Risk of SARS-CoV2-Related Mortality in Non-Small Cell Lung Cancer Patients Treated with First-Line Immunotherapy Alone or in Combination with Chemotherapy

Abstract Background The impact of systemic anticancer treatments on SARS-CoV-2-related mortality is still debatable. Methods By a retrospective analysis of patients with non-small-cell lung cancer (NSCLC) treated with first-line Pembrolizumab or in combination with chemotherapy (ChT) during the first surge of the pandemic. Results The adjusted risk of death was higher in patients treated with ChT + Pembrolizumab (HR 4.6, 1.2–17.4, p = 0.02). The SARS-CoV-2-related mortality rate was higher in patients treated with ChT + Pembrolizumab (p = 0.03), ≥70 years (p = 0.03) and current smokers (p = 0.17). Conclusions The addition of ChT to immunotherapy could be associated with increased risk of mortality and higher SARS-CoV-2-related mortality rate.


Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a novel crown-like enveloped RNA virus responsible for COVID-19 disease, disseminating from Wuhan, Hubei province, China, since December 2019 (1). It shows more virulency though lower fatality than other coronaviruses, such as SARS-CoV and Middle East respiratory syndrome coronavirus (MERS-CoV) (1). COVID-19 seems related to a whole spectrum of clinical phenotypes, displaying more frequently mild severity symptoms and, rarely, respiratory failure, cytokine release syndrome (CRS) and multi-organ failure (2). Overall, it is accountable for a case-fatality rate ranging from 2 to 3% (1). New variants of COVID-19 have been emerging, with potential increased transmissibility, although there is no evidence of increased clinical severity of illness (3).
In this context, patients with cancer are considered a vulnerable category due to the impact of anticancer treatments and the underlying malignancy on immunological impairment, resulting in a higher burden of SARS-CoV-2 infection, disease severity, complication and mortality than in the general population (1). Moreover, although most cancer patients achieve seroconversion after two SARS-CoV-2 mRNA vaccine doses, those with haematologic malignancy and treated with cytotoxic chemotherapy or monoclonal antibodies might have lower antibody titres than non-haematologic cancers and on clinical surveillance or endocrine therapy, respectively (4). To ensure surveillance to this population, the UK Coronavirus Cancer Monitoring Project (UKCCMP) had been launched in March 2020, providing researchers with the largest prospective database of COVID-19 in patients with cancer (5).
The pooled prevalence of cancer in patients with COVID-19 is estimated as 2.0% (95% CI, 2.0-3.0 I2 ¼ 83.2%) compared to an approximately 0.3% incidence of cancer and 0.37% of COVID-19 infection in the overall population (6). Patients with cancer, particularly with thoracic malignancies, have poorer outcomes than the overall population (7). This concept has been challenged by the results of a prospective study conducted by Lee (5). Cytotoxic chemotherapy, immunotherapy, hormonal therapy, targeted therapy and radiotherapy were compared in this dataset. After adjusting for other risk factors, the multivariate analysis concluded that treatment within four weeks before confirmed COVID-19 is neither determining for a more severe disease nor predictive of COVID-19-related death, questioning the actual impact of cancer treatments (5). Moreover, a further researchers' query was to investigate COVID-19 risk according to tumour subtypes. Lee et al. found through a prospective cohort study that cancer type was not a risk factor for increased case-fatality rate, exception made for haematological malignancies (odds ratio [OR] 2.25, 95% CI 1.13-4.57; p ¼ 0.023); this applies to lung cancer too (2).
In particular, this evidence challenges the results coming from the TERAVOLT registry study, collecting patients' data from 42 institutions across eight countries (Italy, Spain, France, Switzerland, Netherlands, USA, UK and China) and conceived to study the impact of COVID-19 in patients with thoracic cancers. According to Garassino et al., patients infected by COVID-19 hit a mortality rate of 33% regardless of their specific systemic treatments (8). This evidence supported the current recommendation of postponing unnecessary active anticancer therapies, such as when the tumour is stable, at least during COVID-19 pandemic surges (9,10). This applies for either chemotherapy, on the premises of a potentially related immunosuppression, and for immune-checkpoint inhibitors (ICIs), due to potential albeit rare and life-threatening immunerelated adverse events (irAEs). Among these, interstitial pneumonia represents possible toxicity from anti-programmed cell death (PD)-1/PDligand(L)1 agents and could be associated with a CRS, with a surge of IL-6 and IFN-c levels (8). According to Luo J et al., in a series of 41 infected patients between those exposed or not to PD-1 blockade, no significant difference in the severity of SARS-CoV-2 and IL-6 peaks was observed, although these data, as for the most of the available evidence, refer to patients infected by COVID-19 without considering the overall population of patients treated with ICIs (11).
This analysis aims to weigh the real impact of ICIs by exploring the risk of SARS-CoV-2-related mortality in patients with NSCLC treated with first-line immunotherapy alone or combined with chemotherapy (ChT) during the first surge of the pandemic.

