In the last few years, standard treatment of un-resectable stage III non-small cell lung cancer (NSCLC) changed considerably after the publication of improved survival results with maintenance 12-month treatment with Programmed Death Ligand 1 (PD-L1) antibody durvalumab following standard concurrent chemoradiotherapy (ChT-RT).1, 2, 3 In the PACIFIC trial, the median progression-free survival (PFS) from randomization was 17.2 months in durvalumab arm versus 5.6 months in placebo arm, while median overall survival (OS) was 47.5 months
We retrospectively analysed unresectable stage III NSCLC patients (according to the 8th TNM classification) who were considered for maintenance treatment with durvalumab (intention to treat population, ITT) after radical ChT-RT and completed treatment until December 2020.7 First 61 patients were included in early access program (EAP) which started in December 2017 and ended in September 2019 when reimbursement was introduced. Afterwards patients were treated with durvalumab as a standard of care. During EAP, patients were treated with durvalumab after at least stable disease with ChT-RT was achieved regardless of PD-L1 expression level, but from September 2019, during standard of care treatment, only patients with PD-L1 ≥ 1% received adjuvant durvalumab according to European Medicines Agency (EMA) registration.
Before treatment, patients underwent a physical examination, computed tomography (CT) of the chest, abdomen and head as well as the 18 F-fluorodeoxyglucose positron emission tomography/computed tomography (18 F-FDG PET/CT) and brain magnetic resonance imaging (MRI) when indicated. All patients had histologically or cytologically confirmed NSCLC from primary tumour or regional lymph nodes, N stage mostly confirmed with endobronchial ultrasound (
Baseline characteristics of patients, including age, gender, pathological features, TNM stage, Eastern Cooperative Oncology Group performance status (ECOG PS), smoking status, PD-L1 expression, mutational status of EGFR, KRAS, ALK, ROS 1, BRAF and NTRK in adenocarcinoma, time to durvalumab start from the end of the RT, treatment completion, PFS and OS from the start of ChT-RT and the start of durvalumab were collected for the analysis. Response rate after ChT-RT was assessed using RECIST 1.1 and during immunotherapy iRECIST.8 Immune related adverse events were assessed by their highest reported grade using Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0.9
PFS was calculated from the beginning of the durvalumab to disease progression or death and OS as the time from the start of the durvalumab to death from any cause. Data from the patients who had not progressed or had not died were censored at the date of last follow-up (February 3, 2021).
The association between the PFS, OS and the basic clinicopathological variables of patients were tested using the log-rank test. OS and PFS curves were estimated using Kaplan-Meier method. The Cox proportional hazards model was used to assess the association between PFS, OS and treatment characteristics. All tests were two tailed. A p-value less than 0.05 was considered statistically significant. All p values reported were based on the two-sided hypothesis. The statistical analyses were calculated using SPSS -21 (IBM Corporation, Armonk, NY, USA).
This study was conducted in accordance with the Declaration of Helsinki. The study was approved by Institutional Review Board Committee and Institutional Ethics Committee (ERIDNPVO-0004/2021).
In total, 118 patients had been identified as candidates for maintenance treatment with durvalumab from December 2017 who completed treatment until December 2020. Of those, 85 (72.0%) patients continued treatment with maintenance durvalumab after ChT-RT and in 33 durvalumab was omitted due to persistent toxicity in 6, treatment refusal in 11 and progressive disease in 16 patients.
Baseline characteristics of 85 patients included in the analyses are detailed in Table 1. Most were male (70.6%) in stage IIIB (56.5%) with squamous cell carcinoma (58.8%), ECOG PS 1 (54.1%). Median age was 63 years (range 36–73 years). PD-L1 expression was positive in 65 (76.5%) patients, in 13 (15.3%) patients was negative and not available in 7 patients (8.2%). No EGFR, ALK, ROS 1, BRAF and TNRK mutations were detected in 65 (76.5%) patients, KRAS mutation was present in 16 (18.8%) patients and for 4 patients mutational status was not available.
