Abstract
Background/Aim: The PACIFIC trial demonstrated improved survival in patients with unresectable stage III non-small cell lung cancer (NSCLC) treated with durvalumab following definitive concurrent chemoradiotherapy (CRT). This study sought to explore real-world outcomes with durvalumab consolidation therapy at our institution. Patients and Methods: We retrospectively identified patients diagnosed with stage III NSCLC at our institution from January 2012 to January 2022. We created two cohorts: one who received durvalumab following definitive CRT and a historical one who did not. Primary outcomes of interest included median progression-free survival (PFS) and overall survival (OS). Additionally, we performed subgroup analysis on the durvalumab cohort to explore the associations between survival and time to durvalumab initiation, PD-L1 expression, and neutrophil-to-lymphocyte ratio (NLR). Results: We identified 79 patients with locally advanced NSCLC who were not surgical candidates. Patients treated with durvalumab (n=44) had significantly improved survival compared to the historical cohort (n=35) including a median PFS of 17.4 months versus 8.0 months (p=0.0019) and a median OS of 37.0 months versus 17.0 months (log-rank p-value=0.07, Wilcoxon p-value=0.02). Within the durvalumab group, outcomes did not significantly differ between those who initiated therapy before or after 42 days of finishing CRT, between various PD-L1 expression levels, or between high or low NLR. Conclusion: Patients who received durvalumab as consolidation therapy following definitive CRT demonstrated significantly improved survival compared to a historical cohort who did not receive durvalumab. Furthermore, durvalumab appears to benefit patients regardless of time to initiation, PD-L1 expression, or NLR.
The management of stage III NSCLC is controversial. While a multimodality approach is recommended, the exact sequence and schedule of each modality, including surgery, radiation and/or chemotherapy, are not yet resolved (1). Historically, the standard of care for unresectable stage III non-small cell lung cancer (NSCLC) has been concurrent chemoradiotherapy (CRT). However, the prognosis remained poor with a five-year survival rate of approximately 15% (2). The PACIFIC trial demonstrated that treatment with durvalumab, a programmed death ligand 1 (PD-L1) immune checkpoint inhibitor, following concurrent CRT led to improved progression-free survival, overall survival, and median time to death or metastasis compared to patients who did not receive durvalumab (3-5). As a result, current guidelines recommend durvalumab as consolidation therapy after definitive concurrent CRT in patients with unresectable stage III NSCLC (6).
Though promising, these results stem from a clinical trial with tightly regulated parameters and strict inclusion and exclusion criteria. For instance, patients were excluded if they had a World Health Organization (WHO) performance status greater than 1 and were only allowed to start durvalumab within 42 days of finishing CRT. Therefore, real-world evidence is required to validate the practicality, effectiveness, and utility of this treatment compared to other options. Given the recency of adjuvant durvalumab approval as treatment, such reports are scarce and limited (7-13). Moreover, most of these reports were not able to adequately comment on overall survival due to the nascency of their data.
Therefore, we sought to analyze the available data at our own institution regarding outcomes of durvalumab therapy. We hypothesized that the cohort at our institution who received adjuvant durvalumab will demonstrate improved outcomes compared to a historical cohort treated with concurrent CRT alone. Furthermore, we expected that this cohort will consist of many patients who would have been ineligible for the PACIFIC trial, yet nonetheless will similarly show such outcomes. Given that our institution implemented durvalumab treatment nearly five years ago, we sought to present a significant amount of longitudinal data to illustrate the impact of durvalumab on progression-free and overall survival.
Patients and Methods
We used the integrated Cancer Repository for Cancer Research (iCaRe2) registry to identify patients diagnosed with stage III NSCLC at our institution from January 2012 to January 2022. iCaRe2 is a multi-institutional resource that collects longitudinal data on consented adult cancer patients (14). We received institutional board review (IRB) approval to investigate medical records for information, such as age, sex, diagnosis dates, cancer stage, cancer histology, smoking status, Eastern Cooperative Oncology Group (ECOG) performance status, clinic visits, laboratory and imaging data, and treatment dates and regimens. Besides requiring a new diagnosis of stage III NSCLC, patients had to be at least 19 years old at the date of diagnosis for inclusion. Patients were excluded if they demonstrated disease progression on concurrent CRT or underwent surgery after CRT.
The study’s main outcomes of interest included median progression-free survival (PFS) and median overall survival (OS). We defined PFS as the time from last CRT treatment to radiographic disease progression, death, or last follow-up. OS was defined as the time from last CRT treatment to death or last follow-up. Survival analysis was performed using the Kaplan–Meier method with those alive and progression-free at last follow-up censored and with log rank testing used to compare outcomes. Multivariate analysis using a Cox regression model was performed to identify hazard ratios (HR) and confidence intervals (CI) for PFS and OS after adjusting for various clinicopathologic features.
