Retrospective study of R/M SCCHN patients that were treated with pembrolizumab-carboplatin-paclitaxel as first- or subsequent lines of therapy, via a compassionate use program for cisplatin-unfit or fluoropyrimidine ineligible patients. Criteria of cisplatin ineligibility were i) ECOG performance status (PS) of 2 and/or ii) creatinine-clearance (ClCr) <60 ml/min and/or iii) CTCAE Gr ≥2 hearing loss and/or iv) CTCAE Gr ≥2 neuropathy and/or v) history of ischemic heart disease and/or vi) history of heart failure. Criteria of fluoropyrimidines ineligibility were i) history of ischemic heart disease and/or ii) history of heart failure and/or iii) complete dihydropyrimidine dehydrogenase (DPD) deficiency and/or iv) severe hepatic insufficiency and/or v) unavailable central venous catheter placement (6,14). In addition, all patients treated with pembro + CP had to have a CPS ≥1.

Other inclusion criteria were: i) ≥18 years of age; ii) confirmed R/M SCCHN of the oropharynx, oral cavity, hypopharynx or larynx progressing on or after a previous line of systemic therapy and not eligible for a curative-intent therapy; iii) Cisplatin-unfit and/or ineligible for fluoropyrimidines as per the above definitions; iv) WHO/ECOG Performance status of 0, 1 or 2 at the time of starting pembro + CP; v) Adequate organ function as defined by: hemoglobin ≥9.0 g/dl, absolute neutrophil count ≥1,500/mm³, platelets ≥75,000/mm³, total bilirubin ≤1.5× upper level of normal (ULN), AST and AST ≤2.5 ULN, CrCl ≥30 ml/min as per the Cockcroft-Gault formula; vi) Body weight >30 kg; vii) a minimum life expectancy of 12 weeks

Exclusion criteria were: i) a histologically confirmed head and neck cancer of any other primary anatomic site, an unknown primary site SCCHN or a non-squamous histology; ii) any unresolved toxicity of CTCAE ≥ grade 2 except alopecia, vitiligo and laboratory values defined in the inclusion criteria; iii) a history or organ transplantation or active or previously documented autoimmune disorders with the exception of vitiligo, alopecia, stable and treated hypothyroidism; iv) active infection including tuberculosis, hepatitis B, hepatitis C or human immunodeficiency virus; v) an uncontrolled intercurrent illness such as ongoing or active infection, congestive heart failure, uncontrolled hypertension, unstable angina pectoris, uncontrolled cardiac arrhythmia, chronic gastrointestinal conditions associated with diarrhea or psychiatric illness or social conditions that would limit the compliance with treatment.

A descriptive study of baseline characteristics was performed. Three different geriatric and comorbidity scores were used: the modified Charlson Comorbidity Index (mCCI) and the Adult Comorbidity Evaluation-27 (ACE-27) with a mCCI ≥2 and ACE-27 ≥3 indicating moderate-to-severe comorbidity. The third scoring system was the Generalized Competing Event Composite Omega Score (GCE-COS) which encompasses different comorbidity scores to predict relative cancer vs. noncancer risk and may be accessed through an online tool, with a GCE-COS ≥0.6 indicating a higher overall survival with oncological treatment (15).

Adverse events (AEs) were recorded using the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0. ORR according to Response Evaluation Criteria in Solid Tumors (RECIST) 1.1, disease control rate (DCR), percentage of tumor change from baseline in target lesions (PCTL), progression-free survival (PFS), duration of treatment (DOT), and OS from the start of pembro + CP and from the start of first-line therapy were studied (16). Dose intensity for each agent and treatment compliance were also recorded.

The primary endpoint was safety. Secondary endpoints were, treatment compliance, ORR, DCR, PCTL, DOT, PFS and OS since pembro + CP (defined as the time from the start of pembro + CP until disease progression or death due to any cause, respectively), and OS since the first line of treatment (defined as the time from the start of first-line therapy until death from any cause). Considering that this is a heterogenous series, with patients receiving pembro + CP at different lines of therapy, efficacy results are presented using a swimmers plot depicting the treatment history from the start of pembro + CP and from the start of first-line therapy following the Trial Reporting in Immuno-Oncology (TRIO) guidelines (17).

Chemotherapy within the Pembro + CP regimen could be administered as a weekly or a three-weekly schedule, as follows: a) Pembrolizumab (200 mg/q3wk IV) plus three-weekly carboplatin (AUC 3–5) and paclitaxel (125–175 mg/m²); b) Pembrolizumab (200 mg/q3wk IV) day 1 plus weekly carboplatin (AUC 1.5-2) and paclitaxel (60–80 mg/m²) days 1, 8 and 15.

