Radiotherapy for laryngeal cancer—technical aspects and alternate fractionation

Abstract

Early laryngeal, especially glottic, cancer is a good candidate for radiotherapy because obvious early symptoms (e.g. hoarseness) make earlier treatment possible and with highly successful localized control. This type of cancer is also a good model for exploring the basic principles of radiation oncology and several key findings (e.g. dose, fractionation, field size, patient fixation, and overall treatment time) have been noted. For example, unintended poor outcomes have been reported during transition from ⁶⁰Cobalt to linear accelerator installation in the 1960s, with usage of higher energy photons causing poor dose distribution. In addition, shell fixation made precise dose delivery possible, but simultaneously elevated toxicity if a larger treatment field was necessary. Of particular interest to the radiation therapy community was altered fractionation gain as a way to improve local tumor control and survival rate. Unfortunately, this interest ceased with advancements in chemotherapeutic agents because alternate fractionation could not improve outcomes in chemoradiotherapy settings. At present, no form of acceleration can potentially compensate fully for the lack of concurrent chemotherapy. In addition, the substantial workload associated with this technique made it difficult to add extra fractionation routinely in busy clinical hospitals. Hypofractionation, on the other hand, uses a larger single fractionation dose (2–3 Gy), making it a reasonable and attractive option for T1–T2 early glottic cancer because it can improve local control without the additional workload. Recently, Japan Clinical Oncology Group study 0701 reprised its role in early T1–T2 glottic cancer research, demonstrating that this strategy could be an optional standard therapy. Herein, we review radiotherapy history from ⁶⁰Cobalt to modern linear accelerator, with special focus on the role of alternate fractionation.

INTRODUCTION

Early laryngeal cancer, especially of the glottis, is primarily a local disease, and radiotherapy is the main treatment used to preserve the larynx. This type of cancer is a good model for testing basic radiotherapeutic hypotheses associated with optimal dose, fractionation, field size, overall treatment time (OTT), and wedge factor usage [1, 2]. With the advent of radiotherapy machines, higher photon energy has enabled us to treat deep-seated tumors. However, higher energy photons worsen dose distribution for shallow-seated tumors, such as in glottic cancer, resulting in poor outcomes [3]. Moreover, precise dose delivery cannot be performed with irresponsible patient fixation, and elevating the dose to the normal tissue worsens the toxicity [4].

The efficacy of radiotherapy is related to several biological factors, including intrinsic cellular radiosensitivity and proliferation during treatment. For instance, accelerated tumor repopulation has been described in head and neck cancer after initiation of radiotherapy [2]. Non-conventional fractionated radiotherapy (alternate fractionation) is an attractive option, as the treatment dose intensity can be increased either by increasing the total dose using a smaller, single fraction (hyperfractionation, HF) or by reducing the OTT (acceleration). Reduced OTT is expected to counteract the tumor growth induced by accelerated fractionation (AF), and thereby improve locoregional control [1, 2]. Such shorter OTT without a dose reduction can be achieved either by applying an accelerated higher dose/fraction (fr) or more fractions per week. In either situation, there is a risk of increased toxicity. Many publications have presented both positive and negative opinions for alternate fractionation, and a summary of the important studies is reviewed below.

FROM COBALT TO LINEAR ACCELERATION: PHOTON ENERGY, FIELD SIZE, SHELL FIXATION, AND WEDGE FACTOR

Increasing X-ray energy did not always improve outcome

The history of external radiotherapy of laryngeal carcinoma can be divided into two periods, characterized by the machines used for external beam radiotherapy. From the 1950s, a ⁶⁰Cobalt external beam was employed, and from the 1960s, linear accelerator (Linac) machines were installed to expand the utility of radiotherapy for deep-seated tumors using higher energy photon beams (8 or 10 MV). Izuno et al. reported an unexpectedly lower local control rate using 10-MV X-rays for T1N0M0 glottic carcinoma [3]. A daily dose of 1.8 or 2 Gy at the midplane was administered in 5 fr/week. The mean total dose was 63.3 Gy (range, 50–70 Gy) in the ⁶⁰Cobalt group and 65.9 Gy (50–80 Gy) in the 8/10 MV group. The 5-year local control was 88% (15/17 patients) in the ⁶⁰Cobalt group compared with only 60% (15/25 patients) in the 8/10-MV group (P = 0.05). This phenomenon may have occurred by ionization build-up and -down effects of photon irradiation, with a decrease in the radiation dose at the interior surface layer of the larynx. This effect can be enhanced in beams with higher photon energies, such as 10-MV X-rays, resulting in poorer local control. Even 6-MV photon beams sometimes resulted in poor dose distribution for anteriorly located tumors that required placing a bolus on the skin surface to compensate for dose distribution [5].

Role of field size, wedge factor, and shell fixation

The standard radiation therapy for T1 glottic carcinoma is 60–66 Gy, with field sizes ranging from 4 × 4 cm to 6 × 6 cm [1]. Harwood et al. reported that field size was the most important factor affecting local control of T1 glottic carcinoma in a free set-up situation [6]. By increasing the treatment field size with the same dose (5500 rad in 5 weeks), local control at 5 years for 160 patients treated with a field size of ≤25 cm² was 82% and for 171 patients treated with a field size of 26 cm² was 91% (P = 0.034). This indicates an underdosed tumor lesion (geographic miss) when using a small field size. Chatani et al. cited the importance of shell fixation using an adequate wedge filter [4]. They compared outcomes between Group A (Styrofoam head holder and band used to ensure immobilization) and Group B (shell fixing device used with wedge filter). The total radiation dose administered was 50–70 Gy over 5–7 weeks. The 5-year progression-free survival (PFS) for Groups A and B was 85 and 90%, respectively. However, in patients with large T1a lesions (total length of one vocal cord) in Groups A and B, the 5-year relapse-free survival rates were 62% and 88% (P = 0.003), respectively. Based on these results, they conducted a randomized controlled trial (RCT) to compare the influence of field width between small-field (5 × 5 cm) and wide-field (6 × 6 cm) sizes in T1N0M0 glottic cancer patients with 60 Gy/30 fr over 6 weeks (Table 2) [4]. Minor chronic complications, such as persistent arytenoid edema lasting more than 6 months or benign polypoid lesion of the vocal cords, was more frequently observed in the wide-field group (23%) than in the small-field group (17%; P = 0.038), while there were no significant differences in acute mucosal and skin reactions. The 5-year PFS was the same in both groups (88%). They concluded that a small field with an appropriately angled wedge filter and shell fixing device is recommended to avoid adverse effects and keep local control [7]. In general, the local control of Tis, T1, T2, T3 and T4 glottic carcinoma treated with radiation alone was reported to be 92–98%, 77–94%, 67–88%, 42–67% and 20–82% (Table 1).

Table 2.

