AVAiLABLE NIS – AVASTIN® in lung cancer treatment in routine oncology practice in Germany

The post-authorization NIS AVAiLABLE (ClinicalTrials.​gov Identifier: NCT02596958) was planned to include 900 patients with non-resectable, advanced, metastatic or recurrent NSCLC other than predominantly squamous cell lung cancer across Germany. Further selection criteria included: (1) age ≥ 18 years, (2) histologically confirmed non-squamous NSCLC, (3) no contraindication to bevacizumab according to current label, (4) therapeutic decision for 1 L treatment with bevacizumab combined with platinum-based chemotherapy made independently from this NIS. The planned observation period was 51 weeks. No study-specific treatments or assessments were scheduled. A study observation interval of maximally six cycles of chemotherapy plus bevacizumab, followed by bevacizumab maintenance therapy until disease progression was recommended. However, according to the non-interventional approach, actual treatment decisions were at the discretion of the treating physicians. Normal merchandise was used and reimbursed by the respective national or private health insurance. The observational plan was evaluated by the Ethics Committee of the Medical Association of Lower Saxony (‘Ärztekammer Niedersachsen’) in Hannover (Germany) and the study performed in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.

The study was conducted across 185 institutions in Germany (medical oncologists and pneumologists in hospitals and private practices) between September 2007 and October 2013. Patients were followed until progression or intolerable toxicity, whichever occurred first. The final documentation was to be performed within 4 weeks after end of bevacizumab treatment, regardless of further therapy options. No long-term follow-up information on deaths was collected. Quality control procedures were applied to each stage of data entry and data handling to ensure that all data were reliable and processed correctly. A data review meeting was held prior to database lock.

Effectiveness endpoints were response rates, PFS, and OS in a large, unselected patient population. Safety endpoints comprised occurrence and frequency of adverse drug reactions (ADRs; also ADRs of special interest; see next section) including seriousness, relatedness to bevacizumab and severity. Additional safety endpoints were the occurrence of any new or rare bevacizumab-related ADRs as well as ADRs leading to treatment discontinuation. Pre-specified subgroup analyses of effectiveness and safety endpoints were performed for patients with/without adenocarcinoma and various age groups. Differences between groups were not tested for statistical significance. Further research questions included patient characteristics at baseline (e.g. demographic data, cancer history) and details on treatment with bevacizumab and chemotherapy (dose, regimen, duration).

Baseline information collected prospectively per patient included demographics and cancer history, current tumor status, previous treatment, and relevant concomitant diseases. The initial diagnosis including adenocarcinoma had to be confirmed histopathologically. Vital signs, standard laboratory assessments, and general condition were also captured at baseline.

During treatment (every 3 weeks), details on the systemic therapy with bevacizumab (daily dose, infusion time, dose deviations and therapy interruptions) and concomitant antineoplastic agents were recorded. Tumor staging and Eastern Cooperative Oncology Group performance status (ECOG PS) were assessed as per clinical routine at the individual center. ADRs and toxicity based on NCI/CTC (National Cancer Institute/Common Terminology Criteria) were reported at each visit retrospectively for the period since the previous visit. Data regarding adverse events, previous and concomitant diseases were coded using the Medical Dictionary for Regulatory Activities (MedDRA) version 15.1; medications were coded using the World Health Organization drug dictionary (WHO DD) version 4.1.

ADRs of special interest included hypertension, hemorrhages, gastrointestinal perforation/fistula, tracheoesophageal fistula, proteinuria, wound healing complications, congestive heart failure, hemoptysis, and thromboembolism. Serious ADRs and ADRs of special interest had to be reported on separate report forms within 24 h of notice. Best tumor response over time, reasons for end of therapy, and further antineoplastic therapy were documented at an end of study visit.

