Prognostic factors of toxicity of immune checkpoint inhibitors in nonsmall cell lung cancer and small cell lung cancer patients: The ToxImmune study

Abstract Background Immunotherapy alone or in combination has clearly improved the survival of patients with lung cancer. However, it may also be responsible for adverse events impacting these patients' quality of life. The ToxImmune study aims to identify prognostic factors that can help to predict immune‐related adverse events. Methods We included all patients aged 18 years and older who had received at least one dose of immune checkpoint inhibitors, with or without other therapy, between June 2015 and December 2020 and were diagnosed with nonsmall cell lung cancer or small‐cell lung cancer. Patients' baseline demographic characteristics, biological blood markers, and imaging by PET‐scanner were collected from electronic medical records. All adverse events (AEs) and immune‐related AEs (irAEs) were recorded (Common Terminology Criteria For Adverse Events V.5.0). Results Sixty‐four patients were included, of whom 60 (94%) presented at least one irAE. The incidence of Common Terminology Criteria for Adverse Events (CTCAE) grade 2 and grade 3–4 was 34% and 8% respectively. Female sex, Primitive Tumor Standardized Uptake Value Max (SUVmax) <5, number of metastases ≥3 and immunotherapy received after the first line were found to be significant risk factors for immune‐related adverse events. Based on the number of risk factors, the ToxImmune score predicts the risk of having a grade ≥2 adverse event (primitive tumor SUV ≥ 5 = 0 vs. primitive tumor SUV <5 = 1, number of metastases <3 = 0 vs. number of metastases ≥3 = 1 and L1 = 0 vs. L1 ≥ 1). The incidence of grade ≥2 adverse events was 20%, 55% and 90% with ToxImmune scores 0, 1 and = 2 respectively (p = .003). Median progression‐free survival (PFS) times were 19.2 months, 6.64 months and 2.63 months for ToxImmune scores 0, 1 and = 2 respectively, p = .13. Median overall survival times were 22.6 months, 16.4 months and 9.8 months for ToxImmune scores 0, 1 and ≥2 respectively, p = .24. The disease control rate (DRR) was 78% in ToxIummune score 0 group, and 50% in ToxImmune score 1 and ≥2 groups (p = .363). Conclusion The ToxImmune score, which is grounded on objective clinical parameters, indicates that cases with a high score had an advanced threat of severe adverse events. The ToxImmune score could therefore be used in clinical practice to identify patients treated for lung cancer with immunotherapy and at risk of severe AE.

Conclusion: The ToxImmune score, which is grounded on objective clinical parameters, indicates that cases with a high score had an advanced threat of severe adverse events. The ToxImmune score could therefore be used in clinical practice to identify patients treated for lung cancer with immunotherapy and at risk of severe AE.

K E Y W O R D S
immune-related adverse events, immunotherapy, lung cancer 1 | INTRODUCTION Worldwide, among cancers lung cancer is recognized as the leading cause of cancer death, responsible for 1.8 million deaths (18%) with 2 206 771 new cases diagnosed in 2020. 1 At the same time, treatment for lung cancers has been revolutionized by the development of Immune Checkpoint inhibitors (ICI), initially in monotherapy and later in combination therapy, which has clearly improved the survival of patients at first in the metastatic stage and today in the localized stage. [2][3][4][5][6][7][8][9][10][11] By interfering with the immune system, immune check point blockade favors the development of autoimmune complications. [12][13][14] Activation of T-cells by ICI can target antigens expressed in normal/nontumoral tissue, or off-target effects, likely a direct result of breaking immune tolerance, and leading to immunotherapy-induced adverse effects (AEs). 5 Immune-related adverse events (irAEs) are defined as any AE associated with drug exposure and consistent with an immune-mediated mechanism of action. 15 Such irAEs are frequent, the incidence can reach 50% in some studies. 6 Various types of irAEs which have been reported include gastrointestinal, hepatic, endocrine, and skin events, as well as pneumonitis, uveitis, infusion-related events and fatigue. 16 The occurrence of irAEs could also be considered as a prognostic factor of response to ICI.
Indeed, patients who experienced at least one irAE showed gains in progression-free survival and overall survival, as well as response rate. [17][18][19][20][21][22][23] At the same time, these irAEs may have an impact on patients' quality of life and can be life-threatening. In the event of severe irAEs, which are potentially life-threatening, the first treatment involves the use of a systemic immunosuppressive treatment, like corticoids, which may counteract the effect of ICIs. Thus, exposure to large amounts of immunosuppressive drugs in the case of high-grade irAEs would be expected to alter the antitumor effect. 24 The potential association between irAEs and outcome in patients with nonsmall cell lung cancer (NSCLC), on the one hand, and identification of patients most likely to experience severe irAEs, on the other, are therefore of utmost importance and deserve closer consideration.
The ToxImmune study thus aims to provide clinical practitioners with a simple tool, the ToxImmune score, suited for routine practice to help identify patients at risk of severe AE and thereby ensure early medical attention. Moreover, the score could also be taken into consideration as prognostic factor for response.

