Phase I Clinical Trial Outcomes in 93 Patients with Brain Metastases: The MD Anderson Cancer Center Experience Free

See commentary by Gounder and Spriggs, p. 3855

Of patients with advanced cancer, approximately 10% to 15% develop brain metastases (1–3). The cancers that mostly frequently metastasize to the brain are breast, non–small cell lung, colorectal, melanoma, ovarian, esophageal, head and neck, testicular, choriocarcinoma, and renal cell (4–13).

Whole-brain radiotherapy and stereotactic radiosurgery are the standard-of-care treatment options for patients with brain metastases from solid tumors. However, the long-term use of these modalities is limited by short- and long-term toxic effects to the brain with repeated use and increasing doses of radiation therapy. These toxic effects, including cognitive impairment, may not become evident until 6 months or longer after cranial radiotherapy (14). Even with the use of standard-of-care doses of brain radiotherapy and stereotactic radiosurgery, disease control is often suboptimal (1).

Patients seeking experimental anticancer therapy for metastatic or advanced cancer management are often excluded from early clinical trials because of their history of brain metastases. To systematically assess the outcomes of patients with advanced cancer and brain metastases in our Phase I Clinical Trials Program, we analyzed the characteristics and outcomes of these patients and compared them with those of patients who presented without brain metastases.

We reviewed the electronic medical records of 1,181 consecutive patients treated in the Phase I Clinical Trials Program at The University of Texas MD Anderson Cancer Center starting in August 2005, and we assessed their associated characteristics and clinical outcomes. Data were collected from transcribed notes in the electronic database. Patient records were reviewed from the time of presentation in the Phase I Clinic.

Patients eligible for phase I clinical trial participation were of various ages and had metastatic or unresectable cancer for which approved curative therapies were no longer effective or for whom there were no appropriate therapies likely to improve their disease status. Patients generally had progressive disease, evidence of evaluable or measurable disease according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria (15), Eastern Cooperative Oncology Group (ECOG) performance status 0–2, and a life expectancy of greater than 3 months. Premenopausal women were required to have a negative pregnancy test, and patients of childbearing potential had to agree to use contraception. Further eligibility criteria varied according to the specific protocol requirements of the studies on which they were enrolled. All patients provided written informed consent prior to enrollment onto a trial. All trials, as well as this analysis, were conducted in accordance with the guidelines of the MD Anderson Cancer Center Institutional Review Board.

Clinical outcomes of patients with a history of brain metastases who were treated in the Phase I Clinical Trials Program were compared with those of (a) all patients treated in the Phase I Clinical Trials Program during the same time period who did not have a history of brain metastases and (b) all patients without brain metastases who were treated on the same clinical trials during the same period of time.

Phase I treatment was determined after clinical, laboratory, and pathologic data were reviewed. The allocation of patients to investigational treatments varied over time according to protocol availability at the time the patients were seen, protocol eligibility criteria, the patient's prior response to therapy, potential toxicity, and patient preference or physician's choice. After the initiation of an investigational therapy, patients were evaluated at least every 3 to 4 weeks.

Selected clinical trials allowed patients with brain metastases to enroll, if the metastases were treated, stable for at least 3 or 6 months, or nonhemorrhagic, depending on the study. Forty-three percent of the clinical trials excluded patients with brain metastases, leptomeningeal involvement, history of bleeding, central nervous system (CNS) metastasis, symptomatic or uncontrolled brain metastasis or epidural disease, active brain metastasis, or a history of brain metastasis.

If a patient was previously treated with radiation therapy, a 3- to 4-week interval was required from completion of radiation therapy to initiation of treatment on a phase I clinical trial, depending on the clinical trial.

During each visit, a history review and physical examination were done and a comprehensive series of metabolic and hematologic tests were conducted. Brain metastases were detected using computed tomography (CT) or MRI studies, which were ordered not as screening tests but when the patient presented to the clinic with symptoms that indicated CNS involvement.

Descriptive statistics were utilized to summarize the patients' characteristics. The χ² test was used to examine the association between 2 categorical variables. The following covariates were analyzed: age; gender; ECOG performance status (16); history of radiation therapy to the brain; number of prior therapies; local recurrence; metastases in the liver; number of metastatic sites; platelet count; and albumin levels.

Best response was assessed by a RECIST measurement team within the Department of Investigational Cancer Therapeutics and verified by an MD Anderson radiologist every 2 cycles of therapy (cycle = 3–4 weeks, depending on the protocol) using RECIST criteria (15). Brain lesions were included as target lesions and were measured at restaging. The number and sites of lesions were determined by an MD Anderson radiologist. Partial response (PR) was defined as a 30% or above decrease in the sum of the longest diameters of target lesions, excluding complete disappearance of disease. Progressive disease (PD) was defined as a 20% or above increase in the sum of the longest diameters of target lesions. Stable disease (SD) was defined as small changes that did not meet the criteria for a PR or PD. Waterfall plot analysis was used to illustrate response, if any, as previously described (17).