A sample size of 900 patients was estimated to allow a 99% probability to record an ADR with a true incidence of 1% at least twice. Data were evaluated using descriptive statistical methods. Missing values were not replaced. Specification of the complete analysis was laid down in detail in the statistical analysis plan, which was finalized prior to database lock. Time-to-event analyses were performed using Kaplan Meier methodology and include the corresponding 95% confidence intervals (CIs).

PFS was defined as time from start of therapy to investigator-assessed disease progression or death from any cause. OS was defined as time from treatment start to death if death occurred within the time window between start of therapy and up to four weeks after last bevacizumab administration (final documentation). Patients without event (progression or death) were censored at the end of study or data cut-off date (whichever occurred first). The disease control rate (DCR), defined as percentage of patients achieving complete response (CR), partial response (PR), and/or stable disease (SD) during the course of the observation, was calculated as further effectiveness endpoint.

For patients with NSCLC other than non-squamous (documented as yes vs. no), subgroup analyses evaluated the effect on PFS by age (< 65 vs. 65 to < 70 vs. 70 to < 75 vs. ≥75 years), sex (male vs. female), presence of adenocarcinoma (yes vs. no/unknown), dose level (7.5 mg/kg vs. 15 mg/kg vs. other doses), TNM-stage (stage III vs. stage IV vs. residual), presence of distant metastases, prior treatment (with vs. without operation, radiation, chemotherapy, other), general condition (ECOG PS), best overall response rate (ORR; SD vs. PR vs. CR), number of cycles with bevacizumab, and number of cycles with maintenance therapy. Incidences and the number of episodes were calculated for ADRs and subgroup analyses were performed for grade ≥ 3 toxicity events by age group (< 65 vs. 65 to < 70 vs. 70 to < 75 vs. ≥75 years), and adenocarcinoma (yes vs. no/unknown).

A total of 996 patients consented to the collection and processing of their personal data, nine patients were excluded due to second-line (2 L) treatment. The analysis population included 987 NSCLC patients who were evaluated in the NIS AVAiLABLE. Mean (± standard deviation, σ) patient age was 61.5 (±9.8) years. 58.3% of the patient population was below 65 years. Male patients accounted for 59.8% of the study population. The average body weight of all patients was 74.8 (±18.6) kg, mean body mass index (BMI) was 25.1 (±4.5) kg/m², and 88.3% of patients had ECOG PS 0 or 1 (i.e. no or only minor restrictions in physical activity). 72.8% of the patients were diagnosed with stage IV disease at the start of observation. A proportion of patients had bone metastases (35.3%), malignant pleural effusion (26.9%), liver metastases (18.3%), adrenal gland metastases (14.7%), and/or brain metastases (10.2%). The majority of patients (90.1%; N = 953 non-missing data) had histologically confirmed adenocarcinoma. Recorded separately, 80.8% (N = 874 non-missing data) met the bevacizumab label ‘NSCLC other than predominantly squamous cell histology’. Baseline characteristics regarding age, gender, ECOG PS, adenocarcinoma/non-squamous cell histology, and tumor stage are displayed in Table 1.

The main therapy combination used was bevacizumab plus carboplatin and paclitaxel (38.0%). In a smaller proportion of patients, bevacizumab was combined with carboplatin and pemetrexed (11.7%) or cisplatin and pemetrexed (10.1%). Besides these three most commonly used regimens, other platinum combinations were used as well. All these regimens are following current recommendations for doublet therapy with platinum including third-generation drugs [3]. Only 9.3% of patients received chemotherapy regimens not containing platinum compounds.

Within the analysis population, the majority of patients achieved PR (43.3%) or SD (29.4%) before progression or intolerable toxicity; CR was reached by 2.3%. The DRC (percentage of patients who achieved CR, PR or SD during observation) was 75.0% and the ORR (CR plus PR) was 45.6% (Table 2).