The final dosing of chemotherapy was allowed to be modified (within the ranges detailed above) in the first or subsequent cycles according to the ECOG performance status, comorbidities, or previous toxicities, as per the investigator's discretion.

RECIST 1.1 criteria were used for response evaluation, with CT scans performed every 8 to 12 weeks, according to the local protocol. Two expert radiologists in head and neck cancer evaluated each case independently. A consensus was achieved in case of discordant cases between the two.

A ≥30% decrease in the sum of the diameters of target lesions (TL) and a ≥20% increase in the sum of diameters of TL or the appearance of new metastatic lesions, defined a partial response (PR) and progressive disease (PD), respectively. Stable disease (SD) was defined as a change in the size of TL ranging between a 30% decrease and a 20% increase in the sum of their diameters (16).

DCR was defined as the percentage of patients achieving CR, PR and SD. Confirmed responses or SD had to be confirmed in at least two consecutive CT scans. Both confirmed and unconfirmed responses, SD and DCR were reported.

PD-L1 membrane expression was measured in tumor cells using an immunohistochemistry assay. The PDL1 IHC 22C3 pharmDx (Dako North America, Carpinteria, CA) immunohistochemistry assay, was used to measure PD-L1 expression in tumor and mononuclear stromal cells. At least a partial membrane staining in ≥1% of cells had to be present to consider a positive expression. The expression of PD-L1 in tumor cells only was used to calculate the tumor proportion score (TPS), as previously described (1). The expression of PD-L1 in tumor cells and mononuclear stromal cells were measured to calculate the CPS value, as previously described (18).

Dose-adapted pembro + CP was hypothesized to be less toxic than the combination of pembrolizumab + three-weekly platinum + 5FU (PF) tested in the Keynote-048 trial (2). The combination of pembrolizumab plus three-weekly platinum + 5FU is associated with a composite 94% of grade ≥3 AEs (summatory of incidence of hematological, gastrointestinal, and respiratory AEs). Dose-adjusted pembro + CP was expected to account for a 45% composite rate of grade ≥3 AEs. With an 80% power and a 95% CI (unilateral test), a total of 10 patients would have to be enrolled to demonstrate a 45% composite rate of grade ≥3 AEs with pembro + CP.

A descriptive analysis of demographic and clinicopathological data was performed. The Kaplan-Meier method was used to estimate PFS and OS, and the Fisher's Exact Test and Pearson's and Kendall's Correlations Tests were used to evaluate any association or correlations, respectively. The software SPSS Statistics for MacOS, version 23.0 (IBM SPSS Statistics for McOS, Version 23.0, Armonk, NY) was used for all statistical analyses.

All patients signed and informed consent form before starting pembro + CP as part of a compassionate use program, approved by the Compassionate Use Therapy Commission of Hospital Clínico Universitario San Carlos. The present study was approved by the Institutional Review Board of Hospital Clínico Universitario San Carlos and was conducted in accordance with the principles of the Declaration of Helsinki. The requirement of informed consent was waived as the study was based on a retrospective analysis of existing administrative and clinical data.

Between March 2020 and August 2021, ten patients were included, that were followed until March 31st, 2022. Baseline characteristics of the patients are summarized in Table I.

The most common primary tumor location was the oral cavity (70%). Seventy percent were stage IVA, 20% stage IVB and 10% stage IVC at initial diagnosis. Seventy percent had been treated with surgery, 80% with adjuvant radiotherapy (RT) or radical chemoradiotherapy (CRT), and 40% had received induction chemotherapy. In the R/M setting, 90% of the patients had received at least one line of therapy prior to pembro + CP, with a median of 3 prior lines (range: 0–6), and up to 60 and 80% had received prior platinum and prior ICIs in the R/M setting, respectively.

Sixty percent of patients were female, median age was 64 years old, and 40% were ≥70 years old. At the start of pembro + CP 20% of patients had ECOG 1 and 80% had ECOG 2. All patients were cisplatin-unfit, and, in addition, 2 patients were not eligible for fluoropyrimidines. Median mCCI score was 9 (range: 7–13), ACE-27 was 3 (range: 3–3) and GCE-COS was 0.812 (range: 0.714-0.867).

Pembro + CP regimens used were as follows: a) Pembrolizumab (200 mg/q3wk IV) plus three-weekly carboplatin (AUC5) and paclitaxel (150 mg/m²): n=1; b) Pembrolizumab (200 mg/q3wk IV) plus three-weekly carboplatin (AUC4) and paclitaxel (150 mg/m²): n=2; c) Pembrolizumab (200 mg/q3wk IV) plus three-weekly carboplatin (AUC3) and paclitaxel (150 mg/m²): n=5; d) Pembrolizumab (200 mg/q3wk IV) plus weekly carboplatin (AUC2) and paclitaxel (80 mg/m²) days 1, 8 and 15: n=2.

