The development of brain metastasis in a patient with non-small cell lung cancer (NSCLC) is an ominous prognostic sign. About 30% of individuals with NSCLC eventually develop brain metastasis.¹ This number increases to about 50% in autopsy series.² When the metastases are multiple, palliative treatment in the form of radiation therapy is recommended. Solitary brain metastasis, however, can be approached surgically. The proportion of NSCLC patients that develops brain metastasis amounts to approximately 40,000 patients per year. The magnitude of this problem can be appreciated by comparing this number with the incidence of new primary cancers of the pancreas (n = 27,000), stomach (n = 24,000), and esophagus (n = 13,000). The median survival rate of untreated lung cancer with brain metastasis is approximately 1 month. Steroid therapy increases the median survival by 2 months. Whole-brain radiation increases survival by 3–6 months.³ Recent reports indicate longer survivals when surgical treatment is combined with whole-brain radiation.^4,5 The experience of several large centers that offer a surgical approach to lung cancer with brain metastasis is discussed herein, with an analysis of factors underlying prolonged survival.

Previous studies have reported 5-year survival rates of 11-21% in patients treated with cranial and thoracic resection.^5,6 Despite this, many patients are offered palliative treatment only with chemotherapy or radiotherapy after a brain metastasis has been detected. There is controversy regarding the ideal management of thoracic disease in this patient population, and there are few series that incorporate patients treated with craniotomy or stereotactic radiosurgery (SRS).⁷

THORACIC RESECTION WITH CRANIOTOMY OR STEREOTACTIC RADIOSURGERY—PATIENTS AND METHODS

In our series from the University of Maryland, 28 patients with solitary brain metastasis were treated by thoracotomy with resection of lung cancer and craniotomy with excision of brain metastasis or SRS. More recent patients in our series have undergone lung resection and gamma knife stereotactic radiosurgery (GK-SRS). The series consisted of 16 men and 12 women ranging in age from 42–70 years, with a mean age of 56.24 years. Complete patient follow-up was accomplished by chart review, telephone call, or letter to the family, referring physician, or other hospital.

The initial presenting symptom was neurologic in 50% of patients. The range of neurologic symptoms included hemiparesis, headaches, monoparesis, ataxia, visual disturbances, seizures, behavioral changes, and mild weakness. In these patients, the onset of the cerebral metastasis was synchronous with the primary in that the pulmonary lesion was identified on the chest x-ray concomitant with the initial presentation of the metastasis. The remainder presented with pulmonary complaints related to their bronchogenic carcinoma, including cough, hoarseness, and chest pain, but later developed symptoms related to both pulmonary and neurologic systems.

Patients with initial neurologic symptoms generally underwent craniotomy or, recently, SRS. Patients who were seen primarily for a pulmonary malignancy initially underwent a pulmonary resection. The types of pulmonary resections performed were lobectomy, bilobectomy, pneumonectomy, and wedge resection. A complete dissection of the mediastinal lymph nodes generally was carried out in conjunction with the pulmonary resection. We now use routine bronchoscopy and mediastinoscopy before deciding to proceed with pulmonary resection. Resections generally were considered curative (R₀), with no gross tumor left behind. All resection margins were tumor-free microscopically, and the mediastinal nodes were removed.

The tumor cell type was most frequently adenocarcinoma, followed by squamous or adenosquamous, large cell undifferentiated, or anaplastic carcinoma. In determining the staging of the tumor, only the status of the primary tumor and lymph nodes was considered because of the known presence of a solitary cerebral metastasis. Based on such consideration, in diminishing frequency, tumors were most often stage I, stage II, and stage IIIA.

Radiation therapy after lung resection or aduvant chemotherapy often was tried. To date, 11 patients are alive and without evidence of recurrent cancer. Seventeen patients have died: 14 died of recurrent cancer, of whom 7 died of widespread systemic metastases, 4 had recurrence in the chest, and 3 had central nervous system metastases. One patient died of multiple-organ failure secondary to sepsis 47 days after lobectomy. One patient died 1 month after a left lower lobectomy of sudden cardiac arrest, and one patient died of respiratory failure complicating severe chronic obstructive pulmonary disease 17 months after craniotomy.

