Long-term outcomes of surgical treatment of non-small cell lung cancer with oligometastatic disease involving only the brain in the era of PET/CT

Abstract

Background

The prognosis of patients with synchronous oligometastatic non-small cell lung cancer (NSCLC) has been improving owing to advancements in imaging techniques and new treatment approaches such as tyrosine kinase inhibitors. This study aimed to investigate the long-term outcomes, including the clinical course after recurrence, of patients with synchronous oligometastatic NSCLC with only brain metastases, treated with bifocal treatment.

Methods

We retrospectively analyzed 22 patients with clinical T1-4 and N0-1 NSCLC with synchronous brain metastases who were diagnosed by preoperative PET/CT and brain CT or MRI and underwent pulmonary resection for the primary site and surgery or radiation therapy for brain metastases at our institution from 2005 to 2019.

Results

The median follow-up period was 60 months. The 5-year recurrence-free survival rate and overall survival rates after pulmonary resection were 31.8% and 58.7%, respectively. In the univariate analysis, pathological N0 status was significantly associated with better recurrence-free survival, but not overall survival. The median survival after recurrence was 24 months. Aggressive brain treatment at sites of recurrence and the use of TKIs after recurrence have significantly prolonged prognosis.

Conclusions

The long-term outcomes in patients with synchronous oligometastatic NSCLC with brain metastases who underwent bifocal treatment, including pulmonary resection, were favorable. In particular, bifocal treatment may provide a chance for cure in patients without lymph node involvement.

Introduction

A large proportion of metastatic non-small cell carcinoma (NSCLC) cases have poor prognosis, and systematic chemotherapy is the primary treatment for these cases. Metastatic NSCLC is a heterogeneous disease that includes cases in which metastases are limited to a few sites, and local treatment is expected to be effective [1]. Many retrospective studies have reported the effectiveness of bifocal treatment, particularly in patients with NSCLC who have only a few brain metastases [2,3,4]. The National Comprehensive Cancer Network Guidelines recommend adding local treatment to primary NSCLC lesions if there is no disease progression after local treatment of brain metastases.

In the European consensus report on the definition and staging of synchronous oligometastatic NSCLC, 18F-fluorodeoxy glucose-positron emission tomography integrated with CT (PET/CT) and brain magnetic resonance imaging (MRI) or computed tomography (CT) is mandatory for staging, and a maximum of five metastases and three organs amenable to radical treatment with acceptable toxicity was proposed [5]. We previously reported that the 5-year survival rate was 21% in 29 patients with brain metastases who underwent bifocal treatment at our institution between 1981 and 2008 [6]. However, in our previous study, preoperative PET/CT was not performed in all patients.

Therefore, we focused on patients with NSCLC with only brain metastases detected using preoperative PET/CT and brain CT or MRI. We retrospectively investigated the long-term outcomes, including the clinical course after recurrence, in patients with synchronous oligometastatic NSCLC with brain metastases who were treated with bifocal treatment.

Patients and methods

Patients

This was a retrospective single-center study. The subjects were patients with synchronous brain metastases from NSCLC who received bifocal treatment, that is, both pulmonary resection and local treatment for brain metastases. Twenty-four patients with brain metastases, diagnosed on the basis of preoperative CT or MRI, received bifocal treatment at the Osaka International Cancer Institute between 2005 and 2019. Of these, two patients were excluded because one patient did not undergo PET/CT before surgery and one patient was lost to follow-up immediately after treatment. Ultimately, 22 patients were enrolled in this study.

Methods

A thorough review of the patient charts was conducted. Data on the patients’ clinical courses were collected. In the present study, tumors were classified and staged according to the 7th edition of the TNM classification of malignant tumors [7].

Indication criteria of bifocal treatment for synchronous brain metastases from NSCLC

At our institute, staging of NSCLC is performed based on chest CT, whole-body fluorodeoxyglucose (FDG)-PET/CT, and brain MRI or CT. FDG-PET/CT was employed to evaluate mediastinal lymph node metastasis, and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) was subsequently performed for patients with a significant FDG uptake in the mediastinal lymph nodes. Because performing EBUS-TBNA on N1 lymph nodes is often associated with technical difficulty, N1 lymph nodes are usually diagnosed using PET/CT alone.

