Abstract
Background/Aim: The aim of this study was to identify patients at high risk of death from neurological cause because these patients may be appropriate candidates for intense brain-directed treatment, in contrast to patients with uncontrollable extracranial disease, inevitably leading to death. In this context, the LabBM score (endpoint: overall survival; five blood test results; often abnormal in patients with widespread disease) may be a relevant tool. Patients and Methods: This was a retrospective single-institution analysis of 101 patients, managed with upfront brain irradiation. Associations between neurological death and different baseline and treatment parameters were assessed. Results: A LabBM score of 0 (five normal blood test results) was present in 32% of patients. Neurological death was recorded in 27%. Seven parameters were associated with neurological death, including the LabBM score (univariate analyses). Three out of the seven were significantly associated with neurological death in the multi-nominal logistic regression analysis. The most important parameter was primary tumor type (colorectal or melanoma), with a hazard ratio of 14.3. Patients without liver metastases were also more likely to die from neurological causes. Finally, patients who did not receive additional systemic therapy were more likely to die from central nervous system progression. The median survival time was 3.9 months (entire cohort). When censoring patients who died from extracranial progression, the median time to neurological death was 17.4 months. Conclusion: The LabBM score was not suitable for prediction of neurological death, in contrast to three other parameters. Interestingly, additional systemic therapy reduced the risk of neurological death, possibly because several new agents have known antitumor activity in the brain.
Ideally, successful treatment of brain metastases would provide long-term local control without high-grade toxicity, i.e. preserved quality of life and ability to live an independent life. If the same aim of disease control could be achieved for all extracranial sites of malignancy, the ultimate goal of long-term survival would be achievable (1, 2). However, survival of patients with brain metastases is still limited in current clinical practice (3-5). Both neurological death [central nervous system (CNS) death, uncontrolled CNS dissemination] and non-neurological death (extracranial disease progression) contributes to the outcome that can be achieved with today’s treatment approaches. After hippocampus-avoidance whole-brain radiation therapy (WBRT) with a simultaneous integrated boost in patients with different types of primary malignancies, the neurological death rate was 27% (6). In a phase II study of surgery followed by hypofractionated radiosurgery (SRS) to the tumor bed in patients with oligometastatic disease with single large brain metastases, 10 out of 50 deaths (20%) were related to neurological causes (7). A different study reported that immunotherapy was associated with improved survival and reduced neurological death after SRS for brain metastases from lung and melanoma primary cancer (8). However, lung cancer is a biologically heterogeneous disease and a different study suggested that patients with epidermal growth factor receptor-mutant non-small cell lung cancer had a high 5-year incidence of neurological death (40%) (9).
Many researchers have developed scores and nomograms that predict different endpoints after local treatment of brain metastases (10, 11). One such analysis tried to dissect neurological and non-neurological death associated with patient, treatment, and systemic disease status in 293 patients receiving SRS after WBRT failure (12). However, the prediction of neurological death at the time of selecting initial treatment has received limited attention. It would be prudent to administer highly efficacious local treatment, e.g. surgical resection/SRS, to patients who are likely to benefit from optimum local control. In contrast, those who are likely to die from uncontrollable extracranial disease would be better candidates for other strategies. Recently, a new prognostic tool has been developed and validated, the LabBM score, which is purely based on five blood test results that mainly reflect extracranial disease burden, inflammatory processes, organ function and cachexia (13). After having validated the LabBM score (14), our group hypothesized that patients with normal blood test results (LabBM score 0) might be more likely to die from neurological than extracranial causes. To test this hypothesis, the present retrospective study was performed.
Patients and Methods
Analogous to our previous validation approach, our single-institution database that includes all patients with parenchymal brain metastases from histologically verified extracranial primary tumors managed with first-line radiotherapy (WBRT, SRS or other fractionated focal radiotherapy; both completed and interrupted treatment courses according to the intention-to-treat principle; no previous brain irradiation) was analyzed. Radiotherapy prescription was individualized, as was further treatment for new or recurrent brain metastases, and systemic progression. The strategies consisted of salvage surgery, SRS, WBRT, systemic therapy or best supportive care. Systemic treatment was usually prescribed as judged appropriate by the patients’ medical oncologists. The patients were treated between January 01, 2012 and December 31, 2019. Extracranial staging consisted of computed tomography. If clinically relevant, other modalities were added to clarify computed tomographic findings, e.g. isotope bone scan, ultrasound, and positron-emission tomography. All five blood tests needed to calculate the LabBM score were routinely assessed approximately 1 week before radiotherapy [normal values: hemoglobin 11.7-15.3 g/dl (females) and 13.4-17.0 g/dl (males); platelets 130-400×109; albumin 34-45 g/l; lactate dehydrogenase <255 U/l; and C-reactive protein (CRP) <5 mg/l]. The LabBM score was calculated as described in the original study (13). Briefly, 1 point was given for lactate dehydrogenase and CRP measurements above the upper limit of normal and 0.5 points for hemoglobin, platelets and albumin below the lower limit of normal. A point sum of 0 indicates a favorable prognosis with regard to overall survival. The maximum point sum is 3.5.
Cause of death was determined from chart review (electronic patient records of all hospitals in the healthcare region). Patients with an unknown cause of death, e.g. because of home death and lack of hospital records, non-cancer death or ongoing follow-up were excluded. Of 212 patients in the database, 101 with known, cancer-related cause of death were included in this study. Overall survival (time to death) from the first day of radiotherapy was calculated employing the Kaplan–Meier method (SPSS 25; IBM Corp., Armonk, NY, USA). Associations between baseline parameters (LabBM score, age, sex etc.) and neurological death were assessed with chi-square tests. Statistically significant parameters, i.e. with a univariate p-value of 0.05 or less, and parameters with strong trend towards significance, i.e. p-value of 0.06-0.09, were included in a multi-nominal logistic regression analysis. As mentioned above, this database created for the purpose of quality-of-care analyses has already been utilized and does not require additional approval by the local Ethics Committee (REK Nord).