Randomized control trials for laryngeal cancer

Study (Tx year)	Site/Stage	NO	Treatment	Fractionation (Gy)	OTT (day)	cBED (Gy)	LC¶	LC¶
Source (PY)	(MF)	PT	Schedule	Single dose: Total dose	reduction	Gain		Gain (%)
Osaka CC (1982–1992) [7]	Glottic T1	137	SF	5 × 5 cm field 2 Gy : 60 Gy		NA	PFS 88%		Late 17% vs 23% (P = 0.038)
Chatani (1996)	(6Y1M)	136	SF	6 × 6 cm field 2 Gy : 60 Gy			88%		Small field (5 × 5 cm) recommended
Poland (1995–1998) [22]	Larynx T1–3	199	SF	2 Gy : 66 Gy			3-y LRC 76%		LC same (AF better tendency)
Hliniak (2002)	(28M)	196	AF	2 Gy : 66 Gy 2 fx Thursday	−7	4.2	81%	5%	Acute AF worse, Late same except skin teleangiectasia
DAHANCA 6 (1992–1999) [23]	GlotticT1–4 (T1–2: 85%)	341	SF	2 Gy: 68 Gy 5 fr/wk			LRC 70.6%		LRC AF better
Lyhne (2015)	(14.5Y)	349	AF	2 Gy: 68 Gy 6 fr/wk	−7	4.2	78.4% (P = 0.04)	7.8%	Acute AF worse, Late same
Osaka CC (1993–2001) [24]	Glottic T1	88	SF	2 Gy: 60–66 Gy			77%		LC AF better
Yamazaki (2006)	(64M)	92	AF	2.25 Gy: 56.25–63 Gy	−7	1.6–2.5	92% (P = 0.002)	15%	Same toxicity
RTOG 9512 (1996–2003) [25]	Glottic T2	119	SF	2 Gy: 70 Gy			70%		LC same (HF better tendency)
Trotti (2014)	(7.9Y)	120	HF	1.2 Gy BID: 79.2 Gy	−2	0.3	78% (P = 0.14)	8%	Acute AF worse, Late same
KORG 0201 (2002–2010) [26]	Glottic T1–2	82	SF	2Gy:T1 66 Gy T2 70 Gy			77.80%		LC Same (AF better tendency)
Moon (2014)	(67M)	74	AF	2.25 Gy: T1 63 Gy T2 67.5 Gy	−7	2.5–3.1	88.5% (P = 0.2)	10.7%	Same toxicity
JCOG0701 (2007–2013) [27]	GlotticT1–2	184	SF	2 Gy:T1 66 Gy T2 70 Gy			3-y 84.1%		LC almost the same
Kodaira (2016)	(4.8Y)	186	AF	2.4 Gy: T1 60 Gy T2 64.8 Gy	−10 to −12	3–3.2	89.7%	5.6%	Same toxicity

Tx = treatment, PY = publication year, Osaka CC = Osaka Medical Center for Cancer and Cardiovascular Diseases, DAHANKA: Danish Head and Neck Cancer Group, RTOG = The Radiation Therapy Oncology Group, KORG = Korean Radiation Therapy Oncology Group, JCOG = Japan Clinical Oncology Group, MF = median follow-up period, SF = standard fractionation, AF accelerated fractionation, HF = hyperfractionation, BID: twice a day, fx = fraction. OTT = Overall treatment time 5 years unless otherwise stated, LC = local control rate, PFS = progression-free survival rate, LRC = locoregional control rate, cBED = the biological equivalent dose (BED) corrected for overall treatment time, Acute = acute toxicity, Late = late toxicity. ¶ = 5 years unless otherwise stated.

Table 1.

Retrospective radiotherapy outcome for glottic cancer according to T category

T category Institution/Author	PY	PT No. subcategory	Tx	Fractionation (Gy)		5 year LC
				Single dose	Total dose
Tis
PMH/Spayne [8]	2001	67	AF	2.55	51	98%
MGH/Wang [9]	1997	60	SF	2	Not available	92%
T1N0
MGH/Wang [9]	1997	665	SF	2	65–66	93%
PMH/Walde [10]	1998	403T1a	AF	~2.5	~50	91%
		46T1b	AF	~2.5	~50	82%
Italia/Cellai [11]	2005	831	SF-AF	2–2.4	60–65	84%
Tohoku U/Nomiya [12]	2008	115 T1a	SF	2	64	92%
		48T1b	SF	2	66	85%
UF/Chera [13]	2010	253T1a	mainly AF	2.25	63	94%
		72T1b	mainly AF	2.25	63	93%
T2N0
MGH/Wang [9]	1997	69T2a	SF	2	66–70	70%
		31T2b	SF	2	66–70	67%
		76T2a	AHF	1.6 BID	70	83%
		61T2b	AHF	1.6 BID	70	72%
PMH/Walde [10]	1998	286	AF	~2.5	~50	69%
UF/Chera [13]	2010	165T2a	mainly HF	1.2BID	74.4	80%
		95T2b	mainly HF	1.2BID	74.4	70%
MDAC/Garden [14]	2003	81	HF	1.1–1.2BID	74–80	*79%
		89	SF	2	32–75	68%
		57	AF	2.06–2.23	66–70	*82%
ChH/Slevin [15]	1993	242	AF	3.3–3.4	50–55	85%
Italia/Frata [16]	2005	256	SF-AF	2–2.4	60–65	73%
T3
MGH/Wang [9]	1997	24	SF	2	70	42%
		41	AHF	1.6BID	70	*67%
PMH/Harwood [17]	1980	112	AF	~2.2–2.5	50–55	51% (3Y)
ChH/Wylie [18]	1999	114	AF	3.3–3.4	50–55	68%
Vancouver/Jackson [19]	2001	70	AF	2.4	60	65%
UF/Hinerman [20]	2007	87	HF	1.2–2	50–79.2	63%
T4
PMH/Harwood [21]	1981	39	AF	~2.2–2.5	50–55	56%
UF/Hinerman [20]	2007	22	HF	1.2BID-2	50–79.2	81%

PMH = Princess Margaret Hospital, MGH = Massachusetts's General Hospital, UF = University of Florida, MDAC = MD Anderson Cancer Center, ChH = Christie Hospital Holt Radium Institute, PY = publication year, SF = standard fractionation, AF = accelerated fractionation, HF = hyperfractionation, AHF = accelerated hyperfractionation, BID = twice a day, Tx = treatment, LC = local control rate. *Statistically significant (vs SF).

ALTERNATE FRACTIONATION: INFLUENCE OF OTT AND FRACTION SIZE

Alternate fractionation can be generally explained in terms of strategy. First, in HF regimens, two to three fractions are delivered each day, with a reduced dose/fr equal to 1.1–1.6 Gy. The reduction of the dose/fr might reduce the risk of late toxicity, despite an increased total dose. These regimens were designed to increase the dose intensity by delivering a higher total dose in the same amount of time. Next, reducing the total treatment time (i.e. accelerating the treatment) should reduce the repopulation of tumor cells between sessions, resulting in improved locoregional control (AF). Those two alternated schedules can be combined, particularly for regimens in which OTT is reduced (accelerated hyperfractionation, AHF). Therefore, it is difficult to clearly differentiate between HF and AF because many trials have used mixed strategies.

