Chemoradiotherapy with high-dose cisplatin compared to weekly cisplatin for locally advanced head and neck squamous cell carcinoma

Abstract

Introduction:

Concurrent chemoradiotherapy (CRT) using high-dose cisplatin (HDC) is standard for patients with locally advanced head and neck squamous cell carcinoma (HNSCC); weekly cisplatin (WC) is an alternative. We aim to compare retrospectively the survival and disease control outcomes between these regimens in our institutional experience.

Methods:

Patients with stage III-IV HNSCC treated with definitive or postoperative CRT between 2012–2018 were identified. Patients were stratified by intent-to-treat CRT. overall survival (OS) and disease-free survival (DFS) were generated and multivariable Cox models were performed.

Results:

193 patients were treated with concurrent HDC (n=69), WC at 40 mg/m2 (WC40, n=88), or WC at <40 mg/m2 (WC<40, n=36). Treatment intent was definitive in 74% and adjuvant in 26%. Baseline differences included age, performance status, and HPV status. Cumulative cisplatin dose ≥200 mg/m2 was achieved in 89% (HDC), 86% (WC40) and 25% (WC<40, p <0.0001). For HDC, WC40 and WC<40, 2-year OS rates were 87%, 77%, 60% and 2-year DFS rates were 75%, 68% and 52%, respectively. Multivariable analysis revealed gender, performance status, primary site, T/N stage and chemotherapy as predictive of OS. Primary site, T/N stage and chemotherapy regimen were associated with DFS. Compared to HDC, no differences in locoregional control (LRC) or distant metastasis were observed between groups.

Conclusions:

Concurrent HDC is associated with increased total cisplatin intensity, OS and DFS compared to weekly cisplatin regimens. LRC was not associated with chemotherapy regimen. HDC remains the standard of care; WC40 is a reasonable alternative that does not appear to sacrifice LRC.

Introduction

Concurrent chemoradiotherapy (CRT) is standard of care for treatment of locally advanced head and neck squamous cell carcinoma (HNSCC). The addition of high-dose cisplatin (HDC) to radiotherapy (RT) has consistently improved survival (1–20%), disease-free survival (10–22%), and locoregional control (11–24%) in the definitive and adjuvant settings (1–6). However, the addition of cisplatin to RT is associated with increase in mucosal, hematologic, renal and neurologic toxicity (1, 3, 7). Other regimens such as weekly cisplatin (WC) or carboplatin/paclitaxel (CP) delivered weekly have been reported as potential alternatives (8–12).

Recent studies comparing HDC and weekly low-dose cisplatin have been published with conflicting results. A randomized study from Tata Memorial Hospital showed significantly reduced locoregional control (LRC) after CRT with WC rather than HDC. CRT was delivered using conventional techniques with total prescription doses of 70 Gy (definitive intent, 7%) or 60 Gy (adjuvant intent, 93%). No differences were observed in progression-free survival or overall survival (OS) between the two groups, and HDC was associated with increased acute toxicity (13). Another study found no differences between the groups after propensity score adjustment to account for selection bias inherent in this study population, as frailer patients with worse outcomes more likely to be treated with WC (14). Importantly, in this study, HDC was also associated with increased hazard of acute kidney injury, neutropenia, dehydration/electrolyte abnormality, and hearing loss on adjusted analyses. Comparison of various dose levels of WC showed inferior OS in patients receiving WC at a dose of less than 40 mg/m² (WC<40) compared to WC at a dose of 40 mg/m² (WC40) or HDC (15). Variations in WC dosing may contribute to prior findings suggesting inferiority of this regimen. We aimed to evaluate and compare the clinical efficacy of WC and HDC in our institutional experience.