Bifocal treatment was indicated for patients with synchronous brain metastasis from NSCLC based on the criteria described below.

• i. Patients with brain metastases only had distant metastases. Patients with brain metastasis and other synchronous distant metastases, such as liver, bone, or adrenal metastases, are contraindicated to bifocal treatment.

• ii. Number of brain metastases is five or fewer.

• iii. Clinical N0-1 disease. Patients with clinical N2 are excluded; clinical N2 disease is considered a good candidate for bifocal treatment [6].

• iv. Complete resection of the primary lesion was considered to be achieved based on preoperative radiological examination with radical intent.

• v. The performance status of the patients was 0–1, and their respiratory function was sufficient to tolerate pulmonary resection.

The indication for pneumonectomy was carefully judged in each case because the perioperative risk was high, and the loss of respiratory function was large in pneumonectomy. The indications for systemic chemotherapy were determined by a multidisciplinary cancer board for each case, based on patient background and performance status.

Selection of therapeutic modalities for brain metastases

The operability criteria for brain metastasis were determined by neurosurgeons. The inclusion criteria for craniotomy were as follows: i) a surgically accessible tumor location and ii) a large tumor (≥ 3 cm in diameter) or a smaller tumor associated with severe neurological symptoms. The exclusion criteria for craniotomy were as follows: post-neurosurgical life expectancy < 3 months and the likelihood of craniotomy resulting in severe neurological dysfunction. Stereotactic radiosurgery (SRS) was indicated for tumors < 3 cm in size with minor or no neurological symptoms. The indication for adjuvant radiotherapy was determined by consensus between neurosurgeons and radiation oncologists based on postoperative imaging studies and intraoperative findings. The timing of treatment for brain tumors depends on the patient’s symptoms. Patients with symptoms caused by brain metastases received treatment for brain tumors before undergoing thoracic surgery. Some patients underwent craniotomy at another hospital before the diagnosis of NSCLC.

Follow-up and treatment after recurrence

Follow-up was performed using CT, brain MRI, and serum carcinoembryonic antigen measurements every 3–6 months. Systemic chemotherapy is generally administered when recurrent disease is detected. If epidermal growth factor receptor (EGFR) mutation or anaplastic lymphoma kinase (ALK) rearrangement was detected, an EGFR-tyrosine kinase inhibitor (EGFR-TKI) or anaplastic lymphoma kinase inhibitor (ALK-TKI) was administered as treatment for recurrence.

Statistical analysis

The Kaplan–Meier method was used to estimate survival, which was calculated from the date of surgical treatment of the primary lesion to the date of the last follow-up examination as the endpoint. Survival curves were compared using log-rank tests. Data were analyzed using R version 4.2.2. Statistical significance was set at P < 0.05.

Results

Table 1 summarizes the characteristics of the 22 patients. The study population consisted of 15 men and 7 women with a median age of 60 years at diagnosis. Eleven patients had neurological symptoms including headache, ataxia, and paralysis. Seventeen patients had solitary brain metastases, and five had multiple (two to five) metastatic lesions.

Table 1 Clinical characteristics of the patients (N = 22)

Table 2 shows the treatment and pathological characteristics of 22 patients. The type of pulmonary resection was lobectomy in 19 patients, pneumonectomy in one patient, and sublobar resection in two patients. Nineteen patients (86%) underwent complete resection of the primary NSCLC. Among the three patients who underwent incomplete resection, two had pleural dissemination and one had positive pleural lavage cytology. No perioperative mortality was observed after the lung or brain surgery. The therapeutic modalities for brain metastasis included craniotomy in nine patients, surgery followed by WBRT in four patients, and SRS in 13 patients. More than half of the patients had pathological T1 or T2 tumors. Six patients were upstaged from the clinical nodal stage to the pathological nodal stage, and one patient was downstaged. Lymph node metastasis was observed in seven patients, but only two had mediastinal lymph node metastases. The histopathologic type was adenocarcinoma in 17 patients and other types in 5 patients. Eight patients received preoperative or postoperative chemotherapy.