Kaplan Meier estimate of time-to-progression resulted in a median PFS of 7.4 months (95% CI: 7.1; 8.4) for the analysis population. 50% of patients were within the range of 3.9 and 13.8 months until estimated disease progression (Fig. 1). In the subgroup of TMN stage IV patients with NSCLC other than predominantly squamous cell histology (N = 492), median PFS was 7.1 months (95% CI: 6.7; 8.0).

The only follow-up was performed four weeks after the end of treatment with bevacizumab. At that time, the mean OS was 18.4 months (±0.5). However, 82.5% of the analysis population were alive four weeks after the end of bevacizumab treatment and had to be censored. Therefore, the OS estimate is based on a low number of events and does not allow for a meaningful interpretation.

Subgroup analyses performed for 706 patients with NSCLC other than predominantly squamous cell histology revealed that longer PFS was associated with the presence of adenocarcinoma at baseline in comparison with other histological NSCLC subgroups (Table 3). For 314 patients receiving maintenance therapy with bevacizumab, the median PFS was 10.9 months (95% CI: 10.0; 12.5) with 50% of the patients within the range of 7.1 and 17.4 months.

A total of 4991 ADRs were observed in 874/987 patients (88.6%). 110/987 patients (11.1%) experienced serious ADRs that led to premature study termination in 44/987 patients (4.6%). In 266/987 patients (27.0%), an ADR was considered treatment-related to bevacizumab.

Overall, ADRs with grade ≥ 3 were reported for 332 patients (33.6%). Table 4 gives an overview of grade ≥ 3 ADRs by Preferred Term occurring in ≥1% of the analysis population and for the subgroups by age. In the age group 70 to < 75 years, a tendency of increased occurrence of grade ≥ 3 ADRs was observed for hematological parameters (anemia, leukopenia, neutropenia, thrombocytopenia), nausea, pain, and pain in extremity; only hypertension grade ≥ 3 occurred more often in the age group ≥75 years than in age classes with patients < 65 years and < 75 years (Table 4). No substantial differences were observed in grade ≥ 3 ADRs between patients with confirmed adenocarcinoma and patients with other histological findings (data not shown).

Bevacizumab-related ADRs grade ≥ 3 occurred in 56 patients (5.7% of the analysis population) with the highest incidence for leukopenia (11 patients, 1.1%), neutropenia (9 patients, 0.9%), and hypertension (10 patients, 1.0%). Patients in the age group ≥75 years and patients with adenocarcinoma were slightly more affected by bevacizumab-related ADRs (data not shown).

Incidences of bevacizumab-related ADRs of special interest were: hypertension 7.0% (grade ≥ 3: 1.0%), proteinuria 3.6% (grade ≥ 3: 0.1%), hemorrhage 0.1% (none grade ≥ 3), hemoptysis 1.4% (grade ≥ 3: 0.2%), gastrointestinal perforation 0.1% (all grade ≥ 3), large intestine perforation 0.3% (grade ≥ 3: 0.1%) and esophagobronchial fistula 0.1% (all grade ≥ 3).

The main reason for the end of therapy was disease progression in 443 (46.4%) patients. 138 (14.5%) patients died from the underlying disease and 37 (3.9%) patients from another cause (without further specification).

Maintenance therapy with bevacizumab was well tolerated and safe, even for more than 20 cycles (median PFS 31.6 months in 12 patients). No new safety signals were observed in this NIS.

To place the present results within the context of previous research, Table 5 gives an overview of the key results of studies investigating bevacizumab 1 L chemotherapy for NSCLC. It compares the NIS AVAiLABLE with the NIS ARIES [15, 18], the phase IV study SAiL [17, 19], and the RCTs AVAiL [8, 20] and E4599 [7, 21].

With regard to the effectiveness outcomes ORR and PFS, the AVAiLABLE results are consistent with previous studies. Of note, no new safety signals were observed in this NIS. The number of ADRs observed during the AVAiLABLE NIS cannot be compared directly with the safety data reported from controlled clinical trials due to varying documentation requirements in different study types. During controlled clinical trials, usually all adverse events are reported regardless of causal relationship to study treatment.