In the whole series of 10 patients, grade 1 or 2 AEs occurred in 100% all of them during the combination phase of pembro + CP. Grade 3–5 AEs developed in 30% of patients (Tables II and SI), being grade 3 in all of them and all of them occurring during the combination phase of pembro + CP as well. There were no treatment-related deaths.

Among patients treated with three-weekly pembro + CP (n=8), AEs of grades 1 or 2 occurred in 100% of the patients, the most common being asthenia, decreased appetite, anemia, peripheral neuropathy, hypomagnesemia, and alopecia. Grade 3–5 AEs occurred in 3 patients (37.5%), with 2 patients suffering grade 3 neutropenia, 1 grade 3 thrombopenia, 1 grade 3 hypomagnesemia, and 1 patient with grade 3 hypertension, the latter considered unrelated to pembro + CP (Tables II and SI).

Among patients treated with pembrolizumab combined with weekly CP (n=2) both patients suffered from AEs of grades 1 or 2 while no grade 3–5 AEs occurred. Tables II and SI summarize adverse events during pembro + CP in the whole population and in each individual patient, respectively.

Among the 8 patients treated with anti-PD(L)1 agents in previous lines in the R/M setting, three patients had suffered from irAEs: Patient 5 suffered from hyperprogressive disease after 3 doses of first-line nivolumab that was successfully rescued with cetuximab-based chemotherapy, Patient 6 suffered from grade 3 pneumonitis after 5 doses of fifth-line nivolumab, and Patient 9 developed grade 2 hypothirodism after 6 doses of first-line nivolumab. None of them, suffered from these or any other irAEs during treatment with pembro + CP.

Median DOT with pembro + CP was 4.5 months (Min-Max: 0.5-24). Median number of combination pembro + CP cycles was 3 (range: 1–11). Five patients also received maintenance pembrolizumab [median 1 cycle (range: 1–25)].

Among patients who received pembro + 3wkCP (n=8) median number of pembrolizumab, carboplatin and paclitaxel cycles 6 (range-1-11). Only three patients suffered delayed administration of pembro + 3wkCP due to toxicity, with a median of 1 cycle delayed (range: 1–2) and a median delay of 7 days (range: 7–14).

Among patients who received pembro + wkCP (n=2) one patient received 3 cycles of three-weekly pembrolizumab and the other patient received 2 cycles. Number of weekly carboplatin and paclitaxel doses were 2 in one patient and 3 the other patient, respectively. None of the patients suffered delays in treatment administration.

All patients discontinued pembro + CP due to radiological and/or clinical progression.

Detailed dosages of each agent received by each patient are summarized in Tables III and SII.

Among 7 evaluable patients, ORR was 14% (1/7), with 1 complete response (CR) in a patient that received pembro + 3wkCP as second line therapy. Three patients showed stable disease (SD) and 3 progressive disease (PD) as best responses during pembro + CP. DCR after the first CT scan evaluation was 43%. Confirmed DCR (after at least 2 consecutive CT scan evaluations) was 14%. The patient achieving a CR to pembro + CP was the only one where ORR was confirmed with at least two consecutive CT scans. Therefore, the confirmed ORR was 14%. Median PCTL was 10% (range: −100% to +80%) (Tables IV and V).

In the whole population, after a median follow-up of 6.5 months (range: 0.5-24), median OS since pembro + CP was 6 months (95% CI 0.5-14) and PFS was 5 months (95% CI 1–9). After a median follow-up of 25.5 m (range: 7–63), median OS since first line was 30 months (95% CI 18–42). Median PFS and OS in the 8 patients treated with pembro + 3wkCP were and 3 months (95% CI 1.1-4.8) and 11 months (95% CI 1.4-20.6), respectively (Tables IV and V, and Fig. 1, Fig. 2, Fig. 3, Fig. 4).

Fig. 4 summarizes the treatment journey since the first line of therapy in each of the 10 patients included in the study.

All patients had a CPS ³1, and 8 patients harbored a CPS ≥20. Patient 3 (CPS=10) and patient 7 (CPS=1) had a CPS <20. In all patients CPS was performed in an archived tumor sample obtained at initial diagnosis, and therefore in all cases other therapies-either in the early or R/M settings-had been administered between the time of biopsy collection and the start of pembro + CP. No statistically significant associations nor correlations were found between PDL1 expression measured through the TPS or CPS and ORR, PFS and OS in the whole series as well as in the 8 patients treated with pembro + 3wkCP (Table SIII).