Table 2 Treatment and pathological characteristics of the patients (N = 22)

The median follow-up time after surgery for the primary lesion was 60 months (range: 4–155 months). The 5-year recurrence-free survival (RFS) rate was 31.8% (median RFS, 60 months; 95%CI, 2–67) and the 5-year overall survival (OS) rate was 58.7% (median OS, 102 months; 95%CI, 21-NA) (Fig. 1). Table 3 shows the results of univariate analyses for RFS and OS. Univariate analyses showed that patients with pN1-2 had worse RFS. None of these factors was significantly associated with OS.

Fig. 1

The long-term outcomes of the whole study population (n = 22). (a) The 5-year recurrence-free survival rate was 31.8%. (b) The 5-year overall survival rate was 58.7%

Table 3 Results of the univariate analysis of prognostic factors for RFS and OS

The clinical characteristics and courses of the seven patients without recurrence are shown in Table 4. Notably, none of the patients showed lymph node involvement. Five of the 7 patients who had no recurrence did not undergo perioperative chemotherapy.

Table 4 The clinical characteristics and courses of the 7 patients without recurrence

During the follow-up period, 15 patients experienced recurrences. The clinical course after bifocal treatment is shown in Fig. 2 and Table 5. The median recurrence-free interval was 2 months (range 0–67) and the median survival time after recurrence was 24 months. The site of the first recurrence was the brain alone in eight patients, extracranial in five, brain and extracranial in one, and unknown in one. All the patients with intracranial recurrence underwent radiotherapy. The results of the univariate analysis of the prognostic factors for a survival after recurrence are shown in Supplementary Table 1. Notably, patients who experienced brain recurrence alone after bifocal treatment tended to have better survival rates than those who experienced extracranial recurrences. The survival time after recurrence according to mutation status is shown in Fig. 3. It is noteworthy that six patients with driver mutations (five of whom were treated with TKIs for recurrent disease) had significantly better survival rates after recurrence (median survival time, 83 months; 95%CI, 62-NA) than nine patients without driver mutations (median survival time, 11 months; 95%CI, 4–24).

Fig. 2

The clinical course of patients with synchronous brain metastasis from NSCLC. DOD died of the disease, DOAD died of another disease, AWD survived with the disease, NED had no evidence of the disease

Table 5 The clinical characteristics and courses of the 15 patients with recurrence

Fig. 3

Overall survival after recurrence according to mutation status (n = 15)

Discussion

The present study demonstrated that bifocal treatment for patients with synchronous oligometastatic NSCLC with brain metastases provided favorable long-term outcomes, with a 5-year OS rate of 58.7%. It has also been demonstrated that patients who receive TKI therapy have favorable outcomes after recurrence.

The 5-year OS of patients with synchronous oligometastatic NSCLC with brain metastasis who underwent bifocal treatment in our study was much better in comparison to the previously reported rates (8–21.4%) [4, 6, 8, 9]. PET/CT is thought to play a large role in selecting patients with oligometastatic NSCLC for whom bifocal treatment is appropriate. PET/CT has been reported to identify extrathoracic disease more accurately than conventional staging and contributes to the improvement of long-term outcomes in patients with oligometastatic NSCLC [10, 11]. In addition to the accurate diagnosis of distant metastasis, PET/CT has been reported to increase the diagnostic accuracy of nodal status assessment, which is a significant prognostic factor for recurrence-free survival [12]. However, 6 of 22 (27%) cases were upstaged from the clinical to pathological nodal stage in this study. Because the diagnostic accuracy of PET-CT for lymph node metastasis is limited, it is important to thoroughly evaluate lymph nodes during surgery in patients with synchronous oligometastatic NSCLC with brain metastases. Although a bifocal treatment strategy is associated with the limitations of the preoperative assessment for lymph node metastasis described above, we believe that this strategy itself can be justified because of its favorable long-term outcomes.