Baseline characteristics of patients treated with durvalumab
N = 85 | ||
---|---|---|
Gender | Female | 25 (29.4 %) |
Male | 60 (70.6 %) | |
Age | Median (range) | |
< 63 | 63 (36 – 73) | |
≥ 63 | ||
ECOG PS | 0 | 37 (43.5 %) |
1 | 46 (54.1 %) | |
2 | 2 (2.4 %) | |
Smoking history | Never | 2 (2.4 %) |
Ex-smokers | 35 (41.6 %) | |
Smoking at diagnosis | 47 (56.0 %) | |
Histology | Adenocarcinoma | 31 (36.5 %) |
Squamous Cell | 50 (58.8 %) | |
Other | 4 (4.7 %) | |
AJCC 8th Edition Stage | IIIA | 26 (30.6%) |
IIIB | 48 (56.5 %) | |
IIIC | 11 (12.9 %) | |
PD-L1 Expression | < 1% | 13 (15.3 %) |
1%-49% | 33 (38.8 %) | |
> 50% | 32 (37.7 %) | |
Unavailable | 7 (8.2 %) | |
Mutational status | No mutations | 65 (76.5 %) |
KRAS | 16 (18.8 %) | |
Unavailable | 4 (4.7 %) |
Abbreviation: N-number, ECOG PS- Eastern Cooperative Oncology Group performance status, PD-L1-programmed dead-ligand 1
Eighty-two (96.5%) patients started treatment with induction ChT and 54 (63.5%) patients received ChT during RT (Table 2). Thirty-one (36.5%) patients were treated with sequential ChT only. Patients received median 3 cycles (range 1–5) of ChT altogether. Seventy-nine patients (92.9%) received gemcitabine and cisplatin as induction ChT and 52 (61.2%) of those switched to etoposide and cisplatin during concurrent treatment with RT. Only 3 patients (3.5%) with adenocarcinoma were treated with pemetrexed and cisplatin as induction as well as concurrent regimen. Patients were treated with the median RT dose of 60 Gy (range 54 Gy–66 Gy), and in most of them (82.3%) partial response was observed after ChT-RT.
Chemoradiotherapy treatment characteristics
N = 85 (%) | ||
---|---|---|
N of ChT | 1 | 3 (3.5%) |
2 | 13 (15.3%) | |
3 | 41 (48.2%) | |
4 | 27 (31.8%) | |
5 | 1 (1.2%) | |
ChT | Gem/cis | 79 (92.9%) |
Etop/cis | 52 (61.2%) | |
Pem/cis | 3 (3.5%) | |
ChT | Induction | 82 (96.5%) |
Sequential only | 31 (36.5%) | |
Concurrent | 54 (63.5%) | |
Concurrent only | 3 (3.5%) | |
RT dose (Gy) | Median (range) | 60 (5 –66) |
V20 (Gy) | Median (range) | 27.2 (7.0–35.6) |
MLD (Gy) | Median (range) | 15.7 (4.0–20.2) |
PTV (cm3) | Median (range) | 416.6 (172.3–1282.6) |
Evaluation after ChT-RT | CR | 10 (11.8%) |
PR | 70 (82.3%) | |
SD | 5 (5.9 %) | |
Time between RT-IT (days) | Median (range) | 57 (12–99) |
ChT = chemotherapy, CR = complete response; etop/cis = etoposide/cisplatin; gem/cis = gimcitabine/cisplatin; IT = immunotherapy; MLD = mean lung dose; N = number of patients; PD = progressive disease; PR = partial response; PTV = planning target volume; RT = radiotherapy; pem/cis = pemetrexed/cisplatin; SD = stable disease; V20 = volume of the lung that receive radiation dose of 20 Gy
Patients started first cycle of durvalumab after a median of 57 days (range 12–99 days) from the end of the RT and were treated with the median of 10.8 months (range 0.5–12 months) (Table 3). Forty-one patients (48.2%) completed treatment with planned 12-month therapy, 25 patients (29.4%) completed treatment early due to the toxicity of durvalumab and 16 patients (18.8%) due to the disease progression. One patient died suddenly after two months of treatment without progression and no known cause of death. One patient stopped treatment after one month due to newly diagnosed prostate cancer and one due to cerebral infarction unrelated to durvalumab treatment.
