Stereotactic body radiotherapy for T1 glottic cancer: dosimetric data in 27 consecutive patients

Abstract

Aim:

Because the clinical feasibility of stereotactic body radiotherapy (SBRT) for early glottic cancer (T1) is controversial, we report dosimetric results in 27 consecutive patients from a prospective phase I and II study that started in 2017.

Methods:

In our approach, only the parts of the true vocal cord containing cancer and those immediately adjacent are planned to be treated to 36 Gy and 30 Gy, respectively, in 3 fractions. Several dosimetric metrics for both target volumes and organs at risk were extracted from individual plans and results were compared to those achieved by other authors in a similar setting.

Results:

Proper coverage was reached at planning in 2/3 of planning treatment volume 30 Gy, but only 4 planning treatment volume 36 Gy; conversely, the maximum dose objective was met for most of the patients on either arytenoid cartilage, but this was not the case for 51.9% and 96.3% of cricoid and thyroid cartilages, respectively. Our dosimetric results are similar to if not better than those achieved by others.

Conclusion:

SBRT in 3 fractions for T1 glottic lesions is dosimetrically challenging. Clinical validation is awaited.

Keywords

SBRT early glottic carcinoma chrondronecrosis

Introduction

Cancer of the larynx is the most common tumor of the head and neck area.¹ Out of laryngeal cancers, ≈2/3 arise in the glottic area, the majority being staged early lesions (T1 or T2) at diagnosis. For these, treatment options include radiotherapy, transoral laser microsurgery, and less frequently open surgery.^2–5 Voice quality is usually better after fractionated radiotherapy than laser surgery,⁶ but this comes at the cost of the higher logistic load of several treatment sessions. Guidelines recommend treatment of the whole larynx to 63–66 Gy in 28–33 fractions even for T1 lesions of the true vocal cords (TVC).⁷

Technological improvements of RT planning and delivery have led to the possibility of reducing treated volumes down to the gross tumor volume (GTV), representing the (portion of the) TVC containing cancer in case of a T1 lesion.⁸ Moreover, volume reduction allows the use of larger doses per fraction, resulting in a smaller than standard number of total treatment sessions. However, hypofractionation increases the risk of late side effects, particularly on the cartilage.⁹ Three experiences addressing at least part of this strategy have been reported with conflicting late toxicity profiles.^8,10,11 The difference in outcome among the three studies reflects at least in part the different biological doses delivered to the organs at risk (OAR) surrounding the target and particularly the cartilaginous framework.

In 2017, we launched a phase I–II study for T1 glottic cancer planning to deliver 36 Gy and 30 Gy to the involved and to the adjacent portions of the TVC, respectively, in 3 fractions. Late laryngeal toxicity results are not mature yet. Therefore, in the present study, we aimed at checking the dosimetric feasibility of stereotactic body radiotherapy (SBRT) in terms of both target coverage and OAR spare in 27 consecutive patients. We compared our dosimetric results with those achieved by others in a similar setting.

Methods

This is a prospective, institutional review board–approved (registry number 897/16), phase I–II study enrolling patients with the diagnosis of early glottic cancer. Detailed inclusion criteria were age between 18 and 80 years, biopsy-proven squamous cell carcinoma (SCC), stage T1a or T1b SCC of the TVC according to American Joint Committee on Cancer 7th ed, no previous radiotherapy to the larynx, no previous surgery except for direct microlaryngoscopy with biopsy, lack of comorbidity that would preclude radiotherapy delivery or tolerance, and specific signed informed consent. Exclusion criteria were biopsy-proven evidence of dysplasia of any grade in other subsites of the larynx, previous local surgery, presence of nodal disease, and impaired cord mobility.

All patients underwent direct microlaryngoscopy under general anesthesia in order to perform an narrow-band imaging–guided biopsy and a simultaneous endoscopy with 0° and 70° rigid telescopes to confirm the extension of the glottic lesion.

SBRT planning and delivery

All patients underwent planning computed tomography (CT) in the treatment position, supine with the head slightly hyperextended, and 1.25-mm-thick slices were obtained during quiet respiration from the vertex to the clavicles. Patients were instructed not to swallow during planning CT and cone beam CT (CBCT) acquisition or treatment delivery. Patients were allowed to swallow in between treatment arcs, though the setup was checked before each arc by CBCT.

A thermoplastic mask with bite block had been preliminarily obtained. The head was aligned in the cephalocaudal axis in order to make it possible to compare symmetric structures. If no proper neck tilt had been obtained, images were reconstructed according to a plane parallel to the vocal cords (or to the intervertebral disk space at the level of C4–C5 or C5–C6).

