Intensive research has shown that PD-L1 inhibitors (programmed death ligand 1) and PD-1 inhibitors (programmed death 1) are effective in first line for advanced or metastatic NSCLC setting. The first approved immune checkpoint inhibitor for second line in metastatic setting of NSCLC was nivolumab. This drug showed a significantly higher overall survival (OS) rate at six months compared to the standard chemotherapy with docetaxel⁽⁵⁾. The therapeutic management of NSCLC has been substantially modified by the development of immune checkpoint inhibitors: nivolumab, pembrolizumab, durvalumab and atezolizumab. These agents have shifted the standard chemotherapy therapies from first line to second line and beyond, marking a new era in the therapeutic management of non-small cell lung cancer⁽⁶⁾.

More recent guidelines have even moved immunotherapy from second line in metastatic NSCLC to front line in the neoadjuvant setting. Still, a high proportion of NSCLC eventually will progress, and the need to establish tailored management options for second-line, third-line and even for fourth-line therapy is becoming increasingly high⁽⁷⁾.

However, randomized controlled studies enroll patients following a precise protocol, and their results may differ in terms of population heterogeneity from clinical practice. Furthermore, the need to corelate and corroborate randomized controlled trial (RCT) data with real-world evidence (RWE) is also becoming a standard of care, for a better understanding of the patients’ profiles, effectiveness of therapy, management of adverse events and cost management.

In Romania, the reimbursement of healthcare costs for immune checkpoint inhibitors started in 2017, with nivolumab for metastatic melanoma therapy, quickly followed by approval and reimbursement in pre-treated metastatic NSCLC⁽⁸⁾.

Thus, the aim of this study is to determine the real-world therapeutic value of nivolumab in metastatic setting, in the second line and beyond. For this purpose, we included and investigated the medical records of thirty-four patients with advanced non-small cell lung cancer from the “Prof. Dr. Alexandru Trestioreanu” National Institute of Oncology, Bucharest. We intend to assess the therapeutic value, safety profile, survival benefit and progression-free interval of disease. Of note, the recommended dosage in 2018 was 1 mg per kilogram every two weeks, not the today standard, which is 240 mg every two weeks.

Methodology

Study design

This has been a single-center retrospective cohort study conducted between January 2018 and December 2020, right after the approval of nivolumab in Eastern Europe for the treatment of metastatic non-small cell lung cancer. The inclusion criteria were: adult patients (over 18 years of age) with advanced or metastatic non-small cell lung cancer. Their therapy must have included immune checkpoint inhibitors (nivolumab) monother­apy or combination with chemotherapy as second line or beyond. The participants were assessed for overall survival and progression-free survival (PFS) for nivolumab and for standard chemotherapy, as well as for the safety profile of nivolumab. The response to therapy was assessed clinically and imagistically, according to the Common Terminology Criteria for Adverse Events 4.0 and the RECIST criteria.

Statistical analysis

For data analysis, we used the R programming language, version 3.6.2, and some graphs and tables were drawn using Excel application in the Microsoft 365 for Enterprise pack. To evaluate the type of distribution, we used Shapiro-Silk test with a significance threshold of 0.05 (if p>0.05, the data is considered to be normally distributed and, therefore, parametric tests can be used). To compare the independent lots, we used two independent sample t-test and Wilcoxon-Mann Whitney test. When analyzing more than two subgroups, we used the ANOVA test (if normal distribution of numerical data) or Kruskal Wallis test (if abnormal distribution of data). To compare proportions, we used either chi-squared test, or Fischer test. For calculating survival values, we used Kaplan-Meier curve, both individually and for independent variables. For comparing Kaplan-Meier curves, we used log-rank test (Cochran-Mantel-Haenszel test). Progression-free survival was calculated as the time from the first initiation of immunotherapy with nivolumab to disease progression or death from any cause. In the medical files, the reports for disease progression were documented, both clinically and radiographically, using RECIST 1.1 criteria. We must mention that the data set was not quite balanced, due to the small sample size.

Results

Patients and tumoral characteristics

A total of 34 patients were included. The majority of them were males (79%), with a median age of 65.5 years old (range: 42-75). We found no statistically significant differences for age regarding one gender or the other (W=115; p=0.3934).