Methods;

Patient Population and Treatment

Patients with HNSCC of the oral cavity, oropharynx, hypopharynx, and larynx treated with definitive or adjuvant CRT between 2012–2018 (n=428) were identified from an Institutional Review Board-approved database using REDCap electronic data capture tools hosted at Wake Forest School of Medicine (16). The Wake Forest School of Medicine Institutional Review Board granted approval for this study. Exclusion criteria were: recurrent or second primary disease (n=50), prior history of RT (n=2), early stage (AJCC 7^th Ed. Stage I-II; n=11) or metastatic (AJCC 7^th Ed. Stage IVC; n=17) disease, induction chemotherapy (n=36), RT at outside facility (n=58), incomplete RT (less than 66 Gy for definitive therapy, less than 60 Gy for adjuvant therapy; n=8), and other non-cisplatin chemotherapy (n=53). Using an intent-to-treat methodology, patients were categorized into three groups based on the first cycle of chemotherapy given concurrent with RT: HDC (100 mg/m2 every 3 weeks for 2–3 cycles at the discretion of the treating oncologist), WC<40, or WC40.

Patients were evaluated in a multidisciplinary head and neck oncology setting including otolaryngologists, radiation oncologists, medical oncologists, radiologists, pathologists, speech language pathologists and dieticians. After CT simulation with thermoplastic mask immobilization, 70 Gy in 35 fractions was delivered to gross disease with 56–63 Gy to low/intermediate risk elective/involved nodal regions for definitive therapy and 60–66 Gy in 30–33 fractions to the tumor bed and involved nodal regions with 54–59.3 Gy to elective nodal regions for postoperative treatment (17). The chemotherapy regimen was determined at the discretion of the treating medical oncologist. Per standard practice at our institution, high dose cisplatin is preferred, and the preferred dose of weekly cisplatin (if used) is 40 mg/m². Generally, we utilized lower weekly doses in patients with comorbidities that affected or could affect the kidney function, in patients with comorbidities that could limit tolerance for intra-venous hydration, in patients with pre-existent neurologic or hematologic conditions, or in patients with poor performance status.

Baseline factors such as age, gender, ECOG performance status, comorbidity measured by the Charlson Comorbidity Index (excluding cancer diagnoses, since all patients in the study cohort have cancer and chemotherapy selection is most often based on non-cancer comorbid disease) (18), tobacco/alcohol use, primary disease site, HPV status, AJCC 7^th Edition TNM stage, and indication for adjuvant CRT (positive margins, extracapsular extension, or both) were abstracted from the medical record. HPV positivity was evaluated using either HPV DNA polymerase chain reaction assay, immunohistochemical staining for the p16 surrogate marker, or HPV DNA in-situ hybridization.

Outcomes

All time-to-event outcomes were calculated from the date of completion of CRT. OS was defined as the duration to death from any cause or last follow-up (right censored). Disease-free survival (DFS) was defined as the duration to any local, regional or distant disease recurrence or death, or last follow-up (right censor). Cancer-specific survival (CSS) was defined as duration to death due to head and neck cancer or (for those who did not experience cancer-specific mortality) to last follow-up or death from any other cause (right censor). Death due to head and neck cancer was defined as any death directly caused by local, regional or distant disease or death in the presence of actively progressive disease. LRC was defined as lack of disease recurrence at the site of the primary tumor or cervical lymph node basins at risk (yes versus no) and distant failure (DF) was defined as the development of distant metastatic disease outside the head and neck (yes versus no). The following adverse events (AEs) that occurred while the patient was on active treatment or within 30 days after the completion of treatment were abstracted from the medical record: any hospitalization, febrile neutropenia, or treatment-related mortality (death deemed at least possibly related to treatment per the National Cancer Institute guidelines for AE reporting) (19). The indication for hospitalization was recorded and graded according to the Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0 (20).

Statistical Analysis

We compared differences between the three chemotherapy groups using chi-square tests for categorical variables and F-tests (ANOVAs, for tests of means) or Kruskal-Wallis tests (for tests of medians) for continuous/ordinal variables. For the entire sample, we estimated median time of follow-up using the reverse Kaplan-Meier method (21). We compared OS, DFS, CSS, LRC, and DF between the three chemotherapy groups using the Kaplan-Meier method and the log-rank test. For bivariate OS and DFS, we also used Cox proportional hazards models to identify clinical factors, in addition to chemotherapy group, associated with OS and DFS. Factors identified through bivariate analyses which were statistically significantly associated with the outcome at p ≤ 0.20 were entered into a multivariable model to assess their impact while adjusting for other potential predictors. The Cox proportional hazards assumptions were tested for all models. Analyses were performed using SAS version 9.4.

