Abstract
Background/Aim: This study aimed to assess the effectiveness and safety of nivolumab versus cetuximab in patients with Recurrent/Metastatic Squamous Cell Carcinoma of the Head and Neck (R/M HNSCC), as well as to analyze possible prognostic factors for response to treatment with nivolumab. Patients and Methods: We conducted an observational, retrospective, descriptive study of patients with R/M HNSCC who initiated treatment with nivolumab or cetuximab monotherapy in two periods of equivalent duration. Overall efficacy was measured in terms of progression-free survival (PFS) and overall survival (OS). Safety was evaluated using the Common Terminology Criteria for Adverse Events classification version 5.0 of the National Cancer Institute. Results: Median OS was 9.1 months with nivolumab (n=34) and 6.3 months with cetuximab (n=12). PFS was 4.3 months for nivolumab and 4.65 months for cetuximab. Any grade adverse events (AEs) were reported in 97% and 100% of the patients treated with nivolumab and cetuximab. Serious AEs were observed in 26% and 58% of the patients, respectively. Elevated albumin values, lymphocytosis, neutropenia, and elevated neutrophil/lymphocyte ratio values were found to have positive prognostic value on the response to nivolumab in R/M HNSCC. Conclusion: Effectiveness of nivolumab in terms of OS remains superior to cetuximab. OS, PFS and severe or any grade AEs were superior in both arms of our study compared to those in clinical trials. The AEs profile of nivolumab differed in our study from that in the clinical trials’ observations. We have identified four statistically significant prognostic variables on the response to nivolumab in R/M HNSCC.
- Nivolumab
- head and neck cancer
- squamous cell carcinoma
- prognostic factor
- cetuximab
- immunotherapy
- oral cancer
Head and neck cancer accounts for 5% of adult cancers, with head and neck squamous cell carcinoma (HNSCC) being the most frequent, accounting for 90% of all head and neck tumors (1). HNSCC is the seventh most common cancer worldwide. In Europe, it has an annual incidence of 90,000 cases; distributed among carcinoma of the larynx (4.6/100,000), oral cavity (3.5/100,000), oropharynx (3.1/100,000), hypopharynx (1.3/100,000), and nasopharynx (0.4/100,000). Nasopharynx carcinoma is considered an independent entity and therefore, is not included in the scope of this study (1-4).
There are multiple risk factors related to the development of this pathology; however, up to 75-85% of HNSCCs are due to smoking and alcohol. Men are more likely to develop HNSCC, with a 10:1 ratio, and the average age of appearance is above 50 years. The presence of the human papillomavirus (HPV) plays an important role in the appearance of oropharyngeal carcinoma, and is responsible for up to 35% of the cases (1, 3).
Five-year relative survival is 61% for laryngeal, 49% for oral cavity, 41% for oropharyngeal, and 25% for hypopharyngeal carcinoma (2). However, those that progress within six months of platinum treatment have a median overall survival (OS) of 6 months (1).
The TNM classification takes into account the location, the size of the tumor mass and its extension to neighboring structures or lymph nodes, the presence or not of metastases, as well as the presence of HPV to establish the stage of the disease, classifying it as: local or locoregional, locally advanced, recurrent, or metastatic (R/M) disease (4-6).
Two-thirds of patients are diagnosed at locally advanced stages. The presence of metastases at the time of diagnosis is rare, not exceeding 10% (3, 7). However, 10-20% of patients in early stages and 50% of those with advanced disease will present with locoregional or distant relapses (3, 8).
Depending on the location and stage of the disease, the therapeutic approach will be decided, based on three fundamental pillars: surgery, radiotherapy, and systemic therapy (2, 3). In the case of R/M HNSCC without the possibility of a surgical or radiotherapeutic approach, the first-line approach has historically been based on chemotherapy. Treatment was decided based mainly on whether the patient had received previous chemotherapy and their general condition (1, 6, 7).
In patients who had not received previous chemotherapy and with good general condition, the treatment of choice was the combination of cisplatin (or carboplatin), 5-fluorouracil and cetuximab (EXTREME protocol); in case of poor general condition or if the patient had previously received platinums, the combination of paclitaxel and cetuximab (ERBITAX scheme) could be considered (1, 6).
Patients who had progressed after more than 6 months of chemotherapy were considered naïve; however, in those who progressed in a period of less than 6 months, treatment options were more limited, and in these cases, second-line treatment with the ERBITAX scheme or with cetuximab, methotrexate and/or taxanes monotherapy should be considered. A progression-free survival of 2-3 months was observed in these cases (6, 9).
