NEW YORK – At the European Society for Medical Oncology Congress on Monday, researchers identified new biomarkers predictive of response to neoadjuvant atezolizumab (Genentech's Tecentriq) plus chemotherapy in triple-negative breast cancer.
The study, led by Giampaolo Bianchini, an oncologist at the San Raffaele Hospital Institute of Hospitalization and Scientific Care in Italy, compared several tests to determine their ability to predict a pathologic complete response, including two tests developed by Oncocyte: a 27-gene immuno-oncology, or IO, score and the 101-gene TNBCtype algorithm, along with tumor intrinsic and extrinsic gene signatures.
They found that certain biomarkers that can be identified pretreatment were associated with complete response on neoadjuvant atezolizumab plus chemo. They also identified biomarkers that appeared to predict how quickly patients would respond. Bianchini noted that these biomarkers could be used to identify patients for treatment de-escalation.
The researchers conducted this analysis on patients enrolled in the 258-patient Phase III NeoTRIP trial, which compared atezolizumab plus chemotherapy against chemotherapy alone in the adjuvant setting. They analyzed RNA sequencing results from pretreatment tissue samples from these patients.
The IO score test, called DetermaIO, measures the presence of subtypes of infiltrating inflammatory cells and the presence or absence of a differentiated stromal microenvironment. The test gives patients either a positive or negative IO score based on the analysis of the tumor microenvironment. The algorithm then is able to predict response to immune checkpoint inhibitors, like atezolizumab.
The researchers found that when assessed pretreatment, the IO score was predictive of complete response in the atezolizumab arm but not in the chemotherapy-alone arm. They found that 43 percent of patients had a positive IO score, and a positive IO score was associated with a 20 percent increase in complete response in patients treated with atezolizumab.
The researchers also analyzed patients based on the TNBCtype algorithm. The six TNBC subtypes were initially identified in 2011 based on a 2,188-gene algorithm. In 2016, researchers further narrowed that algorithm to 101 genes. The TNBCtypes algorithm could also be used to predict response to neoadjuvant chemotherapy. Oncocyte has also developed TNBCtypes assay based on the 101-gene algorithm.
The six subtypes were characterized by certain gene or immune expression. BL1 and BL2 are basal-like types driven by cell cycle and DNA damage response genes; M and MSL are mesenchymal types characterized by genes involved in cell differentiation and growth factor pathways; the IM, or immunomodulatory, subtype is driven by immune-related genes; and the LAR subgroup is defined by androgen signaling.
When Bianchini's group used TNBCtype, they found that among the six subtypes, patients with the LAR subtype, which is defined by high androgen receptor expression, had the lowest complete response rate, 22 percent for atezolizumab and 19 percent for chemo. Patients with the BL1 subtype, characterized by abnormal expression of cell cycle-regulating and DNA repair-related genes, had the best response with a 70 percent complete response rate in the atezolizumab arm and 54 percent in the chemo arm. However, the associations with complete response were not statistically significant based on interaction tests.
The researchers also analyzed the complete response rate with 151 tumor intrinsic and extrinsic gene signatures, including the HALLMARK gene collection, NanoString Technologies' nCounter Breast Cancer 360 Panel, and others. For these signatures, researchers analyzed samples collected pretreatment and on the first day of the second cycle of treatment.
In the atezolizumab arm, three features were associated with complete response: apoptosis, miotic spindle, and B-cell memory. They also identified several features in the tumor microenvironment associated with resistance to chemotherapy, including angiogenesis, stromal markers, adipogenesis, epithelial-mesenchymal transition, androgen response, fatty acid metabolism, and cholesterol homeostasis. For three of those features — angiogenesis, cholesterol homeostasis, and fatty acid metabolism — treatment with atezolizumab increased the complete response rate by about 20 percent.
The researchers also explored the biological reasons that may be enabling some patients to respond quickly to atezolizumab, a group of super-responders. Based on the pretreatment samples, they found these super-responders tended to have activated immune-related pathways, indicating a more inflamed and immunologically active microenvironment, along with high immune exhaustion signaling, Bianchini said.
"We observed these very quick responses, two times more frequently on atezolizumab [than chemotherapy]," Bianchini said. "This is related to the mechanism of action of atezolizumab and demonstrates that these patients who [respond quickly] after just one cycle are strongly enriched for characteristics that [make them] very likely to benefit the most from immune checkpoint [inhibition]. It makes a lot of sense to target this group for a de-escalation strategy."
Bianchini hopes to identify similar predictive biomarkers based on liquid biopsy samples. There are already several predictive biomarkers that can be identified with liquid biopsy sequencing tests, like T-cell receptor and clonality valuation, he said.
However, for a certain subgroup of patients who have low complete response rates, there may not be a liquid biopsy test that can fully assess all biomarkers to identify the best treatments, he added.
"The effort to move biomarker [sequencing] from tissue to liquid is good, but to understand the biology and potentially identify new drug combinations for patients in need, you cannot do these in liquid [biopsy] alone, you need both liquid and tissue," Bianchini said.