Optimized radiotherapy treatment strategy for early glottic carcinoma

ABSTRACT

The local control rates of T1 bulky and T2 glottic carcinoma treated via radiation therapy alone are unsatisfactory; thus, we aimed to evaluate the efficacy and safety of our treatment protocol for early glottic carcinoma. Patients with early glottic squamous cell carcinoma treated via radiation therapy from January 2007 to November 2019 were reviewed. Patients were treated with: 63–67.5 Gy/28–30 fractions of radiation therapy alone for T1 non-bulky; concurrent chemoradiotherapy with S-1 and 60 Gy/30 fractions for T1 bulky and T2 favorable; and concurrent chemoradiotherapy with high-dose cisplatin and 66–70 Gy/33–35 fractions for T2 unfavorable glottic carcinoma. Local failure rates were estimated using the cumulative incidence function, overall and disease specific survival rates were estimated using Kaplan-Meier analysis, and adverse events were evaluated. Eighty patients were analyzed; the median age was 69.5 (range, 26–90) years, the median follow-up time for survivors was 40.1 (range, 1.9–128.4) months, and the 3-year local failure, disease specific survival, and overall survival rates were 5.8%, 98.3%, and 94.4%, respectively. In T1 bulky and T2 cases, the local failure rate was significantly lower in the concurrent chemoradiotherapy than in the radiation therapy alone group. Grade 3 acute dermatitis and mucositis were noted in nine and four patients, respectively. There were no acute adverse events of Grade 4 or higher, or late adverse events of Grade 2 or higher. The treatment protocol was effective and well-tolerated; thus, the efficacy of concurrent chemoradiotherapy was suggested in T1 bulky and T2 cases.

INTRODUCTION

The goals of treatment for early glottic carcinoma (GC) are tumor eradication and preservation of larynx functions, including speaking and swallowing. The recommended treatment strategy for early GC is partial laryngectomy, including endoscopic and open resection, and definitive radiation therapy (RT); generally, the efficacies of resection and RT are comparable.^1,2 Previous studies have shown that the local control (LC) rate of early stage GC varies widely, while that of RT alone for T2 GC is unsatisfactory, ranging from 65% to 80%.^3-5 Reddy et al⁶ reported that tumor size is an important prognostic factor for the LC rate in T1 GC, and that the LC rate is lower in T1 bulky tumors. Several studies have subclassified T2 GC into T2a and T2b, reporting lower LC rates in T2b.^7-9 In this study, we subclassified T1 and T2 GCs into T1 nonbulky, T1 bulky, T2 favorable, and T2 unfavorable, following previous reports.

We first designed a concurrent chemoradiotherapy (CCRT) protocol using S-1 (tegafur, gimeracil, and oteracil) for T1 bulky and T2 favorable GCs with the aim of improving the LC rate; the efficacy and safety of this protocol was demonstrated in our previous phase I/II study.^10,11 S-1 is an orally administered antineoplastic agent shown to be effective against a variety of solid tumors including head and neck cancer. Due to the inadequate results observed in our previous experiments regarding CCRT with low-dose cisplatin (CDDP)/5-fluorouracil for T2 GC,^12,13 we were concerned that CCRT with S-1 may be inadequate in patients with T2 unfavorable GC; thus, CCRT with high-dose CDDP was selected. Furthermore, we changed the RT dose from 2.0 Gy/fraction to 2.25 Gy/fraction, thereby reducing the number of fractions from 35 to 28 for patients with T1 nonbulky GC. This was based on reports, from Japan as well as overseas, that the LC rates for T1 tumors were higher than 90% when using 2.25 Gy/fraction.^14,15 Our research group also reported comparable efficacy and acceptable safety of the 2.25 Gy/fraction method, compared with the conventional 2.0 Gy/fraction method, in a multicenter survey of the Tokai Study Group for Therapeutic Radiology and Oncology conducted in Japan from 2011 to 2015.¹⁶ A recent meta-analysis by Benson et al,¹⁷ including several randomized controlled trials, also reported that hypofractionation for early GC is effective for improving LC rates.

The overall results of our optimized treatment strategy were reported in 2017, demonstrating that our protocol, which has been used in our institution since 2007,¹⁸ was both effective and well tolerated.¹⁹ The purpose of this study was to evaluate the clinical efficacy of our optimized treatment strategy in early GC in a larger cohort.