There are many publications related to alternate fractionation, and it can generally improve radiotherapy outcomes, as shown in retrospective analyses (Table 1) and multiple RCTs (Table 2). Of these, large variations in the OTT have been reported, from 1 week of reduction (moderate acceleration) to >3 weeks of reduction (considerable acceleration). For comparison of outcomes, the biological equivalent dose (BED) was calculated using α/β = 10 Gy for tumor and early reaction, and α/β = 3 for chronic reaction:

$$\mathrm{BED}=\left(\mathrm{total\; dose}\right)\times \left[1+\text{fraction dose}/\left(\mathrm{\alpha }/\mathrm{\beta }\right)\right]$$

In addition, the corrected BED for OTT (cBED) was determined by:

$$\text{cBED}=(\text{BED2 Gy}–0.6)\times (\text{OTT}–T_{lag}),$$

where T_lag is the assumed lag period of 28 d for a burst of accelerated repopulation of tumor clonogenic cells to occur, 0.6 is the rate of dosage loss in 2-Gy fractions, and:

$$\begin{array}{c}\mathrm{BED2}\mathrm{Gy}=\left(\mathrm{number\; of\; fractions}\right)\times \left(\mathrm{fraction\; doses}\right)\hfill \\ \times \left[\left(\mathrm{\alpha }/\mathrm{\beta }+\mathrm{fraction\; doses}\right)/\left(\mathrm{\alpha }/\mathrm{\beta }+2\right)\right].\hfill \end{array}$$

We hypothesized that radiotherapy started on Monday with no holidays during the radiotherapy period, except for Sunday and Saturday, if elapsed days were not described in the literature.

Randomized controlled trials examining alternate fractionation for laryngeal cancer

Table 2 shows the major RCTs. A Polish trial studied the effect of moderate AF (66 Gy/33 fr over 38 d; 2 fr every Thursday) compared with sequential fractionation (SF; 66 Gy/33 fr over 45 d) in 395 T1–T3 glottic cancer patients. Patients’ ages were ≤75 years, and they had World Health Organization Grade 0–1 T1–T3, N0, M0, glottic, or supraglottic laryngeal cancers were randomized. The AF and SF groups showed a 5% difference in locoregional control after 3 years (AF: 81%; SF: 76%; P = 0.37) [22]. At the end of radiotherapy, patients treated with AF complained of more severe reactions than those treated with SF. Eight weeks after treatment, AF patients had more severe pain upon swallowing (P = 0.027). Four months after treatment, all types of toxicity symptoms, except for skin telangiectasia (P = 0.001), were similar in both treatment groups.

The Danish Head and Neck Cancer Group (DAHANCA) tested a unique AF strategy (5 or 6 fr/week) [23]. They randomized 694 non-metastatic glottic cancer patients into groups that received 5 or 6 weekly fractions of the same total dose. The locoregional control was 78.4% in the 6-fr/week group and 70.7% in the 5-fr/week group, with a corresponding hazard rate (HR) of 0.72 (95% confidence interval (CI): 0.53–0.97; P = 0.04). The effect of AF on loco-regional failure (LRF) was especially evident in well-differentiated tumors (HR = 0.42; 95% CI: 0.23–0.75) and in T1–T2 tumors (86% of included patients; HR = 0.60; 95% CI: 0.41–0.89). Severe acute mucositis was found in the 6-fr/week group, but the incidence of late morbidity was the same. In addition, a Dutch retrospective analysis of 1050 T1–2N0 glottic cancer patients showed that 6 fr/week resulted in 6% improvement in local control after 5 years compared with 5 fr/week, which is in line with the DAHANCA RCT (7.8%) [28].

The University of Florida reported excellent outcomes, not only by HF (1.2 Gy bis in die; BID/twice a day), but also using hypofractionated AF (2.25 Gy/d) in a retrospective analysis (Table 1) [13]. They reported a local control of 100% when carefully selected patients with tumors limited to one vocal cord and measuring 5–15 mm were treated with similar total doses of 61–67 Gy in 2.25 Gy/fr. In contrast, the local control was only 80% for patients treated with 2–2.2 Gy/fr [1]. Accordingly, the Osaka group conducted a RCT (2 versus 2.25 Gy) and found that hypofractionated AF (2.25 Gy/fr) improved 5-year local control by 15% compared with SF in 180 T1 patients [24]. The total radiation dose administered was 60 Gy/30 fr in 6 weeks for minimal tumors (two-thirds of vocal cord or less) or 66 Gy/33 fr over 6.6 weeks for larger than minimal tumors (more than two-thirds of vocal cord) in the 2-Gy arm, and 56.25 Gy/25 fr in 5 weeks for minimal tumors or 63 Gy/28 fr in 5.6 weeks for larger than minimal tumors in the 2.25-Gy arm. The 5-year local control was 77% for the 2-Gy arm and 92% for the 2.25-Gy arm (P = 0.002). No significant differences were found between these two arms in terms of acute mucosal, skin, or chronic adverse reactions.

The Radiation Therapy Oncology Group (RTOG 9512) study compared HF (1.2 Gy BID; 79.2 Gy/66 fr) versus SF (70 Gy/35 fr) for 239 T2N0 vocal cord carcinoma patients. HF increased local control by 8% (HF: 78%; SF: 70%; P = 0.14), corresponding to a 30% HR reduction [26]. The 5-year disease-free survival (DFS) rate in HF and SF patients was 49 and 40% (P = 0.13), respectively, and the overall survival was 72 and 63% (P = 0.29), respectively. HF had higher rates of acute skin, mucosal, and laryngeal toxicity. Late effects experienced by patients with Grade 3–4 carcinomas were 8.5% (95% CI: 3.4–13.6%) after SF and 8.5% (3.4–13.5%) after HF at 5 years.

The Korean Radiation Oncology Group (KROG 0201) compared SF (T1: 66 Gy/33 fr; T2: 70 Gy/35 fr) and hypofractionated AF (T1: 63 Gy/28 fr; T2: 67.5 Gy/30 fr) in 156 patients with T1–T2 glottic cancer [26]. As 282 patients were required, the study was closed prematurely due to poor accrual (only 156 patients). The 5-year local control rate was 77.8% for SF and 88.5% for AF (HR = 1.55; P = 0.213). No significant difference was observed in the toxicity profile between the two arms. In a subgroup exploratory analysis for T1a disease, the 5-year local PFS trended positively in the AF arm (76.7% versus 93.0%; HR = 3.65; P = 0.056).

