 years of age had better prognosis. Yang et al.  found a considerably poorer prognosis in patients > 70 years than in those 40 years of age. Quan et al.  also defined three age groups among CRC patients: <60, 60–74, and ≥ 75 years. Similar to our findings, patients ≥ 75 years of age had the poorest prognosis, and those < 60 years had the best prognosis. In another study, older age was also identified as an unfavorable prognostic factor for BMs in CRC . Because the prognosis of patients with CRC and synchronous BMs worsens with age, younger individuals should be treated more aggressively than older patients.
The tumor marker CEA is frequently used to monitor CRC patients during treatment. Among patients with metastatic disease, CEA levels are increased (> 5 ng/mL) in ~ 70% [29, 30]. Several studies have reported higher CEA levels at the time of brain metastasis diagnosis in CRC patients [31–34], but only three reported a putative predictive effect [4, 15, 35]. Consistent with those reports, our study identified CEA as an independent prognostic predictor, with a 1-year survival rate of 26.8% for CEA-positive patients and 40.3% for CEA-negative patients. Thus, CEA positivity is a poor prognostic indicator in CRC patients with brain metastases.
According to our findings, the location of the primary tumor had no prognostic impact in CRC patients with brain metastases. In a univariate analysis, Huerta et al.  found that patients with RCC had a poorer survival rate than those with LCC (4.6 vs. 10.7 months; HR 3.5, p = 0.025), but in a multivariate Cox regression analysis the difference was not significant. Another study found that patients with a left-sided primary tumor had a 1.5-fold better prognosis than those with a right-sided primary . This may reflect differences in the division of the left and right colon. In the univariate Cox regression performed in our study, extracranial metastases were significantly related to a poor prognosis but this result was not confirmed in the multivariate Cox regression.
Our study had several limitations. First, only patients with BMs at first presentation with CRC were included, because the SEER database only records the status of BMs at the time of initial diagnosis. Second, the SEER database does not contain information on Karnofsky performance status, the number of BMs, comprehensive treatment of BMs, or molecular markers; hence, these parameters were not included in our analysis.
In conclusion, primary tumor resection (PTR) offers survival benefits to patients with BMs from CRC. Age, CEA level, and PTR were identified as independent risk factors affecting the survival of CRC patients with synchronous BMs.