High Circulating Tumor Cell Count at Baseline and Increases in Count During Treatment Linked to Reduced Survival in Metastatic Breast Cancer
In a pooled analysis of individual patient data reported in The Lancet Oncology, Bidard et al found that baseline circulating tumor cell counts and increases in counts during treatment were highly prognostic in patients with metastatic breast cancer beginning new treatment and that addition of circulating tumor cell count improved predictive strength of established clinicopathologic models.
Study Details
The study involved reported and unreported data from 1,944 patients with metastatic breast cancer from 20 studies in 17 European centers. The eligible studies had patients who were starting a new line of therapy, data on progression-free survival or overall survival, and circulating tumor cell counts using the CellSearch method prior to the start of the new treatment (baseline).
The association of circulating tumor cell counts and progression-free survival and overall survival was assessed by Cox regression models stratified by study, and the prognostic value of circulating tumor cell count during treatment was assessed using a landmark method. The added value of circulating tumor cell counts to predictive clinicopathologic models was assessed by a resampling procedure using likelihood ratio (LR) χ2 statistics.
Effect of Baseline Count and Changes During Treatment
Median follow-up for the pooled population was 23 months. Of the 1,944 patients, 911 patients (46.9%) had baseline circulating tumor cell counts ≥ 5/7.5 mL. This higher vs lower count was associated with significantly reduced progression-free survival (hazard ratio [HR] = 1.92, P < .0001) and overall survival (HR = 2.78, P < .0001).
Further, when adjusted for baseline circulating tumor cell count, increased circulating tumor cell counts (≥ 5 per 7.5 mL) at 3 to 5 weeks and at 6 to 8 weeks after the start of treatment was associated with significantly reduced progression-free survival (HR = 1.85, P < .0001, at 3–5 weeks; HR = 2.20, P < .0001, at 6–8 weeks) and overall survival (HR = 2.26, P < .0001, at 3–5 weeks; HR = 2.91, P < .0001 at 6–8 weeks).
Improves Clinicopathologic Models
Clinicopathologic predictive models for progression-free survival included tumor histologic subtype, histologic grade, number of previous lines of chemotherapy and hormone therapy for metastatic disease, performance status, presence of liver or visceral metastasis, and synchronous metastases at diagnosis of breast cancer. The model for overall survival included tumor histologic subtype, number of previous hormone therapies, performance status, liver metastases, brain or leptomeningeal metastases, metastasis-free interval, and age.
The addition of baseline circulating tumor cell count to these models significantly improved predictive value for both progression-free survival (LR = 38.4, P < .0001) and overall survival (LR = 64.9, P < .0001). The addition of circulating tumor cell count change at 3 to 5 weeks and at 6 to 8 weeks after start of treatment produced further significant improvements in models for progression-free survival (LR = 8.2, P = .004, at 3–5 weeks; LR = 15.3, P < .0001, at 6–8 weeks) and overall survival (LR = 11.5, P = .0007, at 3–5 weeks; LR = 14.6, P < .0001, at 6–8 weeks).
Elevated baseline cancer antigen 15-3 and carcinoembryonic antigen were associated with significantly shorter survival on univariate analysis, but only carcinoembryonic antigen added predictive information when added to the best clinicopathologic plus baseline circulating tumor cell model. Changes in these tumor markers during treatment (at 3–5 and 6–8 weeks) did not add predictive information to either the best baseline model or to the models including circulating tumor cell count changes during treatment.
The investigators concluded, “These data confirm the independent prognostic effect of [circulating tumor cell] count on progression-free survival and overall survival. [Circulating tumor cell] count also improves the prognostication of metastatic breast cancer when added to full clinicopathological predictive models, whereas serum tumour markers do not.”
Jean-Yves Pierga, MD, of Institut Curie, is the corresponding author for The Lancet Oncology article.
The study was funded by the Nuovo-Soldati foundation for cancer research and by Janssen Diagnostics. For full disclosures of the study authors, visit www.thelancet.com.
The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.