The Japan Clinical Oncology Group (JCOG 0701) tried to confirm the non-inferiority of the efficacy of AF (2.4 Gy/fr) compared with SF in patients with T1–2N0M0 glottic tumors [27]. Those in the 2.4-Gy AF arm did not seem to show an increase in late events or worsening voice changes compared with the SF arm. Although non-inferiority was not confirmed, the similar efficacy (3-year local control, 89.7% versus 84.1%) and toxicity of AF to SF, as well as its practical convenience, indicate that AF has potential as a treatment option for early glottic cancer. Overall, the 5-year local control was modestly higher (3–10%) with alternate radiotherapy compared with SF for laryngeal/glottic carcinoma (Table 2), but the difference was borderline statistically significant. Yamoha et al.’s meta-analysis of three trials [24–26] yielded an HR for local control of 0.59 (95% CI: 0.43–0.81; P = 0.001), strongly supporting the use of AF in this setting [29].

The cBED calculation included OTT and repopulation of tumor cells. All trials used a higher cBED (0.3–4.2 Gy) for the AF arm (almost all trials used a 1 week shorter experimental arm), which resulted in 5–15% improvement in local control. As a result, their clinical data could be partially explained with the cBED equations presented above, which take the repopulation of tumor cells after T_lag into account. Therefore, the number of elapsed days is an important and dominant factor in laryngeal, especially glottic, cancer radiotherapy.

Randomized controlled trials examining alternate fractionation of locally advanced head and neck tumors, including non-laryngeal cancers

In 2006, a pivotal MARCH meta-analysis mainly dealing with locally advanced head and neck carcinoma was published [30]. The MARCH meta-analysis concluded that there was significant improvement in local control and overall survival using AF radiotherapy based on a collection of individual patient data from 15 RCTs with a total of 6515 patients and a median follow-up of 6 years. Tumor sites were mostly oropharyngeal (larynx 36.5%), and 5221 (74%) patients had Stage III–IV disease. AF improved locoregional control versus SF (6.4% at 5 years; P < 0.0001), particularly in local failure (HR = 0.82; 95% CI: 0.77–0.88; P < 0.001) and survival benefit (3.4% at 5 years; HR = 0.92; 95% CI: 0.86–0.97; P = 0.003), whereas the benefit to nodal control was less pronounced. The survival benefit was significantly higher with HF (8% at 5 years) than with AF radiotherapy (2% without total dose reduction and 1.7% with total dose reduction at 5 years; P = 0.02). The benefit was significantly higher in the youngest patients [HR: <50 years old, 0.78 (0.65–0.94); 51–60 years old, 0.95 (0.83–1.09); 61–70 years old, 0.92 (0.81–1.06); >70 years old, 1.08 (0.89–1.30); P = 0.007]. Therefore, there was little merit for the use of AF in patients aged 60 years or more, which should be kept in mind when treating elderly patients. The major trials are depicted in Table 3.

Table 3.

Randomized control trials for locally advanced head and neck cancer, including cancer of the larynx

Study (Tx year)	Site/Stage	NO	Treatment	Fractionation (Gy)	OTT(day)	cBED (Gy)	LC¶	LC	Survival¶	Toxicity
Source (PY)	% of larynx (MF)	PT	Schedule	Single dose: total dose	reduction	Gain		Gain
Moderate accelerated AF
RTOG 79–13 (1979–83) [31]	Stage III–IV or T2N0 BT, NP, MS	93	SF	1.8–2 Gy: 66–73.8 Gy			2-y LRC 29%		32%	A13%	LRC same
MARCIAL (1987)	Larynx 9–12% (NA)	94	HF	1.2 Gy BID: 60 Gy	−12 to –24	−2.2 to −2.8	30% (NS)	1%	28%	23% (P = 0.12)	Acute worse tendency, Late same
RTOG9003 (1991–1997) [32]	Stage II–IV	268	SF	2 Gy: 70 Gy			LRC 29.3%		19.3%	AG3–29.1%	LRC OS HF better
Beitler (2014)	Larynx 15–16%(14.1Y)	263	HF	1.2 Gy BID: 81.6 Gy	−1	6.8	37.1% (P = 0.05)	7.8%	26.1% (P = 0.05)	27.90%	Acute AF-C worse tendency
		274	AF-S	1.6 Gy BID: 67.2 Gy/6wk*1	−6	−0.8	29.0%	0.3%	22.4%	28.80%	Late same
		268	AF-C	1.8 (+1.5 Gy) : 72 Gy/6wk*2	−7	11.8	33.7%	4.4%	25.3%	36.6% (P = 0.09)
India (1998–2004) [33]	Stage III–IV	142	SF	2 Gy: 66 Gy			2yLRC 55%		2-y DFS 52%	AG3 mucositis19%	LRC AF better
Ghoshal (2008)	Larynx 25.6% (24M)	143	AF-C	1.8 Gy(+1.5 Gy)/67.5 Gy/5wk#3	−12	4.0	74% (P = 0.0006).	19%	72% (P = 0.0007)	35% (P = 0.01).	Acute AF-C worse
IAEA (1999–2004) [34]	Stage I–IV	450	SF	2 Gy: 66–70 Gy			LRC 30%		28%	ACM5%, ADermatitis 11%	LRC AF better
Overgaard (2010)	Larynx 24% (99M)	458	AF	Same dose: 6/w	−7	4.2	42%(P = 0.004)	12%	35% (P = 0.07)	10% (HR = 2.15), 20% (1.91)	Acute AF worse, Late same
ARTSCAN (1998–2006) [35]	Except glottic T1–2, N0	367	SF	2 Gy: 68 Gy			LRC 64.9%		MST 5.4y	ACM**62%/Dysphasia** 40%	LC same outcome
Zackrisson (2015:Sweden)	Larynx 21% (5.3Y)	366	AF-C	2(+1.1 Gy) BID: 68 Gy/4.5 wk*4	−15	5.6	65.50%	0.6%	5.1 y	40%**(P < 0.01)/18%** (P < 0.01)	Acute AF worse, Late same
CAIR (1993–1996) [36]	T2–4N0–1M0	49	SF	1.8 or 2 Gy:72 (60–76) Gy			33%		20%**	5y LG3–4 18%	2 Gy AF too toxic (22% necrosis)
Skladowski (2006:Poland)	Larynx 42% (37 M)	51	AF	Same dose: 7/wk	−12	7.2	75% (P < 0.01)	43%	62%**	19% (1.8 Gy/fr: Late same)	(2 Gy ⇒1.8 Gy/fr) LC OS AF better
CAIR-2 (1996–2006) [37]	T2–4N0–1M0	172	AF	1.8 Gy: 66.6–72 Gy			60%		40%	AG3–4 6%	LC OS Toxicity same
Skladowski (2013:Poland)	Larynx 47.8% (90 M)	173	AHF	Same dose: Tu and Fr BID	−19	11.4	65%	5%	44%	6%	weekend on = weekend off
Very accelerated AF
EORTC 22851 (1985–1995) [38]	T2–4 except HP	253	SF	2 Gy: 70 Gy			LRC 46%		MST 24m	3y LSevere 15%	LRC AHF better, OS Same
Horiot (1997)	Larynx 13.5% (4Y9M)	247	AHF	1.6 Gy TID: 72 Gy/5wk*5	−16	9.2	59% (P = 0.02)	13%	21 m	37% (P < 0.001)	AHF too toxic (two myelitis)
CHART (1990–1995) [39, 40]	Except T1N0 oral, OP, HP, L	366	SF	2 Gy: 66 Gy			10-y LR 43%		26%	LMucosal necrosis 9%	LC OS same
Saunders (2010:UK)	Larynx 46% (NA)	552	AHF	1.5 Gy TID: 54 Gy	−29	3.2	50%	7%	29%	5% (P = 0.02)	Acute Late, AHF better (see text)
Vancouver trial (1991–1995) [41]	Stage III/IV	41	SF	2 Gy:66 Gy			3 y 44.3%		3y 56.8%	Late G3/4 = 24%/5%	Early closure for toxicity (G4)
Jackson (2001)	Larynx 50% (NA)	41	AHF	2 Gy BID: 66 Gy	−22	13.2	49.1%	4.8%	59.4%	17%/19% (P = 0.1)	AHF too toxic
TROG (1991–1998) [42]	Stage III/IV	171	SF	2 Gy: 70 Gy			LRC 47%		DFS 40%	ACM 94%	LRC same
Poulsen (2001)	Larynx 13% (53M)	172	AHF	1.8 Gy BID: 59.4 Gy	−25	3.4	52%	5%	46%	71% (P < 0.001)	Acute AHF worse, Late better
GORTEC (1994–1198) [43]	Unresectable T3–4, N0–3	129	SF	2 Gy: 70 Gy			6y 58%**		17%**	AG3 Mucositis 23%	LC AHF better, OS same
Bourhis (2006)	Larynx 4% (6Y<)	137	AHF	2 Gy BID:62–64 Gy	−25	9	82% (P < 0.01)	24%	22.0%	75% (P < 0.0001)	Acute AHF worse, Late same