| PATIENTS AND METHODS
The ToxImmune study was a retrospective study conducted in the Oncology Department at Begin Military Hospital, France, where the study was declared and given prior approval by the local ethics committee.

| Patients
We included all patients aged 18

| Radiological parameters
Baseline SUVmax of primitive tumors before starting ICI treatment was taken into account. We reported the best response on ICI, according iRECIST.

| Treatment
Data were collected on: type of ICI received, alone or in combination, dates of start and of end of treatment, number of cycles, reasons for stopping treatment, best response. These data were then analyzed with the primary aim of establishing a score, called ToxImmune, which could be used routinely to identify patients at risk of immune AE among those with lung cancer receiving ICI. We defined the ToxImmune score (0; 1; 2) for immune-oncology (IO) related AE respectively as absence; 1; or ≥2 prognostic factors associated with grade ≥2 AE. The secondary objectives were to assess the impact of these AEs on survival and assess the potential of the ToxImmune score as a prognostic factor for survival and disease control rate (DCR, complete response + partial response and stable disease).

| Statistical analysis
Patient and tumor characteristics were compared using the Chisquared and Student's t tests. The optimal cutoff values as a prognostic variable for ICI toxicities were chosen from a receiver operating characteristic (ROC) curve computing area under the curve (AUC) with the criterion variable and "grade ≥2 AE during ICI treatment" as condition variables.
Survival times were calculated from the time of histological diagnosis and survival rates were estimated using the Kaplan-Meier method. Survival curves were compared using the log-rank test for the univariate analysis. Time-to-event endpoints were defined as the time between the date of first ICI injection and the last follow-up or first event, recurrence or death for PFS, and estimated by the Kaplan-Meier method. Time-to-event curves were compared using the logrank test. Hazard ratios (HR) and 95% confidence interval (CI) were estimated with univariate analysis. A p value of less than .05 was considered as significant.
Statistical analyses were performed using R (version 3.3.2).