Patients’ distribution regarding gender and smoker status was 79% for male smokers and 21% for non-smokers. The proportion of smokers was statistically higher than the one for non-smokers (p=0.02044).

The histopathologic distribution of cases identified the majority of non-squamous NSCLC, with 56% of patients, and 44% with squamous carcinoma (X-squared=0.52941; df=1; p=0.4669) – Figure 1.

Distribution of metastases

The patients were grouped by types of metastatic sites that have been identified in the lot. We can observe that most of the participants had lung metastases (Table 1).

Comorbidities

In the real-world setting, most patients have significant comorbidities, mostly cardiovascular and meta­bolic, given the fact that lung cancer is strongly associated with an unhealthy lifestyle, obesity, smoking habits and with high blood pressure, according to Table 2.

Therapy sequencing characteristics

All the included participants received first-line ther­apy (34 patients) followed by second-line (29%), third-line (9%) and fourth-line therapy (6%) – Figure 2.

First-line therapy

The distribution of first-line therapy choices is shown in the following pie chart diagram (Figure 3).

Most patients have received gemcitabine + carboplatin (GMZ+Carbo), followed by Alimta® + carboplatin (Alimta®+Carbo) and paclitaxel + carboplatin (Pacli+Carbo).

Vinorelbine + carboplatin (VNR+Carbo) and vinorelbine monotherapy (VNR) were equally used as presented in Figure 3.

Immunotherapy

All analyzed patient files and records have documen­ted the administration of immunotherapy for advanced and metastatic NSCLC, second line and beyond. Immunotherapy with nivolumab was relatively well tolerated, 24 of the enrolled patients have not had any type of adverse events. The remaining 10 patients reported at least one type of adverse event, as follows: three patients reported pulmonary adverse events, followed equally by cardiovascular events, endocrine, rheumatological events and, in a lesser extent, hepatic adverse events. Of note, one patient reported a cardiovascular adverse event that required therapy with steroids and the permanent discontinuation of immunotherapy.

Therapeutic results

The therapeutic result with nivolumab was heterogeneous. None of the patients experienced complete res­ponse (CR). Half of the patients had progressive disease (PD), followed by those with stable disease (SD), and one patient with partial response (PR) – Figure 4. 

Number of nivolumab cycles

All of the enrolled patients received nivolumab in the second line and beyond, but the number of cycles varied from only one administration to thirty-one cycles of nivolumab. Most patients had a mean number of administrations between one and 10 cycles, followed by a mean number of administrations between 10 and 20 cycles, and followed by the next cut-off of 20-30 cycles. Only one patient received more than 30 cycles of nivolumab (31, to be exact). After assessing the disease status according to RECIST criteria, only 30 of the 34 patients could be evaluated, four of them had either been lost to follow-up or had inconclusive imagistic interpretations (X-squared=5.9276; df=NA; p=0.1129) – Figure 5.

Progressive disease after nivolumab therapy

Seventeen of the enrolled patients had progressive disease under nivolumab immunotherapy. After disease progression, most of these could not benefit anymore after any other form of therapy due to a declined health status (ECOG 3). Of the patients who were able to tolerate further the therapy, two received vinorelbine (VNR) and two received gemcitabine (GMZ). Other types of therapy for each patient were administered, as presented in Figure 6. Of note, we included eight patients (n=8; 47%) who were unable to tolerate further the therapy after progression under nivolumab, for a better understating of the subsequent lines of therapy.

Adverse events for patients with stable disease and/or partial response

To sum up, we found that thirteen patients had either stable disease or partial response to nivolumab therapy, according to RECIST 1.1 criteria. Regarding these adverse events, we designed the following table to reflect the general situation with types of reactions noted in the medical files by the caring physicians (Figure 7).

General overall survival

For estimating the overall survival (OS), we had to use a surrogate marker. This surrogate marker is represented by the difference between the date of permanent dismissal from the records (registered in the medical documentation) and the date of enrolment for therapy initiation (also registered in the medical records). The median OS was 32.57 months, but with extreme high values. Most patients (75%) had a median OS below 45.55 months. We grouped the participants depending on OS values into four quartiles:

• Very small (<12.46) – quartile 1

• Small (<24.43) – median or quartile 2

• Medium (<45.55) – quartile 3

• Significantly high – the rest.