Materials and methods
An Excel database was generated for the scope of this analysis. Extracted raw data from hospitals electronic medical records (EMR) of two European centres, the Lincoln County Hospital, United Lincolnshire Hospitals NHS Trust, Lincoln, UK and University Hospital, Geneva, Switzerland, were entered into the database. The investigator led the process of patient identification in each of the two centres. The following three search criteria were applied: diagnosis of NSCLC (by entering the C39 code of the International Statistical Classification of Diseases and Related Health Problems [ICD] 10th edition), first-line treatment with Pembrolizumab alone or in combination with chemotherapy (by manual selection of all the records returned by the EMR under the term Pembrolizumab), the period between January 2020 and May 2020, any histology, age, performance status, known PD-L1 status, clinical stage according to the TNM classification v8.0. If the record met all the three mentioned criteria, the patient was eligible for the analysis and relevant clinical and treatment-related information were extracted from the EMR and entered into the database, including: age, sex, smoking status, country, disease stage, Eastern Cooperative Oncology Group (ECOG) Performance Status (PS), autoimmune disease, PD-L1 expression, histology, chemotherapy, neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation index (SII) (or the NLR x platelets). The baseline NLR and SII were calculated from reports of routine blood yielded within seven days from treatment start according to local laboratories. To ensure patients' confidentiality, data were anonymized, and only essential information was collected and retained. Data were centrally reviewed, checked for quality and consistency, and analyzed. The cause of mortality was that reported in the death certificate. To define a SARS-CoV-2-related death, a SARS had to be reported as the cause of death including radiological and/or microbiological evidence of SARS-CoV-2-related infection. This retrospective analysis was conducted following the Declaration of Helsinki and involved a real-world series of patients treated according to the clinical practice. The ethical approval was waived because all patients in each centre signed a specific consent form for their data collection and sharing with other institutions and this analysis was part of a retrospective institutional audit project.
We described the overall mortality rate, the death-related causes, and the clinical features of dead patients. Descriptive statistics were used to summarize the patient demographic and treatment characteristics, with medians and proportions as appropriate. Statistical significance of variables across the two groups was investigated by Chi-square and Mann-Whitney U tests for dichotomous and continuous variables, respectively, with an acceptable significance value of p < 0.05. The OS was calculated from the date of treatment start until death or date of the last follow-up and was estimated using the Kaplan-Meier method reported as medians with confidence limits (95% CI) and compared using the two-sided log-rank test, with an acceptable significance value of p < 0.05. Patients who did not have events at the time of the analysis were censored. Univariable logistic regression analysis was used for the association of treatment and death rate by computing Odds-ratio (OR) with 95% confidence intervals (CIs) and alpha level set to p < 0.05.