The TVCs were divided in thirds along the axial plane as illustrated in Figure 1. The anterior commissure was considered a separate part. The parts containing biopsy-proven SCC were prescribed 36 Gy in 3 fractions; the next ones in each direction (anterior and posterior) were planned to receive 30 Gy in 3 fractions. Treatment had to be delivered every other day. The anterior commissure was electively treated to 30 Gy for lesions of the anterior third. Other examples are provided in Figure 1. Laterally, clinical target volume (CTV)36 was limited to the cord, while CTV30 included a variable portion of the paraglottic space.¹² Posteriorly, the vocal process of the arytenoid cartilage was included within the CTV for tumors involving the posterior third. The thyroid and cricoid cartilages, as well as the air cavity, were always excluded from any CTV. In a craniocaudal direction, the CTV included both the supraglottic and the subglottic aspects of the TVC.

CTVs were expanded to planning target volume (PTV) by an anisotropic margin of 3 mm laterolateral and anteroposterior and 5 mm craniocaudal.

OAR included the following: spinal cord, whole larynx,¹³ contralateral TVC (cTVC), ipsilateral and contralateral arytenoid cartilages (iAC and cAC),¹⁴ thyroid cartilage (TC),¹⁴ cricoid cartilage (CC),¹⁴ esophagus, hypopharynx (including also the inferior constrictor muscle), thyroid gland, and ipsilateral and contralateral carotid arteries (iCA and cCA, respectively).

The purpose of planning was adequate PTV coverage (V95 ⩾ 95%) while satisfying the dose/volume objectives reported in Table 1. The cTVC was constrained only for T1a lesions. The maximum dose inside PTV36 was tentatively limited to 107% of the prescription dose. The highest planning priority was to meet dose volume objectives on the spinal cord and TC/CC/iAC (Table 1). However, on a case-to-case basis, allowance was made to violate cartilage maximum dose up to 36 Gy in order to increase PTV coverage.

Table 1.

Dose volume objective at planning.

Region of interest	Metric	Dose objective	Priority
PTV36	V_95%	>95%	2
PTV36	D_0.1cc	3852 cGy	2
PTV30	V_95%	>95%	2
Spinal cord	D_0.1cc	900 cGy	1
Ipsilateral carotid artery	D_1cc	1100 cGy	3
Contralateral carotid artery	D_1cc	750 cGy	3
Thyroid cartilage	D_0.1cc	3000 cGy	1
Cricoid cartilage	D_0.1cc	3000 cGy	1
Ipsilateral arytenoid cartilage	D_0.1cc	3000 cGy	1
Contralateral arytenoid cartilage	D_0.1cc	2000 cGy	3
Contralateral true vocal cord	D_mean	2500 cGy	3
Whole larynx	D_mean	700 cGy	3
Hypopharynx	D_0.1cc	1500 cGy	3
Extralaryngeal tissues	D_0.1cc	2000 cGy	3

PTV: planning target volume.

Plans were generated by the Eclipse™ treatment planning system (Varian Medical Systems, Palo Alto, CA) and calculated with the analytical anisotropic algorithm (AAA) (version 15.6) at a grid size ⩽1.5 mm. Two to four arcs and 6 MV photons in flattening filter-free (FFF) mode of a Varian TrueBeam linear accelerator (LINAC) equipped with a Millennium multileaf collimator (120 leaves) were used. Dose rate was maximized up to 1400 MU/min to shorten treatment delivery time. Setup was verified on the LINAC by CBCT before each arc matched on the cartilage framework and corrected if shift was >2 mm in any direction.

Dosimetric data

For the purpose of the present study, dose volume data for each region of interest (ROI) were extracted from the planning system at dose bins of 10 cGy. Various metrics were obtained for individual ROI. Of note, plans were obtained following the metrics implemented for other head and neck cancers treated with intensity-modulated radiation therapy (IMRT) at our institution (Table 1), while we extracted several other dosimetric features per International Commission on Radiation Units & Measurements (ICRU) report 83 on prescribing, recording, and reporting photon-beam IMRT.¹⁵ Therefore, for target volumes, we extracted the volume of the ROI, the percent of ROI receiving at least 95% of the prescription dose (V_95%), and the dose to 2%, 50%, 95%, and 98% of ROI (D_2%, D_50%, D_95%, D_98%, respectively). D_2%, D_50%, and D_98% are defined per ICRU 83 as the near maximum dose (D_near-max), the median absorbed dose (D_median), and the near-minimum dose (D_near-min), respectively.¹⁵ For ROIs larger than or equal to 2 cc, the volumes near-max/near-min represent 2%/98% of the ROI, while for smaller ROIs, the near-max/near-min is an absolute volume of 0.035 cc. Interestingly, the same metrics are recommended also for SBRT with photons as per ICRU report 91.¹⁶ For each metric, we retrospectively defined the following thresholds as adequate: 95%, 100%, and 107% of the prescription dose for D_near-min, D_median, and D_near-max, respectively.