The Mann-Whitney test was used to compare age, performance status, number of chemotherapy cycles, and total radiation dose. Fisher’s exact test was used to compare smoking status, cancer stage, histology, and chemotherapy regimen, and the Chi-square test was used to compare sex and incidence of radiation pneumonitis. To explore the association between PD-L1 expression and outcomes, three groups were created based on PD-L1 levels: >50%, 1-49%, and <1%. We also analyzed whether outcomes varied between those who initiated durvalumab treatment within or after 42 days following completion of CRT and thus, dichotomized data on number of days from end of CRT to durvalumab as ≤42 versus >42 days. Additionally, we investigated the association between pre-durvalumab neutrophil-to-lymphocyte ratio (NLR) by dichotomizing values into cutoffs of <3 versus ≥3 based on previous studies (15-18). All statistical tests were two-sided, and the threshold for statistical significance was set at a p-value <0.05. The statistical analyses were carried out using the SAS (SAS Institution Inc., Cary, NC, USA) Studio, R (R Core Team, 2022) and RStudio (RStudio Team, 2022).
Results
In total, 79 patients met the inclusion criteria, including 44 patients in the durvalumab cohort and 35 in the historical cohort (Table I). There were no significant differences between the durvalumab and the historical cohorts in terms of demographic characteristics or chemoradiation approaches. For the entire population, median age was 65.3 years, and 57.0% were female. Patients were most commonly staged as IIIA and typically had histologic subtypes of adenocarcinoma or squamous cell carcinoma. Additionally, the median ECOG was 1, most patients were former smokers, and the most frequent chemotherapy regimen used was carboplatin/paclitaxel. The mean radiation dose received was 6,020.8 cGy with a 25.3% incidence of radiation pneumonitis.
Patient characteristics.
Median PFS for the durvalumab cohort was 17.4 months compared to 8.0 months for the historical cohort (p=0.0019) (Figure 1A). Similarly, median OS for the durvalumab cohort was 37.0 months compared to 17.0 months in the historical cohort (Figure 1B). This difference was more prominent in the early course of treatment (log-rank p-value=0.07; Wilcoxon p-value=0.02). On multivariate analysis after controlling for age, sex, smoking status, performance status, histology, and stage, the significant impact of durvalumab on PFS and OS persisted (Table II and Table III). Our outcomes were similar to the five-year outcomes seen in the PACIFIC trial, which demonstrated a median PFS of 16.9 months versus 5.6 months in the durvalumab and placebo groups, respectively, and a median OS of 47.5 months versus 29.1 months, respectively (5).
Kaplan–Meier survival curves for durvalumab and historical cohorts. A) Progression-free survival curve for durvalumab and historical cohorts. B) Overall survival curve for durvalumab and historical cohorts.
Multivariate analysis of prognostic factors for progression-free survival (PFS).
Multivariate analysis of prognostic factors for overall survival (OS).
Finally, we performed various subgroup analyses of the durvalumab cohort. These patients received 16.4 cycles of durvalumab on average, and 25% experienced at least one immune-related adverse event (irAE) (Table I). In the PACIFIC trial, patients initiated durvalumab within 42 days of finishing CRT. In our population of 44 patients who received durvalumab, half began therapy within 42 days of finishing CRT and the other half started after. No difference in median PFS was seen between these groups (10.5 months in patients who received durvalumab <42 days after CRT versus 22.7 months in patients who received it >42 days after CRT, p=0.75) (Figure 2A). Similarly, median OS did not significantly differ between these groups (32.8 months versus 52.7 months, respectively, p=0.61) (Figure 2B).
Kaplan–Meier survival curves based on time of durvalumab initiation. A) Progression-free survival curve based on time of durvalumab initiation. B) Overall survival curve based on time of durvalumab initiation.
We also examined the effect of PD-L1 expression on outcomes in the durvalumab cohort. We separated patients into three categories based on PD-L1 expression level: >50%, 1-49%, and <1%. Of the 44 durvalumab patients, 14 did not have PD-L1 levels available, leaving 30 for analysis. An exactly even distribution between the three groups was seen (Table IV). Moreover, they did not significantly differ in median PFS or median OS. The median PFS was 19.5 months in the >50% PD-L1 level group, 17.1 months in the 1-49% group, and 19.6 months in the <1% group (p=0.72). The median OS for these groups were 26.0 months, 27.9 months, and 53.5 months, respectively (p=0.25) (Table IV). Additionally, we explored the potential impact of pre-durvalumab NLR. As described above, we stratified patients into two groups: those with an NLR <3.0 (“low NLR”) and those with an NLR ≥3.0 (“high NLR”). 21% of patients had a low NLR while 79% of patients had a high NLR. No difference was seen in median PFS or median OS between these two groups (Figure 3A and B). The median PFS in those with low NLR was 14.0 months and 17.4 months in those with high NLR, and the median OS was 31.7 months and 40.8 months, respectively.