Survival was measured from the date of presentation to the Phase I Clinical Trials Program until death from any cause or last follow-up. Time-to-treatment failure (TTF) was measured from the first day of treatment on a clinical trial (the first trial entered) in our Phase I Clinic to the date the patient went off study because of toxicity, disease progression, or death. Toxicities were assessed using the National Cancer Institute Common Terminology Criteria (NCI CTC) for Adverse Events, version 3.0 (18).

The χ² test was used in univariate analyses for response. Survival and hazard functions were estimated using the Kaplan–Meier method, and survival between groups was compared using the 2-sided log-rank test. The multivariate Cox proportional hazards regression model was used to examine risk factors related to survival, TTF, and time to development of brain metastases after adjusting for other factors (P ≤ 0.05 was considered to be statistically significant). Statistical analyses were conducted by a statistician and were carried out using SAS 9.1 (SAS Institute, Inc.) and S-Plus, version 7.0 (Insightful Corp.) software.

Overall, the records of 1,181 consecutive patients were examined. Ninety-three (7.87%) of the treated patients had brain metastases at the time of initial treatment in the Phase I Program and 79 (6.68%) patients developed brain metastases after their initial treatment in the program. The distribution of 93 patients with brain metastases by diagnosis was as follows: melanoma (24 of 81; 29.63%) and breast (21 of 112; 18.75%), lung (17 of 97; 17.53%), endocrine (9 of 97; 9.28%), head and neck (3 of 52; 5.77%), gynecologic (4 of 82; 4.88%), gastrointestinal (2 of 391; 0.51%), or other cancers (13 of 269; 4.8%).

Distribution of patients with and without brain metastases according to primary tumor type is described in Table 1.

The median age of the 93 patients with a history of brain metastases at the time of presentation to the clinic and who were subsequently enrolled on study was 54 years (range: 12–76 years). There were 36 (39%) men and 57 (61%) women.

For the 93 patients who presented to the Phase I Clinic with brain metastases, the median time from the date of last radiation treatment to the first day of phase I treatment was 5.5 months (90% CI: 1.3–26.6). The median time from the diagnosis of brain metastases to the date of first radiation treatment was 0.69 months (90% CI: 0.1–6.7). The median time from the date of diagnosis of brain metastases to the first day of phase I treatment was 10.2 months (90% CI: 2.1–36.0). Of the 82 (88%) patients who had prior radiation therapy, 64 patients had whole-brain radiation, 17 had stereotactic radiosurgery, and 1 patient had an unknown type of radiation outside MD Anderson.

The baseline characteristics of the 93 patients who enrolled on study and had a history of brain metastases at the time of presentation to the Phase I Clinical Trials Program were compared with those of the 1,088 patients who enrolled on study and presented during the same period without brain metastases (Table 2). Patients with brain metastases who were seen in the Phase I Clinic were more frequently younger and women and had lower rates of liver metastases and lower bilirubin levels compared with patients who presented without brain metastases.

Of 93 patients, 11 did not receive radiation treatment prior to phase I therapy for the following reasons: 6 patients refused to receive radiation therapy and elected to participate in a clinical trial in our department; 4 patients were asymptomatic and their brain metastases were stable; and 1 patient decided to be treated at another institution with an experimental dendritic cell vaccine.

Of the 93 patients who presented to the Phase I Clinic with brain metastases and were enrolled on study, 79 (84.9%) patients were treated with targeted therapy alone, 8 (8.6%) patients with cytotoxic chemotherapy alone, and 6 (6.5%) with targeted therapy plus cytotoxic chemotherapy. Of the 1,088 patients without brain metastases at presentation who were treated in the Phase I Program, 65% were treated with targeted therapy alone, 11.5% were treated with chemotherapy alone, and 23.5% were treated with targeted therapy plus chemotherapy.

For the 93 patients with brain metastases, the rates of SD lasting at least 4 months and PR/CR were 15% and 2%, respectively (Table 3). In these 93 patients, the response in the CNS metastatic sites after phase I treatment was as follows: 54 patients had progression (13 in the CNS only and 41 in both CNS and other organs), 37 patients had stable lesions in the CNS, and 2 patients had brain metastases that actually responded to treatment. For the remaining 1,088 patients without brain metastases, the rates of SD ≥4 months and PR/CR were 22% and 5%, respectively (Table 3).

In multivariate analysis for response in 93 patients with brain metastases, independent factors predicting higher rates of combined SD for ≥4 months and PR/CR were normal serum lactate dehydrogenase (LDH) levels (P = 0.05) and no prior brain radiotherapy (P = 0.02).