For OARs, besides the ROI volume, we extracted the mean dose (D_mean) and the dose to 1 cc (D_{1 cc}), 0.1 cc (D_{0.1 cc}), or 0.035 cc (D_{0.035 cc}), as appropriate. Moreover, for cartilages, the nominal dose was converted to the biological one at 2 Gy per fraction through the linear quadratic formula¹⁷ assuming an α/β value of 3.8 Gy for chondronecrosis.¹⁸ Similarly, the dose to both the contralateral TVC and the whole larynx was converted with the same approach using an α/β value of 3 Gy.¹⁹ For selected OARs, the volume of ROI receiving a linear quadratic equivalent dose at 2 Gy (LQED2) higher than a certain amount was extracted as well. For TC, CC, and iAC, the total dose cut point at 2 Gy per fraction was set at 70 Gy; for the cAC, at 60 Gy; and for both carotid arteries, at 40 Gy.²⁰ For the whole larynx we used the D_mean and a threshold of 43.5 Gy.¹³

Statistics

Spearman rho test was used as a nonparametric method to assess the correlation between variables; the Mann-Whitney U test was used as nonparametric method to investigate the distribution of values between 2 groups. The Wilcoxon rank test was used to compare the dosimetric data from the present study to those extracted from the literature. Statistical significance was claimed for p values < 0.05, though a p value less than 0.1 was considered a trend towards significance. All analyses were conducted with SPSS v.25.0 and GraphPad v8.0.

Results

From May 2017 to January 2020, 27 patients were accrued. All patients were planned and treated; none of them was excluded due to dosimetric results at planning. T-stage was T1a and T1b in 17 (63.0%) and 10 (37.0%) patients, respectively. Fourteen patients (51.9%) had involvement of the anterior commissure. The anterior, middle, and posterior thirds were involved in 2 (77.8%), 18 (66.7%), and 9 (33.3%) patients, respectively. The median number of involved subsites was 2 (range 1–4).

SBRT plans were generated by 9 planners, for a median number of 2 plans each (range 1–10). All patients but one were treated with a volumetric modulated arc therapy (VMAT) approach; 2, 3, or 4 arcs were used in 18, 5, and 3 patients, respectively. Overall, the median (interquartile range [IQR]) duration of each arc was 0.57 minutes (0.48–1.04); median (IQR) overall beam on time per session was 1.56 minutes (1.07–2.15 minutes). For the patient treated with 7-field IMRT, the median beam on time for the whole treatment session was 1.18 minutes. Regarding plan acceptance, 21 were approved by a single radiation oncologist; a second radiation oncologist accepted the remaining ones. All plans were dosimetrically verified and median gamma passing rate (3%, 3 mm, th10) was 93.7% (IQR 91.0%–96.8%).

CTV/PTV dosimetric data

Median (IQR) values for selected dosimetric metrics of target volumes are reported in Table 2. Figure 2 shows individual dose-volume histogram for both CTVs and PTVs.

Table 2.

Selected dosimetric metrics at planning for the various regions of interest.