Durvalumab treatment characteristics and influence on overall survival
Treatment characteristics | N (%) | P | |
---|---|---|---|
Time between RT-IT | Median (days) | 57 (12–99) | 0.689 |
< 57 | |||
≥ 57 | |||
Treatment time of IT |
Median (months) | 10.8 (0.5–12.0) | |
< 10.8 | |||
≥ 10.8 | |||
Treatment with IT | Completed | 41 (48.2%) | 0.095 |
Early stopped due to AE | 25 (29.4%) | ||
Progression | 16 (18.8%) | ||
Other | 3 (3.6%) | ||
Response after IT** | CR | 29 (34.1%) | 0.213 |
PR | 11 (12.9%) | ||
SD | 15 (17.6%) | ||
PD | 10 (11.8%) | ||
Progression*** | Loco-regional | 21 (24.7%) | 0.217 |
Metastatic | 10 (11.7%) | ||
Metastatic and local | 5 (5.9%) | ||
Metastatic spread | CNS | 5 (5.9%) | 0.101 |
Other | 10 (11.7 %) |
*difference in overall survival between patients treated with immunotherapy less or more than median time;
** including evaluation up to 4 months after completed immunotherapy in patients with 12-month therapy as well as in early stopped due to adverse events, later progression is not included. Two patients were not evaluable;
*** observed progression until the last evaluation date;
AE = adverse events; CNS = central nerve system; CR =complete response; IT = immunotherapy;
PD = progressive disease; PR = partial response; RT = radiotherapy; SD = stable disease
Twenty-five patients (29.4%) discontinued durvalumab early due to the toxicity after the median treatment time of 6.0 months (range 0.5–11 months). Twelve patients (14.1%) had pneumonitis that started significantly earlier after introduction of durvalumab than other AE (2.0 months
In total, 42 (49.6%) patients experienced 47 AE, 38 patients had 1 AE, three patients had 2 and one patient had 3 AE. Pneumonitis was found in 15 (17.6%), arthralgia in 5 (5.9%), skin toxicity in 13 (15.3%), colitis in 6 (7.1%) and hypothyroidism in 8 (9.4%) patients.
Median PFS from the durvalumab start was 22.0 months, estimated 12- and 24-month PFS were 71% (95% CI: 61.2–80.8%) and 45.8% (95% CI: 32.7– 58.9%) (Figure 1). During durvalumab treatment, 16 patients progressed after the durvalumab treatment median time of 6.1 months (range 0.5–11.0 months), 7 with loco-regional and 9 with distant metastases. Altogether, 36 (42.4%) patients have progressed until the last follow-up date, 21 (24.7%) patients with loco-regional failure only and 15 (17.6%) patients with distant metastatic disease. Of those, 5 patients had also local progression.
Age, gender, ECOG PS, stage, histology, PD-L1 expression, mutational status, smoking status, RT dose, time between end of RT and start of durvalumab, time of durvalumab treatment in non-progressive patients during durvalumab, ChT sequential
Univariate and multivariate analysis of predictors for progression free survival
Variable | Univariate analysis | Multivariate analysis | |||
---|---|---|---|---|---|
HR (95% CI) | p | HR (95 % CI) | p | ||
Age | < 63 years | 0.67 (0.35–1.29) | 0.231 | ||
≥ 63 years | |||||
Gender | Male | 0.62 (0.32–1.20) | 0.156 | ||