Results from clinical trials are prone to selection bias due to specific inclusion and exclusion criteria. Moreover, their pre-specified diagnostic and follow-up measures are not always representative approaches of day-to-day practice [22]. In comparison, a NIS is an appropriate method of gathering real-world data for effectiveness and tolerability of a listed drug under routine conditions, although a potential bias related to data being incorrectly reported, transcribed, or missing and a remaining selection bias cannot be ruled out.

In the study at hand, only 706 out of 859 patients included with adenocarcinoma at baseline actually met the bevacizumab label ‘NSCLC other than predominantly squamous cell histology’ (Table 1).

A single follow-up visit was planned at four weeks after the last bevacizumab administration; no further documentation of survival data was scheduled as per protocol. 804 (82.5%) patients were still alive at the time point of last documentation and had to be censored for OS analysis. Due to the short follow-up period, this NIS does not provide reasonable conclusions regarding OS.

In the context of other major studies investigating bevacizumab in combination with 1 L chemotherapy for patients with advanced non-squamous NSCLC, the NIS AVAiLABLE confirms the effectiveness and safety of bevacizumab therapy. Although the study started in 2007, the platinum-doublet chemotherapy used for 1 L treatment in advanced NSCLC is still the current first choice and has become routine clinical practice. According to recent guidelines, combination with bevacizumab and other platinum-based chemotherapies may be considered in eligible patients (non-small-cell cancer and ECOG PS 0–1) and bevacizumab may be added to carboplatin plus paclitaxel if no contraindications exist [5, 13]. Subgroup analyses for patients with NSCLC other than predominantly squamous cell histology show effectiveness for bevacizumab in all chemotherapy combinations containing platinum. As this NIS was conducted under real-life conditions with choice of the chemotherapy regimen at the investigator’s discretion, the number of 90 (9.7%) patients not receiving the recommended platinum treatment may not come as a surprise although specified otherwise in the inclusion criteria. It is in accordance with the results of the observational ARIES and phase IV SAiL studies [15, 17] and may be attributed to patients unsuitable for platinum therapy due to age, concomitant diseases, or performance status.

Pre-specified subgroup analyses focused on patients with/without adenocarcinoma and different age groups (Table 3). Median PFS was 3 months longer in patients with adenocarcinoma compared to other histological NSCLC subgroups. All age groups benefit from bevacizumab treatment. Overall, maintenance therapy with bevacizumab was well tolerated and safe, even for more than 20 cycles (median PFS 31.6 months in 12 patients).

Other targeted therapies with VEGFR antibodies (aflibercept, ramucirumab) and tyrosine kinase inhibitors with selectivity for VEGFRs (sorafenib, sunitinib, nintedanib, cediranib, motesanib, pazopanib, axitinib, vandetanib) showed responses and improved PFS but no survival advantages in patients with advanced NSCLC [4]. Therefore, the established standard of care of bevacizumab in combination with carboplatin/paclitaxel in the 1 L setting, and ramucirumab in combination with docetaxel in the 2 L setting are recommended by recent guidelines. In addition, substantial progress resulted from cancer immunotherapies targeting either the programmed death ligand 1 (PD-L1) or the programmed death-1 (PD-1) pathway in patients with NSCLC. Based on new data from clinical trials in NSCLC, immunotherapies form a new standard in 2 L treatment (nivolumab, pembrolizumab) [23, 24], or 1 L treatment (pembrolizumab) [25] of patients expressing PD-L1, and possibly even for those who do not.

The combination of cancer immunotherapies, including bevacizumab, may amplify the immune system’s ability to eliminate cancer [26]. Current and future roles of bevacizumab in the 1 L therapy of NSCLC include combination therapy with erlotinib and bevacizumab in populations with EGFR-mutations [27, 28] as well as combination therapy with atezolizumab and bevacizumab [29].