Phase 2 data showed atezolizumab plus bevacizumab led to limited responses in patients with metastatic colorectal cancer and an MSI-like gene signature.
Treatment with atezolizumab (Tecentriq) plus bevacizumab (Avastin) generated responses in patients with chemotherapy-resistant metastatic colorectal cancer (mCRC), but these responses were not upheld in a subgroup of patients with a microsatellite instable (MSI)–like gene signature, according to findings from a multicenter, single-arm, open-label, phase 2 trial (NCT0298269).1
Efficacy was assessed in the intention-to-treat population and showed an objective response rate (ORR) of 38.6% (95% CI, 24.4%-54.5%; P = .053). Responses were enriched in patients with MSI-real tumors by standard diagnostic assays, with an ORR of 65.2% (95% CI, 42.7%-83.6%) compared with 9.5% (95% CI, 1.2%-30.4%) among patients with microsatellite stable (MSS) tumors (P = .001). The presence of liver metastases was associated with lower response rates, with an ORR of 50.0% (95% CI, 31.3%-68.7%) in patients without liver metastases vs 14.3% (95% CI, 1.8%-42.8%) in those with liver metastases; across both strata, MSI-real tumors consistently achieved higher ORRs than MSS tumors. The ORR was 15.4% (95% CI, 1.9%-45.4%) in patients with MSS tumors without liver metastases.
“As expected, the response to atezolizumab immunotherapy combined with the anti-VEGF agent bevacizumab in patients with mCRC harboring MSI ascertained by conventional polymerase chain reaction [PCR] or immunohistochemistry [IHC]-based methods was satisfactory,” lead study author Maria Elena Elez Fernandez, MD, PhD, and coauthors expressed in their discussion of the results. “However, this benefit did not transpose to a broader set of patients ascertained by a gene expression classifier that identifies tumors harboring downstream consequences of deficient DNA mismatch repair mechanisms. The response within the subgroup of MSS CRCs was small and comparable to that reported in a recent systematic review.”
Elez is a medical oncology consultant at Vall d’Hebron University Hospital and senior investigator at Vall d’Hebron Institute of Oncology in Barcelona, Spain.
This multicenter, single-arm, open-label phase 2 trial evaluated atezolizumab plus bevacizumab in patients with chemotherapy-refractory mCRC classified as MSI-like by a gene expression signature derived from formalin-fixed, paraffin-embedded primary tumor tissue using Agendia microarray analysis. Patients were enrolled across seven European sites (Spain, Italy, the Netherlands, and Belgium) within the European Commission–funded MoTriColor program and could have had MSS or MSI status by standard diagnostic assays (PCR and IHC).2 Treatment consisted of atezolizumab 1200 mg plus bevacizumab 7.5 mg/kg every 21 days and continued until disease progression, unacceptable toxicity, or withdrawn consent.
Key eligibility criteria included being at least 18 years or older and having histologically confirmed, unresectable mCRC with at least 1 measurable lesion per RECIST 1.1 criteria, progression or relapse following at least 1 prior chemotherapy line in the metastatic setting, MSI-like gene expression signature, adequate hematologic and organ function, life expectancy greater than 12 weeks, and World Health Organization performance status of 1 or less. Patients with prior exposure to CD-137 agonists, anti–CTLA-4, anti–PD-1, anti–PD-L1, or other immune pathway–targeting agents were excluded.
The primary end point was investigator-assessed ORR per RECIST 1.1, with the design assuming a minimum response rate of 25% and a target efficacy rate of 45%. Secondary end points included progression-free survival (PFS), overall survival (OS), time to response (TTR), and duration of response (DOR), all assessed by RECIST 1.1, as well as safety and tolerability measured by the incidence and severity of adverse effects (AEs) per NCI CTCAE 4.0.
In the MSI-real cohort (n = 24), the median age was 64.0 years (IQR, 57.3-73.0). Sex distribution was balanced, with 12 male patients (50.0%) and 12 female patients (50.0%). Median height was 167.0 cm (IQR, 161.0-177.5), with height missing for 1 patient.1 The median time from diagnosis to study entry was 18.2 months (IQR, 14.2-31.0). Primary tumors originated in the right colon (n = 17; 70.8%), followed by the left colon (n = 3; 12.5%), rectum and/or sigmoid (n = 3; 12.5%), and transverse colon (n = 1; 4.2%). Histologic grade was low (grade 1/2) in 10 patients (47.6%) and high (grade 3/4) in 10 (47.6%); grade was not evaluable in 1 patient (4.8%), with 3 patients missing data on grade. Liver metastases were present in 6 patients (25.0%). Prior lines of chemotherapy included none in 3 patients (12.5%), 1 in 15 (62.5%), and 2 or more in 6 (25.0%); 13 patients (54.2%) had received prior bevacizumab.
In the MSS cohort (n = 21), the median age was 62.0 years (IQR, 59.0-73.0). The majority of patients were male (n = 11; 52.4%) and 10 were female (47.6%). Median height was 168.0 cm (IQR, 158.0-175.0). The median time from diagnosis to study entry was 26.1 months (IQR, 23.9-39.1). Primary tumor location was most frequently the right colon (n = 10; 47.6%), followed by the rectum and/or sigmoid (n = 7; 33.3%) and left colon (n = 4; 19.1%); no transverse colon primaries were reported. Histologic grade was low (grade 1/2) in 11 patients (61.1%) and high (grade 3/4) in 4 (22.2%); grade was not evaluable in 3 patients (16.7%), with 3 patients missing data on grade. Liver metastases were present in 8 patients (38.1%). Prior lines of chemotherapy included 1 in 5 patients (23.8%) and 2 or more in 16 (76.2%); 17 patients (81.0%) had received prior bevacizumab.
Across the overall study population, the median PFS was 6.4 months (95% CI, 4.1-21.2) and the median OS was 26.1 months (95% CI, 9.1-not estimable [NE]). Outcomes differed significantly by MSS. Patients with MSI-real tumors achieved a markedly longer median PFS of 23.2 months (95% CI, 8.1-NE) vs 4.0 months (95% CI, 2.0-8.1) in patients with MSS tumors (P < .001). MSI-real patients also experienced significantly longer OS vs their MSS counterparts (P < .001).
The presence of liver metastases was associated with substantially worse PFS and OS across both MSI subgroups. Among responders, median TTR was 2.1 months (95% CI, 2.0-5.1) in the MSI-real subgroup vs 9.8 months (95% CI, 2.2-NE) in the MSS subgroup. The DOR was longer in patients with MSI-real disease (median not reached) vs those with MSS disease (median, 5.9 months; 95% CI, 5.9-NE), although these between-group differences were not statistically significant for TTR (P = .200) or DOR (P = .400).
From a safety standpoint, 97.8% of patients experienced AEs, most commonly fatigue (48.9%), hypertension (28.9%), abdominal pain (24.4%), and diarrhea (22.2%). Grade 3 or higher AEs occurred in 48.9% of patients and were attributed to atezolizumab in 11.1% of cases and to bevacizumab in 22.2%. Two grade 5 AEs were reported: 1 MSI-real patient died from colonic hemorrhage considered possibly related to bevacizumab, and 1 MSS patient died from severe acute respiratory distress deemed unrelated to either study drug.