The overall survival for nivolumab therapy was calculated as the time from the first initiation of nivolumab to the date of death or loss to follow-up, whichever came first. From the total of 34 patients, 13 deceased during the course of our study. The median time was 72.3 weeks (18.07 months). The confidence interval (CI) was 95%, between 12.5 months and infinite (NA). This fact is common when dealing with a small sample size (number) of patients – Figure 8.

Overall survival for nivolumab by gender

We have summarized the survival graph and the Kaplan-Meier (KM) overview for each of the two variables. The median value for women was 15.3 weeks (3.8 months). For men, this value could not be reached (not available for calculation; NA). There were registered significative results between women and men (Chisq=611.1 on 1 degrees of freedom; p=0.0009) – Figure 9.

Overall survival for nivolumab by smoking status

The median value for non-smokers was 32.5 (8.12 months), and for smokers the median value was 93 weeks (23.25 months; p=0.1) – Figure 10.

Overall survival for nivolumab by histopathology

We divided the patients by squamous and non-squamous histology. The median value for squamous carcinoma was 72.3 weeks (18.1 months), and for non-squamous NSCLC the histology was 93 weeks (23.25 months; Chisq=0.7 on 1 degree of freedom; p=0.4) – Figure 11.

Progression-free survival for nivolumab

Of the total of 34 patients, four couldn’t be evaluated; therefore, we have excluded them from analysis and 17 patients had progressive disease (56.67%); 13 patients (43.33%) had positive therapeutic response (noted as stable disease or partial response).

The median was 17.3 weeks (4.3 months), with 95% CI, between 14.9 weeks (3.7 months) and infinity (NA), because of the small sample size studied here – Figure 12.

Nivolumab progression-free survival by gender. The median for women was 14.5 weeks (3.6 months) and for men, the median was 17.3 weeks (4.3 months; Chisq=3.7 on 1 degree of freedom; p=0.05) – Figure 13.

Progression-free survival for nivolumab by smoking status. The median for non-smokers was 12.6 weeks (3.1months). For smokers, we found this value to be 20.1 weeks (5 months; Chisq=6.6 on 1 degree of freedom; p=0.01) – Figure 14.

PFS for nivolumab by histology. The median for the non-squamous type was 17.3 weeks (4.3 months) and for the squamous type, it was not reached (NA; Chisq=0 on 1 degree of freedom; p=0.9) – Figure 15.

Discussion

In terms of survival (OS and PFS), we found similar results with well-known studies such as Check Mate 017, whose results were median OS of 9.2 months (95% CI; 7.3 to 13.3), and Check Mate 075, with median overall survival of 12.2 months (95% CI; 9.7 to 15)^(9,10). More favorable responses were associated with a good performance status, male gender, smoker status and non-squamous histology and fewer comorbidities, similar to the findings of Zhao et al.⁽¹¹⁾ We did not find hyper-progressive disease and none of the patients included reached the complete response (CR) to immunotherapy. This is somewhat consistent with the findings of literature reviews for Check Mate 017, 057, 003 and 003, where only seven complete responses have been achieved in the entire study population of 664 patients⁽¹²⁾. We found that nivolumab was relatively well tolerated, with just one serious adverse event that needed the permanent discontinuation of immunotherapy. Like in other findings⁽¹³⁾, we can confirm that immunotherapy with nivolumab mostly induced pulmonary and cardiovascular adverse events, but manageable and well tolerated. What is interesting is that seventeen of the enrolled patients (50%) had progressive disease under nivolumab immunotherapy, and almost all could not benefit further from any other form of therapy. We ask ourselves the question of how to correctly identify patients who will mostly benefit from immunotherapy. And, furthermore, after disease progression, could these patients still be able to tolerate other lines of therapy?
 

Conflict of interest: none declared

Financial support: none declared

This work is permanently accessible online free of charge and published under the CC-BY.