In a phase II study reported in The Lancet Oncology, Shu et al found that neoadjuvant treatment with atezolizumab plus nab-paclitaxel/carboplatin produced a major pathologic response in 57% of patients and a complete pathologic response in 33% of patients with resectable, predominantly stage IIIA non–small cell lung cancer (NSCLC).
The authors noted that studies of neoadjuvant chemotherapy alone have shown pathologic complete response rates of 5% to 8% in this setting.
The investigator-initiated trial enrolled 30 patients at Columbia University Medical Center, Massachusetts General Hospital, and Dana-Farber Cancer Institute with stage IB to IIIA disease, Eastern Cooperative Oncology Group performance status of 0 or 1, and a history of smoking. Patients received atezolizumab at 1,200 mg on day 1; nab-paclitaxel at 100 mg/m² on days 1, 8, and 15; and carboplatin at AUC5 on day 1 of 21-day cycles. Patients without disease progression after two cycles received two additional cycles followed by surgical resection. The primary endpoint was major pathologic response (defined as ≤ 10% of residual viable tumor at surgery) in the intention-to-treat population (n = 30).
Patients had a median age of 67 years (range = 62–74 years); 15 (50%) were women; 17 (57%) had adenocarcinoma, 12 (40%) had squamous cell carcinoma, and 1 had large-cell neuroendocrine carcinoma; disease stage was IIA in 4 (13%), IIB in 3 (10%), and IIIA in 23 (77%); and programmed cell death ligand 1 (PD-L1) expression (per tumor proportion score) was ≥ 50% in 8 (27%), ≥ 1% in 16 (55%), < 1% in 12 (40%), and unknown in 2 (7%).
Surgery and Responses
No treatment-related surgical delays occurred, with the median time between the last administration of systemic chemotherapy or atezolizumab being 26.5 days. Of the 30 patients, 29 were brought to the operating room with the intention of surgery, and 26 (87%) patients had an R0 resection. Among the four patients not receiving R0 resection, one developed brain metastases during neoadjuvant therapy and was ineligible for surgery, and three had disease considered to be unresectable at exploration.
Among all 30 patients, 23 (77%) received treatment on at least day 1 of all four cycles, with 29 (97%) completing three or more cycles. Five patients (17%) went to surgery early due to treatment-related adverse events (myelosuppression in 4, and neuropathy in 1).
Major pathologic response was observed in 17 (57%, 95% confidence interval [CI] = 37%–75%) of the 30 patients, and pathologic complete response was observed in 10 (33%, 95% CI = 17%–53%). Of the 10 patients with a pathologic complete response, 6 had stage IIIA disease at baseline. Among the patients with R0 resection, those with a major pathologic response included 8 (53%) of 15 with adenocarcinoma (5 [33%] with pathologic complete response) and 8 (80%) of 10 with squamous cell carcinoma (5 [50%] with pathologic complete response).
The median pathologic response among the patients who underwent R0 resection was −92.5%. Analysis by PD-L1 expression status of < 50% vs ≥ 50% showed major pathologic response in 10 (63%; pathologic complete response in 5) of 16 patients vs 6 (75%; pathologic complete response in 4) of 8.There were no significant associations between major pathologic response or pathologic complete response and PD-L1 expression using a ≥ 50% cutoff (P = .67).
On Response Evaluation Criteria in Solid Tumors, 19 (63%) of the 30 patients had objective responses (all partial responses) and stable disease was observed in 9 patients (30%), with objective responses observed irrespective of PD-L1 status. Median best change in tumor size was −34% in patients with PD-L1 expression of < 1% and −40% in those with PD-L1 expression of ≥ 1% (P = .18). Two patients (7%) had suspected disease progression during treatment.
At a median follow-up of 12.9 months from the first day of treatment, 19 patients—representing 63% of all 30 patients and 73% of 26 who had undergone R0 resection—were alive and had no evidence of disease. Disease recurrence was observed in nine (30%), including the four patients (13%) who did not undergo resection.
In the entire population, median disease-free survival was 17.9 months (95% CI = 14.3 months–not reached) and median overall survival was not reached (95% CI = 27.6 months–not reached). Post hoc analyses showed a median disease-free survival of 14.3 months vs 34.5 months in patients without vs with a major pathologic response.
The most common treatment-related adverse events of any grade among all 30 patients included neutropenia (26 patients [87%]), anemia (21 [77%]), thrombocytopenia (19 [63%]), fatigue (17 [57%]), alopecia (14 [47%]), and nausea (13 [43%]). The most common treatment-related grade 3 or 4 adverse events were neutropenia (15 [50%]), increased alanine aminotransferase (2 [7%]), increased aspartate aminotransferase (2 [7%]), and thrombocytopenia (2 [7%]).
Serious treatment-related adverse events consisted of grade 3 febrile neutropenia in one patient (3%), grade 4 hyperglycemia in one patient (3%), and grade 2 bronchopulmonary hemorrhage in one patient (3%). No treatment-related deaths were reported.
The authors concluded, “Atezolizumab plus carboplatin and nab-paclitaxel could be a potential neoadjuvant regimen for resectable non–small cell lung cancer, with a high proportion of patients achieving a major pathological response, and manageable treatment-related toxic effects, which did not compromise surgical resection.”
Naiyer A. Rizvi, MD, of the Division of Hematology/Oncology, Columbia University Irving Medical Center, is the corresponding author for The Lancet Oncology article.
Disclosure: The study was funded by Genentech and Celgene. For full disclosures of the study authors, visit thelancet.com.