Region of interest and metric	Dose objective^a	Median	IQR		Patients, n (%)^a
CTV30
Volume, cc		1.3	1	1.6
V_95%, %	⩾95	100	100	100	26 (96.3)
D_{0.035 cc}, cGy	⩽3210	3762	3718	3795	0 (0)
D_50%, cGy	3000	3620	3535	3650	27 (100)
D_{V−0.035 cc}, cGy	>2850	3130	3000	3190	23 (85.2)
PTV30
Volume, cc		6.5	5.6	7.5
V_95%, %^a	⩾95	97.5	94.2	98.6	19 (70.4)
D_2%, cGy	⩽3210	3750	3720	3780	0 (0)
D_50%, cGy	3000	3350	3190	3425	27 (100)
D_98%, cGy	>2850	2840	2780	2880	13 (48.1)
CTV36
Volume, cc		0.4	0.2	0.6
V_95%, %	⩾95	100	98.9	100	23 (85.2)
D_{0.035 cc}, cGy	⩽3852	3741	3713	3790	25 (92.6)
D_50%, cGy	3600	3690	3635	3720	24 (88.9)
D_{V−0.035 cc}, cGy	>3420	3580	3445	3605	21 (77.8)
PTV36
Volume, cc		2.9	1.6	4.0
V_95%, %^a	⩾95	90.2	74.2	92.8	4 (14.8)
D_2%, cGy	⩽3852	3780	3750	3810	26 (96.3)
D_50%, cGy	3600	3600	3560	3620	16 (59.2)
D_98%, cGy	>3420	3260	3100	3355	1 (3.7)
D_{0.1 cc}, cGy^a	⩽3852	3760	3730	3780	27 (100)
Spinal cord
Volume, cc		18.3	15.5	20.2
D_{0.1 cc}, cGy^a	⩽900	590	525	685	27 (100)
D_2%, cGy	⩽900	490	445	585	27 (100)
Ipsilateral carotid artery
Volume, cc		4.2	3.5	5.3
D_{1 cc}, cGy^a	⩽1100	940	845	970	26 (96.3)
D_2%, cGy	⩽1100	850	795	890	26 (96.3)
V_{LQED2–40 Gy}, cc		0	0	0
Contralateral carotid artery
Volume, cc		4.1	3.6	5.3
D_{1 cc}, cGy^a	⩽750	720	685	770	17 (63.0)
D_2%, cGy	⩽750	670	630	710	24 (88.9)
V_{LQED2–40 Gy}, cc		0	0	0
Thyroid cartilage
Volume, cc		8.9	7.2	10.6
D_{0.1 cc}, cGy^a	⩽3000	3420	3205	3600	1 (3.7)
D_2%, cGy	⩽3000	3360	3180	3575	3 (11.1)
D_mean, Gy	⩽3000	2028	1830	2127	27 (100)
V_LQED2–70Gy, cc		1.2	0.6	2.2
Cricoid cartilage
Volume, cc		5.3	4.1	5.8
D_{0.1 cc}, cGy^a	⩽3000	3020	2900	3130	13 (48.1)
D_2%, cGy	⩽3000	3030	2925	3125	11 (40.7)
D_mean, Gy	⩽3000	1406	1293	1658	27 (100)
V_{LQED2–70 Gy}, cc		0.12	0.06	0.27
Ipsilateral arytenoid cartilage
Volume, cc		0.23	0.15	0.29
D_{0.1 cc}, cGy^a	⩽3000	2870	2443	3065	19 (70.4)
D_{0.035 cc}, cGy	⩽3000	3084	2918	3296	12 (44.4)
D_mean, Gy	⩽3000	2898	2525	3039	18 (66.7)
V_{LQED2–70 Gy}, cc		0.07	0.03	0.13
Contralateral arytenoid cartilage
Volume, cc		0.21	0.16	0.29
D_{0.1 cc}, cGy^a	⩽2000	1650	1387	1900	23 (85.2)
D_{0.035 cc}, cGy	⩽2000	1890	1572	2119	16 (59.3)
D_mean, Gy	⩽2000	1611	1452	1886	22 (81.5)
V_{LQED2–60 Gy}, cc		0.12	0.07	0.20
(Whole) larynx
Volume, cc		21.6	18.5	25.3
D_mean, cGy^a	⩽700	1770	1664	1942	0
D_mean (LQED2, cGy)	⩽4350	3152	2845	3678	24 (88.9)
Contralateral true vocal cord
Volume, cc		0.37	0.31	0.47
D_{0.1 cc}, cGy		3150	2915	3510
D_{0.035 cc}, cGy		3358	3086	3636
D_mean, cGy^a	⩽2000	2821	2642	3147	1 (3.7)
Hypopharynx
Volume, cc		7.8	6.4	9.5
D_{0.1 cc}, cGy^a	⩽1500	1850	1555	2385	4 (14.8)
D_2%, cGy	⩽1500	1740	1495	2240	8 (29.6)
Esophagus
Volume, cc		6.7	4.7	8.8
D_2%, cGy		90	55	290
Thyroid gland
Volume, cc		14.6	10.1	22.8
D_2%, cGy		865	610	1055
D_mean, cGy		104	60	141

CTV: clinical target volume; IQR: interquartile range; LQED2: linear quadratic equivalent dose at 2 Gy; PTV: planning target volume.

Metrics constrained at planning.

The objective of 95% PTV covered by at least 95% of the prescription dose was reached in more than 2/3 of PTV30, but only 4 (14.8%) PTV36. Overall, the results show that the median and near-minimum absorbed doses for CTV30/36 were adequate in the majority of patients. However, the expansion to PTV resulted in decreased coverage, especially for PTV36, with only one patient (3.7%) having adequate D_98%. Near-maximum doses (D_2%) for PTV36 were for all cases but one within 107% of the prescription dose. However, as expected, this was never the case for PTV30 that included PTV36.