MATERIALS AND METHODS

Patients

This study was approved by the Institutional Review Board of Nagoya University Hospital (reference number: 2020-0018) and conducted in accordance with the 1964 Declaration of Helsinki and subsequent amendments. Informed consent was obtained from all study participants prior to initial treatment.

Patients with T1 or T2N0M0 (stage I–II) glottic squamous cell carcinoma—treated via definitive RT at Nagoya University Hospital between January 2007 and November 2019—were included in the study. The histological diagnoses were confirmed by biopsy, and all tumors were staged according to the eighth edition of the Union for International Cancer Control staging classification system.²⁰ In addition, T1 and T2 GC were substaged. T1 bulky tumors were defined as tumors extending over the entire true vocal fold, and/or horseshoe-shaped tumors extending over the anterior commissure and one-third of both true vocal folds; T1 nonbulky tumors were defined as tumors that did not extend over these boundaries, based on the report by Reddy et al⁶ T2 favorable tumors were defined as tumors with normal vocal fold mobility; T2 unfavorable tumors were defined as tumors with impaired vocal fold mobility, based on previous studies.^7-9 The treatment policy for each patient was established in a conference attended by both otolaryngologists and radiation oncologists, and the final treatment modality was selected based on the patient’s wishes.

Treatment

Figure 1 shows the details of the treatment protocol for early GC at our institution; RT was performed for all patients once a day on weekdays, ie, five times a week, with no irradiation on holidays. Patients with T1 nonbulky lesions were treated with 28–30 fractions of RT alone at 2.25 Gy/fraction per day. Patients with T1 bulky or T2 favorable GC were treated with CCRT using S-1. S-1 was taken once daily, 3–6 h prior to RT; RT was initiated on the same day as S-1 administration, and the patients were treated with 30 fractions at 2.0 Gy/fraction per day. S-1 was not administered on days when RT was not performed; the dose of S-1 was 55.3 mg/m² per day, the recommended dose determined in our phase I/II study.^10,11 Patients with T2 unfavorable lesions were treated with CCRT using high-dose CDDP. CDDP was administered intravenously at 80mg/m² tri-weekly concurrently with RT for three courses. Patients were treated with 33–35 fractions of RT at 2.0 Gy/fraction per day, initiated on the same day as CDDP administration. Patients who were not eligible for CCRT due to advanced age or renal dysfunction were treated with 28–30 fractions of RT alone at 2.25 Gy/fraction per day.

Fig. 1.

Our optimized treatment strategy

RT: radiation therapy

fr: fraction

CCRT: concurrent chemoradiotherapy

CDDP: cisplatin

All patients were immobilized in the supine position wearing a thermoplastic mask and underwent computed tomography (CT) simulation without contrast enhancement. RT was delivered using a 4 MV photon beam with 3D-CT-based technology. Two parallel-opposed lateral fields were used to set up a rectangular irradiation field of 6 × 6 cm in most cases, and a pair of 15° or 30° wedge filters were used to optimize the dose distribution. In patients with T2 GC, we first used a slightly larger field size for irradiation upto 40 Gy than in T1 GC patients; thereafter, the field size was reduced depending on the reduction in tumor size.

Outcome measures

The outcomes of this study were the evaluations of efficacy and safety of our optimized treatment protocol in early GC; this included the assessment of clinical response, local failure (LF) rate, overall survival (OS) rate, disease specific survival (DSS) rate, and adverse events (AEs) in all patients. One to two months after completing RT or CCRT, the clinical response of each patient was evaluated by combining the fiberscopic, CT, and magnetic resonance imaging findings. A clinical response was defined as the complete disappearance of all measurable lesions (determined by the otolaryngologist via fiberscopy), as well as no evidence of progression or lymph node metastases (determined by the radiologist via CT or magnetic resonance imaging). LF was defined as including both local recurrence alone and local recurrence with regional lymph node recurrence. Patients were monitored for AEs at 1-week intervals during treatment, 1-month intervals during the first year, 2-month intervals during the second year, every 3 months during the third year, every 4 months during the fourth year, and every 6 months thereafter. AEs were classified according to the Common Terminology Criteria for Adverse Events Version 4.0.