The median survival duration for all 1,181 patients from the first visit to the Phase I Clinic to the date of death or last follow-up was 10 months (95% CI: 9.1–10.9). Of the 93 patients with brain metastases who were enrolled on studies in the Phase I Clinic, 19 patients remained alive. The median survival durations of patients with brain metastases and those without brain metastases were 7.5 months (95% CI: 6.1–10.3 months) and 10.3 months (95% CI: 9.4–11.3 months), respectively (P = 0.002; Table 3; Fig. 1A). In patients with brain metastases, there was no difference in survival between those treated with targeted therapy alone and those treated with targeted therapy plus chemotherapy (Fig. 1B).

In multivariate analysis for survival in the 93 patients with brain metastases, independent factors predicting longer survival were normal levels of hemoglobin (P = 0.01), normal levels of serum LDH (P = 0.014), and no prior brain radiation therapy (P = 0.02).

The median TTFs on the first clinical trial after referral for patients with and without brain metastases were 1.74 months (95% CI: 1.6–1.8 months) and 1.84 months (95% CI: 1.8–1.9 months), respectively (P = 0.28; Table 3; Fig. 1C).

Of the 93 patients with brain metastases who were treated in the Phase I Clinic, 91 patients had treatment failure. Of these 91 patients, 13 had progressive disease in the CNS alone, 41 had progressive disease in the CNS as well as in other organs, and 37 had progressive disease in organs other than the CNS. Of these 93 patients, there was no difference in TTF in patients treated with chemotherapy with or without targeted therapy versus targeted therapy alone (P = 0.221; Fig. 1D).

The rates of grade 3 and 4 toxicity in patients who enrolled on study and who presented to the Phase I Clinical Trials Program with and without brain metastases were 12% and 10%, respectively (P = 0.77; Table 3).

The median survival of 93 patients with brain metastases treated on phase I clinical trials (7.5 months) was shorter than that of 527 patients without brain metastases treated on the same protocols (9.8 months; P = 0.01; Fig. 1E). However, there was no difference in TTF between the 2 groups (1.74 months vs. 1.84 months, respectively; P = 0.61; Fig. 1F).

In multivariate analysis for survival of 620 patients with and without brain metastases treated on the same clinical trials in which pretreatment characteristics were considered, independent factors predicting shorter survival were performance status greater than 0 (P = 0.0002), number of prior therapies (P = 0.0002), liver metastases (P < 0.0001), thrombocytosis (platelets > 440 × 10⁹/L; P = 0.005), and hypoalbuminemia (P < 0.0001); brain metastases were not an independent factor predicting survival (P = 0.22; Table 4).

We also compared the treatment-related neurologic adverse events ≥ grade 3 between patients with (n = 93) and without (n = 527) brain metastases enrolled on the same clinical trials during the same period of time. The main neurologic adverse events related to treatment were peripheral neuropathy, confusion, and seizure. Among 93 patients with brain metastases, 2 patients had grade 3 and 4 peripheral neuropathy, 1 patient had confusion, and 1 patient had seizures. Among 527 patients with no brain metastases treated on the same protocol, 4 had peripheral neuropathy ≥ grade 3, 6 experienced confusion, and 2 had seizures.

There were no statistically significant differences in grade 3 and 4 peripheral neuropathy (P = 0.22), grade 3 and 4 confusion (P = 0.99), or grade 3 and 4 seizures (P = 0.39) between patients with and without brain metastases. Neurologic adverse events were seen in 4 (4.3%) of 93 patients with brain metastases versus 12 of 527 (2.3%) without brain metastases treated on the same protocols. Although this difference was not statistically significant, a more robust analysis with a greater number of events and, therefore, patients would be required to draw meaningful conclusions.

After a median follow-up of 8 months (range: 0–40 months) after phase I treatment, 79 (7.3%) of 1,088 patients who enrolled on study and initially presented without brain metastases subsequently developed them (Fig. 2A). The event rates for new brain metastases from time of initial Phase I Clinic visit were 6% at 1 year, 13% at 2 years, and 19% at 3 years (Fig. 2A). Of the 1,181 patients treated in the Phase I Clinic, a total of 172 patients presented with or developed brain metastases during phase I treatment. The event rates for brain metastases from the time of diagnosis of the primary cancer were 2.9% at 1 year, 5.9% at 2 years, and 8.9% at 3 years (Fig. 2B).

In multivariate analyses (Cox regression model), the only independent factor predicting the development of brain metastases from time of presentation to the Phase I Clinic was elevated serum levels of LDH (>1 × upper limit of normal; P = 0.02). A history of thrombosis and leukocytosis trended toward an association with the development of brain metastases, but these variables were not statistically significant for brain metastasis development (P = 0.08 and P = 0.085, respectively; Table 5).