Nivolumab was approved in 2016 by the Food and Drugs Administration (FDA) and in 2017 by the European Medicines Agency (EMA) for the treatment of patients with R/M HNSCC that had progressed within 6 months of platinum treatment. The approval was based on the phase III Checkmate-141 study, which compared nivolumab versus the investigator’s monotherapy treatment of choice (methotrexate, docetaxel or cetuximab) (10-12).
Nivolumab is a fully human monoclonal immunoglobulin G4 (IgG4) antibody that binds to the programmed death receptor (PD-1) on T cells, blocking its interaction with its ligands (PD-L1 and PD-L2) thereby restoring antitumor immunity (11, 12). After its approval, nivolumab has been positioned in clinical practice guidelines as the treatment of choice in refractory or relapsed patients after less than 6 months of platinum treatment and who have not received previous immunotherapy (2, 6, 11, 13). In our center, nivolumab is not included in the pharmacotherapeutic guidelines for use in R/M HNSCC and requires authorization on an individualized basis.
Immunotherapy is a generally well-tolerated treatment compared to alternatives such as chemotherapy (7, 14). In the clinical trials of nivolumab for HNSCC, percentages of AEs of any grade were observed in 58.9% of patients in the nivolumab arm versus 77.5% in the control arm. Serious AEs occurred in 13.1% of patients in the nivolumab arm and 35.1% in the control arm. The main AEs observed with nivolumab were fatigue, nausea, rash, lack of appetite, and pruritus (10, 15).
The introduction of immune checkpoint inhibitors has been a revolution in the approach to this and other types of neoplasms (14). It has been observed that there are groups of patients who will achieve, with immunotherapy, long therapeutic responses (16, 17). It is interesting to find predictors that allow us to identify which patients will benefit most from immunotherapy. A study focused on metastatic colorectal cancer with microsatellite instability has recently proposed a nomogram for predicting response, based on six variables: regimen used (doublet with anti CTLA4 or anti PD-1 in monotherapy), Eastern Cooperative Oncology Group Performance status (ECOG) (0 or higher), number of previous lines received, neutrophil/lymphocyte ratio and platelet count (16). A recent study focused on head and neck cancer showed a relationship between the inflammation-based prognostic score (IBPS) after treatment initiation with nivolumab and OS (18).
The objectives of this study were: to evaluate the overall efficacy, measured in progression-free survival (PFS) and overall survival (OS) of nivolumab compared to cetuximab in monotherapy; to evaluate the safety of nivolumab in the treatment of R/M HNSCC compared to cetuximab; and to analyze possible prognostic factors for good response to immunotherapy.
Patients and Methods
An observational, retrospective, descriptive study of patients with R/M HNSCC who initiated treatment with nivolumab at a Spanish Third Level Hospital between January 2018 and June 2021 was performed. This study was approved by the Principado de Asturias Ethics Committee (approval no. 2022.020). The Ethics Committee of the Principado de Asturias waived the need to obtain consent for the collection, analysis, and publication of the retrospectively obtained and anonymized data for this non-interventional study.
Patients with R/M HNSCC treated with cetuximab monotherapy between June 2014 and January 2018, a period of equivalent duration and prior to the inclusion of nivolumab as a therapeutic alternative in these patients, were used as comparators. The cut-off date for data collection was 11/31/2021.
Demographic variables, clinicopathological subtype, date of progression or death due to any cause, reason for discontinuation of treatment, presentation of mutations in the biopsy (PDL-1, p16, p53, p63) and occurrence of AEs were obtained from the electronic medical record (Cerner Millenium®).
These parameters were recorded the day before and/or closest to the start of treatment, or, if not available, the day immediately after the start of treatment; these values where never more than one month apart from the first cycle with nivolumab. Missing values were replaced by the sample median. The parameters collected and evaluated were: lactate dehydrogenase (LDH), aspartate dehydrogenase (AST), alanine aminotransferase (ALT), albumin, hemoglobin and neutrophil, lymphocyte, and platelet counts.
The number of patients, dosing regimen, number of cycles, duration of treatment and previous chemotherapy treatment were obtained from the oncohematologic patient management program used at our hospital (Farmis_Oncofarm®). Continuous variables were presented as median and interquartile range; categorical variables as number of observations and percentage.