Statistical analysis

OS, DSS and LC rates were analyzed using the Kaplan-Meier method with the log-rank test. The LF curves were estimated using the cumulative incidence function—considering the competing risk of death without LF—and compared using Gray’s test. All statistical analyses were performed using EZR (The R Foundation for Statistical Computing, Vienna, Austria), a graphical user interface for R.²¹

RESULTS

A total of 80 patients were analyzed. The median follow-up time for survivors was 40.1 (range, 1.9–128.4) months, the median duration of RT treatment was 42 (range, 37–66) days, and the median number of days off irradiation was 2 (range, 0–21) days. There were two patients who discontinued irradiation during the treatment period: one patient had a 21-day irradiation pause due to the onset of unstable angina during RT, the other had a 9-day irradiation pause due to fever caused by side effects of chemotherapy. Patient characteristics are shown in Table 1; the median age was 69.5 (range, 26–90) years, 77 (96%) patients were male, and 14 (18%) had another primary cancer prior to or concurrent with RT. Anterior and posterior commissure invasion occurred in 38 (48%) and 10 (13%) patients, respectively.

Table 1.

Characteristics of the patients

		n	Percent (%)
Total no. of patients		80	100
Age, median (range), y	69.5 (26-90)
Gender
Male		77	96
Female		3	4
Performance status (ECOG)
0		41	51
1		37	46
2		2	3
T stage
T1a		27	34
Nonbulky		22	28
Bulky		5	6
T1b		13	16
Nonbulky		9	11
Bulky		4	5
T2		40	50
Favorable		33	41
Unfavorable		7	9
A-com involvement
Yes		38	48
No		42	52
P-com involvement
Yes		10	13
No		70	87
Macroscopic classification of tumor
Superficial		36	45
Exophytic		41	51
Unclassified		3	4
Another primary cancer prior to or concurrent with RT
Yes		14	18
No		66	82
Smoking
Yes		73	91
No		7	9
Alcohol
Yes		66	83
No		14	17

ECOG: Eastern Cooperative Oncology Group

A-com: anterior commissure

P-com: posterior commissure

Details of the treatment are shown in Table 2; 51 (64%) patients were treated via RT alone through 28–30 fractions of 2.25 Gy/fraction per day. Exceptionally, one patient was treated with 35 fractions of RT alone at 2.0 Gy/fraction per day, as a subglottic skip lesion was suspected via fiberscopy; although deemed pathologically negative, the irradiation area was prophylactically expanded. Twenty-nine patients were treated with chemotherapy, receiving 30–35 fractions of RT at 2.0 Gy/fraction per day; 22 (76%) patients received S-1, six (21%) received CDDP, and one (3%) received UFT which components are tegafur and uracil. Patients treated with UFT received sequential chemotherapy before and after RT at other hospitals. One patient who received S-1 was treated with 25 fractions of 2.25 Gy/fraction per day; this was performed according to our new CCRT protocol²² aimed at shortening the treatment time. The remaining patients who received S-1 or CDDP were treated according to the study protocol. The initial response to treatment was complete response in all cases, except for one patient who was transferred to the hospital before assessment.

Table 2.

Treatment details

Radiation [1 dose/ fr]	Total dose/ fr	Number of patients (%) (n = 80)
[2.25Gy/ fr]	67.5Gy/ 30fr (RT alone)	14 (18)
	63Gy/ 28fr (RT alone)	36 (45)
	56.25Gy/ 25fr (CCRT with S-1)	1 (1)
[2Gy/ fr]	60Gy/ 30fr (CCRT with S-1)	21 (26)
	66Gy/ 33fr (CCRT with CDDP)	3 (4)
	70Gy/ 35fr (CCRT with CDDP or RT alone)	5 (6)
Chemotherapy	Types of chemotherapy	Number of patients (%) (n = 29)
	S-1	22 (76)
	CDDP	6 (21)
	UFT	1 (3)

fr: fraction

RT: radiation therapy

CCRT: concurrent chemoradiotherapy

CDDP: cisplatin

UFT: tegafur and uracil

Overall outcomes are shown in Figure 2. Among all patients, the 3-year LF rate was 5.8% (95% CI, 1.8–13.1); the 3-year OS and DSS rates were 94.4% (95% CI, 83.4–98.2) and 98.0% (95% CI, 86.6–99.7), respectively. In the univariate analysis, the LF, OS and DSS rates did not show any statistically significant association with either clinical or tumor characteristics. Recurrence and death occurred in five patients each. All five recurrent patients were treated with RT alone: one was a T1 nonbulky case and four were T2 cases; of these, three patients developed only local recurrence, one developed only regional lymph node recurrence, and one developed both local recurrence and regional lymph node recurrence. All patients underwent either salvage surgery, salvage radiotherapy, or both. Four recurrent patients achieved progression-free survival, and one patient died of aspiration pneumonia after salvage surgery; the remaining four of five patients died from secondary primary cancers.