Results
The clinical characteristics of the 111 patients included in the analysis are summarized in Table 1. Sixty-six patients (59%) were on Pembrolizumab and 45 (41%) on combination therapy with chemotherapy (ChT) and Pembrolizumab. There was an expected significant difference between treatment cohorts and PD-L1 expression, as Pembrolizumab alone was offered only to those patients with PD-L1 tumour expression >50% (100%). In contrast, only two (4%) in the ChT þ Pembrolizumab cohort had a PD-L1 > 50% (p < 0.001). Other significant differences between the two treatment cohorts regarded a higher prevalence of PS 2 in the Pembrolizumab group (9 vs. 0%, p ¼ 0.04) and non-squamous histology in the ChT þ Pembrolizumab group (84 vs. 58%, p ¼ 0.003); a nonsignificant trend toward to older patients' age in the Pembrolizumab cohort was observed (47 vs. 31%, p ¼ 0.09).
The median follow-up and OS were significantly longer for the Pembrolizumab as compared to the Pembro þ ChT cohort (p < 0.001 and p ¼ 0.009, respectively) ( Table 2). The risk of death, adjusted for differences between treatment populations, or ECOG PS, PD-L1 and histology, was significantly higher in the group of patients treated with ChT þ Pembrolizumab than with Pembrolizumab alone (hazard ratio [HR] 4.6, 1.2-17.4, p ¼ 0.02).

Discussion
Since the SARS-CoV-2 global pandemic outburst has begun, a relevant question has been whether or not immunotherapy plays an impactful role in the risk of SARS-CoV-2 related mortality in patients with NSCLC.
Our retrospective analysis was based on a population of patients with NSCLC treated with first-line immunotherapy or chemoimmunotherapy and unselected for the COVID-19 infection. Immunotherapy seemed not to impact the risk of SARS-CoV-2-related mortality, during the surge of the pandemic, whilst the addition of ChT was associated with an increased risk of mortality and higher SARS-CoV-2-related mortality rate. A not negligible SARS-CoV-2-related mortality of 8% was observed.
This information may be relevant in the context of the complex interrelationship between the individual vulnerability of cancer patients related to their cancer type, comorbidities and different anticancer therapies. Based on a meta-analysis of 26 studies involving 5571 cancer patients infected with SARS-CoV-2, any anticancer treatment including surgery, chemotherapy, immunotherapy and targeted therapies was associated with disease severity or mortality (12). In contrast, by a previous meta-analysis of 16 studies involving 3558 patients, active chemotherapy was associated with a higher risk of death than no active chemotherapy in cancer patients infected with SARS-CoV-2. In contrast, no significant association with the risk of death was related to other treatments, like active targeted therapies, immunotherapy, chemoimmunotherapy or recent surgery (13). However, a possible limitation of those studies relied on the inclusion of only patients infected by SARS-CoV-2. By a prospective observational study on 293 consecutive patients with cancer on active chemotherapy, ICIs or targeted therapies, of whom 89 (30%) screened as SARS-CoV-2 positive, the incidence of serious adverse events was higher for SARS-CoV-2-positive patients but not related to either of the anticancer treatments (14). Patients with lung cancer may have an increased risk of death than other cancers, possibly due to a combination of the underlying pulmonary compromise due to smoking history, the increased use of steroids, advanced age and frequent comorbidities (9,(15)(16)(17).
Similarly, the present study has the advantage to look at the impact of the SARS-CoV-2 in terms of mortality on a homogenous group of patients for cancer type and treatment setting unselected for their SARS-CoV-2 status treated during the first pandemic surge. Besides the addition of the chemotherapy to the immunotherapy, we confirmed the co-existence of other clinical factors such as older age (>70 years), heavy smoking history and a pro-inflammatory systemic condition (with high NLR and SII) as possible contributing risk factors to the SARS-CoV-2related mortality. Older age, male gender and comorbidities (i.e. hypertension and cardiovascular disease) were specific features of patients with cancer and symptomatic COVID-19 who died by the analysis conducted by Lee et al. (5) Moreover, a systemic pro-inflammatory response, lymphocytopenia and high LDH, and chemotherapy-associated neutropenia, may represent key drivers of mortality from SARS-CoV-2 in cancer patients (16)(17)(18). Specifically, a pro-inflammatory response may lead to the deterioration of COVID-19 via the CRS and cytokine stormrelated acute respiratory distress syndrome (ARDS) (19). The most relevant limitations of this study are the limited sample of patients (including the small number of SARS-CoV-2related deaths), the retrospective nature of the analysis and the lack of a screening of patients for COVID-19. Our findings deserve confirmation in larger series to outline the real risk of SARS-Cov-2 in terms of mortality from the use of ICI and/or ChT in patients with NSCLC and the role of other clinical features.