The efficacy criteria were OS and PFS calculated by nonparametric analysis using the Kaplan–Meier method. The influence of the variables on efficacy was measured, based on their influence on OS, by Cox proportional hazards analysis. All analyses were performed using R Statistical Software (v4.0.0; R Core Team April, 2020).
The evaluation of the development of toxicities was performed using the Common Terminology Criteria for Adverse Events (CTCAE) classification version 5.0 of the National Cancer Institute (NCI). Serious AEs were considered to be those of grade 3 or higher.
Results
During the study period, 34 patients started treatment with nivolumab and 12 patients with cetuximab. In the nivolumab group (82% male), the median age was 64.5 years (range=60-71 years). The main tumor sites were larynx (24%, 8), oral cavity (18%, 6), hypopharynx (15%, 5), tongue (12%, 4), glottis (6%, 2), supraglottis (6%, 2), oropharynx (6%, 2), tonsil (3%, 1) and pharynx (3%, 1). In 3 cases (9%) the neoplasm was of unknown origin. In the cetuximab group (83% male), the median age was 67 years (range=61.5-69.5 years). The main tumor sites were oral cavity (33%, 4), oropharynx (33%, 4), larynx (17%, 2) and hypopharynx (8%, 1).
All patients treated with nivolumab had received previous chemotherapy treatment, with nivolumab being used as second line in 65% [22] of the cases, as third line in 29% [10], and as fourth line in 6% [2] of cases. In addition, 84% [27] had received previous radiotherapy and 74% [23] had undergone surgery.
Of the patients treated with cetuximab, 83.3% [10] had received previous radiotherapy, and 75% [9] had undergone surgery. Similarly, as in the nivolumab group, 100% had received previous systemic chemotherapy, with cetuximab being used as second (50%, 6), third (33%, 4) or fourth (17%, 2) line of treatment.
The median duration of treatment in the nivolumab group was 3 months (range=1.5-8.6 months) and the median number of cycles received was 5 (range=3-11.75 cycles). At the cut-off date, one patient was still on treatment with nivolumab; in the remaining patients, the main reasons for discontinuation were disease progression (52%, 17), appearance of AEs (23%, 8), deterioration of general condition (12%, 4), or death (12%, 4).
In the cetuximab group, the median duration of treatment was 1.4 months (range=0.95-5.72 months). The median number of cycles received was 7 (range=4.75-23 cycles). Cetuximab was discontinued for the following reasons: disease progression (58%, 7), poor general condition (17%, 2), death (17%, 2) or AEs onset (8.3%, 1).
Overall survival. At the cut-off date, 100% of patients treated with cetuximab and 74% [25] of those treated with nivolumab had died. Median overall survival was superior in the nivolumab group, with a median of 9.1 months (95%CI=8-24) versus the cetuximab group, in which a median overall survival of 6.3 months (95%CI=3.4-NA) was observed (HR=0.5; 95%CI=0.24-1.03; p=0.058) (Figure 1).
Kaplan–Meier overall survival (OS) curve stratified by treatment used: cetuximab (red) or nivolumab (blue).
Progression-free survival. At the study cut-off date, 31 (91%) of patients on nivolumab treatment and 12 (100%) of patients on cetuximab treatment had progressed. PFS in the nivolumab-treated group was 4.3 months (95%CI=3.6-9.7) vs. 4.65 months (95%CI=2.8-NA) in the cetuximab arm (HR=0.59; 95%CI=0.29-1.19; p=0.14) (Figure 2).
Kaplan–Meier progression-free survival (PFS) curve stratified by treatment used: cetuximab (red) or nivolumab (blue).
Toxicity. Regarding the onset of AEs, a total of 106 AEs were recorded in the nivolumab group, affecting 97% [33] of patients. Of these, only 10 (9%) were serious. Serious AEs affected 26% [9] of patients.
The main AEs reported were related to blood and lymphatic system (35%, 37), metabolism and nutrition (27%, 29), endocrine (13%, 14), analytical (10%, 11), gastrointestinal (3%, 3), nervous system (3%, 3), skin and subcutaneous tissue (3%, 3), general disorders (2%, 2), hepatobiliary disorders (1%, 1), respiratory system (1%, 1), immune system (1%, 1) or renal and urinary system (1%, 1).
