More than 25% of patients with locally advanced breast cancer (LABC) will relapse during their first five years after treatment, but Canadian researchers have developed a viable list of five risk factors that may predict relapse after neoadjuvant therapy. The University of Ottawa researchers presented the results of their analyses on Thursday, December 8, during the San Antonio Breast Cancer Symposium, noting that several other groups have also researched prediction tools. 

The retrospective study included 546 patients with LABC who were treated with neoadjuvant therapy over a 10-year period (2005–2015). Researchers assessed the following factors.

  • Demographics
  • Tumor size
  • Nodal and receptor status
  • Grade
  • Human epidermal growth receptor 2 (HER2)
  • Disease stage
  • Cancer treatment
  • Clinical outcomes

The researchers used the Cox regression model to formulate a prediction tool, with the primary endpoints being local or distant relapse rate and time to relapse during the first five years after treatment.

Neoadjuvant therapy had been prescribed in the 546 patients as follows.

  • 5-fluorouracil–epirubicin–cyclophosphamide followed by docetaxel in 91 patients (17%)
  • Doxorubicin and cyclophosphamide (AC)-docetaxel in 330 patients (60%)
  • Other regimens in 124 patients (23%)
    • AC
    • AC-paclitaxel
    • Taxotere and cyclophosphamide (TC)
    • TC-trastuzumab 

The overwhelming majority of patients (440, 81%) underwent a mastectomy compared to only 12% (67 patients) who underwent breast-conserving surgery. Almost 90% of the patients received adjuvant radiotherapy, and all patients had trastuzumab: 173 patients (34%) for HER2-positive disease and endocrine treatment (tamoxifen and/or aromatase inhibitors) and 356 (44%) patients for endocrine-sensitive disease. The relapse rate during the first five years of follow-up was 17.3%, with the highest rate in those with distant disease (13.2%). 

After analyzing more than 60 variables, researchers found only five factors that showed significant influence on risk for relapse during the first five years. 

  • Residual disease (yes = 4, no = 0; hazard ratio [HR] = 4.25; p = 0.000]
  • Lymph nodes status (positive = 3, negative = 0; HR = 2.27; p = 0.006)
  • Inflammatory histology (yes = 2, no = 0; HR = 1.90; p = 0.003)
  • Estrogen receptor status (positive = 2, negative = 0; HR = 2.07; p = 0.001)
  • Adjuvant radiotherapy (yes = 0, no = 1; HR = 1.76; p = 0.036)

When these factors are combined the following Risk Prediction (RP) Score can be constructed (see Table 1). With an internal validation showing a sensitivity of 75%, the rate of relapse was seven times higher in patients with an RP score of 8–12 compared to patients with an RP score of 0–5.

“Patients with high risk may require additional treatment or more active follow-up strategies; this simple model may be used to design unique studies in locally advanced breast cancer based on the RP score,” the researchers concluded. Future plans include further validation of the model in a larger, multicenter or provincial population.

Table 1. Risk Prediction Score

Score Risk of Five-Year Relapse Total Number of Patients Number of Patients With Relapse
0–5 Low (7%) 153 (28%; C: 77, A: 76) L: 3 (4%), D: 2(3%), L+D: 0
6–7 Intermediate (26%) 220 (40%; C: 96, A: 124) L: 5 (4%), D: 27 (22%), L+D: 0
8–12 High (51%) 172 (32%; C: 59, A: 113) L: 9 (8%), D: 43 (38%), L+D: 5 (4.5%)

A= analyzed; C = censored; L = local; D = distant; L+D = local and distant