Bold represents too toxic schedule. Tx = treatment, PY = publish year, OTT = overall treatment time, RTOG = The Radiation Therapy Oncology Group, IAEA = International Atomic Energy Agency, ARTSCAN = Accelerated RadioTherapy of Squamous cell Carcinomas in the head and Neck, CAIR = 7-days-a-week fractionation Continuous Accelerated IRradiation, EORTC = European Organization for Research and Treatment of Cancer, CHART = Continuous, Hyperfractionated, Accelerated RadioTherapy, TROG = Trans-Tasman Radiation Oncology Group, GORTEC = Groupe d'Oncologie Radiothérapie Tête et Cou, MF = median follow-up period, NA = not available, BT = base of tongue, OP = oropharyngeal cancer, NP = nasopharyngeal cancer, MS = maxillary sinus, L = larynx, SF = standard fractionation, HF = hyperfractionation, AF = accelerated fractionation, AHF = accelerated hyperfractionation, AF-S = split-course HF, AF-C = AF with concomitant boost, BID = twice a day, TID = three times a day LC = local control rate, LRC = locoregional control rate, MST = median survival time (5 years unless otherwise stated), ON = osteoradionecrosis, CM = confluent mucositis, HR = hazard ratio. Toxicity G = grade, A or Acute = acute toxicity, L or Late = late toxicity. Details of radiotherapy: *1 = 1.6 Gy BID × 12 d ⇒2 week split ⇒ 1.6 Gy BID × 9 d; *2 = 1.8 Gy × 18 d ⇒(1.8 + 1.5 Gy) BID × 12 d; *3 = 1.8 Gy × 10 d ⇒1.8 Gy + 1.5 Gy BID × 15 d; *4 = (1.1 Gy + 2 Gy) BID × 20 fr⇒ 2 Gy × 3 d; *5 = 1.6 Gy TID × 8 d ⇒12–14 d split ⇒1.6 Gy TID × 17 d. **Estimated from figure, OTT reduction and cBED gain are calculated from represented case of each study arm. ¶ = 5 years unless otherwise stated.

Moderate AF and HF schedules

RTOG 9003 tested several AF radiation schemes: HF (1.2 Gy BID; 81.6 Gy/68 fr over 7 weeks), continuous AF (AF-C: 72 Gy/42 fr over 6 weeks), and split-course AF (AF-S: 67.2 Gy/42 fr over 6 weeks with a 2-week rest after 38.4 Gy) compared with SF (70 Gy/35 fr over 7 weeks) [32]. Patients with Stage III or IV disease (Stage III–IV oral cavity, oropharynx, or supraglottic larynx or Stage II–IV base of tongue or hypopharynx; N = 1076) were randomized to four treatment arms. The three experimental AF arms were based on the radiation fractionation schedules developed at three leading academic institutions in the USA. The University of Florida has championed HF since 1978 [1], including a split-course HF schedule at the Massachusetts General Hospital [9] and an AF with concomitant boost schedule (AF-C) wherein a second fraction of 1.5 Gy was delivered in the afternoon of the last 12 treatment days at MD Anderson Cancer Center [1]. AF-C was designed with two additional premises: (i) the boost dose would be delivered to a smaller volume, making it more tolerable, and (ii) accelerated repopulation after initial radiation could best be overcome by treatment intensification when the tumor was growing at its fastest rate, which was at the end of the treatment course. At 5 years, HF was significantly superior in locoregional control (HR = 0.79; 95% CI: 0.62–1.00; P = 0.05) and overall survival (HR = 0.81; P = 0.05). Toxicity did not differ in the experimental arms compared with SF. This is the largest trial to test alteration in RCT fashion.

In Poland, a unique 7-d/week fractionation with continuous accelerated irradiation (CAIR) schedule intended to reduce the total AF treatment duration by 2 weeks was implemented [36]. The 5-year local control was 75% in the CAIR and 33% in the SF group (P < 0.00004), with significant improvement in DFS and overall survival. Notably, CAIR patients encountered severe toxicity with 2-Gy fractions, and radiation necrosis developed in five patients (22%) as a consequential late effect. Then, the single dose was reduced from 2 to 1.8 Gy, which improved toxicity so that it was no longer severe. This result implied that while 14 Gy/week could be too toxic, 12.6 Gy/week is a feasible schedule if a total dose reduction is not intended.

Next, the CAIR-2 study compared the original CAIR schedule (once a day, 7 d/week) and AF-C (once a day, 3 d/week, and BID, 2 d/week) in 345 patients with squamous cell carcinoma of the oral cavity, larynx, and oro- or hypopharynx (Stage T2–4N0–1M0) [37]. The total dose ranged from 66.6–72 Gy (T2: 66.6–68.4 Gy; T3–T4: 70.2–72 Gy), with 1.8 Gy/fr, and the number of fractions ranged from 37–40 fr over 37–40 d. Locoregional control at 5 years was 63% for CAIR versus 65% for AF-C, and the corresponding overall survival was 40 and 44%, respectively. Confluent mucositis was the main acute toxicity, with an incidence of 89% in CAIR and 86% in AF-C patients. The 5-year rate of Grade 3–4 late radiation morbidity was 6% for both regimens.

