Purpose: The HER2 extracellular domain (ECD) is enzymatically cleaved from the cell membrane. Shed ECD in serum has been studied as both prognostic and predictive markers. Lapatinib is a dual inhibitor of HER2 and epidermal growth factor receptor kinases. We examined the prognostic and predictive role of HER2 ECD in a randomized trial of paclitaxel with placebo or lapatinib in women with HER2-negative or -unknown breast cancer.

Patients and Methods: Patients (n = 579) with newly diagnosed metastatic breast cancer (MBC) were randomly assigned to paclitaxel with placebo or lapatinib. HER2 status was determined centrally. ECD was centrally measured by enzyme linked immunoassay in available samples at baseline (b; n = 472), week 9, and every 12 weeks thereafter. Results were correlated to overall response rate (ORR) and progression-free survival (PFS).

Results: Elevated baseline ECD (bECD) levels ( 16 ng/mL) did not predict HER2 tumor status (sensitivity, 62%; specificity, 75%). In HER2-negative tumors, elevated bECD was not correlated with improved efficacy for lapatinib plus paclitaxel versus placebo plus paclitaxel (ORR: odds ratio, 1.6; 95% CI, 0.1 to 3.8; P = .365; PFS: hazard ratio, 0.94; 95% CI, 0.60 to 1.47; P = .797). ECD levels tended to decrease over time when bECD was elevated. ECD conversion from low to high was associated with worse PFS. Converting from high to low was associated with a better PFS. A consistently low ECD level had better PFS than a consistently elevated ECD. All associations were found to be independent of lapatinib.

Conclusion: HER2 bECD does not predict lapatinib benefit in patients with HER2-negative MBC. Changes in ECD status correlates with patient outcome regardless of treatment given. Measuring HER2 ECD is not currently recommended for predicting benefit to lapatinib.