Survival after craniotomy often exceeds 5 years, with a 37% 5-year survival reported in prior studies. The difference in survival between those who received brain irradiation after craniotomy and those who did not was not significant in our series, although this finding has been challenged recently in reports of other similar series. Relief of neurologic symptoms after craniotomy is usually immediate.

In prior series, the following factors were analyzed to determine the effects on survival: age, sex, order of presentation (i.e., cerebral, pulmonary, or synchronous presentation), interval between thoracotomy and craniotomy if the thoracotomy was done first, interval between craniotomy and thoracotomy if the craniotomy was done first, cell type of tumor, type of pulmonary resection (i.e., pneumonectomy, bilobectomy, lobectomy, or wedge resection), curative or palliative resection, T classification of the lung tumor, nodal status (N₀, N₁ versus N₂) of the lung primary, lung tumor stage, duration of neurologic symptoms prior to craniotomy, location of brain metastasis, brain irradiation after craniotomy versus no irradiation, and use of chemotherapy or radiation therapy for the lung lesion. By univariate analysis, three factors were found to correlate with longer survival: curative pulmonary resection (P = 0.001), nodal status (P = 0.001), and age less than 55 years (P = 0.006). However, when all the factors were analyzed by the Cox multivariate model, only curative resection remained a significant factor for prolonged survival (P < 0.01).

DISCUSSION

A solitary brain metastasis associated with primary brochogenic carcinoma can occur without producing any neurologic symptoms. In a recent study, 42 patients with a solitary brain metastasis were treated with GK-SRS from 1993–2006. There were 27 men and 15 women, and the median age was 58 years (range 38–74 years). The median Karnofsky performance status (KPS) was 90 (range 70–100). Thirty-eight patients (90.5%) presented with symptoms of a solitary brain metastasis or were found to have brain metastasis on staging brain MRI within 1 month of histologic diagnosis of their primary NSCLC. The maximum diameter of the single brain metastasis was between 0.5–3.5 cm (median 1.5 cm). Brain lesions were located as follows: parietal lobe (12), frontal lobe (10), temporal lobe (9), occipital lobe (7), cerebellum (3), and thalamus (1). Initial staging to evaluate the extent of thoracic and extracranial disease included CT scans of the chest and abdomen (n = 42) and PET scans (n = 13).

Surgical staging was performed on 27 of 42 patients using mediastinoscopy, mediastinal dissection, or transbronchial needle aspiration to identify positive hilar and mediastinal lymph nodes. Twenty-two patients (52.4%) had radiographically or pathologically involved hilar (N₁) and/or mediastinal (N₂/N₃) lymphadenopathy; the thoracic disease thus was stage I, stage II, or stage III in 14, 9, and 19 patients, respectively.

The median dose prescribed was 18 Gy to the 50% isodose line (range 11-25 Gy). Additional whole-brain radiation therapy (WBRT) was delivered to 33 of 42 patients based on physician and/or patient preference. Twenty-one patients had WBRT after GK-SRS and 12 before. WBRT preceded thoracic therapy or chemotherapy in 21 patients, whereas 12 patients received it after thoracic therapy or chemotherapy or at the time of CNS progression.

Patients were considered to have definitive thoracic therapy if they underwent surgical resection or received sequential or concurrent chemotherapy and external beam radiation with definitive intent. Twenty-six patients (62%) completed definitive thoracic therapy: 9 patients had sequential or concurrent chemotherapy and radiation, 12 patients underwent surgical resection with or without preoperative or postoperative therapy, and 5 patients underwent a planned trimodality approach with preoperative chemoradiation followed by surgical resection. The median dose of thoracic radiation delivered to patients treated definitively was 61.2 Gy (range 45–68.4 Gy). Nondefinitive thoracic therapy (n = 16) included chemotherapy alone, palliative radiation therapy at doses greater than 2 Gy per fraction for an abbreviated course, radiation therapy followed by chemotherapy, and no therapy in 6, 4, 3, and 3 patients, respectively.

The median overall survival for the 42 patients was 18 months (range 1.5–150 months). The 1-, 2- and 5-year actuarial overall survival rates were 71.3%, 34.1%, and 21%, respectively. Currently, there are 8 patients alive with a median active follow-up of 64.5 months (range 9–150 months). The cause of death was identified in 20 of 34 patients. Neurologic progression was determined to be the cause of death in 5 of 20 patients (20%). The sites of progression in these 5 patients were CNS alone (3), CNS and distant (1), and CNS and thoracic (1). Symptomatic radiation necrosis requiring intervention (resection) in the absence of intracranial progression was documented in 1 patient.

