The EAP population comprised patients aged ≥18 years with histologically or citologically confirmed stage IIIB or IV NSCLC (squamous and non-squamous) that had progressed or recurred after at least one prior systemic treatment for advanced or metastatic disease. Patients were required to have an ECOG Performance Status of 0, 1 or 2; completion of prior chemotherapy, tyrosine kinase inhibitors or palliative radiotherapy, 2 weeks before starting nivolumab; resolution of all adverse events (AEs) to grade 1 (NCI CTCAE: National Cancer Institute Common Terminology Criteria for Adverse Events, version 4) at baseline; adequate organ function and a life expectancy of at least 6 weeks. Exclusion criteria included active brain metastasis, positive test for hepatitis B or C virus indicating acute or chronic infection, known history of testing positive for human immunodeficiency virus (HIV), a severe autoimmune disease and patients with a condition requiring systemic treatment with either corticosteroids (>10mg daily prednisolone equivalent) or other immunosuppressive medications.

The treatment plan was explained in detail and all patients signed and dated a written informed consent form provided by Bristol-Myers Squibb. The study was approved by the INFARMED (Portuguese National Authority of Medicines and Health Products) and was executed in accordance with the Declaration of Helsinki, Good Clinical Practice, and local ethical and legal requirements.

We retrospectively collected clinical data for the eligible patients enrolled in the Nivolumab EAP at 30 Portuguese cancer centres from June 2015 to December 2016. All patients included were followed until the end of data collection on July 2018. The evaluation included a review of demographic and tumour characteristics: age, gender, smoking status, performance status, clinical stage, histology and number of previous treatments. All patients included in the analysis received at least one cycle of nivolumab at the standard dose of 3mg per kilogram of body weight (60-minute intravenous infusion) every 2 weeks for ≤24 months. The treatment was continued depending on disease progression, unacceptable toxicity, completion of permitted cycles (24 months) or withdrawal of consent. Dose escalation or reduction was not allowed, but treatment could be delayed in the event of toxicity. Treatment beyond progression was allowed if, in the investigator’s opinion, the patient was deriving clinical benefit, in the absence of rapid disease progression, severe toxicity or performance status escalation. Computed tomography (CT) based evaluation was performed around every 12 weeks.

Demographic analysis included all patients in EAP. Safety was monitored by AE assessment, physical examination and laboratory tests. Investigators graded AEs according to the NCI CTCAE, version 4, and determined their causal relationship to treatment; for specific immune-related AEs, the protocol specific treatment guidelines were used. The EAP had no pre-specified end-points. Safety analysis included all patients treated (those who received at least one dose of study drug). Collected efficacy data included investigator-assessed objective tumour response, date of progression and survival information. Tumour response was assessed using the Response Evaluation Criteria in Solid Tumours (RECIST) version 1.1. Efficacy analysis included all evaluable patients and the exploratory assessments analysed were: Overall response rate (ORR; defined as the combined rates of complete and partial responses), disease control rate (DCR; defined as the combined rates of complete response, partial response and stable disease), progression-free survival (PFS) and overall survival (OS). PFS was defined as the time from date of first nivolumab treatment until date of first evidence of disease progression or death (whichever occurred first). The OS was also calculated from the first day of administration of nivolumab treatment until the patient’s death or last date of follow-up.

A descriptive statistical analysis was performed, including central tendency and dispersion parameters for the quantitative variables and absolute and relative frequencies for categorical variables. Categorical variables were presented by numbers and percentiles, medians and ranges were reported for continuous variables. The Kaplan-Meyer method with a 95% Confidence Interval (CI) was used to estimate PFS and OS. Patients who were alive, without disease progression were censored at the time of last follow-up. Patients lost to follow-up were censored at the time of last contact. Patients who died because of their lung cancer before any radiological response evaluation were considered to have progressive disease at the time of death. A univariate analysis to test for differences between groups was performed using the log-rank test, followed by a multivariate analysis by Cox regression model to identify prognostic factors associated with OS. Univariate and multivariate hazard ratios were reported with a 95% CI. Statistical significance was set at p less than 0.05. All statistical analyses were carried out with SPSS Statistical Software, version 21.0 (IBM_SPSS). We used BiostatXL MIX 2.0 (available at: http://www.meta-analysis-made-easy.com) to compute sub-group analysis estimates.