Very accelerated AF schedule

The European Organization for Research and Treatment of Cancer (EORTC 22851) tested an interesting experimental AHF applying 3 fr/d (6 Gy, 3 fr/d, 72 Gy/45 fr over 5 weeks; first course: 28.8 Gy/18 fr over 8 d, 12–14 d split; second course: 43.2 Gy/27 fr/7 d) to SF (70 Gy/35 fr over 7 weeks) in T2, T3, and T4 head and neck cancers (hypopharynx excluded) in 512 patients younger than 75 years [37]. AF showed a significantly better 5-year locoregional control (P = 0.02), with a 13% gain (95% CI: 3–23) over SF. This improvement was of larger magnitude in patients with poorer prognoses (N2–N3, any T; T4 any N) than those in more favorable stages. Specific survival tended to favor the AF arm (P = 0.06). Acute and late toxicity were increased in the AF arm. However, late severe damage occurred in 14% of patients in the AF arm versus 4% in the SF arm, and two cases of radiation-induced myelitis occurred after doses of 42 and 48 Gy to the spinal cord. Therefore, this regimen was considered too toxic, and a less toxic scheme should be investigated.

Continuous, hyperfractionated, accelerated radiotherapy (CHART: 1.5 Gy, 3 fr/d, 54 Gy/36 fr over 12 d) was tested by 11 centers in the UK in 918 patients (except those with TlN0 tumors) compared with SF (66 Gy/33 fr over 6.5 weeks) [39, 40]. Severe and early acute mucositis increased with CHART, but healed by 8 weeks. Locoregional control, local control, nodal control, disease-free interval, freedom from metastasis, and overall survival showed no differences between the two arms. In exploratory subgroup analyses, there was evidence of a greater response to CHART in younger patients (P = 0.041) and patients with poorly differentiated tumors (P = 0.030). In the larynx, there was evidence of a trend towards increasing benefit with more advanced T-stage (P = 0.002). CHART reduced severity in a number of late morbidities, most strikingly for skin telangiectasia, mucosal ulceration, and laryngeal edema. Ten-year estimates of severe xerostomia were 23% for CHART and 31% for conventional radiotherapy (HR = 0.8414; 95% CI: 0.7238–0.9788; P = 0.03). In the CHART arm, 50% of patients presented with signs of laryngeal edema compared with 60% in the conventional arm (P = 0.05), whereas mucosal necrosis was observed in 5 and 9%, respectively (P = 0.02). Osteoradionecrosis occurred in 0.4% of patients after CHART and 1.4% of patients after SF.

The Vancouver trial demonstrated severe toxicity with AF (2 Gy BID, 66 Gy/33 fr over 22–25 d) in patients with Stage III/IV head and neck cancer (50% larynx) [41]. Grade 4 toxicity caused the trial to be discontinued after 82 of the planned 226 patients had been randomized; these 82 patients included patients with tumors of the oral cavity, oropharynx, hypopharynx or larynx. The initial improvement in local control of AF was not sustained for 3 years. PFS, DFS and overall survival were also similar in both groups. The PFS (DFS) was 49.1% (59.4%) in AF and 44.3% (56.8%) in SF. Grade ≤3 late effects were similar, but Grade 4 reactions were significantly higher in AF.

It is interesting to note that the Vancouver trial (14 Gy/week, up to 66 Gy total) stopped due to severe toxicity, while the GORTEC trial completed their exploration without excessive toxicity (14 Gy/week, up to 62–64 Gy total). It is speculated that GORTEC used a relatively long interval between fractions (>8–9 h) and a reduced total dose, which limited late toxicity to the safe range. CAIR also showed that 14 Gy/week (7 fr/week, 2 Gy/fr every day, up to 72 Gy total) is too toxic a schedule, but 12.6 Gy/week (7 fr/week, 1.8 Gy/fr every day, up to 72 Gy total) could be tolerated.

Alternation radiotherapy adds to efficacy in chemoradiotherapy settings?

Chemotherapies, both induction chemotherapy (ICT) followed by radiotherapy or concurrent chemoradiotherapy (CRT) improved outcomes compared with radiotherapy alone [44]. Furthermore, several studies tested the role of adding to chemotherapy (Table 4). EORTC 24954 reported that alternation radiotherapy had little influence if combined with ICT [45]. In their study, resectable advanced squamous cell carcinoma of the larynx (T3–T4) or hypopharynx (T2–T4), N0–N2, were randomized. In the ICT-SF arm (n = 224), patients with a ≥50% reduction in primary tumor size after two cycles of cisplatin and 5-fluorouracil received another two cycles, followed by radiotherapy (70 Gy/7 weeks). In the alternation ICT-radiotherapy arm (n = 226), a total of four cycles of cisplatin and 5-fluorouracil (in weeks 1, 4, 7 and 10) were alternated with 20 Gy of radiotherapy during the three 2-week intervals between chemotherapy cycles (60 Gy total). All non-responders underwent salvage surgery and postoperative radiotherapy. Survival with a functional larynx was similar in ICT-SF and alternation ICT-radiotherapy (HR of death and/or event = 0.85; 95% CI = 0.68–1.06), as were median overall survival (4.4 and 5.1 years, respectively) and median PFS (3.0 and 3.1 years, respectively). Grade 3–4 mucositis occurred in 64 (32%) patients in the SF arm and 47 (21%) patients in the alternation radiotherapy arm. Late severe edema and/or fibrosis was observed in 32 (16%) patients in ICT-SF and in 25 (11%) in the alternation ICT-radiotherapy.

Table 4.

Randomized control trials for combination of alternated fractionation and chemotherapy for locally advanced head and neck cancer, including cancer of the larynx