Patients who had definitive thoracic therapy (n = 26) versus those who had nondefinitive therapy (n = 16) had a median overall survival of 26.4 months (95% confidence interval 16.2-36.6 months) versus 13.1 months (95% confidence interval 4.3–21.8 months) and a 5-year overall survival rate of 34.6% versus 0% (P < 0.0001), respectively. There was no statistical difference between patients treated definitively with (n = 18) or without (n = 8) surgery (P = 0.369). Patients with a KPS of 90 or greater had a median overall survival of 27.8 months compared with 13.1 months for those with a KPS of less than 90 (P < 0.0001). The prognostic factors significant on multivariate analysis were definitive thoracic therapy (relative risk = 2.97, P = 0.020) and KPS (relative risk = 5.85, P = 0.001).

Since the brain is affected by metastatic disease in 30-50% of patients, routine CT scan or MRI of the brain is recommended by our group in all cases of bronchogenic carcinoma. The diagnosis of brain metastasis in the past was made by nuclear isotope brain scanning or arteriography or both early in the study. CT scanning has been used in all patients since 1976. MRI has been used since 1985. For patients suspected of having cerebral metastases, double dose-delayed CT has proved significantly more sensitive than CT scans obtained immediately after the administration of a lesser dose of iodinated contrast material. Davis and colleagues⁸ reported that MRI with enhancement proved superior to double dose-delayed CT for lesion detection, anatomic localization of lesions, and differentiation of solitary versus multiple lesions.

Brain metastasis has been considered an advanced progression of the disease and has been treated historically with corticosteroids and irradiation. Although corticosteroids produce rapid improvement in the neurologic symptoms, they prolong life for a median of 2 months only. Radiation therapy provides 80% relief of symptoms, but the median survival rate is only 3-6 months.³ Ballantine and Byron⁹ in 1948 and Flavell¹⁰ in 1949 were the first to carry out staged surgical excision of a solitary non-small cell intracranial metastasis with the primary intrathoracic lesion. Magilligan and colleagues⁴ in 1976 introduced the modern approach of combined lung/brain resection with a 5-year survival rate of 21% and a low mortality of 3%. Subsequently, large series of patients treated with the combined modality of resection of cerebral metastasis followed by brain radiation were reported. Burt and colleagues⁵ reported 185 consecutive patients undergoing combined therapy. The overall survival rate was 55% at 1 year, 27% at 2 years, 18% at 3 years, and 13% at 5 years, with a median survival of 14 months. Vecht and colleagues¹¹ reported 63 patients receiving combined treatment of neurosurgery and WBRT with a median survival rate of 10 months. Lonjon and colleagues¹² reported 36 patients receiving such treatment with a median survival of 9.6 months.

Our past studies of combined treatment confirm these results. The survival rate of 28 patients undergoing this treatment was 58% at 1 year and 37% at 5 years, with a median survival of 1.60 years. Most of these patients received postoperative WBRT in the range of 3000-4500 rads. In 10 patients, a small-field boost of 900-2500 rads to the tumor-bearing area was added after completion of the WBRT. Two patients developed radiation fibrosis of the brain, one with incapacitating ataxia and the other with deterioration of memory. The advisability of postoperative WBRT remains unanswered.

Armstrong and colleagues¹³ evaluated 185 patients with non-small cell lung cancer who underwent resection of brain metastases. Forty-two patients who received preoperative WBRT (23%) were excluded. Sixty-four patients were equally divided into two groups, one (n = 32) received no WBRT; the other was prognostically matched to the first group (n = 32). The third group consisted of all other WBRT patients (n = 79). Most patients received 3000 rads in 10 fractions. Overall brain failures occured in 38% of the first group, 47% of the second group, and 42% of the third group. The use of WBRT had no apparent impact on survival or on overall brain failure rates. The only impact of WBRT was the reduction of focal failure, defined as failure within the brain adjacent to the site of resected brain metastasis.