Results

Patient Cohort & Treatment Delivery

In total, 193 patients with locally advanced HNSCC treated with concurrent CRT were included in the analysis. Sixty-nine patients received HDC, 88 patients received WC40, and 36 patients received WC<40. Overall, patients receiving WC<40 were older, had worse performance status, and had higher comorbidity scores (Table 1). The majority of patients in each chemotherapy group were treated with definitive CRT using intensity-modulated RT or volumetric modulated arc therapy (Table 2). Median prescribed RT dose was 70 Gy (range, 60–70.2) for all groups. The median duration of elapsed time on treatment (in days) did not differ between groups (HDC: 49, WC40: 49, WC<40: 50). As computed with the reverse Kaplan-Meier method, median follow-up for the entire sample was 41.9 months (95% CI 36.6–48.4 months).

Table 1:

Patient Baseline Characteristics

	High-dose Cisplatin (n=69)	Weekly Cisplatin, 40 mg/m2 (n=88)	Weekly Cisplatin, <40 mg/m2 (n=36)	p-value
Age in years	57 (52–63)	60 (53–64.5)	63 (58.5–70)	<0.00
≤65	54 (78.3)	66 (75.0)	19 (52.8)	01*
>65	15 (21.7)	22 (25.0)	17 (47.2)	0.02
Gender
Male	55 (79.7)	73 (83.0)	29 (80.6)	0.87
Female	14 (20.3)	15 (17.0)	7 (19.4)
ECOG Performance Status	62 (89.8)	84 (95.4)	23 (63.9)	<0.00
0–1	4 (5.8)	4 (4.6)	12 (33.3)	01
2–3	3 (4.4)	0 (0)	1 (2.8)
Missing
Charlson Comorbidity Index	12 (17.4)	25 (28.4)	0 (0)	0.03
0	26 (37.7)	24 (27.3)	15 (41.7)
1–2	18 (26.1)	23 (26.1)	12 (33.3)
≥3	13 (18.8)	16 (18.2)	9 (25.0)
Missing
Alcohol use
Yes	31 (44.9)	36 (40.9)	15 (41.7)	0.87
No	38 (55.1)	52 (59.1)	21 (58.3)
Current/former smoker
Yes	50 (72.5)	73 (83.0)	26 (72.1)	0.22
No	19 (27.5)	15 (17.0)	10 (17.9)
Primary Site
Oropharynx HPV+	38 (55.1)	42 (47.7)	12 (33.3)	0.17
Oropharynx HPV−	4 (5.8)	13 (14.8)	8 (22.2)
Larynx/Hypopharynx	23 (33.3)	25 (28.4)	12 (33.3)
Oral Cavity	4 (5.8)	8 (9.1)	4 (11.1)
T Stage
T1	9 (13.0)	10 (11.4)	4 (11.1)	0.90
T2	25 (36.2)	29 (33.0)	10 (27.8)
T3	23 (33.3)	25 (28.4)	12 (33.3)
T4a	11 (15.9)	20 (22.7)	8 (22.2)
T4b	1 (1.5)	4 (4.6)	2 (5.6)
N Stage
N0	5 (7.3)	15 (17.1)	7 (19.4)	0.36
N1	12 (17.4)	9 (10.2)	3 (8.3)
N2a	3 94.4)	2 (2.3)	1 (2.8)
N2b	27 (39.1)	27 (30.7)	16 (44.4)
N2c	21 (30.4)	32 (26.4)	7 (19.4)
N3	1 (1.5)	2 (2.3)	2 (5.6)
Missing	0 (0)	1 (1.1)	0 (0)
Stage
III	13 (18.8)	17 (19.3)	9 (25.0)	0.73
IV	56 (81.2)	71 (80.7)	27 (75.0)
Positive Marginb	4 (25)	9 (39.1)	2 (18.2)	0.34
Missing	0 (0)	0 (0)	1 (9.1)
Extracapsular Extensionb	8 (50.0)	14 (60.9)	3 (27.3)	0.23
Missing	2 (12.5)	5 (21.7)	2 (18.2)

All data are summarized using n (%) or median (interquartile range) unless otherwise specified.