Female | |||||
ECOG PS | 0 | 0.66 (0.34–1.28) | 0.226 | ||
1–2 | |||||
Stage | IIIA | 0.89 (0.54–1.48) | 0.663 | ||
IIIB | |||||
IIIC | |||||
Histology | Adenocarcinoma | 0.84 (0.41–1.70) | 0.635 | ||
Squamous Cell | |||||
Smoking status | Never | 0.59 (0.33–1.07) | 0.084 | 1.72 (0.26–11.00) | 0.567 |
Ex-smokers | |||||
Smoking at diagnosis | |||||
Mutational status | No | 0.99 (0.87–1.12) | 0.881 | ||
KRAS | |||||
PD-L1 | < 1% | 1.03 (0.67–1.59) | 0.882 | ||
1%–49% | |||||
> 50% | |||||
Time to durvalumab | < 57 days | 0.63 (0.32–1.24) | 0.186 | ||
≥ 57 days | |||||
RT dose | < 60 Gy | 1.10 (0.43–2.84) | 0.838 | ||
≥ 60 Gy | |||||
No of ChT | Up to 3 | 2.06 (0.94–4.51) | 0.069 | 0.83 (0.23–2.99) | 0.783 |
4–5 | |||||
ChT | Sequential | 1.51 (0.79–2.85) | 0.209 | ||
Concurrent | |||||
Response after IT | CR | 0.066 (0.008–0.518) | 0.010 | 0.067 (0.008–0.535) | |
PR/SD | |||||
Durvalumab treatment time | < 10.8 months | 3.16 (1.62–6.17) | 0.001 | 1.18 (0.227–6.17 | 0.841 |
≥ 10.8 months |
CR =complete response; ECOG PS = Eastern Cooperative Oncology Group performance status; ChT =chemotherapy, RT = radiotherapy, IT = immunotherapy; PD = progressive disease; PD-L1 = programmed dead-ligand 1; PR =partial response; SD = stable disease
Regarding the pattern of progression, we found more loco-regional only progression in squamous cell carcinoma (79.4%) and distant metastases in adenocarcinoma (81.8 %), the difference was significant (p = 0.002) (Table 5). In addition, all patients with KRAS mutation that progressed, had distant metastases only (p < 0.001). On the contrary, all patients that discontinued treatment early due to AE and progressed later (n = 8), had loco-regional failure only (p = 0.024).
Pattern of progression
Progression (N of patients) | Loco-regional only | Metastatic | p | |
---|---|---|---|---|
Gender | Male | 12 | 10 | 0.563 |
Female | 9 | 5 | ||
Stage | IIIA | 7 | 3 | 0.663 |
IIIB | 11 | 9 | ||
IIIC | 3 | 3 | ||
Histology | Adenocarcinoma | 2 | 9 | |
Squamous cell | 19 | 5 | ||
Other | 0 | 1 | ||
Mutation | KRAS | 0 | 7 | |
No mutation | 21 | 8 | ||
PD-L1 | < 1% | 3 | 4 | |
1%–49 % | 8 | 6 | 0.435 | |
50 % | 10 | 4 |
N = number; PD-L1 = programmed dead-ligand 1
Median OS from the durvalumab start has not been reached after the median follow-up time of 23 months (range 2–35 months). Twelve- and estimated 24-month OS were 86.7% (95% CI: 79.5–93.9%), and 68.6% (95% CI: 57.2–79.9%), respectively (Figure 3). In total, 25 patients (29.4%) have died until the last follow-up date.
Age, gender, ECOG PS, stage, histology, PD-L1 expression, mutational status, smoking status, RT dose, ChT sequential
After progression, six patients had no additional treatment, mostly due to progressive deterioration of patient’s performance. Fourteen patients were treated with RT only and eleven with ChT, of those six received aditional RT and one of them also immunotherapy. Two patients had surgery of brain metastases and three patients had salvage surgery of primary tumour.