Fig. 2.

Overall outcomes

Fig. 2A: Cumulative incidence curves of local failure among all patients.

Fig. 2B: Kaplan–Meier curves of overall survival among all patients.

Fig. 2C: Kaplan–Meier curves of disease specific survival among all patients.

RT: radiation therapy

The cumulative incidence curves of LF by tumor classification are shown in Figure 3. The 3-year LF rates in the T1 and T2 groups were 2.9% (95% CI, 0.2–13.2) and 8.1% (95% CI, 2.0–19.8), respectively (p=0.39). The 3-year LF rates in the T1 nonbulky, T1 bulky, T2 favorable, and T2 unfavorable groups were 3.7% (95% CI, 0.3–16.2), 0% (95% CI, 0–0), 6.8% (95% CI, 1.2–19.8), and 14.3% (95% CI, 0.5–49.1), respectively.

Fig. 3.

Cumulative incidence curves

Fig. 3A:Cumulative incidence curves of local failure in the T1 (black) and T2 (red) groups.

Fig. 3B:Cumulative incidence curves of local failure in the T1 bulky (black), T1 nonbulky (red), T2 favorable (green), T2 unfavorable (blue) groups.

RT: radiation therapy

We subcategorized T1 bulky and T2 patients into the CCRT and RT alone groups to evaluate the efficacy of CCRT; four patients who discontinued chemotherapy due to AEs, and one patient who received sequential UFT chemotherapy, were excluded. Finally, there were 24 patients in the CCRT group and 20 patients in the RT alone group. The cumulative incidence curves for LF in the two groups are shown in Figure 4. The 3-year LF rate was significantly lower in the CCRT than the RT alone group (0% [95% CI, 0–0] vs 19.9% [95% CI, 4.4–43.3]; p=0.025). The patients treated via RT alone who could not receive CCRT had a single or multiple reasons for deviating from our protocol, including advanced age; underlying diseases, such as renal dysfunction or heart failure; treatment strategy considering secondary primary cancers; and patient convenience. Two patients discontinued chemotherapy due to renal failure, one due to laryngeal bleeding, and one due to fever as a side effect. The patient with fever discontinued S-1 on day 29; only RT was restarted on day 34, administered to 70 Gy. The other three patients who discontinued chemotherapy continued RT as per our protocol.

Fig. 4.

T1 bulky/T2 cases

Cumulative incidence curves of local failure in the RT alone (black) and CCRT (red) groups.

RT: radiation therapy

CCRT: concurrent chemoradiotherapy

Table 3 shows the AEs. Grade 3 acute AEs were noted in 10 patients overall: three patients (30%) treated with CCRT, and seven (70%) treated with RT alone. Grade 3 dermatitis was noted in nine patients (11%), and mucositis in four patients (5%). There were no acute AEs of grade 4 or higher, or late AEs of grade 2 or higher.

Table 3.

AEs (CTCAE v4.0)

Grade (n = 80) Number of patients (%)
	no AEs	grade 1	grade 2	grade 3
Acute
Dermatitis	1 (1)	31 (39)	39 (49)	9 (11)
Mucositis	5 (6)	49 (61)	22 (28)	4 (5)
Laryngeal edema	32 (40)	48 (60)	0 (0)	0 (0)
Laryngeal hemorrhage	74 (93)	6 (7)	0 (0)	0 (0)
Laryngeal pain	7 (9)	68 (85)	5 (6)	0 (0)
Late
Hoarseness	50 (63)	30 (37)	0 (0)	0 (0)
Soft tissue necrosis	80 (100)	0 (0)	0 (0)	0 (0)
Laryngeal edema	79 (99)	1 (1)	0 (0)	0 (0)
Laryngeal hemorrhage	80 (100)	0 (0)	0 (0)	0 (0)
Laryngeal pain	75 (94)	5 (6)	0 (0)	0 (0)

AEs: adverse events

CTCAE v4.0: Common Terminology Criteria for Adverse Events version 4.0

DISCUSSION

We have been treating early GC with our optimized treatment strategy—shown in Figure 1—since 2007; its safety and efficacy were reported in our initial report in 2017.¹⁹ In this study, we increased the number of patients and reevaluated the safety and efficacy of this treatment strategy. Mostly, the 3-year OS and DSS rates of all patients were similar or slightly better than observed in our previous report. The 3-year LC of all patients was also good; however, the occurrence of relapses slightly increased compared with our previous report. Four of the five relapses were T2 cases where the treatment strategy was changed from CCRT to RT alone; thus, the strategy deviated from our treatment protocol. The LF rates in this study were generally comparable to those reported in several randomized controlled trials regarding patients treated via altered-fractionation radiotherapy,^15,17,23,24 whether in the T1-2, T1, or T2 group. AEs were generally comparable to our previous report and other randomized controlled trials.^17,19,23,24