Severe AEs occurred mainly at the level of the blood and lymphatic system, with 4 cases of grade 3 lymphopenia and 3 cases of grade 3 anemia. There was also one case of grade 3 liver failure, one grade 3 vasovagal reaction, and one grade 4 hypoglycemia in a patient with a history of diabetes mellitus. In the cetuximab arm, AEs were observed in all patients, with a total of 65. Of these, 14% [9] were serious and occurred in 58% [7] of patients.
The main AEs reported with cetuximab were metabolic and nutritional (29%, 19), analytical alterations (15%, 10), skin and skin tissue alterations (12%, 8), gastrointestinal (11%, 7), general disorders (11%, 7), blood and lymphatic system (11%, 7), occurrence of infections (3%, 2), respiratory disorders (3%, 2), vascular (2%, 1), psychiatric (2%, 1) or immune system disorders (2%, 1).
Serious AEs were observed at the analytical level with one event of lymphopenia grade 3 and one of increased GGT grade 3; at the respiratory level with one event of dyspnea grade 3; at the hematological level with one event of anemia grade 3; one event of dysphagia grade 3; one case of hyperuricemia grade 3. Two grade 4 AEs were observed: one episode of confusion and one case of anaphylaxis.
Predictors of response to nivolumab. In the statistical analysis according to ECOG at baseline, the median OS was 17.2 months (95%CI=9.73-NA) in the ECOG 0 group (n=4); 9 months (95%CI=5.93-NA) in ECOG 1 (n=28), and 4.4 months (95%CI=2.27-NA) in ECOG 2 (n=2). Figure 3 shows the probability of OS stratified according to ECOG.
Kaplan–Meier overall survival (OS) curve stratified by ECOG: 0 (red), 1 (green), 2 (blue).
The median PFS was 7 months (95%CI=3.96-NA) in the ECOG 0 group (n=4); 4.3 months (95%CI=3.10-12.17) in ECOG 1 (n=28), and 2.63 months (95%CI=2.26-NA) in ECOG 2 (n=2).
Based on the line of treatment in which nivolumab was used, OS was superior in those cases in which nivolumab was used in second-line (n=22), with a median of 9.7 months (95%CI=8.6-NA) versus those in which it was used in third-line or later (n=12) in which the median OS was 7 months (95%CI=3.1-NA) (HR=1.32; 95%CI=0.57-3.01; p=0.52) (Figure 4).
Kaplan–Meier overall survival (OS) curve stratified by treatment lines: second line (red) or subsequent (blue).
Median PFS was 4.15 months (95%CI=3.63-NA) in the group in which nivolumab was used in second-line versus 4.93 months (95%CI=2.3-NA) in cases in which it was used in subsequent lines.
OS as a function of neutrophil-to-lymphocyte ratio (NLR) was 11.7 months (95%CI=4.5-NA) in the arm with NLR greater than 3 (n=11) and 9.03 months (6.53-NA) in the arm with NLR less than 3 (HR=1.24; 95%CI=0.52-2.94; p=0.63) (Figure 5). PFS was 4.3 months (2.8-NA) in the NLR >3 group, compared to 4.27 (3.1-12.13) in the NLR <3 group (HR=0.91; 95%CI=0.42-1.98; p=0.81).
Kaplan–Meier overall survival (OS) curve stratified by neutrophil/lymphocyte ratio: greater than 3 (red) or less than 3 (blue).
The analysis of the different variables studied on OS, performed by Cox regression, is detailed in Figure 6. A statistically significant influence of albumin, neutrophil and lymphocyte count and NLR on OS can be observed.
Cox proportional hazards regression model. Influence of variables on overall survival (OS). LDH: Lactate dehydrogenase; AST: aspartate dehydrogenase; ALT: alanine aminotransferase; NLR: neutrophil/lymphocyte ratio.
Discussion
Our study, performed under real-life conditions, shows results that exceed both OS and PFS compared to those observed in the clinical trials of nivolumab (15). No increase in PFS was observed in patients treated with nivolumab; however, a benefit in terms of OS was observed with nivolumab compared to cetuximab.
With regard to AEs, we observed much higher incidences than those described in the literature, both of any grade or severe AEs. These serious AEs appeared more frequently in the cetuximab than in the nivolumab arm.