Durvalumab cohort stratified by PD-L1 expression level and outcomes.
Kaplan–Meier survival curves based on neutrophil-to-lymphocyte ratio (NLR). A) Progression-free survival curve for patients based on NLR. B) Overall survival curve for patients based on NLR.
Discussion
The PACIFIC trial marked an important change in the landscape of treatment for patients with unresectable stage III NSCLC (3). Its findings led to the approval by the United States Food and Drug Administration (FDA) in 2018 of durvalumab as consolidation therapy for these patients following concurrent platinum-based chemotherapy and radiation therapy barring disease progression (7). However, given the controlled nature of the PACIFIC trial, real-world evidence of outcomes with durvalumab is needed, which this study aimed to delineate.
At our institution, patients who received durvalumab demonstrated improved survival outcomes compared to a demographically similar historical cohort who did not receive immunotherapy, including longer median PFS (17.4 months versus 8.0 months) and median OS (37.0 months versus 17.0 months). While the log-rank test for overall survival did not reach significance, the Wilcoxon test did, indicating a large difference between groups in early survival. These differences held true even after controlling for factors, such as age, sex, smoking status, histology, and performance status. Since the initial PACIFIC trial findings, five-year survival outcomes have been reported, which also revealed longer median PFS and OS in its durvalumab cohort compared to its placebo cohort (5). Additionally, outcomes in our durvalumab cohort were comparable to those in the PACIFIC study (median PFS and OS of 16.9 months and 47.5 months, respectively). Though almost none of the current literature has been able to fully report on overall survival due to the recency of durvalumab’s incorporation into clinical practice, Sankar et al. compared veterans who did and did not receive durvalumab and likewise found longer median PFS and OS in the former (9). Moreover, our results prove that durvalumab has effectiveness outside of the tightly controlled parameters in the PACIFIC trial. For example, 23 of our 44 (52.3%) patients who received durvalumab did not fall under the study design of the PACIFIC trial based on performance status and/or time to durvalumab initiation. When we compared patients who started durvalumab within 42 days of finishing CRT, which was the design in the PACIFIC trial, with patients who did so after 42 days, we saw no differences in PFS or OS.
Another aspect of our durvalumab subgroup analysis involved comparing outcomes based on PD-L1 expression. We stratified patients into three groups based on their PD-L1 tumor proportion scores: ≥50%, 1-49%, and <1%. We did not observe a statistically significant difference in survival between these three groups. Though higher tumor PD-L1 expression is associated with shorter survival, it has also been shown to correlate with improved treatment benefit from anti-PD-1/anti-PD-L1 agents in NSCLC (19-23). However, the impact of PD-L1 expression on response to immunotherapy remains controversial, as many studies have demonstrated no difference in survival based on PD-L1 level, including in patients who receive durvalumab (24-26). While no statistically significant difference between the three PD-L1 groups was seen in our results, those whose expression levels were <1% had the longest median OS.
Similarly, we did not identify a difference in outcomes between patients with high versus low NLR, using 3.0 as a cutoff. NLR has been used as a marker for inflammation and described in multiple studies as a prognostic marker in various cancers with higher values correlating with worse survival (27-29). While the mechanism for this association has not been fully elucidated, it is known that inflammation promotes a favorable microenvironment for cancer cell growth and spread (30, 31). We chose 3.0 as a cutoff based on previous studies (15-18). Surprisingly, patients in our study with high NLR had longer PFS and OS, though again this was not statistically significant. Also, although we used 3.0, the optimal cutoff for prognostic purposes has not been firmly established as several other values have been used (30, 32, 33).
Study limitations. As a retrospective study, it relies on accurate documentation of medical records, and we are not able to comment on any causal relationship between treatments and outcomes. Additionally, multiple patients had missing data or were lost to follow-up. Finally, since the data were obtained from a single institution, our sample size was relatively small, and a multicenter investigation would enhance generalizability.
In conclusion, the present study demonstrates that consolidation therapy with durvalumab following concurrent CRT is associated with improved survival in patients with unresectable stage III NSCLC. These findings align with those reported in the PACIFIC trial. Moreover, survival benefit was similar regardless of the time to durvalumab initiation, PD-L1 expression level, or neutrophil-to-lymphocyte ratio.
Footnotes
Authors’ Contributions
JQT – conceptualization, investigation, methodology, project administration, visualization, writing – original draft, writing – review and editing; YX – formal analysis, visualization, writing – review and editing; LMS - formal analysis, visualization, writing – review and editing; OA – resources, writing – review and editing; ASM – resources, writing – review and editing; AKG – conceptualization, methodology, project administration, resources, supervision, writing – review and editing.
Conflicts of Interest
The Authors have no conflicts of interest to declare in relation to this study.
- Received December 22, 2023.
- Revision received January 8, 2024.
- Accepted January 9, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