OAR dosimetric data

Selected dosimetric data are summarized in Table 2 and illustrated in Figure 3. The D_{0.1 cc} constraint was met at planning for most of the patients on either AC. However, this was not the case for 51.9% and 96.3% of CC and TC, respectively. Moreover, on average, the near maximum dose absorbed by the TC was higher than the one received by the CC (p < 0.001) and the iAC (p = 0.001). The median (IQR) absolute volumes of cartilage receiving a dose higher than the LQED2 of 70 Gy are summarized in Table 2. The difference of volumes exposed to 70+ LQED2 among cartilages are highly significant atp = 0.001 (CC vs iAE) and p < 0.001 (both CC vs TC and iAE vs TC).

Within each selected cartilage, the D_{0.1 cc} was strongly correlated to the near-maximum dose (D_2% for both TC and CC and D_{0.035 cc} for AC) (TC: ρ = 0.990, p < 0.001; CC: ρ = 0.866, p < 0.001; iAC: ρ = 0.805, p < 0.001; cAC: ρ = 0.985, p < 0.001). Moreover, ipsilateral arytenoid cartilage and cricoid cartilage D_2% were highly correlated to each other (ρ = 0.706, p < 0.001), while neither one was correlated to TC D_2% (ρ = 0.145/p = 0.472 and ρ = 0.059/p = 0.770 for CC and iAC, respectively).

The larynx D_mean exceeded the dose objective at planning in all the patients, even if laryngeal D_mean was always ⩽43.5 LQED2. Similarly, the cTVC D_mean dose objective was violated in all but 1 patient. Regarding the hypopharynx, only a minority of patients were able to meet the dose volume objective (Table 2) but the dose objective of 15 Gy in 3 fractions (Table 1) was excessively conservative, being equal to 24 Gy at 2 Gy assuming an α/β ratio of 3 Gy.

Comparison with literature data

Al-Mamgani et al.⁸ published on 30 patients with T1a glottic cancer treated with a novel hypofractionated schedule in 16 fractions targeting the involved vocal cord only. Patients were treated with IMRT after daily image guidance with CBCT at Erasmus Medical Center (EMC). The next selected experience is the one from the Seoul National University (SNU), in which the authors tested two schedules to cover the whole larynx (to 47.6 Gy/17 fractions and 40.7 Gy/11 fractions) while simultaneously boosting the primary tumor (to 59.5 Gy and 55 Gy, respectively).¹⁰ Both T1 and T2 glottic cancers were eligible, with 6 and 7 patients accrued in each treatment schedule. Treatment was planned in quiet respiration and delivered through 2 arcs. Sher et al.¹¹ reported the outcome of a phase I fractional dose escalation study for patients with early stage glottic cancer including Tis, T1, and T2 at University of Texas Southwestern (UTSW), Dallas. Patients were enrolled in 1 of 3 schedules, which were progressively hypofractionated. A total of 4, 13, and 12 patients were planned to receive 50 Gy/15 fractions, 45 Gy/10 fractions, and 42.5 Gy/5 fractions, respectively. The target was identified at 4D planning CT and expanded isotropically by 5 mm, of which 2 mm were from internal gross tumor volume to CTV and 3 mm from CTV to PTV.¹¹

D_mean was the dose metric more consistently reported across institutions. Only UTSW and EMC reported maximum doses.¹¹ All doses have been normalized at 2 Gy per fraction in order to allow direct comparison across the various schedules.

Figure 4 illustrates available results.

Both the D_mean and the D_near-max on the contralateral arytenoid cartilage were significantly higher at our institution compared to UTSW, while for the ipsilateral AC this was the case only for the more fractionated schedule (15 fractions). For all selected OARs but the cTVC, the median biologically equivalent D_mean achieved in the present study was significantly lower than the one achieved at SNU in either 11 or 17 fractions.

Discussion

Under the conditions set here (Table 1), the present article shows that planning SBRT for T1 glottic carcinoma is a challenging task with a delicate trade-off between target coverage and OAR sparing. It is unclear whether SBRT for early carcinoma of the glottic larynx is clinically feasible in terms of long-term laryngeal toxicity, with one study being prematurely closed for severe toxicity¹⁰ and another one providing encouraging, albeit criticized,²¹ results.¹¹ Besides patient-related characteristics, the (biological) dose to the OAR (surrounding the tumor) is a major determinant of the risk of toxicity after radiotherapy,²² and the present data show that the dose tolerance has to be overcome at least on one cartilage in order to increase target coverage in this particular setting. Therefore, after normalization at 2 Gy per fraction, we compared the dose metrics achieved here to those reported by other mature studies in order to predict the clinical outcome of our patients. Our SBRT approach for early laryngeal cancer is unique under several aspects, including the fractionation of the total dose in only 3 treatment sessions. Since the risk of complications of slow responding tissues, such as the cartilaginous and the connective tissues, strictly depends on the size of the dose per fraction, with larger doses being associated with a higher risk of complications,²³ despite the adjustment in total dose, we came from a less favorable setting compared to the one of more fractionated schedules. When we planned the present study, we favored iAC/TC/CC sparing over PTV coverage, but soon we realized that a higher than ideal dose had to be accepted on the cartilages (Figure 3) in order to minimize cold spots within the target.