However, Vecht and colleagues¹¹ compared the effect of neurosurgical excision plus radiotherapy with radiotherapy alone in a prospective, randomized test of 63 patients. WBRT was given in two fractions per day for a total of 4000 rads. The combined treatment compared with radiotherapy alone led to a longer survival. Median survival was 10 months in patients treated with the combined approach and 6 months in patients treated with radiotherapy alone (P = 0.04).

The factors contributing to prolonged survival have been addressed by various authors. Magilligan and colleagues⁴ found a wedge resection to be a significant predictor of improved survival; because this type of resection generally is reserved for small peripheral tumors with no hilar or mediastinal adenopathy, it suggests that the size of the primary tumor directly influenced survival. Rossi and colleagues¹⁴ found that the vigor of the patient, as assessed by Karnofsky and Zubrod scales and absence of nodal disease, influenced survival rate. Burt and colleagues⁵ found no significant difference in age, locoregional stage (TN), or histologic features in patients with synchronous versus metachromous lesions. However, multivariate analysis demonstrated that complete resection of the primary disease significantly prolonged survival. Lonjon and colleagues¹² found that the postoperative clinical status (Karnofsky score) and the postoperative neurologic grading were significant factors to determine survival. Nakagawa and colleagues¹⁵ found that the variables significantly associated with a favorable prognosis included surgical excision of the primary lesion, adenocarcinoma as the histologic diagnosis, the use of adjuvant treatment, a preoperative score of over 80% on the Karnofsky scale, and metastasis confined to the brain. Additional but nonsignificant contributors to a good prognosis included age under 65 or 70 years, early-tumor stage, curative lung cancer surgery, a single metastatic brain tumor, a solid versus cystic tumor, and a supratentorial location of the brain metastasis. Our series agrees with those of Hankins and colleagues⁶ and Burt and colleagues⁵, namely, that the most significant factor in prolonged survival following combined surgery and radiation for solitary brain metastasis was curative excision of the primary lung tumor.

However, despite prolonged survival and improvement in the quality of life after surgery and radiation therapy, recurrence of the brain metastasis contributes to the death of these patients. Patchell and colleagues¹⁶reported that the recurrence at the site of the original brain metastasis was 20% in the surgery group and 52% in the radiation group. Nakagawa and colleagues¹⁵ reported that 19% of patients treated with surgery or radiation died directly because of the brain metastasis, and 3.6% died of treatment-related complications.

Nakagawa and colleagues¹⁵ recommended that adjuvant treatment generally should follow excision of brain metastasis, considering that metastatic lesions smaller than 1.0 cm, which are not seen on CT scan, can be shown by MRI postoperatively. Radiation-insensitive tumors might disappear on MRI after combined chemotherapy and irradiation owing to enhancement of the radiation effect by chemotherapy. A significantly longer survival was found in patients who received adjuvant treatment than in those who did not. Chemotherapeutic regimens were divided into those involving cis-diamine dichloroplatinum, nitrosoureas, and other anticancer agents. Patients given cis-diamine dichloroplatinum had a significantly longer mean survival time (468 days) than patients given other anticancer agents (243 days) (P < 0.05).

Another recent approach to solitary brain metastasis is the use of a gamma knife with precise localization of the tumor by stereotactic method, which is promising, especially in patients who are not good surgical candidates.¹⁷

Multiple series have demonstrated that thoracic therapy and extent of thoracic disease may have an impact on survival. In a series by Bonnette and colleagues, 99 of 103 patients had surgical resection of their synchronous solitary brain metastasis and primary NSCLC. The median overall survival was 12.4 months, and the 5-year overall survival was 11%.¹⁸ Moreover, Billings and colleagues reported a median and 5-year overall survival of 24 months and 21.4%, respectively, for 28 patients who underwent surgical resection for their brain and thoracic disease. The superior overall survival in the series of Billings and colleagues may be attributed to the 15 patients (53.6%) with thoracic stage I disease. Contrary to the series of Bonnette and colleagues, Billings and colleagues reported a significant improvement in overall survival if the there was no pathologic evidence of lymph node metastasis (5-year overall survival 35% versus 0%, P = 0.001).¹⁹ Hu and colleagues reviewed 84 patients who underwent surgical resection or SRS for their brain metastasis, but only 44 patients received any therapy for their thoracic disease. The median overall survival of 15.5 months was significantly better for those who had thoracic therapy versus 5.9 months for those who did not (P = 0.046).²⁰

A different approach to minimize brain atrophy and mental deterioration following radiotherapy is the use of intraoperative radiation therapy at the time of surgical intervention. Nakamura and colleagues²¹ reported 1-year survival of 59% in 14 patients undergoing surgery and intraoperative radiation therapy, which is similar to the result obtained in 71 patients receiving surgical excision and whole-brain irradiation. The frequency of remote recurrence after the new therapy was 20% in 1 year, which was almost the same as that of the usual therapy (surgery plus whole-brain irradiation).