From June 2015 to December 2016, a total of 229 patients with advanced NSCLC were enrolled at 30 Portuguese centres. Of the 229 patients, data was available for analysis in 219 patients (3 patients did not start treatment, and unavailable data in 7 patients); all patients included in the analysis received at least one cycle of nivolumab. The median number of nivolumab cycles administered was 14 (range, 1–52). The median age at the start of the treatment was 64 years (range 37–83) and the majority of patients were males (n=154; 70.3%). Most cases were former/current smokers (n=152; 69.4%). Patients with non-squamous histology predominated (n=193; 88.1%). All patients had previously received at least one platinum-based therapy and 84 (38.4%) patients received 2 prior systemic lines of therapy; 25 (11.4%) three lines, 21 (9.6%) four lines; 18 (8.2%) more than four lines of treatment. According to the Eastern Cooperative Oncology Group Performance Status score (ECOG-PS), at the start of treatment 26 (11.9%) patients had an ECOG-PS 0, 164 (74.9%) had ECOG-PS 1 and 29 (13.2%) had ECOG-PS 2. In our cohort, no routine tumour staining for PD-L1 was done, because at the time of the EAP it was not part of routine clinical practice nor was it an inclusion criteria for EAP. See Table 1 for a comprehensive list of population baseline characteristics, stratification factors and prior therapy.

With a median follow-up of 17.1 (1–34.1) months, of the 219 patients enrolled, the best overall response in the total patient population was complete response (CR) in 3 (1.4%) patients, partial response (PR) in 46(21%) patients, stable disease (SD) in 68 (31%) patients and progressive disease (PD) in 68 (31.1%) patients. No evaluation could be carried out on 34 patients (15.5%) because of poor general condition or early death. Thus, the ORR was 22.4% and DCR was 53.4%. At the time of survival analysis, the median PFS was 4.91 months (95% CI, 3.89–6.11) and median OS was 13.21 months (95% CI, 9.89–16.53) in the overall population (Fig. 1). The 6, 12- and 18-month OS rates were 71%, 56.5% and 43% respectively.

Figure 1.

Kaplan-Meyer curves: Progression free survival (A); Overall survival (B).

(0.13MB).

Of the 178 patients (81.3%) who discontinued treatment, the reasons for discontinuation were progressive disease in 116 patients (52.2%), death in 23 patients (10.5%), adverse events in 28 (12.5%) and other reason in 22 (10%). A total of 136 patients had at least one adverse event in a total of 175 treatment–related AE’s of any grade, among which 28 led to treatment discontinuation, listed in Table 2. The most common toxicities reported were asthenia/fatigue, 41 events (23.4%), and endocrinopathy, 25 events (14.3%). Discontinuation of the study drug due to treatment-related AEs occurred in 28 patients (16%) in total, 10 of them (5.7%) due to pneumonitis and 6 (3.4%) due to colitis. AEs were managed using protocol-defined toxicity management algorithms.

Subgroup analyses were performed for OS (Fig. 2). In the univariate analysis, ECOG PS score was the only factor which significantly correlated with overall survival (Fig. 2A) by comparing patients who had an ECOG PS of 0–1 with those with an ECOG PS of 2, with a Hazard Ratio (HR) of 3.80 (95%CI: 2.30 to 6.07), p<0.0001). There were no statistically significant differences in gender (p=0.320), age (patients ≥75 vs. <75 years; p=0.821), histology (squamous vs. non-squamous, p=0.14), or number of prior therapy lines (≥2 vs. 1, p=0.15) and smoking status (never vs. former/smoker, p=0.43). In the multivariate analysis, ECOG PS score remained the only factor significantly related to overall survival (p<0.0006). PS 0-1, Stage IIIB and Disease Control (DC) with nivolumab increase the probability of surviving at least one year, unlike PS 2 and Stage IV (Fig. 2B).

Figure 2.

Association between baseline characteristics and survival.

(0.38MB).