Overall our results show that PTV30 coverage was achievable in the majority of patients; when it was not achieved, this was due to underdosage of the portions outside the CTV. Conversely, PTV36 coverage was rarely achieved but, again, CTV36 was properly dosed in the majority of patients. Overall, these data suggest that the target volumes should be kept as small as possible. Regarding CTV delineation, similarly to Sher et al.,¹¹ we restricted the high dose volume to the gross tumor volume only. Moreover, we contoured uninvolved parts of the TVC but, compared to Al-Mamgani et al.,⁸ we avoided including the whole TVC and subdivided the target in two dose levels. Our approach is also different from the one of Kang et al.,¹⁰ who contoured the whole larynx in the lower dose volume. Our median CTV volumes, 1.3 cc and 0.4 cc for CTV30 and CTV36, are both significantly (p < 0.001) smaller than the ones from Erasmus MC (3.1 cc)⁸ and UTSW (3.93 cc),¹¹ respectively, and we doubt that they can be safely shrunk further.

The expansion margin from CTV to PTV should be kept as small as possible. We planned our patients under quiet (free) breathing and we added an anisotropic margin to CTV of 3 mm (laterolateral and anteroposterior) and 5 mm (craniocaudal) derived from the EMC experience with daily CBCT before IMRT.²⁴ Others have used isotropic margins of 3 mm from CTV to PTV.^10,11 The study of Kwa et al.²⁴ on tumor motion in 42 patients undergoing single vocal cord irradiation at EMC shows that CBCT at set up allows to correct for most of the interfraction motion with residual error justifying a margin size in the order of ≈1.5 mm (laterolateral), ≈2 mm (anteroposterior), and ≈3 mm (craniocaudal). However, based on CBCT at the end of treatment, there is evidence that intrafraction motion requires, on average, an additional ≈1.4 mm in the craniocaudal direction, while the increase in both the anteroposterior and laterolateral directions is almost negligible (0.1 mm).²⁴ Moreover, intrafraction deterioration of margins will depend also on the delivery time, with the LINAC-based VMAT approach being the shortest, as supported by the median delivery time per arc less than 1 minute reported in the present study. We hypothesize that, despite the delivery of a significantly larger dose per fraction, the combination of VMAT and CBCT before each arc and a dose rate up to 1400 MU/min may allow an even smaller margin size than the one from Kwa et al.²⁴ and a study analyzing the displacements within each arc at our institution is undergoing.

Regarding OARs, TC, CC, contralateral TVC, larynx, and hypopharynx were not spared in more than 50% of plans (Table 2). We have already commented on both the hypopharynx and larynx, the former subjected to an excessively tight objective at planning and the latter biologically spared in the majority of the patients. Indeed, all but 3 patients treated at our institution had a laryngeal D_mean equal to or less than 43.5 LQED2 compared to none of those treated at Seoul.¹³ All reported cTVC mean doses are very close, with the lowest median value achieved at Erasmus MC, which accrued only unilateral lesions and whose functional results on voice quality were excellent.⁸ One may argue that in our experience the mean absorbed equivalent dose at 2 Gy per fraction to the cTVC was around 70 Gy and thus no dosimetric sparing was performed over a standard treatment targeting the whole larynx. On the other hand, this could also be viewed as the opportunity to treat the disease in only 3 fractions while delivering a standard dose to the unaffected part. A prospective study assessing the function of the unaffected cTVC at stroboscopy is ongoing.