The use of hyperthermia plus nitrosoureas has been reported in 17 patients with non-small cell lung cancer with brain metastasis. Sixteen (94%) responded with clinical improvement, radiologic regression, or disease stabilization. The survival time of the improved patients was 12.7 months.²²

There is a concern regarding the true cause of the single brain lesion because the majority of patients were diagnosed on MRI without stereotactic-guided biopsy. Patchell and colleagues found that 11% of patients with abnormal imaging had intracranial disease other than metastasis; however, this study included a heterogeneous collection of malignancies.¹⁶ The frequency of false-positive MRI findings with more modern imaging in patients with NSCLC is probably lower. In addition, KPS, age, extent of thoracic disease, and other patient characteristics may have influenced the decision to offer GK-SRS and/or definitive thoracic therapy in our series.

Despite the potential for long-term survival, many patients are offered only chemotherapy or palliative radiation therapy for their thoracic disease without considering their thoracic stage and performance status. At our center, an aggressive staging and treatment paradigm has been instituted when approaching patients with a synchronous solitary brain metastasis. Patients undergo a brain MRI and CT/PET scan to appropriately determine the extent of intracranial and extracranial disease. Studies have shown that CT scans often underestimate the extent of intracranial disease and that PET scans may identify metastases in approximately 25% of patients thought to have thoracic disease only.²³ Thus overall survival actually may be improved with PET scanning in all patients.²⁴ In addition, surgical candidates will undergo surgical mediastinal staging. Patients then are selected for definitive brain and thoracic management based on the extent of intracranial and thoracic disease, presence of involved lymph nodes, and physiologic/performance status. The timing of brain and thoracic therapy also depends on these factors. Patients with a good KPS are often recommended to receive GK-SRS or surgical resection and WBRT. If patients have neurologic symptoms or a large brain metastasis and are not surgical candidates, we recommend GK-SRS and WBRT prior to thoracic therapy or chemotherapy. WBRT may be delivered prior to GK-SRS in order to decrease the volume of the lesion. This may allow a higher GK-SRS dose to be delivered. If the brain lesion is small and asymptomatic, we perform GK-SRS prior to thoracic therapy. If the patients do not progress extracranially, we proceed with WBRT after thoracic therapy or chemotherapy. If patients require further evaluation over a 2- to 3-week period to assess their surgical candidacy and extent of thoracic and extrathoracic disease, we may proceed with WBRT before thoracic therapy for logistical reasons.

SUMMARY

The use of SRS or craniotomy and resection in the treatment paradigm of patients with synchronous solitary brain metastasis from NSCLC is recommended. The median overall survival of 18 months and 5-year overall survival of 21% are similar to surgical series and stage III patients treated with concurrent chemoradiation. Improved KPS at diagnosis and definitive thoracic therapy significantly affected survival. This potential for long-term survival has influenced our treatment approach. Thus patients should be considered for definitive thoracic therapy after undergoing SRS or surgical resection of their brain metastasis.

EDITOR'S COMMENT

This is a unique subgroup of lung cancer patients with metastatic disease. Several reports of aggressive treatment of these so-called oligometastatic lesions have shown good local control with reasonable long-term survivals. In the era of stereotactic radiosurgery for brain lesions, up to 3 lesions have been treated aggressively, followed by stage-specific local lung cancer treatment. In general, this treatment is offered only to stage I or II patients. On the other hand, stage III patients who recur commonly in the brain after a reasonable time interval also are treated aggressively. The role of prophylactic brain radiation in patients with non-small cell lung cancer will be clarified with the results of the Radiation Therapy Oncology Group (RTOG protocol 0214).

–MJK

REFERENCES