| RESULTS
A total of 64 patients were included in our study and received at least one dose of ICI. The median age was 66 years (range 50-97), 72% were male and 16% presented at least one comorbidity. Fifty patients (72%) had a smoking history. Sixty patients (94%) were diagnosed with nonsmall cell lung cancer and 73% from the entire NSCLC cohort had a metastatic disease. Twenty patients (31%) had PDL1 > 1%. Thirteen patients (20%) presented more than 4 metastatic sites.
The majority of patients (98%) presented an ECOG performance status of 0-1 before starting ICI.
Patients' characteristics are summarized in Table 1.
The irAE profile is given in Table 2.
The percentage of patients who stopped immunotherapy due to toxicity was 14% in our population.
The median Progression Free Survival (PFS) was 10.6 months and the median Overall Survival (OS) was 19.3 months. 42% of patients received later-line treatments after the failure of ICI therapy ICI.
Based on this analysis, we established the ToxImmune score to predict the risk of having a grade ≥ 2 adverse event (primitive tumor  Female sex, although significantly associated with higher grades of AE, was not included in the calculation because, statistically, the Chi2 test of independence did not indicate that it could be dissociated from the number of metastases (p = .03). Given that, of the two, median number of metastases was more strongly associated with the risk of grade ≥2 AE than sex and more reliably so (p = .004 vs. p = .028 respectively), the number of metastases, rather than sex, was taken into account in the ToxImmune score, although female sex can still be considered as a complementary indicator.
PLR <250 was also significantly associated with the risk of AE grade ≥2, but given the infrequent use of PLR in clinical routine, it was not included in the ToxImmune score.
The DCR was 78% in the ToxIummune score = 0 group, 50% in ToxImmune score 1 and 2 groups, (p = .363) ( Figure 2). Likewise, there was a discernible connection between the ToxImmune score and the risk of "progressive disease" at the first assessment of the disease after initiation of ICI (at week 6): 24% for score = 0, 48% for score = 1, and 56% for score = 2, (p = .114).
In addition to predicting the incidence of grade ≥2 AEs, the Tox-Immune score thus appears as a more accurate prognostic factor for survival, both PFS and OS, than AEs alone.

| DISCUSSION
To the best of our knowledge, this is the first study establishing a predictive AE score, which can be used in clinical practice to identify patients at high risk of AE and monitor them more closely.
ICI treatment may cause irAEs in any part of the body, thus presenting a serious challenge for clinicians. The reported incidence of any-grade irAEs associated with ICI treatment ranges from 15% to 90%, depending on the therapeutic class of the ICIs and the type of cancer. 25 In our study, the incidence of irAEs was 90%, including 58% grade 1. The incidence of severe adverse event was 8%, similar to incidence rates reported elsewhere in the literature. The main AEs and irAEs reported were musculoskeletal disorders followed by digestive disorders, pneumonitis and dermatological disorders.
In accordance with our own findings, Valpione et al confirmed the impact of female sex on the occurrence of adverse events. 26 They reported that female sex was significantly and independently associated with higher risk of severe AE (OR = 1.5, 95% CI 1.06-2.16). In contrast, however, while both studies point to female sex as a prognostic factor for AE, our data do not show it to be independent from the number of metastases.
Other studies have shown a correlation between irAE and survival. Patients who experienced at least one irAE survived longer compared to patients with no irAE. [8][9][10][11][12][13][14][27][28][29][30][31][32][33] Our analysis suggests that the severity of irAE has an even greater impact on overall survival.  41 These biomarkers, however, have only a limited role in current practice and were therefore not included in the ToxImmune score to ensure that it could be calculated simply using commonly available information.
We acknowledge the limitations of our study, mainly because of its retrospective nature the sample size with an heterogenous population (different agents used, a mixture of locally-advanced and metastatic patients, and SCLC and NSCLC patients).
Nevertheless, our data came from a real-world scenario and provide a basis on which to calculate the ToxImmune score in patients receiving ICI treatment for lung cancer. A validation cohort will be used to assess the suitability and importance of the ToxImmune score as a prognostic factor for immune-mediated AE and prognostic factor for response to ICI and survival.

| CONCLUSION
ICI has clearly changed the prognosis for patients with metastatic lung cancer and has also shown an impact on survival in localized stages without metastasis. However, the risk of AE impacts patients' quality of life and may prove life-threatening. It is therefore essential to identify patients at risk of serious AEin order to prevent, identify and manage adverse events early on. The ToxImmune score provides a helpful tool for routine clinical practice and may prove useful in discussing the management of at-risk patients. ACKNOWLEDGMENT Daniel Henkel.

FUNDING INFORMATION
This research received no external funding.

CONFLICT OF INTEREST
The authors declare no conflict of interest with this work.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.