Regarding laryngeal cartilages (Figure 4), our median biologically equivalent mean doses are significantly better than those reported by both SNU and EMC,^{8, 10} while near-maximum doses are not different from those planned at UTSW on the iAC in 5 to 10 fractions.¹¹ One may argue that for small OARs, such as the arytenoid cartilages, the D_mean is a reasonable surrogate for the near-maximum dose (both D_{0.1 cc} and D_{0.035 cc}), while for larger ones, such as TC and CC, the D_mean may be significantly lower than both the D_{0.1 cc} and the D_2% (Figure 3). As a result, for large OARs, the mean dose may be underestimating the true clinical effect and the near-maximum dose would be a more appropriate metric, though the clinical impact of tiny hot spots within the cartilage is unclear. Since the arytenoid is the one with the highest risk of chondronecrosis among laryngeal cartilages,²⁵ our strategy was to favor the dose volume objective on the iAC among the other cartilages and this resulted in a significantly lower volume of iAC exposed to 70+ LQED2 than both CC and TC (Figure 3). Moreover, the risk of GR3+ late laryngeal toxicity was 0 and 33% (95% confidence interval [CI] 9.5%–70.0%) after 17 and 11 fractions, respectively, in Seoul,¹⁰ while it was 7.7% (95% CI 1.4%–33.3%) and 8.3% (95% CI 1.5%–35.4%) after 10 and 5 fractions, respectively, at UTSW.¹¹ Based on the dosimetric data reported in Figure 4, we should expect a rate of late GR3+ laryngeal toxicity close to or slightly higher than those reported at UTSW.

The reader should be aware of limitations of the present study. First, for the delineation of the larynx and its substructures, we followed consistently the available atlases,^13,14 but we cannot exclude that differences exist in contouring among institutions. Second, dose calculation accuracy might be an issue when small volumes and tissue dishomogeneities are involved. Shiraishi et al.²⁶ reported that for a fully air-embedded PTV volume of 3 cc, the gamma passing rate (2%, 2 mm, th10%) was only 89.7% for AAA. However, in our settings, median PTV30 volumes were larger (6.5 cc), only partially surrounded by air, and treated by 6MV photons in FFF mode, all representing characteristics associated with a high passing rate. Indeed, in the experience of Zhang et al.²⁷ on SBRT for early glottic cancer, the AAA algorithm provided a gamma passing rate (2%, 2 mm, th10%) of 94.7%. Third, we assume that planned and delivered dose distributions are the same, but this may not be the case.²⁸ Even if it has been shown that Cyberknife- and LINAC-based planning are equally effective in planning SBRT for early laryngeal cancer,²⁷ Cyberknife is associated with longer treatment delivery times and thus a higher risk of intrafraction organ motion that is only partially compensated by real-time tracking.²⁷ On the other hand, it has been hypothesized that the faster treatment delivery may increase incidence of normal-tissue toxicity, presumably by increasing conversion of sublethal to lethal damage in cells.²⁹ Finally, the conversion at 2 Gy per fraction is associated with inherent uncertainties in estimating the individual radiosensitivity, even if in the particular case of the laryngeal cartilages the alpha/beta value has been derived from clinical data.¹⁸

In conclusion, SBRT in 3 fractions for T1 glottic lesions is dosimetrically challenging due to the close proximity of the targets and the OARs. Nevertheless, the CTV can be properly dosed and the arytenoid cartilages spared in the majority of patients, but at the cost of exceeding the tolerance on the thyroid cartilage and, less often, on the cricoid cartilage. The implications of these dose distributions are uncertain and await clinical validation.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Giuseppe Sanguineti

References

Ferlay

Shin

Bray

, et al. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010; 127: 2893–2917.

Silver

Beitler

Shaha

, et al. Current trends in initial management of laryngeal cancer: the declining use of open surgery. Eur Arch Oto-Rhino-Laryngol 2009; 266: 1333–1352.

Hartl

Ferlito

Brasnu

, et al. Evidence-based review of treatment options for patients with glottic cancer. Head Neck 2011; 33: 1638–1648.

Mendenhall

Werning

Hinerman

, et al. Management of T1-T2 glottic carcinomas. Cancer 2004; 100: 1786–1792.

van Loon

Sjogren

Langeveld

, et al. Functional outcomes after radiotherapy or laser surgery in early glottic carcinoma: a systematic review. Head Neck 2012; 34: 1179–1189.

Aaltonen

Rautiainen

Sellman

, et al. Voice quality after treatment of early vocal cord cancer: a randomized trial comparing laser surgery with radiation therapy. Int J Radiat Oncol Biol Physics 2014; 90: 255–260.

Merlotti

Alterio

Vigna-Taglianti

, et al. Technical guidelines for head and neck cancer IMRT on behalf of the Italian Association of Radiation Oncology: head and neck working group. Radiat Oncol 2014; 9: 264.

Al-Mamgani

Kwa

Tans

, et al. Single vocal cord irradiation: image guided intensity modulated hypofractionated radiation therapy for T1a glottic cancer: early clinical results. Int J Radiat Oncol Biol Physics 2015; 93: 337–343.