^*Kruskal-Wallis test of difference in medians

Table 2:

Treatment Delivery Details

	High-dose Cisplatin (n=69)	Weekly Cisplatin, 40 mg/m2 (n=88)	Weekly Cisplatin, <40 mg/m2 (n=36)	p-value
Treatment
Definitive CRT	53 (76.8)	65 (73.9)	25 (69.4)	0.71
Surgery+RT	16 (23.2)	23 (26.1)	11 (30.6)
RT Dose	60 (60–66)	66 (60–66)	60 (60–70)	0.86*
RT Modality
IMRT/VMAT	64 (92.8)	83 (94.3)	26 (72.2)	<0.000
3D-IMRT	5 (7.2)	4 (4.6)	8 (22.2)	1
3D	0 (0)	1 (1.1)	2 (5.6)
RT Elapsed Days	43 (41.5–45)	45 (42–50)	44 (41–56)	0.75*
Prescribed Cisplatin Dose (mg/m2)	100 (100–100)	40 (40–40)	33 (30–33)	<0.0001*
Cumulative Cisplatin Dose (mg/m2)	200 (200–200)	240 (190–250)	180 (120–198)	<0.0001*
Patients Treated with Cumulative Cisplatin Dose ≥ 200 mg/m2	61 (89.4)	76 (86.4)	9 (25.0)	<0.0001
Cycles Completed at Prescribed Dose	2.0 (2–2)	5 (4–6)	6 (4–6)	<0.0001*
Cycles Completed with Dose Reduction	2.0 (2–2)	6 (5–7)	6 (4–6)	<0.0001*

All data are summarized using n (%) or median (interquartile range) unless otherwise specified.

^*Kruskal-Wallis test of difference in medians

The prescribed dose for HDC was 100 mg/m² and 40 mg/m² for WC40. For patients in the WC<40 group, the median initial prescribed cisplatin dose was 33 mg/m² (33 mg/m²: n=23, 68.9%; 30 mg/m²: n=12, 33.3%; 20 mg/m²: n=1, 3%) (Table 2). In the HDC group, 13 patients received three every-third-week cycles of cisplatin, 40 patients received 2 cycles, and of those, 13 received a third administration at a reduced dose. A significantly higher proportion of patients in the HDC (89.4%) and WC40 (86.4%) groups received a total cumulative dose ≥200 mg/m² compared to the WC<40 group (25%). A similar pattern of difference was observed when patients were stratified by definitive versus postoperative CRT.

Survival and Disease Control

Two-year OS was 86.5%, 76.5%, and 60.2% for those treated with HDC, WC40, and WC<40, respectively (Figure 1). In bivariate Cox models, chemotherapy regimen, gender, ECOG status, smoking status, alcohol use, postoperative CRT, primary site, and T and N staging were all associated with OS at p ≤ 0.20 (Table 3). When we included the above variables in a multivariable Cox model (HPV status was folded into the primary site variable), gender, ECOG status, tumor site, T and N staging, and chemotherapy regimen remained significant predictors of OS at p<0.05. Males, patients with worse ECOG status, patients with primary site of larynx/hypopharynx or oral cavity (relative to those with primary site HPV+ oropharynx), patients with higher T staging and N staging, and patients receiving WC40 (HR 3.30) or WC<40 (HR 4.13) versus HDC had significantly elevated mortality risk in the multivariable setting (Table 3). Two-year CSS was 91.7% (HDC), 82.1% (WC40), and 76.2% (WC<40).