Study (Tx year)	Site/Stage	NO	Treatment	Fractionation	LC¶	OS (PFS)¶	Toxicity
Source (PY)	% of larynx (MF)	PT		Schedule
EORTC 24954 (1996–2004) [45]	HPC & Larynx	224	ICT-SF	#1PF × 4→RT(2 Gy:70 Gy)	69.2%	48.5%	AG 3–4 mucositis 32%/Lfibrosis 16%	LC, OS same
Lefebvre (2009)	Larynx 48% (6.5Y)	226	Alt ICT-RT	#2PF × 4/Alt RT(2 Gy:60 Gy)	67.7%	51.9%	21%/11%	Toxicity same
RTOG0129 (2002–2005) [46]	Stage III–IV	361	SF/CRT	2 Gy:70 Gy+#3CDDP × 3	LRC 72%	56%	AG3–5 82.3%/LG3–5 36.5%	LRC, OS same
Nguyen-Tan (2014)	Larynx 26% (4.8Y)	360	AF-C/CRT	1.8 (+1.5 Gy):72 Gy/6 wk*1+#4CDDP × 2	69%	59%	77.2%/37.9%	Toxicity same
GORTEC99–02 (2000–2007) [47]	Stage III–IV	279	SF/CRT	2 Gy;70 Gy+#5CF × 3	3-y LRC 58.3%	3 y 42.6%	AG3–4: 69%/Atubing 60%	LRC, OS AF worse
Bourhis (2012)	Larynx 6% (5.2Y)	280	AF-C/CRT	1.8 (+1.5 Gy):70 Gy/ 6wk*2 +#6CF × 2	54.6%	39.40%	76%/64%	Tubing*tub AF worse
		281	AF	1.8 Gy BID: 64.8 Gy	50.1% (P = 0.045)	36.5% (P = 0.04)	84% (P = 0.0001)/70% (P = 0.045)	Other, Late same
India (2000–2007) [48]	Stage II–IV	57	SF	2 Gy: 66–70 Gy	32%	36%	LG2–3 Skin 19%/LG2–3 Xerostomia 23%	Early closure poor accrual
Ghosh–Laskar (2016)	Larynx 16% (54M)	65	SF/CRT	2 Gy: 66–70 Gy+ #7CDDP × 6	49% (P = 0.049)	56%	23%/42%	LC SF/CRT better
		64	AF	2 Gy: 66–70 Gy 6 fr/wk	27%	41%	20%/31%	Acute SF better, Late same
Thailand (2003–2007) [49]	Stage III–IV	48	SF/CRT	2 Gy: 66 Gy+#5CF × 3	70%	76%	AG3–4 mucositis 41.7%	LC same, OS AF-C worse
Chitapanarux (2013)	Larynx 48% (43M)	37	AF-C	2 (+1.2 Gy): 70 Gy/6 wk*3+#6CF × 2	55% (P = 0.18)	63.5% (P = 0.05)	67.6% (P = 0.01)	Acute AF-C worse
CONDOR (2009–2012) [50]	Stage III–IV	27	ICT-SF/CRT	#8TPF × 4→2 Gy:70 Gy+ #9CDDP × 3	12 wks RR 81.5%	2 y 78%	AG3–4 mucositis 26%	Early closure for non feasibility
Driessen (2016)	Larynx 8% (38M)	29	ICT-AF/CRT	#8TPF × 4→2 Gy:70 Gy/6 wks+#10CDDP × 6	72%	79%	59% (P < 0.05)	Only 32% receive planned CDDP

Tx = treatment, PY = publication year, LC = local control rate, OS = overall survival time, PFS = progression-free survival rate, MF = median follow-up period (5 years unless otherwise stated), ICT = induction chemotherapy, SF = standard fractionation, AF = accelerated fractionation, Alt = alternating, RT = radiotherapy, CRT = concurrent chemoradiotherapy, AF-C = AF with concomitant boost, BID = twice a day, EORTC = European Organization for Research and Treatment of Cancer, RTOG = The Radiation Therapy Oncology Group, GORTEC = Groupe d'Oncologie Radiothérapie Tête et Cou, *tub = tubing dependency continued worse in AF until 5 years, A or Acute = acute toxicity, L or Late = late toxicity, toxicity G = grade, RR = response rate. Details of chemotherapy: #1 PF = cisplatin 100 mg/m² + 5-FU 1000 mg/m² × 5 d × 4 cycles followed by SF 70 Gy; patient (Pt) underwent surgery if response dose did not reach partial response after two cycles of PE; #2 PF = cisplatin 20 mg/m² + 5-FU 200 mg/m² × 1 week ⇒ RT 20 Gy/10 fr/2 wk ⇒ PE ⇒ RT 20 Gy/10 fr⇒ PE ⇒RT 20 Gy/10 fr ⇒ PE; #3 = CDDP = 100 mg/m² q3W; #4 = 30 mg/m²/wk; #5 = carboplatin70 mg/m² + fluorouracil 600 mg/m² × 4 d; #6 = carboplatin 70 mg/m² + fluorouracil 600 mg/m² × 5 d; #7 = cisplatin 30 mg/m²/wk; #8 = TPF = docetaxel 75 mg/m2 d1, cisplatin 75 mg/m2 d, fluorouracil750 mg/m² d-5; #9 = 100 mg/m², #10 = 40 mg/m². Details of radiotherapy: *1 = 1.8 Gy × 18 d+ (1.8 + 1.5 Gy)BID/last 12 d = 72 Gy/6 wk; *2 = 2 Gy × 20 d + 1.5 GyBID × 10 d = 70 Gy/6 wk; *3 = 2 Gy × 20 d+ (1.8 Gy+1.2 Gy)BID × 10 d = 70 Gy/6 wk. ¶ = 5 years unless otherwise stated.

The RTOG 0129 conducted a comparison between 360 AF/CRT (72 Gy/42 fr over 6 weeks) and 361 SF/CRT (70 Gy/35 fr over 7 weeks) cases [46]. They concluded that AF-C combined with concurrent cisplatin did not improve outcome or increase late toxicity. The patients had Stage III–IV carcinoma (T2N2–3M0, T3–4 any N M0, no T1–2N1 or T1N2–3) of the oral cavity, oropharynx, hypopharynx, or larynx, and a Zubrod performance status of 0–1. Prescribed cisplatin doses were 100 mg/m² q3W for two and three cycles, respectively. No differences were observed in 5-year overall survival (59% versus 56%; P = 0.18), DFS (45% versus 44%; P = 0.42), locoregional control (69% versus 72%; P = 0.76) or metastasis (18% versus 22%; P = 0.06). There were also no overall differences in Grade 3–4 acute mucositis (33% versus 40%) and worst Grade 3–4 late toxicity (26% versus 21%).

In summary, multiple meta-analyses have reported that alternated radiotherapy could not appeal its merit if combined with chemotherapy [29, 51]. For example, Gupta et al. carried out a meta-analysis including five RCTs, involving 1117 patients and 627 deaths [50]. The overall HR of death was 0.73 (95% CI; 0.62–0.86), which significantly favored CRT over AF (P < 0.0001): DFS (HR = 0.79, 95% CI = 0.68–0.92; P = 0.002) and LRC (HR = 0.71, 95% CI = 0.59–0.84; P < 0.0001). CRT did not elevate the ratio of severe acute toxicity (dermatitis and mucositis); however, hematological toxicity, nephrotoxicity and late xerostomia were significantly increased with CRT. At present, no form of acceleration can potentially compensate fully for the lack of concurrent chemotherapy [29, 51].

On the other hand, RTOG 9011 unveiled the result of a long-term decrease in the survival rate in the CRT arm (one possibile explanation is toxicity of the chemotherapy) [51]; however, no deterioration in survival was found in the HF arm of RTOG 9003 for 5 years or later [32], which imply the safety of alternation strategy than CRT strategy over the long term. In addition, only the GORTEC trial provided data comparing CRT and alternated radiotherapy, and no data is available for comparing HF and CRT. Therefore, there is a room for examine the role of HF if its benefit would overcome the institutional and patient inconvenience.

DISCUSSION

For definitive radiotherapy, there is an extensive body of published data regarding the management of early glottic cancer with ⁶⁰Cobalt or 2–4-MV photons. Factors studied for prognostic importance related to local failure included pretreatment hemoglobulin, sex, T-category, histology differentiation, anterior commissure involvement, subglottic extension, tumor bulk, and treatment factors prescribed earlier in the current review [1]. Those important factors could be translated into principles of modern radiotherapy, even when advanced technologies are utilized. In this paper, we focused on radiological treatment factors versus other factors (e.g. tumor and patient factors).