In addition to the advancement of staging modalities, the improvement in treatment after recurrence may have contributed to a better prognosis in the present study. Although the recurrence rate remained high (i.e., 15 of 22 patients [68%]), the median survival time after recurrence was relatively long (24 months). Frost et al. reported that repeated local ablative therapy for recurrent brain metastases resulted in favorable long-term outcomes [13]. In the present study, two patients had no relapse after local ablative therapy, indicating the importance of aggressive local treatment for brain relapses to provide long-term disease control. The use of TKIs is another factor contributing to prolonged post-recurrence survival. In the present study, it was observed that the five patients who used TKIs after recurrence survived longer than those who did not use TKIs, in accordance with previous reports [14, 15]. Recently, the effectiveness of radiotherapy for primary tumors in patients with oligometastatic NSCLC who underwent TKI therapy was assessed in a randomized controlled trial. In their study, the addition of upfront radiotherapy for primary tumors significantly improved the progression-free survival and OS for EGFR-mutated NSCLC compared with first-line TKI alone [16]. Based on this report, resection of the primary tumor may have contributed to an improved prognosis, even in patients with EGFR mutations.

Based on these findings, aggressive local treatment for brain relapses and appropriate use of EGFR-TKIs are important for the treatment of relapse after bifocal treatment for NSCLC with synchronous brain metastases.

It is important to select cases for which bifocal treatment is effective. In previous studies on synchronous oligometastatic NSCLC with brain and other metastatic sites, several factors, such as old age, multiple metastatic lesions, and lymph node metastasis, were associated with poor prognosis [3, 17]. In the present study, pathological N0 status was the only factor that was significantly associated with better recurrence-free survival. The lack of significance of the pathological N0 status as a prognostic factor for OS may be attributed to the effective treatment of recurrent disease with TKIs in patients with lymph node metastasis. Since all cases that eventually experienced no recurrence had a pathological N0 status, it is thought that patients with a pathological N0 status have a particularly high chance of long-term recurrence-free survival and potential for curative treatment. This supports the proactive application of bifocal treatment in patients with a pathological N0 status.

Whether systemic chemotherapy should be added to bifocal treatment is controversial. In previous reports on oligometastatic NSCLC, patients who underwent chemotherapy first and had no disease progression were allowed to receive additional local treatments [18,19,20]. The National Comprehensive Cancer Network Guidelines also recommend chemotherapy and confirm the absence of disease progression prior to local treatment. In contrast, in our study, most patients underwent thoracic surgery without chemotherapy, and their prognoses were relatively good. Whether chemotherapy should be administered prior to local treatment remains a topic for future investigation. Regarding postoperative chemotherapy, an analysis of the effect of systemic chemotherapy after surgery in patients with NSCLC with synchronous brain metastasis showed that postoperative chemotherapy improved the prognosis in patients with pathological N1, but not in those with pathological N0 [21]. In our study, all patients with pathological nodal involvement experienced recurrence, which led us to consider adjuvant chemotherapy for patients with pathological nodal involvement. Determining whether postoperative chemotherapy should be administered to patients with pN0 disease is a future task.

Immune checkpoint inhibitors (ICIs) have recently become the standard therapy for NSCLC without driver gene mutations. Takamori et al. conducted a propensity-matched analysis using the National Cancer Database to analyze the survival benefit of ICIs in patients with stage IV brain metastases [22]. In their study, multivariable Cox modeling after propensity score matching confirmed the survival benefits of ICIs. In Japan, ICIs have been available since 2017, but only three cases in this study were treated with ICIs; as such, a sufficient analysis has not been conducted. Determining the indication for bifocal treatment for brain metastatic NSCLC without driver gene mutations is a future task.

The present study has several limitations. First, this was a single-center retrospective study with a limited number of patients. The favorable outcomes may have been influenced by selection bias. Second, not all cases had pathologically confirmed brain metastases [23]. This may have influenced the outcome of the present study. Third, the analysis involved patients who received treatment for over a decade, during which there were alterations in therapeutic modalities. In particular, the development of second- and third-generation TKIs has had a great impact on outcomes in patients with EGFR mutation-positive tumors. These factors may have influenced the study outcomes. Fourth, the group receiving bifocal treatment was not compared to the control group treated with chemotherapy alone. A multicenter randomized phase III study is needed to answer the question of whether initial surgery plus metastasis-directed therapy or initial systemic therapy is better.

In conclusion, the long-term outcomes in patients with synchronous oligometastatic NSCLC with brain metastases who underwent bifocal treatment, including pulmonary resection, were favorable. In particular, bifocal treatment may provide a chance for a cure in patients without lymph node involvement.