Figure 1:

Kaplan-Meier plot of overall survival by type of chemotherapy concurrent with radiotherapy

Table 3:

Bivariate and Multivariable Analyses Clinical Predictors of Overall Survival and Disease-free Survival

Overall Survival	Bivariate		Multivariable
	HR (95% CI)	p-value	HR (95% CI)	p-value
Age ≥65	1.17 (0.69–2.00)	0.56
Male Gender	1.84 (0.90–3.75)	0.09	2.61 (1.12–6.09)	0.026
ECOG Performance Status 2–3 (v. 0–1)	4.39 (2.43–7.91)	<0.0001	2.22 (1.12–4.40)	0.023
Postoperative Radiotherapy (v. Definitive)	1.95 (1.16–3.28)	0.01	1.54 (0.82–2.88)	0.18
Ever smoker	1.81 (0.86–3.80)	0.12	0.98 (0.41–2.35)	0.97
Alcohol use	0.89 (0.54–1.48)	0.20	0.65 (0.37–1.15)	0.14
Comorbidity Index (v. 0)			-	-
1–2	1.21 (0.53–2.78)	0.31
≥3	1.26 (0.55–2.88)	0.31
Primary Site (v. Oropharynx HPV+)
Oropharynx HPV−	1.88 (0.84–4.20)	0.003	1.57 (0.65–3.81)	0.001
Larynx/Hypopharynx	2.55 (1.39–4.68)		3.07 (1.45–6.48)
Oral Cavity	4.08 (1.82–9.19)		7.36 (2.59–20.87)
T3–4 (v. T1–2)	2.36 (1.36–4.08)	0.002	2.34 (1.26–4.35)	0.007
N2–3 (v. N0–1)	1.64 (0.87–3.09)	0.12	4.64 (2.15–10.01)	<0.0001
Chemotherapy (v. HDC)		0.0006*		0.002*
WC40	2.53 (1.26–5.07)	0.009	3.30 (1.53–7.10)	0.002
WC<40	4.20 (2.02–8.72)	0.0001	4.13 (1.79–9.53)	0.0009
Disease-free Survival	Bivariate		Multivariable
	HR (95% CI)	p-value	HR (95% CI)	p-value
Age ≥65	1.49 (0.91–2.38)	0.10	1.12 (0.66–1.91)	0.67
Male Gender	1.48 (0.80–2.74)	0.22	-	-
ECOG Performance Status 2–3 (v. 0–1)	3.69 (2.11–6.45)	<0.0001	1.79 (0.93–3.45)	0.08
Postoperative Radiotherapy (v. Definitive)	1.61 (1.0–2.60)	0.05	1.37 (0.79–2.35)	0.26
Ever smoker	2.04 (1.05–3.97)	0.04	0.93 (0.43–2.02)	0.85
Alcohol use	0.89 (0.56–1.40)	0.60	-	-
Comorbidity Index (v. 0)			-	-
1–2	0.93 (0.46–1.90)	0.77
≥3	1.15 (0.58–2.31)
Primary Site (v. Oropharynx HPV+)				0.002
Oropharynx HPV−	2.47 (1.21–5.06)	<0.0001	2.14 (0.98–4.67)
Larynx/Hypopharynx	3.44 (1.99–5.95)		3.78 (1.90–7.53)
Oral Cavity	3.55 (1.61–7.80)		3.73 (1.53–9.12)
T3–4 (v. T1–2)	2.51 (1.54–4.10)	0.0002	1.94 (1.11–3.39)	0.02
N2–3 (v. N0–1)	1.51 (0.87–2.62)	0.14	3.23 (1.67–6.23)	0.0005
Chemotherapy (v. HDC)		0.01*		0.03*
WC40	1.50 (0.86–2.62)	0.15	2.00 (1.67–6.23)	0.03
WC<40	2.45 (1.34–4.46)	0.004	2.52 (1.22–5.22)	0.01

CI, confidence intervals; HR, hazard ratio

^*Global p-value for the variable of interest

Two-year DFS was 75.4%, 67.7%, and 52.0% for those treated with HDC, WC40, and WC<40, respectively (Figure 2). In bivariate Cox models, age >=65, ECOG status, postoperative CRT, smoking status, primary tumor site, T and N staging, and chemotherapy regimen were all associated with DFS at p<=0.20 (Table 3). When we included the above variables in a multivariable Cox model (HPV status was again folded into the primary site variable), site, T and N staging, and chemotherapy regimen remained significantly associated with DFS at p<0.05. Patients with larynx/hypopharynx (HR 3.78) or oral cavity (HR 3.73) primary site (relative to those with primary site HPV+ oropharynx), patients with T3-4 (HR 1.94) or N2-3 (HR 3.23) disease, and patients receiving WC40 (HR 2.00) or WC<40 (HR 2.52) compared to HDC had significantly elevated risk of disease progression in the multivariable setting (Table 3).