The main AEs observed in our study differed from those of the clinical trials (10, 15). A higher incidence of AEs of any grade was observed at the hematologic level, as well as alterations of metabolism and nutrition, and endocrine disorders that were not described in the clinical trials of nivolumab in HNSCC. A lower incidence of cutaneous AEs was also observed. Similarly, the serious AEs observed in our study were blood and lymphatic system disorders, affecting 18% [3] of patients treated with nivolumab, a figure that differs from the 1.3% observed in the clinical trials (15). The number of patients studied needs to be increased to be able to verify this trend and its relevance in clinical practice. Furthermore, when evaluating AEs, we must take into account that some of them may not have been recorded in the electronic medical record because they are well-known AEs of the drug or because they are not considered relevant for the medical oncologist responsible of the patient. It would be interesting to have a patient reported outcomes (PRO) system to evaluate more subjective AEs or those of a lesser degree.
Although clinical guidelines recommend evaluating the presence of PD-L1 mutations in R/M HNSCC and base the therapeutic approach decision on their expression or not (2, 6, 7), most of the patients in our study did not have this determination or other anatomic-pathological tumor markers, therefore it was not possible to determine the influence of these mutations on the response to nivolumab.
The presence of ECOG 0 has been proposed as a predictor of good response to immunotherapy (16); however, in our study it was not possible to make this comparison because there were not enough patients in the ECOG 0 group. This may be due to the functional status of the patients in real life conditions, which will be influenced by age, previous treatment with platinums and the etiology of the cancer. Despite this, there is a trend towards longer survival in terms of OS and PFS at lower ECOG, although these are not statistically significant.
Nivolumab’s technical data sheet and clinical guidelines recommend its use in second line after lack of response or recurrence to platinum therapy (2, 11-13); however, in our study, we observed a high percentage of patients in whom it was used in subsequent lines; in these cases we observed a worse response in terms of OS, although this difference was not statistically significant.
The NLR has been proposed as a predictor of response to immune checkpoint inhibitors (16). In our study, using the Kaplan–Meier model, a trend towards greater OS was observed in the group with a NLR less than three; however, this difference was not statistically significant. Regarding PFS, no differences were observed based on the NLR.
We wanted to evaluate the potential of various analytical values as predictors of response to immunotherapy, and we observed that the effectiveness of nivolumab in R/M HNSCC measured in terms of OS was significantly conditioned by the value of albumin, neutrophils, lymphocytes, and NLR. Elevated values of albumin, lymphocytes and NLR, as well as neutropenia, were found to be predictors of a better response to treatment with nivolumab, with an increase in OS. No statistically significant differences were observed for the other parameters studied.
A recent study focused on head and neck cancer showed that the inflammation-based prognostic score (IBPS) had an important role as prognostic response factor, but that it should be considered after treatment initiation. The IBPS post-nivolumab had a significant relationship with OS, but this relationship was not observed with the pre-nivolumab IBPS values (18). These results contrast with those observed in our study, reinforcing the need of larger cohort studies to verify the validity of these prognostic response factors.
When evaluating the results, it should be taken into account that the sample size was not calculated before the study was carried out, since all the patients in the selected population were included. Having a larger population in both the nivolumab and cetuximab arms would have increased the statistical power of the results.
In conclusion, both nivolumab and cetuximab have demonstrated OS and PFS values in our study that are superior to those observed in the clinical trials. However, the effectiveness of nivolumab in terms of OS remains superior to that of cetuximab. A much higher incidence of severe or any grade AEs was observed in both arms of our study compared to that observed in the clinical trials. The AEs profile of nivolumab differs in our study from the clinical trials’ observations. We have identified four positive statistically significant prognostic variables on the response to nivolumab in R/M HNSCC.
Footnotes
Authors’ Contributions
Guarantor of integrity of the entire study: Lola Macía-Rivas; Study concepts and design: Lola Macía-Rivas, Mónica Carbajales-Álvarez, Cristina Álvarez-Asteinza; Literature research: Lola Macía-Rivas, Cristina Álvarez-Asteinza; Experimental studies/data analysis: Lola Macía-Rivas, Clara Luz Fernández-Laguna; Statistical analysis: Lola Macía-Rivas, Iván Maray-Mateos; Manuscript preparation: Lola Macía-Rivas, Mónica Carbajales-Álvarez; Manuscript editing: Lola Macía-Rivas, Ana Lozano-Blázquez, Cristina Álvarez-Asteinza, Clara Luz Fernández-Laguna, Iván Maray-Mateos.
Conflicts of Interest
The Author(s) declare that there are no conflicts of interest in relation to this study.
- Received January 23, 2023.
- Revision received February 6, 2023.
- Accepted February 7, 2023.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