The results of PFS and OS in our series of patients with stage III NSCLC treated with durvalumab after ChT-RT confirmed improved survival compared to our historical data of treatment before durvalumab introduction.10, 11, 12 After the median follow-up time of 23 months, 12- and 24-month PFS (71% and 45.8%, respectively) and OS (86.7% and 68.6%, respectively) in our series were comparable with those in the PACIFIC trial (PFS 55.3% and 44.8% and OS 83.1% and 66.3%).1, 2, 3, 4 In addition, data from other real-world reports have confirmed the advantage of maintenance treatment with durvalumab over ChT-RT only.13, 14, 15, 16 In a series of 62 patients in the report of Offin
The broad usage of induction ChT before RT, and ChT selection of gemcitabine in our group of stage III NSCLC patients offered comparable survival rates and safety as reported with other schedules in recent publications.14, 15, 16 Some patients with stage III NSCLC are not candidates for concurrent ChT-RT due to the age and comorbidity.5,6 Most patients (96.5 %) in our series started treatment with ChT, majority of them with platinum-based ChT including gemcitabine and 63.5% of all continued platinum-based ChT during RT. In the PACIFIC study, only 25.8% of patients were treated with induction ChT and 99.8% of patients with concurrent platinum-based ChT with etoposide, vinblastine, vinorelbine, taxans or pemetrexed.1,17 Only few patients in PACIFIC trial were treated with gemcitabine, and additionally, gemcitabine was not used in none of the recently published real-world durvalumab treatment reports. In the real-world reports of durvalumab treatment, platinum-based ChT with etoposide was used in 11.0% to 21.8% of all patients concurrent with RT, and induction ChT was used in up to 32.5% of patients.14, 15, 16
Treatment with induction gemcitabine in our historical analysis had not revealed excessive AE 10, 11, 12. Induction, sequential and concomitant regimes were well tolerated also in the present series (data not shown). In present analysis, there were no differences in PFS and OS between the patients treated with sequential ChT only or concurrent ChT. Additionally, the effectiveness of platinum-based ChT with induction gemcitabine was not inferior to other ChT schedules when comparing PFS and OS. Comparing the best response to ChT-RT, we observed a higher rate of CR and PR (11.8% and 82.3%) than the PACIFIC trial (1.9% and 48.7%).1, 2, 3, 4 The median time to first cycle of durvalumab from the end of RT in our series was 57 days (range 12–99 days) which is considerably longer than in PACIFIC trial (range 1–42 days), and longer as reported by Offin with a median time of 1.5 months (range 0.3–7.7 months), and Desilets with 33 days (range 1–94 days).14,15 Early completed durvalumab treatment due to progression in our analysis was observed in 18.8% of patients as compared to 30.2% in PACIFIC trial and 34.1% reported by Faehling.1,16 At the evaluation up to 4 months after the completion of durvalumab treatment (planned 12-month therapy or early completed due to AE), we observed CR, PR and SD in 34.1%, 12.9% and 17.6% of patients. Notably, the only predictor for improved PFS in our series was CR compared to PR/SD after durvalumab treatment that was confirmed in multivariate analysis (p = 0.032).
Baseline characteristics of patients in our series differed from patients in PACIFIC trial. In our analysis, more patients in stage IIIB and IIIC (69.4%
The salvage treatment for most patients with loco-regional progression was reirradiation with or without reinduction ChT. Three patients in our series had salvage surgery due to progressive primary tumour with observed regression of the lymph nodes. In all, the histology revealed down-staging of the lymph nodes and persistent malignant cells in the primary tumour. The high proportion of loco-regional failures only opens the important emerging issue how to deal with the patients after completion of ChT-RT and maintenance durvalumab with PR or SD. Regarding our results, surgery might be an appropriate additional treatment option in selected patients with PR, especially in patients with squamous cell carcinoma. Further clinical trials are investigating incorporation of immunotherapy at different time point in treatment with ChT-RT in stage III NSCLC patients. Additionally, the research in modulating the immune response by interfering with specific alternative immune receptors, pathways and mediators is ongoing and might offer additional knowledge that would affect the treatment of stage III NSCLC patients.21 However, as demonstrated in advanced NSCLC, one treatment might not be suitable for all and in the future, it could be revealed that personalized multimodality approach for selected stage III NSCLC patients might enable better survival results.
Our results presented here were collected as a single institution experience. Due to small number of patients, this series might be underpowered to detect significant impact on survival for different treatment regimens and probable prognostic variable. Also, some information in statistical analysis might be lost, due to dichotomisation of continuous data. Due to retrospective nature of the analysis, some data were not available for all patients. However, despite more advanced stage III NSCLC and squamous cell histology, our results are consistent with the PACIFIC trial. Additionally, further studies are warranted assessing management of patients with loco-regional SD or PR after durvalumab treatment.
The survival data in present analysis confirmed the advantage of maintenance durvalumab in the treatment of unresectable stage III NSCLC patients over ChT-RT only and our results are in line with the PACIFIC trial as well as with recently published real-world reports. Additionally, with mostly gemcitabine as induction platinum-based ChT, the survival outcomes confirmed our treatment regimen as efficient and well tolerated.