Results
Thirty-six patients (36%) received radical brain-directed treatment, e.g. surgical resection/SRS. The others were treated with palliative WBRT, largely 10 fractions of 3 Gy. Fifty-three percent received additional systemic therapy after brain-directed treatment, including immunotherapy, tyrosine kinase inhibitors, endocrine therapy and classical cytotoxic chemotherapy. As shown in Table I, most patients had primary lung cancer (49%) and active extracranial disease (85%). A LabBM score of 0 was present in 32%. Neurological death was recorded in 27%. The latter endpoint was not associated with age, Karnofsky performance status, primary tumor control, presence of synchronous brain metastases, presence of symptomatic brain metastases, and administration of radical brain-directed therapy (all p>0.2). However, seven other parameters were associated with neurological death, including the LabBM score (Table II). Three out of seven were significantly associated with neurological death in the multi-nominal logistic regression analysis. The most important parameter was primary tumor type (colorectal or melanoma), with a hazard ratio of 14.3. Patients without liver metastases were also more likely to die from neurological causes, with a hazard ratio of 6.5. Finally, patients who did not receive additional systemic therapy were more likely to die from CNS progression, with a hazard ratio of 3.5.
Patient characteristics.
Associations between baseline parameters and neurological death.
Median survival time for the entire cohort was 3.9 months (95% confidence interval=2.6-5.2 months). When censoring patients who died from extracranial progression (non-neurological), the median time to neurological death was 17.4 months (95% confidence interval=0.7-34.1 months) and the 5-year survival rate was 19%.
Discussion
Treatment of brain metastases would be simple and straightforward if an anticancer drug providing long-term brain control including eradication of micrometastases existed, or if the established local treatment approaches did not cause symptomatic side-effects, such as radiation necrosis or cognitive decline. Reality is however different and actual decision-making and multimodal treatment approaches are quite complex (15). Despite notable advances, the literature consistently reports that 20-47% of patients die from uncontrolled CNS disease (6, 7, 9, 12, 16, 17). It has also been reported in two very large series that more than 50% of patients harbored extracranial metastases [ranging from 52% in lung non-adenocarcinoma to 85% in renal cell carcinoma in the study by Sperduto et al. (10)] (18). Thus, the presence of multiple sites of metastatic disease can explain why the long-term survival rates of patients with brain metastases still are disappointing.
The neurological death rate in the present study (27%) was within the range of other studies. For the first time, the LabBM score was tested as a predictor of neurological death, based on a sound rationale. Anemia, high CRP, low albumin and its other components are more closely related to extracranial than intracranial disease burden. However, the multi-nominal logistic regression analysis failed to confirm that the LabBM score adds value to the prediction of neurological death. Other predictors were identified, e.g. a colorectal or melanoma primary tumor. These tumor types have long been considered relatively radioresistant and difficult to control, especially with WBRT (19, 20), which was often utilized in the present study. In line with a previous study, limited to immunotherapy after SRS for lung cancer and melanoma (8), our results indicate that additional systemic therapy (not stratified for drug subgroup because of limited study size) reduced the risk of neurological death. In the present era of clearly improved systemic therapies (21), this result appears plausible.
In a previous nomogram study, the cause of death was non-CNS-related in 45% and CNS-related in 47%, whereas 7.5% were alive at last follow-up (12). The median time to CNS-related death was 9.3 months, shorter than in the present study and likely due to the different setting, i.e. salvage SRS after initial WBRT. Factors that predicted CNS-related death included melanoma histology (confirmed here), brainstem location (not included here), and lower dose margin of SRS delivered to the metastasis (not included here since not all patients received SRS). Colorectal tumors were not evaluated as a separate group but rather lumped together with “other”. The concordance index of the predictive model was 0.66. External validation is not yet available. Factors identifying patients at increased risk from death due to systemic disease included lung histology (relative to breast), declining Karnofsky performance status, and progressive systemic disease at the time of SRS salvage. We are not aware of any validated models predicting the risk of neurological death in the initial treatment setting. Given the large biological heterogeneity, which impacts on prescription of targeted drugs passing the blood–brain barrier and also immunotherapy, e.g. in different subtypes of non-small cell lung cancer (22), separate analyses for molecularly-defined groups of primary tumors are needed. Such efforts require much larger databases than typically available in single institutions.
Limitations of the present study include the unexpectedly small number of eligible patients, statistical power of subgroup analyses, and retrospective design. Relatively few patients received up-front treatment other than WBRT, a fact that affects the likelihood of intracranial tumor control at the treated site. However, distant brain failure (new brain metastases rather than local growth of an initial lesion) is a well-known disadvantage of primary focal treatment because large parts of the brain remain untreated (7, 16, 17). Not all of these failures can be salvaged successfully (23). In the present study, neurological death was equally common in patients treated with more efficacious lesion-directed approaches aiming at local control, e.g. surgery/SRS (25%), and palliative WBRT (28%). Even if we were unable to develop a straightforward, clinically applicable predictive model, the present results provide relevant insights and a basis for additional research, which may result in individualized treatment approaches with limited side-effects despite high CNS efficacy.
Footnotes
Authors’ Contributions
CN participated in the design of the study and performed the statistical analysis. CN, BM and RY conceived the study and drafted the article. All Authors read and approved the final article.
Conflicts of Interest
The Authors declare that they have no conflicts of interest.
- Received November 15, 2020.
- Revision received December 7, 2020.
- Accepted December 8, 2020.
- Copyright© 2021, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