Figure 2:

Kaplan-Meier plot of disease-free survival by type of chemotherapy concurrent with radiotherapy

Locoregional failure occurred in 34 of 193 patients; 2-year LRC was 81.3%, 81.6%, and 84.0% for those treated with HDC, WC40, and WC<40, respectively (Figure 3). Of 22 regional failures, 19 (86%) occurred within the RT treatment volumes. Resection margins were positive in 3 of 4 postoperative patients with local failure and extranodal extension was present in 6 of 7 of postoperative patients with regional failure. The median dose to the region with locoregional failure was 70 Gy (range, 54–70). Thirty-six patients experienced distant failure at one or more sites: 8 of 69 in the HDC group, 19 of 88 in the WC40 group, and 9 of 36 in the WC<40 group. Two-year freedom from DF was 89.7%, 78.5%, and 72.6% in the HDC, WC40, and WC<40 groups.

Figure 3:

Kaplan-Meier plot of locoregional control by type of chemotherapy concurrent with radiotherapy

Treatment Toxicity

Adverse events including hospitalization, febrile neutropenia, and treatment-related mortality are summarized in Table 4. Hospitalization occurred in 36.3% for primarily infectious and gastrointestinal indications. These hospitalizations were for the management of potentially life-threatening processes (CTCAE Grade 4) in 7.1% while the remainder were Grade 3. Febrile neutropenia (3.1%) and treatment-related mortality (1.0%) were rare with no statistically significant differences between groups.

Table 4:

Hospitalization, Febrile Neutropenia, and Treatment-related Mortality

	High-dose Cisplatin (n=69)	Weekly Cisplatin, 40 mg/m2 (n=88)	Weekly Cisplatin, <40 mg/m2 (n=36)	p-value
Hospitalization	23 (33.3)	32 (36.4)	15 (41.7)	0.701
Indication for Hospitalization
- Acute kidney injury	2 (8.7%)	1 (3.1%)	0 (0.0%)
- Anemia	0 (0.0%)	0 (0.0%)	1 (6.7%)
- Dehydration	0 (0.0%)	2 (6.2%)	0 (0.0%)
- Delirium	0 (0.0%)	0 (0.0%)	1 (6.7%)
- Dyspnea	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Enterocolitis infectious	0 (0.0%)	(3.1%)	0 (0.0%)
- Febrile neutropenia	2 (8.7%)	0 (0.0%)	0 (0.0%)
- Fever	1 (4.3%)	0 (0.0%)	1 (6.7%)
- Gastrocutaneous fistula	0 (0.0%	1 (3.1%)	0 (0.0%)
- Hyperglycemia	1 (4.3%)	0 (0.0%)	0 (0.0%)
- Hyponatremia	1 (4.3%)	0 (0.0%)	2 (13.3%)
- Hypotension	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Laryngeal edema	1 (4.3%)	1 (3.1%)	0 (0.0%)
- Lung infection	3 (13.0%)	3 (9.4%)	5 (33.3%)
- Nausea/Vomiting and/or Diarrhea	7 (30.4%)	7 (21.9%)	1 (6.7%)
- Neutropenic fever	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Neutrophil count decreased	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Oral hemorrhage	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Oral mucositis	0 (0.0%)	0 (0.0%)	1 (6.7%)
- Pharyngeal mucositis	0 (0.0%)	1 (3.1%)	1 (6.7%)
- Respiratory failure	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Seizure	0 (0.0%)	1 (3.1%)	0 (0.0%)
- Sepsis	1 (4.3%)	2 (6.2%)	1 (6.7%)
- Skin infection	0 (0.0%)	2 (6.2%)	1 (6.7%)
- Sore throat	1 (4.3%	0 (0.0%)	0 (0.0%)
- Syncope	3 (13.0%)	2 (6.2%)	0 (0.0%)
- Thromboembolic event	0 (0.0%)	2 (6.2%)	0 (0.0%)
Febrile Neutropenia	2 (2.9)	1 (1.1)	3 (8.3)	0.134
Treatment-related Mortality	0 (0)	2 (2.3)	0 (0)	0.672