The concept of alternation radiotherapy depends on fractionation and OTT, but it was difficult to determine which was the dominant factor. Several previous studies have confirmed the importance of fraction size for local control. They treated Stage T1 laryngeal cancer with 63 Gy using fraction sizes of 1.75, 1.9 and 2.1 Gy [52]. The 1.75-Gy arm was discontinued when local control was obtained for only 1 out of 3 patients (33%), compared with 12 out of 14 (86%) patients treated with 1.9 Gy fractions and 4 out of 4 (100%) patients treated with 2.1 Gy fractions. Rudoltz et al. found that the local control of patients treated with <2 Gy/fr was 62%, compared with 87% for those treated with ≥2 Gy/fr (P = 0.006) [53]. The local control was 100% if treatment was completed within 42 d, 91% if within 43–46 d, 74% if within 47–50 d, 65% if within 51–54 d, and 50% if within 55–66 d (P = 0.001) [53]. On the other hand, decreasing the dose/fr (≤1.2 Gy) can protect late-responding normal tissues more than tumor cells, leading to a differential effect that allows the delivery of a higher total dose than that of SF in the same overall time and constitutes the basis of HF radiotherapy.

Simultaneously, accelerated radiotherapy schedules have been developed with the aim of overcoming tumor cell repopulation during the treatment course. Accelerated radiotherapy is related to the rapid tumor cell kinetics of these cancers, with a potential doubling time of 2–6.8 d in the majority of cases [54–56]. Several RCTs have compared SF with AF radiotherapy to examine the importance of OTT. Of these, large variations in the OTT have been reported, from 1 week (moderate acceleration) to >3 weeks of reduction (high acceleration).

Moderate AF has been examined in many RCTs (Tables 2, 3), without dose reduction. The DAHANCA [23, 28] improved 5-year locoregional control rate by 10% using a 6-daily fr/week group (OTT = 39 d) versus the SF 5-daily fr/week group (OTT = 46 d). Those schedules decreased the OTT by 1 week (12 Gy/week), during the whole course of therapy, yielding a 10% increase in the 5-year locoregional control rate, with good early and late tolerance (Table 2) [23, 28]. HF with commonly used regimens of 80.5–81.6 Gy given in 1.15–1.2 Gy/fr BID with a 6-h interval, over the conventional 7 weeks, also provided a 10% improvement in locoregional control [1], owing to the total dose escalation. Generally, all of these alternate fractionations offer a similar gain in locoregional control, and it appears that an average dose accumulation of ~12 Gy/week was a key to the success of definitive radiation therapy alone, although timing of dose delivery was different. These figures (i.e. a 1–2 week reduction) concur with many other trials and imply that the important threshold of elapsed days was ~40–45 d. Overall, the evidence of a positive effect from reduced treatment time on locoregional control is convincing. The effect of AF on LRF was especially evident in well-differentiated tumors in the DAHANCA trial, which is in agreement with other studies [43]. This indicates that AF repopulation demands a certain degree of cellular differentiation in order to respond adequately to the induced trauma [2, 21, 22]. To secure such a response, the tumor needs to have a functional mechanism capable of regeneration, which is most likely to happen in well-differentiated tumors. Furthermore, the reaction might be controlled by signaling from the surrounding normal mucosa, and the response is, therefore, seen only in the T-site and not the nodal metastases [23, 28].

Very accelerated schedules were used in several RCTs (Table 3) [39–43]. CHART and TROG showed a marginal benefit in favor of the AF arm [39, 40, 42], suggesting that the total dose might have been too low (CHART: 54 Gy in 12 d; TROG: 59.5 Gy in 4 weeks) to obtain a significant benefit in terms of tumor control. In contrast, GORTEC (2 Gy BID, 62–64 Gy total) showed superior local control when the total dose was maintained at higher levels [28, 42]. Therefore, when using very AF, it may be important to keep the total dose as high as is reasonably achievable. In contrast, using higher total doses (16–25.2 Gy, weekly) in association with strong AF schedules, such as in the EORTC 22851 (72 Gy in 5 weeks, between moderate and very accelerated) and Vancouver trials (66 Gy in 3.5 weeks) [38, 41], increased acute toxicity, as seen in GORTEC, and may also increase late effects. It is important to note that those studies left lessons of severe adverse reactions [43]. On the other hand, the harmful effects of larger doses/day used in AF can be compensated for by delivering a lower total dose in this situation, and the gain made in reducing the overall time compensates for the small reduction in total dose administered.

cBED is a useful indicator when speculating the alternation radiotherapy schedule, considering repopulation has occured 28 day or later radiotherapy. A higher cBED implied better local control in multiple RCTs (Table 2). Again, the gain from the reduction in overall time exceeded the loss resulting from a reduction in the total dose (or BED10), which reflects the cBED concept. Table 2 shows that a 0.5–6.2-Gy cBED gain yielded a 5–15% improvement in local control. A meta-analysis in 2006 indicated a 6% gain in locoregional control using alternate fractionation, which was reproduced in several succeeding RCTs published thereafter. On the other hand, cBED did not apply in very AF and CRT and several forms of moderate AF-C (acceleration in late period of treatment) or AF-S. Additionally, there was no useful indicator of toxicity if the alternation radiotherapy schedule was applied (i.e. BED3 because it does not include the OTT concept), even though several RCTs set their schedule according to BED3 for late toxicity. In future, dose–volume analyses including time factors using 3D data could be fruitful for improving local control and reducing toxicity.

In some of these trials, the AF and HF regimens were feasible and were beneficial in locoregional control but had a modest effect on survival [43]. This discrepancy between the marked benefit observed in locoregional control and the limited effect on survival has already been reported [30]. Since the combination of chemotherapy has already been established as a standard regimen, AF could not be used as a new standard regimen. One more limitation is that almost all prior trials were performed with 2D radiation therapy. However, it seems unlikely that we would compare IMRT-based HF (or AF-C) with SF/CRT [32]. Improvements in radiation techniques and the development of new biological agents and biomarkers force us to reconsider all therapeutic combinations [32].

In addition, from the point of view of health economics, the moderate hypofractionated schedule can be expected to win out over the SF or HF schedule, because the shorter OTT achieved with the hypofractionation scheme reduces the socio-economic burden for patients and radiotherapeutic institutions; patients benefit from the reduced costs and treatment time. This would enable radiotherapeutic institutions to maintain the mechanical and human resources required to meet the increasing demand for radiotherapy. In conclusion, properly modified alternate fractionation could be a good option for improving local control without elevating late toxicity for head and neck squamous cell carcinoma. However, in locally advanced cancer, chemotherapy overcomes the efficacy of alternate fractionation and limits its role in early glottic cancer and/or situations where chemotherapy cannot be used.

CONFLICT OF INTEREST

The authors declare that they have no competing interests.