HER2-Enriched Subtype Strong Predictor of Response to Dual HER2 Blockade

Aleix Prat, MD, PhD

TheHER2-enriched subtype of HER2-positive breast cancer is a strong predictor of sensitivity to dual HER2 blockade without the use of chemotherapy, according to findings presented at the 2016 San Antonio Breast Cancer Symposium.

In the non-randomized, open-label PAMELA study, theHER2-enriched subtype predicted pathologic complete response (pCR) following neoadjuvant lapatinib and trastuzumab among patients with HER2-positive, hormone receptor (HR)-positive disease, said Aleix Prat, MD, PhD, of the August Pi i Sunyer Biomedical Research Institute, in Barcelona, Spain.

The finding provides independent confirmation that characterization of PAM50 breast cancer subtype can identify molecular profiles of patients who may safely be spared chemotherapy compared with HR status, which until now was the only molecular marker that consistently predicted pCR inHER2-postive disease in the absence of chemotherapy, said Prat.

On the basis of the PAMELA results, “We believe that the studies evaluating the long-term survival outcomes of chemotherapy-free dual HER2 blockade seem justified after selecting patients based on variables such as intrinsic subtyping, HR status, response at week 2, and pathological response at surgery,” Prat said.

Global gene expression analysis has shown that all the intrinsic molecular subtypes of breast cancer can be identified within clinicallyHER2-positive breast cancer, including theHER2-enriched, luminal B, luminal A, and basal-like subtypes. In addition, intrinsic subtypes are not fully recapitulated by HR status.

Moreover, protein expression data coming from the Cancer Genome Atlas has identified a group of breast tumors that are both clinicallyHER2-positive andHER2-enriched by intrinsic subtyping, and that these breast cancers have the highest expression of the proteinsHER2and phospho-HER2, andEGFRand phospho-EGFR, “suggesting that within clinicallyHER2-positive disease, tumors ofHER2-enriched subtype are likely to have the highest activation of the HER2/EGFR signaling pathway,” Prat said. With this background, the PAMELA trial was designed.

The primary goal of PAMELA was to identify profiles predictive of clinical benefit from targeted therapy, using the PAM50 intrinsic subtype predictor, compared with traditional clinicalHER2classification. Participants were recruited from 19 sites, and had untreated stage I-IIIA primary HER2-overexpressing and/or amplified breast tumors >1 cm. Patients were treated with dual HER2 blockade consisting of lapatinib, 1,000 mg/day, and trastuzumab, 6 mg/kg every 3 weeks, for 18 weeks. HR-positive patients were also treated with endocrine therapy (letrozole or tamoxifen) according to their menopausal status, for the same 18 weeks.

Intrinsic subtyping was performed at baseline and at week 2, and a mandatory ultrasound was performed at week 6. With progressive disease—defined as any increase in tumor size—paclitaxel was added to dual HER2 blockade at 80 mg/m²weekly for 12 weeks, with trastuzumab maintained at the original dosage and the lapatinib dosage reduced to 750 mg/day for safety reasons. If the patient was receiving endocrine therapy prior to progression, it was withdrawn before chemotherapy was administered.

Of the 151 patients enrolled, 77 were HR-positive, of whom 88.3% completed treatment, and 74 were HR-negative, of whom 93.2% completed treatment. About 60% of women were postmenopausal at baseline (mean age: 55 years), 92% had either stage I or stage II tumors, and 65% were node-negative. The median tumor size was 24 mm. Of the patients with HR-positive disease, 48% received letrozole and 52% received tamoxifen. The primary outcome measure was pCR at the time of surgery, predicted by PAM50HER2-enriched subtype.

At baseline, 66.9% of all tumor tissue samples wereHER2-enriched. Luminal tumors were identified in 49% of HR-positive disease, and basal-like subtype was identified in 12.2% of HR-negative samples. However, a substantial proportion ofHER2-enriched tumors were of theHER2-enriched subtype by gene expression in both HR-positive samples (49.3%) and HR-negative samples (85.1%).

The pCR rate was 40.6% in women withHER2-enriched tumors versus 10.0% in the non-HER2-enriched (bivariate analysis overall response (OR): 4.5;P= .005). When the axilla were included, the pCR rate was 34.7% in the women with HER2-enriched tumors and 10% in those with non-HER2-enriched tumors. In the non-HER2-enriched group, the pCR rate was 0% in the 22 patients with luminal A subtype tumors, 12.5% in the 16 patients with tumors of the luminal B subtype, 11.1% in the 9 patients with basal-like subtype, and 66.7% among the 3 patients with normal-like subtype. In addition, tumors that converted to normal-like 2 weeks after starting dual HER2 blockade were more likely to experience a pCR, compared with those that remained non-normal-like.

The pCR rates were 18.2% in patients with HR-positive breast cancer and 43.1% in those with HR-negative disease (bivariate OR: 2.3;P= .038).

The pCR rates in the breast according to changes in intrinsic subtype between baseline and week 2 were also assessed. The distribution of normal-like tumors increased from 2.1% at baseline to 48.6% at week 2. “Those tumors that became normal-like at week 2 had a probability of achieving a pCR of 48.6% versus 13.5% in those tumors that are called non-normal, with a delat od 35.1%,” said Prat. “The fact that we see more normals at week 2, and that normals at week 2 are associated with more pCRs, suggested that this is a biomarker of early tumor response.” As evidence, mean tumor cellularity was low, at a mean of 7.1%, in the samples that were normal-like, compared with 37% in the non-normal-like samples.

Reference:

Prat A, Cortes J, Pare L, et al. PAM50 intrinsic subtype as a predictor of pathological complete response following neoadjuvant dual HER2 blockade without chemotherapy in HER2-positive breast cancer: First results of the PAMELA clinical trial. Presented at: 2016 San Antonio Breast Cancer Symposium; December 6-10, 2016; San Antonio, TX. Abstract S3-03.