Discussion

HDC is the accepted standard of care chemotherapy regimen for concurrent CRT in the treatment of HNSCC with definitive or post-operative intent. Some patients are ineligible for HDC due to comorbidity, performance status and/or patient/provider preference. In this setting, alternative regimens are needed. A randomized study of predominantly postoperative oral cavity squamous cell carcinoma (SCC) patients failed to show non-inferiority of WC (30 mg/m² weekly) compared to HDC for LRC (13). Additionally, the utilized dose of cisplatin (30 mg/m²) is lower than the standard weekly dose of 40 mg/m² supported by consensus guidelines, though this regimen had been studied previously with benefit compared to RT alone at the same institution (22–24). Early (abstract-only) results of a randomized study for postoperative patients (JCOG1008) indicated non-inferiority of WC40 relative to HDC (25).

Bauml and colleagues performed a population-based study of the Veteran’s Health Administration database, employing propensity-score adjusted analyses to account for sociodemographic and baseline disease-related factors (14). No survival differences were observed after propensity score adjustment and HDC was associated with increased hazard of acute and chronic adverse events. In that report, 49% in the HDC group received a total cisplatin dose ≥200 mg/m² compared to 25% in the WC group. In our cohort, 89% (HDC) and 86% (WC40) of patients reached 200 mg/m² compared to 25% in the WC<40 group. This similarly low proportion of patients achieving a cisplatin intensity of 200 mg/m² was found despite similar rates of patients treated with adjuvant/postoperative intent (for whom the target cisplatin intensity is different) between groups and persisted after stratifying by definitive or adjuvant CRT. These results are consistent with a recent study identifying statistically significant differences in OS for patients receiving WC<40, WC40 and HDC (15).

With regard to survival: in our study, chemotherapy regimen (using HDC as a reference) was significantly associated with OS on multivariable analysis with both WC40 and WC<40 associated with an increased hazard of death. With regard to DFS, all regimens (compared to HDC) were associated with worse outcomes. There was no significant impact of chemotherapy regimen noted on LRC, with 2-year LRC rates for WC<40 (84%) comparing favorably with HDC (81%) and WC40 (82%). This may be related to limited follow-up or selection bias. Freedom from distant metastasis, on the other hand, was numerically (though non-significantly) lower in the WC<40 group (73% versus 90% in the HDC group and 79% in the WC40 group). Considering that cancer-specific survival was lowest for patients treated with WC<40, these findings suggest that efficacy or eligibility for salvage therapy (if needed) in patients initially treated with WC<40 may affect the observed survival differences more so than higher rates of disease failure.

Concurrent chemotherapy with RT in locally advanced head and neck cancer is thought primarily to improve DFS and OS by increasing LRC rates and preventing the often-profound consequences of local and/or neck failure. Improvements in distant failure have not been observed to parallel the LRC benefit in multiple randomized trials (3, 7, 26, 27). In our study, rates of LRC and distant metastasis were not significant between groups. This differs from a prior study of almost exclusively oral cavity SCC patients but is consistent with a recently reported non-inferiority study that included a broader spectrum of primary sites and that identified weekly cisplatin at a dose of 40 mg/m² as non-inferior to HDC (13, 25). Planned subgroup analyses did not identify primary site-specific OS benefits (oral cavity/hypopharynx v. oropharynx/larynx cancers), but a similar analysis was not reported for LRC. Given our relatively small sample size, subgroup analysis of patients with locoregional failure in our study was not feasible. However, among those who did experience recurrence, positive margins or extranodal extension were frequent, and failures mostly occurred within high-dose regions, consistent with commonly seen patterns of failure in HNSCC (28, 29). The fact that there were no differences in LRC seen between chemotherapy regimens may indicate that the dosing of cisplatin does not significantly impact pattern of failure for these patients (though it may also reflect a lack of statistical power given the limited number of locoregional failure events in our sample.).Though some patients at high risk of failure may warrant consideration of treatment escalation, other methodologies (i.e. RT dose escalation, additional radiosensitizing agents, immunotherapy) could be considered.

