Emerging research is suggesting that outcomes from neoadjuvant chemotherapy may be correlated with two genetic mutations that are common in breast cancer—PIK3CA and TP53. Their presence may affect response to treatment, and mutational shift after treatment may affect survival, according to studies presented at the 2013 San Antonio Breast Cancer Symposium.

GeparSixto Study

Women with HER2-positive, hormone receptor–positive tumors with mutations in the PI3K/AKT pathway responded poorly to neoadjuvant therapy in the German Geparsixto study reported at the meeting.¹

The PIK3CA mutation is the second most common in breast cancer, observed in about 20% of tumors, and alterations in this pathway are associated with resistance to anti-HER2 agents. But the data have been discrepant with regard to its prognostic or predictive value, especially in HER2-positive tumors. The researchers, therefore, investigated the frequency and prognostic associations of PIK3CA mutations in HER2-positive and triple-negative primary breast cancer treated with neoadjuvant therapy.

The GeparSixto study was a prospective analysis of participants of the GeparSixto and GeparQuinto neoadjuvant trials. Patients received either trastuzumab (Herceptin) or lapatinib (Tykerb) in the GeparQuinto trial, or both inGeparSixto, in addition to 18 to 24 weeks of chemotherapy.

Mutational analysis (exons 9 and 20) of tissue samples was available on 360 women with HER2-positive tumors and on 285 women with triple-negative tumors. Within these groups, the PIK3CA mutation was identified in about 20% and 7%, respectively. By hormone receptor status in the HER2-positive patients, mutations were detected in about 21% of both receptor-negative and receptor-positive patients.

“We found that very few women with HER2-positive and hormone receptor–positive breast cancer with a PIK3CA mutation experienced a pathologic complete response after neoadjuvant therapy,” said Sibylle Loibl, MD, Associate Professor at the University of Frankfurt and investigator with the German Breast Group in Neu-Isenburg, Germany.

PIK3CA Mutation Associated With Less Response

The combined analysis from both GeparQuinto and GeparSixto showed that, in HER2-positive patients, pathologic complete response rates after dual HER2 blockade were significantly lower in the PIK3CA mutant group compared to patients with wild-type ­PIK3CA (17% vs 37%). In HER2-positive patients who were hormone receptor–positive and harbored a PIK3CA mutation, only 6.3% of patients achieved pathologic complete response, Dr. Loibl reported.

“By hormone receptor status, this effect was exclusively confined to the hormone receptor–positive subgroup,” she said. “There was no difference in the hormone receptor–negative group.”

“The difference in [pathologic complete response] rate between mutant and wild-type patients was largest among patients who received double HER2 blockade,” she added. “The [pathologic complete responses] were numerically higher without the mutation, and this seems confined to patients receiving trastuzumab. There was no difference in the lapatinib-treated cohort, who had the same [pathologic complete response] rate irrespective of mutation status,” she said.

“The results are in concordance with NeoALTTO and Neosphere,” she noted. “Patients with PIK3CA-mutant HER2-positive and hormone receptor–positive breast cancer are resistant to chemotherapy and dual anti-HER2 treatment, and other treatment options need to be tested in this group.”

Dr. Loibl and her team are now conducting the randomized phase II neoadjuvant NeoPHOEBE trial of the pan-PI3K inhibitor buparlisib.

Loss of PIK3CA and TP53 Mutations

Chinese investigators reported in San Antonio that the loss of TP53 and PIK3CA mutations after neoadjuvant chemotherapy heralds a favorable prognosis.² Small studies have shown these mutations can be lost after neoadjuvant chemotherapy, and the researchers hypothesized that this “genetic evolution” could impact prognosis.

“We hypothesized that loss of TP53 and PIK3CA mutation status might be a common phenomenon after neoadjuvant chemotherapy and could serve as a prognostic biomarker,” said Zhi-Ming Shao, MD, of Fudan University Shanghai Cancer Center.

The study included two neoadjuvant cohorts: 206 locally advanced breast cancer patients in the training cohort and 158 in the validation cohort. All received four cycles of weekly paclitaxel and carboplatin, and underwent mutational analysis by exome sequencing before and after neoadjuvant treatment. A third cohort included 81 patients with prior surgical resection who received adjuvant chemotherapy, and who underwent mutational analysis of microdissected tumor foci.

The study excluded patients who achieved a pathologic complete response, since this could confound the association with survival; instead, it correlated mutation status with partial response (Miller-Payne grade 3 or 4).

“We aimed to identify changes in mutation status, and the association between mutational loss and chemotherapy response and survival,” Dr. Shao said.

Favorable Prognostic Sign

At baseline, the distribution of the mutations differed by breast cancer subtype. TP53 mutations were most common in basal-like tumors, while PIK3CA mutations were most frequent in luminal tumors. Mutation status for either gene had no influence on patient survival, but loss of the mutation did, Dr. Shao reported.

After neoadjuvant chemotherapy, mutation rates overall decreased significantly in both cohorts, from approximately 25% to approximately 12% (P < .001). Loss of mutation was significantly associated with better pathologic response.

In cohort 1, a partial response or greater was observed in 77.8% of patients who shifted from mutated status to wild-type, vs 45.9% who did not shift or who shifted from wild-type to mutated status. Similarly, in cohort 2 the response rates were 80% and 50.8%, respectively (P < .001 for both).

Furthermore, mutation shift, from mutated to wild-type, was associated with better disease-free survival (P = .033) and overall survival, which exceeded 90% at 5 years (P = .045). By Cox proportional hazards model, a shift from mutation to wild-type, vs remaining wild-type, conferred hazard ratios of 0.62 (P = .004) in the training set and 0.72 (P = .011) in the validation set, Dr. Shao reported.

Genetic analysis of cohort 3 revealed strong intratumoral heterogeneity in TP53 or PIK3CA mutations in 28% of the samples. “Both mutant and nonmutant cancer cells coexist in the same tumor. This partially explains the genetic basis of the mutation shift after chemotherapy,” he said.

Dr. Zhao noted that chemotherapy sensitivity differs between mutant and wild-type cells. Loss of mutations may make cells more sensitive to chemotherapy, conveying a favorable prognosis. Patients with mutated status who remain mutated are more resistant to chemotherapy and have a worse prognosis, he suggested.

Ben Park, MD, PhD, Associate Professor of Oncology at Johns Hopkins, who moderated a poster session on PI3K pathway mutations, commented that “the next challenge will be to understand how tumor biology changes with tumor progression, or over the course of therapy, and how to develop methods to understand the whole cancer burden so that we can treat the clones that will be the bad actors.” ■

Disclosure: This work was conducted with a grant from the EU FP 7 Programme. Drs. Loibl and Park reported no potential conflicts of interest.

References

1. Loibl S, Denkert C, Schneeweis A, et al: PIK3CA mutation predicts resistance to anti-HER2/chemotherapy in primary HER2-positive/hormone-receptor-positive breast cancer—prospective analysis of 737 participants of the GeparSixto and GeparQuinto studies. 2013 San Antonio Breast Cancer Symposium. Abstract S4-06. Presented December 12, 2013.

2. Shao Z-M, Jiang Y, Yu K-D: Exome sequencing identifies shift in TP53 and PIK3CA mutation status after paclitaxel-based neoadjuvant chemotherapy in breast cancer. 2013 San Antonio Breast Cancer Symposium. Abstract S4-05. Presented December 12, 2013.