Harrison

Crennan

Cruickshank

, et al. Hypofractionation reduces the therapeutic ratio in early glottic carcinoma. Int J Radiat Oncol Biol Physics 1988; 15: 365–372.

10.

Kang

Kim

, et al. Early closure of a phase 1 clinical trial for SABR in early-stage glottic cancer. Int J Radiat Oncol Biol Physics 2019; 105: 104–109.

11.

Sher

Timmerman

Nedzi

, et al. Phase 1 fractional dose-escalation study of equipotent stereotactic radiation therapy regimens for early-stage glottic larynx cancer. Int J Radiat Oncol Biol Physics 2019; 105: 110–118.

12.

Gregoire

Evans

, et al. Delineation of the primary tumour clinical target volumes (CTV-P) in laryngeal, hypopharyngeal, oropharyngeal and oral cavity squamous cell carcinoma: AIRO, CACA, DAHANCA, EORTC, GEORCC, GORTEC, HKNPCSG, HNCIG, IAG-KHT, LPRHHT, NCIC CTG, NCRI, NRG Oncology, PHNS, SBRT, SOMERA, SRO, SSHNO, TROG consensus guidelines. Radiother Oncol 2018; 126: 3–24.

13.

Sanguineti

Adapala

Endres

, et al. Dosimetric predictors of laryngeal edema. Int J Radiat Oncol Biol Physics 2007; 68: 741–749.

14.

Choi

Refaat

Lester

, et al. Development of a standardized method for contouring the larynx and its substructures. Radiat Oncol 2014; 9: 285.

15.

ICRU Report 81 on prescribing, recording, and reporting photon-beam intensity-modulated radiation therapy (IMRT). J ICRU 2010; 10: 1–93.

16.

ICRU Report 91 on prescribing, recording, and reporting of stereotactic treatments with small photon beams. J ICRU 2017; 14: 1–160.

17.

Wheldon

Deehan

Wheldon

, et al. The linear-quadratic transformation of dose-volume histograms in fractionated radiotherapy. Radiother Oncol 1998; 46: 285–295.

18.

Maciejewski

Taylor

Withers

HR.

Alpha/beta value and the importance of size of dose per fraction for late complications in the supraglottic larynx. Radiother Oncol 1986; 7: 323–326.

19.

Rancati

Fiorino

Sanguineti

NTCP modeling of subacute/late laryngeal edema scored by fiberoptic examination. Int J Radiat Oncol Biol Physics 2009; 75: 915–923.

20.

Zumsteg

Riaz

Jaffery

, et al. Carotid sparing intensity-modulated radiation therapy achieves comparable locoregional control to conventional radiotherapy in T1-2N0 laryngeal carcinoma. Oral Oncol 2015; 51: 716–723.

21.

Mendenhall

Holtzman

Dagan

, et al. In regard to Sher et al. Int J Radiat Oncol Biol Physics 2020; 106: 220–221.

22.

Bentzen

Constine

Deasy

, et al. Quantitative analyses of normal tissue effects in the clinic (QUANTEC): an introduction to the scientific issues. Int J Radiat Oncol Biol Physics 2010; 76: S3–S9.

23.

Withers

Thames

HD.

Dose fractionation and volume effects in normal tissues and tumors. Am J Clin Oncol 1988; 11: 313–329.

24.

Kwa

Al-Mamgani

Osman

, et al. Inter- and intrafraction target motion in highly focused single vocal cord irradiation of T1a larynx cancer patients. Int J Radiat Oncol Biol Physics 2015; 93: 190–195.

25.

Keene

Harwood

Bryce

, et al. Histopathological study of radionecrosis in laryngeal carcinoma. Laryngoscope 1982; 92: 173–180.

26.

Shiraishi

Fong de Los Santos

Antolak

, et al. Phantom verification of AAA and Acuros dose calculations for lung cancer: do tumor size and regression matter? Pract Radiat Oncol 2019; 9: 29–37.

27.

Zhang

Chiu

Dubas

, et al. Benchmarking techniques for stereotactic body radiotherapy for early-stage glottic laryngeal cancer: LINAC-based non-coplanar VMAT vs. Cyberknife planning. Radiat Oncol 2019; 14: 193.

28.

Mao

Rozario

, et al. Online dosimetric evaluation of larynx SBRT: a pilot study to assess the necessity of adaptive replanning. J Appl Clin Med Physics 2017; 18: 157–163.

29.

Mueller

Karam

SD.

SBRT for early stage larynx: a go or no go? it’s all in the delivery. Int J Radiat Oncol Biol Physics 2019; 105: 119–120.