There is a dearth of data from systematic analyses of the clinical outcomes of patients with brain metastases referred to an early clinical trials clinic. Although many pharmaceutical company–sponsored and NCI-sponsored phase I clinical trials exclude patients with brain metastases or limit their access to new drugs, our results suggest that brain metastases should not be an exclusion criterion. Patients with brain metastases did have lower overall rates of SD lasting at least 4 months and PR/CR compared with patients without brain metastases (17% vs. 27%, respectively; P = 0.03). Overall survival rates were also lower for patients with brain metastases than for patients without brain metastases (7.5 vs. 10.3 months; P = 0.002). Independent factors predicting longer survival in patients with brain metastases were low levels of LDH, hemoglobin ≥ 11 g/dL, and no prior brain radiation therapy (the latter probably reflecting the size of brain metastases). However, when patients with and without brain metastases were compared, brain metastases were not an independent factor predicting survival. Other covariates were more significant: Independent factors predicting shorter survival were performance status > 0 (P = 0.0002), number of prior therapies ≥ 2 (P = 0.0002), liver metastases (P < 0.0001), thrombocytosis (platelets >440 × 10⁹/L; P = 0.005), and hypoalbuminemia < 3.5 g/dL (P < 0.0001; Table 4). Furthermore, the TTF on a clinical trial was not different for patients with brain metastases versus those without (1.74 vs. 1.84 months; P = 0.28). The rates of grade 3 and 4 toxicity also did not differ (12% in patients with brain metastases vs. 10% in patients without brain metastases; P = 0.77). In addition, when the 93 patients with brain metastases were compared with 527 patients treated on the same protocols who did not have brain lesions, there was no difference in treatment-related serious neurologic adverse events.

Interestingly, the current analysis shows that the event rates for new brain metastases from time of first visit in the Phase I Clinic increased from 6% during the first year to 13% at 2 years and 19% at 3 years. In multivariate analyses (Cox model), the only independent factor predicting the development of brain metastases was elevated serum levels of LDH (P = 0.02) at the time of referral.

Other investigators have reported a 1.4% incidence of brain metastases in patients with colorectal carcinoma seen within a 13-year period (5). The median time from diagnosis of metastatic colorectal cancer to brain metastases was 9.0 months. With a median follow-up of approximately 6 months, the median survival duration after diagnosis of brain metastases was 5.4 months (5). The difference in the results of that study compared with the current study may be associated with differences in the patient population such as diverse diagnoses and prior therapies.

In our patient population, the tumor types that were associated with brain metastases were similar to those reported by other investigators and included melanoma, breast cancer, and lung cancer (19). Other investigators also found that patients treated with surgery and radiation therapy had a median survival of 10 months compared with 5 months for patients treated with radiation therapy alone, or 12 months for patients treated with surgery, radiation therapy, and chemotherapy(19).

Limitations of the current study include the small numbers of patients with brain metastases, the inclusion of nonuniform therapy, and the retrospective nature of the analysis. Furthermore, patients with brain metastases were not, in general, treated with similar classes of agents as those without brain metastases as evidenced by the greater proportion of patients with brain metastases treated with targeted therapies (84.9% vs. 65% for patients with and without brain metastases, respectively) versus those treated with targeted plus cytotoxic agents (6.5% vs. 23.5% for patients with and without brain metastases, respectively). These differences may also influence some endpoints such as response and toxicity.

Some caution is needed in the interpretation of results about neurotoxicity. Four (4.3%) of 93 patients with brain metastases versus 12 (2.3%) of 527 patients without brain metastases treated on the same protocols showed neurotoxic side effects, and although this difference was not statistically significant (P = 0.435), a more robust analysis with a greater number of events and patients would be required to draw more precise conclusions. However, toxicity data were prospectively collected during the period of patient treatment.

In conclusion, our study suggests that brain metastases occur in a significant subset of patients referred for early clinical trials and that their presence is associated with shorter survival and lower response rates. However, the reduced survival may be due to other factors that coexist with brain metastases, as the presence of brain metastases themselves was not an independent factor predicting survival. Patients with brain metastases did not have a compromised TTF, nor did they suffer from an increased rate of serious toxicities, including those related to the neurologic system. Taken together with the grave nature of brain metastases and the urgent need to find new treatments for them, our data suggest that enrolling patients with brain metastases on early clinical trials is safe and should be encouraged.

Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NCRR or NIH. Information on NCRR is available at http://www.ncrr.nih.gov/. Information on Re-engineering the Clinical Research Enterprise can be obtained from http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp.

This publication was made possible by grant number RR024148 from the National Center for Research Resources (NCRR), a component of the NIH and NIH Roadmap for Medical Research.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.