GENEVA, SWITZERLAND—Overall survival (OS) was improved across all subgroups of patients with non-small cell lung cancer (NSCLC) who received atezolizumab compared with chemotherapy, including poor prognostic factors that were evaluated in an analysis of data from the OAK trial. Additionally, this study observed no evidence of quicker disease progression was observed with the PD-L1 monoclonal antibody atezolizumab compared to chemotherapy, suggesting this may not be a phenomenon solely related to cancer immunotherapies as has been previously suggested. These findings were presented at the 2018 ESMO Immuno-Oncology Congress (Abstract LBA1).
David R. Gandara, MD, Director of the Thoracic Oncology Program at the UC Davis Comprehensive Cancer Center in Sacramento, presented data on behalf of colleagues from a study designed to determine whether fast progression was a factor in patients treated with atezolizumab in the OAK trial (NCT02008227). This analysis used fast progression, which was defined as an increase of 50 percent or more in the sum of the longest diameters of tumors (SLD) from baseline to first assessment at 6 weeks or upon death due to disease progression, as evaluated by the investigator within 12 weeks of treatment without a post-treatment scan.
Fast progression was used as a surrogate for the rare phenomenon, hyper-progressive disease (HPD), which has emerged in some patients treated with immunotherapy in clinical trials. HPD is characterized by rapid tumor growth rate and accelerated disease progression following PD-L1 checkpoint inhibitor therapy. Assessment of HPD requires an evaluation of pre-treatment tumor growth rates, therefore the surrogate of fast progression was used in this study. HPD has been associated with age more than 65 years (Clin Cancer Res 2017;23(8):1920-1928), and epidermal growth factor receptor (EGFR) mutation (Clin Cancer Res 2017;23:4242-4250), and is characterized by a poorer OS.
OAK Study Details
The phase III OAK study compared checkpoint inhibition with atezolizumab versus chemotherapy with docetaxel in the second- and third-line settings in 850 patients with NSCLC. In OAK, superior median OS of 13.8 months was observed with atezolizumab compared to 9.6 months with docetaxel (HR 0.73) in the intent to treat population.
To examine fast progression as a surrogate for HPD, the investigators first categorized fast progressors in the two treatment arms. Similar numbers of patients in both treatment arms met the criteria for fast progression and/or HPD.
The investigators then reviewed the baseline characteristics of fast progression and non-fast progression patients in OAK for selected pre-treatment factors that could be correlated with fast progression for a comparison between the atezolizumab and docetaxel treatment arms. The OS per arm was evaluated according the selected factors, which included early treatment failure with the previous therapy, lactate dehydrogenase (LDH) levels ≥225 units/L, SLD ≥80 mm, and ≥3 metastatic sites.
As noted, the proportion of patients having the selected baseline factors for fast progression was equivalent in the two treatment arms; 44 (10.4%) patients on atezolizumab compared to 41 (9.6%) patients on docetaxel met the criteria for fast progression. The overall baseline characteristics were generally similar between arms except the atezolizumab arm contained more male patients, more non-smokers, and fewer (10 [31%] versus 15 [44%], atezolizumab versus docetaxel) patients that had demonstrated early failure to prior treatment. In both arms, the majority of patients had been heavily pre-treated.
Regarding the cohort of patients having baseline indicators for fast progression, 24 (55%) patients in the atezolizumab arm versus 21 (51%) patients in the docetaxel arm had baseline LDH ≥225 units/mL, and 21 (48%) versus 18 (44%) patients had baseline SLD ≥80 mm, respectively. In the respective cohorts, 32 (73%) versus 28 (68%) patients had three or more metastatic sites at baseline and 10 (31%) versus 15 (44%) had experienced failure to their previous treatment within 180 days of initiation.
In addition, 21 (48%) versus 20 (49%) patients were aged ≥65 years, and three (7%) versus two (5%) patients were EGFR mutation positive, therefore, having factors associated with HPD. Another variable associated with poor prognosis is PD-L1 expression on tumour or immune cells; PD-L1 was expressed in 25 (57%) patients on atezolizumab versus 24 (59%) of patients on docetaxel, and PD-L1 was as not expressed in 19 (43%) versus 17 (42%) of patients, respectively.
OS Favored Atezolizumab
An evaluation of OS according to each pre-treatment risk factor for fast progression indicated median OS was improved overall with atezolizumab. Median OS was 8.9 months with atezolizumab versus 6.2 months with docetaxel in patients showing early progression on prior treatment. Patients with high baseline LDH levels had median OS of 11 versus 8.9 months, median OS was 9.4 versus 6.9 months in patients with high SLD, and patients with ≥3 metastatic sites demonstrated median OS of 11.7 versus 8.6 months with atezolizumab versus docetaxel, respectively.
Twenty (45%) patients receiving atezolizumab and 12 (29%) patients on docetaxel demonstrated an SLD change ≥50 percent from baseline within 6 weeks. In the respective treatment arms, 24 (55%) versus 29 (71%) patients died due to disease progression within 12 or fewer weeks of treatment.
Equivalent proportions of patients in each treatment arm experienced fast progression, suggesting that rapid post-baseline progression is not specific to atezolizumab or anti–PD-L1 therapy in general. The superior benefit with atezolizumab over docetaxel was consistent across all subgroups defined by baseline factors associated with aggressive disease, according to the investigators. Further study is required to understand the underlying biological mechanisms of why certain patients experience tumor growth acceleration, as well as consistent means to measure this phenomenon, especially in the era of CIT combination therapies.