Predicting Clinical Response to HER2-Targeted Therapy

The first of two major research projects in Dr. Thompson's Hormones and Cancer Laboratory is predicting clinical response to HER2-targeted therapy in breast cancer.

The EGFR family member HER2 (ERBB2) is overexpressed in 15 to 20 percent of breast tumors. Although the prognosis for such patients is generally poor, the use of humanized monoclonal anti-HER2 antibodies, such as trastuzumab, has resulted in long-term (10 years) disease-free survival in about 80 percent of patients with early-stage HER2-positive disease.

However, the breast tumors of about 20 percent of such patients are refractory to adjuvant trastuzumab-containing regimens, and patients develop tumor relapse within 10 years or less.

The challenge is to identify this subset of patients, who could then be enrolled in clinical trials to test the efficacy of emerging HER2-targeted therapies.

In short, although the lab team may know how to cure most HER2-positive patients, work remains to identify and focus efforts on the patients whose cancer is unresponsive to current standard of care.

Whole-transcriptome analysis of patients

The Hormones and Cancer Lab has carried out whole-transcriptome analysis of patients enrolled in the North Central Cancer Treatment Group N9831 clinical trial.

The lab's analysis of these data was informed by three central concepts:

• That the clinical behavior of the tumors was linked to biological functions
• That these biological functions could be identified using a systems biology approach that focused on networks of genes with significant associations with clinical outcome (Cox hazard ratios)
• That different subsets of these genes might govern the relevant biological processes in individual tumors

Using a network-based analysis of genes with significant hazard ratios, the lab identified a biological process (immune function) that predicts response to trastuzumab. This process is defined by a panel of 14 immune function genes that contribute to an immune enrichment score that can be calculated for individual tumors.

Immune-enriched and nonenriched cohorts

Using response surface mapping tools, Dr. Thompson's Hormones and Cancer Lab defined cut-off points that stratify HER2-positive tumors into immune-enriched and nonenriched cohorts.

Patients with immune-enriched tumors, defined by abundance of any nine or more of the 14 transcripts, benefit from adjuvant trastuzumab (HR = 0.35, p < 0.0001).

Patients with nonenriched tumors seem to receive no benefit (HR = 0.89, p = 0.53). Furthermore, immune gene enrichment conveys no benefit to patients with HER2-positive breast cancer who receive chemotherapy alone (HR = 0.90, p = 0.64); the p-value for the immune-enrichment status-treatment arm interaction term was < 0.0001.

Thus, the lab's immune enrichment score is predictive rather than prognostic, and identifies the patient population upon which future clinical trials must focus.

The research team's immediate goal is to carry out analytical validation of a gene panel that interrogates the immune status of primary HER2-positive tumors before initiation of adjuvant therapy.

Dr. Thompson's team is using a NanoString custom code set that reads out the activity of about 800 immune function genes. These data will be used to build genomic landscape models that predict the manner in which an individual tumor will respond to trastuzumab.

Clinical implementation of the predictive model

The lab's initial efforts focus on samples from the N9831 clinical trial. The research team is assessing the specificity, sensitivity, accuracy and reproducibility of the immune gene function model on the NanoString platform as a first step toward clinical implementation of the predictive model.

Clinical validation is being carried out using samples from the NSABP B31 and FinHER adjuvant trastuzumab clinical trials to determine if the team's observations from N9831 are confirmed in other early-stage HER2-positive patient cohorts.

The lab is also testing its immune enrichment model as a predictive marker of pathological complete response (pCR) to neoadjuvant HER2-targeted therapy in early-stage HER2-positive patients.

The team is testing its hypothesis that immune enrichment predicts pCR in the Cancer and Leukemia Group B 40601 and TBCRC026 neoadjuvant HER2-targeted clinical trials.

The lab's long-term aim is to determine if its immune enrichment score predicts long-term relapse-free survival in patients who receive neoadjuvant trastuzumab alone or in combination with lapatinib or pertuzumab.

The lab's central objectives are to carry out analytical and clinical validation studies to establish the efficacy of its immune enrichment model for identifying patients who are unlikely to benefit from adjuvant trastuzumab.

Emerging HER2-targeted therapies

It is this group of patients who should be enrolled in future clinical trials to test the efficacy of emerging HER2-targeted therapies, including combinations of antibodies (trastuzumab plus pertuzumab), T-DM1, antibodies plus small molecule receptor tyrosine kinase inhibitors (trastuzumab plus lapatinib or neratinib), or immunotherapy. Several clinical trials testing these therapies are in progress, and samples from these trials will be used in the lab's clinical validation studies.

Elucidation of a link between the immune status of the tumor and response to anti-HER2 antibody therapy raises a significant number of mechanistic questions that are likely to impact understanding of the mechanism of action of such biological drugs.

The immediate clinical impact of the lab's findings is on the ability to predict trastuzumab response and to identify patients who will require other, perhaps new, therapeutic approaches to achieve long-term disease-free survival.