It is worth noting that favorable LRC in all three groups provides rationale for the utility of limited weekly cisplatin in patients unable to tolerate a more intensive regimen, particularly in patients with advanced age, poor performance status, or those treated with palliative intent. Additionally, recent efforts to de-intensify therapy for patients with HPV-associated oropharyngeal SCC employ WC at doses of 30–40 mg/m² rather than HDC (30, 31). In the multivariable model controlled for several factors including chemotherapy regimen, HPV-positive oropharyngeal patients experienced favorable survival compared to HPV-negative oropharyngeal cancer or other non-oropharyngeal patients. Considering these favorable outcomes and the lack of significant differences in LRC between cisplatin groups, further study of de-intensified regimens utilizing WC40 are warranted, though with caution so as not to sacrifice the high rates of survival and disease control with HDC.

No differences between severe toxicity (hospitalization, febrile neutropenia or treatment-related mortality) were noted between groups in this study. Febrile neutropenia and life-threatening hospitalization indications were more common in the WC<40 group, though this did not reach statistical significance and the numbers within this sample are small. These findings may be related to small sample size, limited event numbers, or perhaps clinical practices meant to reduce AEs in the groups treated with less intensive concurrent cisplatin regimens. Dose reductions that occurred during treatment, which are evident in the significantly reduced cumulative cisplatin dose in the WC<40 group, indicate the regimen was likely tailored to ensure patient tolerance and compliance with the prescribed course of radiotherapy.

This analysis is limited by its retrospective nature is indeed subject to selection biases such as those based on patient age and comorbidity that are evident in our cohort. We attempt to adjust for this using the Charlson Comorbidity Index, which offers insight into comorbidity-related selection biases. However, neither comorbidity score nor age were significant predictors of survival, suggesting a limited role of these metrics in assessing general health-related confounders in patients with HNSCC. While patients in the WC<40 group were older, had worse performance status and higher comorbidity, multivariable models controlling for ECOG performance status (age and comorbidity index not included due to lack of significance on bivariate models) showed persistently increased hazard of death or disease recurrence. Though it may drive therapeutic biases, comorbidity may indeed be less predictive in this cohort of advanced head and neck cancers (32). As a result, differences in survival may be related to suboptimal therapy or underdosing of cisplatin rather than competing comorbidity (33, 34).

Additional limitations of this study include the duration of study period, which may include shifting disease states and practice patterns, such as an increased prevalence of HPV positive oropharyngeal carcinoma. Categorization of groups based on chemotherapy regimen and subgroup analyses based on primary site and HPV status also contributed to small sample sizes, limiting our statistical power. The inclusion of only locally advanced (AJCC 7^th Edition Stage III-IVB) disease, in an effort to evaluate a population for which the indication for concurrent CRT is uniform, further limited the sample size. Locoregional failure rates were relatively low in this cohort and the low number of events precluded complex multivariable modeling. Our findings are also limited by the small sample size of patients treated postoperatively and are mostly generalizable to definitive patients. As such, these findings are limited to hypothesis generation and require prospective validation.

Conclusion

In conclusion, HDC remains the optimal choice for CRT for locally advanced HNSCC. WC at a dose of 40 mg/m² allows for the delivery of similar cumulative cisplatin dose intensity but is associated with worse OS and DFS in this cohort of definitive and postoperative patients. Lower dose WC (<40 mg/m²) was associated with older age, worse performance status, and higher comorbidity and worse disease control and survival which appeared to be independent of comorbidity. Prospective study is warranted, particularly in the definitive setting, and further efforts to measure and account for selection biases in head and neck cancer patients are needed.