In this study, we performed sub-analyses to evaluate the efficacy of CCRT in T1 bulky/T2 cases, finding that the LF rate was significantly lower in the CCRT group than in the RT alone group. The efficacy of CCRT in T2 laryngeal cancer was previously reported in several studies.^25,26 Overall, the low LF observed in this study may have been due to the use of CCRT for T1 bulky/T2 cases as risk groups.

The RTOG 91-11 by Forastiere et al²⁷ reported that use of CCRT for advanced laryngeal cancer increased the mortality rate not attributable to GC or treatments. Among the five patients who died in our study, one died from relapse, and four from secondary primary cancers; of these four patients, only one patient was treated with CCRT. There were no late AEs of grade 2 or higher in the toxicity profile of this study. The CCRT group appears to have attained good long-term results; however, longer follow-up is needed to verify the long-term effects.

In T1 bulky/T2 cases, patients who could not receive CCRT due to performance status, age, and comorbidities such as renal dysfunction were included in the RT alone group; we therefore need to explore alternative treatments to improve LC in T1 bulky/T2 patients not eligible for treatment via CCRT. We recently planned a new protocol for CCRT using S-1 (UMIN000023416) and have since started a new clinical trial.²² Our new protocol improves the current one by reducing the S-1 administration period to 5 weeks while maintaining the daily dose, and instead changing the RT dose to 2.25 Gy/fraction for a total of 25 fractions. With these improvements, older patients may be more likely to undergo CCRT because the total S-1 dose is reduced and the hospital visit duration is shortened. However, even with this new protocol, there still are patients who are not eligible for treatment via CCRT; therefore, new treatment options need to be devised.

Twenty-two patients received CCRT with S-1, and six received CCRT with CDDP in this study; neither of these patients experienced a recurrence. Although various CCRT regimens with several antineoplastic agents have been adopted to improve the LC rate for T2 GC, their doses and schedules have not been determined. S-1 and CDDP have been reported to act as antitumor medication, as well as radiosensitizers.^28,29 S-1—an orally available antineoplastic agent—is easy to use, as patients can receive the S-1 regimen in an outpatient setting. A recent meta-analysis by Shih et al³⁰ reported that CCRT with S-1 for head and neck squamous cell carcinoma resulted in a good tumor response, favorable survival rate, and low number of AEs. Considering the inadequate results of our previous CCRT treatment regimen with low-dose CDDP/5-fluorouracil for T2 GC,^12,13 we prescribed high-dose CDDP for T2 unfavorable cases in this study. Forastiere et al reported the efficacy and safety of CCRT with high-dose CDDP for advanced laryngeal cancer³¹; however, the dose used in our regimen (80 mg/m²) was lower than the one they used for advanced GC (100 mg/m²); nevertheless, the dosing schedule was the same. Our dose reduction is expected to be acceptable for early GC and to reduce side effects. However, additional cases are required to evaluate the efficacy and safety of our CCRT regimen with high-dose CDDP.

Limitations

Due to the retrospective nature of the study, the selection of patients based on age and comorbidities may have affected the results of the LF rate analysis when comparing the CCRT and RT alone groups. Our results seem to confirm the efficacy and safety of our original optimized treatment in early GC; however, although the number of patients in this study was greater than that in our previous report,¹⁹ the small sample size remains a limitation of this study. We therefore need to increase the sample size over a longer duration to further investigate the efficacy and safety of our protocol.

CONCLUSION

Treating early GC according to our optimized protocol resulted in low LF rates, high OS and DSS rates, and acceptable AEs. Our results also suggest that CCRT is more effective than RT alone in T1 bulky/T2 cases.

ACKNOWLEDGMENTS

This study was possible thanks to the cooperation among many researchers. We would like to express our sincere gratitude to each of them.

CONFLICT OF INTEREST

The authors declare that they have no competing interests.

Abbreviations

GC

glottic carcinoma

RT

radiation therapy

LC

local control

CCRT

concurrent chemoradiotherapy

CDDP

cisplatin

LF

local failure

OS

overall survival

DSS

disease specific survival

AEs

adverse events