Clinical Outcomes for Larynx Cancer Patients Treated with Refinement of High-Dose Radiation Treatment Volumes

Abstract

Background:

To evaluate disease control, toxicities, and dose to dysphagia/aspiration risk structures (DARS) using a direct gross tumor volume (GTV_70Gy) to planning target volume expansion (dPTV_70Gy) for patients with squamous cell carcinoma of the larynx (LSCC).

Methods:

A retrospective review was performed on patients with LSCC treated between 2003-2018. Clinical outcomes, toxicities, and dosimetric data were analyzed.

Results:

Seventy-three patients were identified. Overall survival at 5-years was 57.8%. Five-year local and regional control was 79.8% and 88.2%, respectively. Distant metastatic-only failure was 2.7%. Eighty percent of failures were 95% contained within the dPTV_70Gy. Mean dose and the volume of DARS receiving 70 Gy was significantly lower for dPTV_70Gy compared to a consensus-defined PTV_70Gy.

Discussion:

Judicious reduction in high-dose target volumes can preserve high tumor control rates while reducing dose to normal surrounding structures underscoring the potential benefit of this approach in enabling local therapy intensification to improve locoregional control.

Introduction

Radiotherapy (RT)-based larynx preservation for patients with laryngeal squamous cell carcinoma (LSCC) has been established over the past several decades through a step-wise progression of randomized controlled trials^1–6. Except for T1N0 and T2N0 glottic SCC primaries, rates of 5-year larynx preservation and overall survival in appropriately selected patients treated with RT are approximately 65.8%-83.6% and 53.8%-58.1%, respectively⁷. Randomized trial outcomes have been confirmed by national cancer registry analyses supporting the importance of appropriate patient selection and efficacy of this approach^8,9.

Randomized data have demonstrated increased locoregional control with concurrent radiochemotherapy (RCT) compared to sequential regimens or RT alone². However, this approach results in increased acute and long-term toxicity^10,11. Indeed, retrospective analyses have demonstrated a relationship between irradiation of structures within and adjacent to the larynx and severe toxicities^12–15. Further, a relationship of these toxicities has been correlated with increasing dose^12,16,17. Given these findings, techniques to spare these critical structures from high dose RT could improve the acute and long-term toxicities associated with RT-based larynx preservation. As the dose of RT required to sterilize gross disease in patients with LSCC has remained relatively consistent over time at approximately 70 Gy in 2 Gy fractions, other methods for reducing toxicity warrant evaluation.

RT techniques used throughout the development of current day RT-based larynx preservation regimens have been refined from 2D approaches using bony anatomy to highly conformal techniques, such as intensity modulated RT (IMRT) and volumetric modulated arc RT (VMAT), to treat gross tumor (GTV), clinical target (CTV) and planning target volumes (PTV). Unlike the uniformity in dose used to sterilize high-, intermediate-, and, low-risk treatment volumes, the techniques used to create these volumes are variable^18–27. Consensus guideline recommendations for the design of planning target volumes for head and neck primary tumors continue to emerge²⁸ although practice heterogeneity remains a significant confounder across institutions²⁹. In brief, the consensus guideline approach expands the GTV_70Gy by 5 mm to create a CTV_70Gy. An additional 5 mm is added to the CTV_70Gy to create an intermediate dose CTV_60Gy. CTVs are trimmed from areas unlikely to contain subclinical disease such as air and bone and then expanded by 3-5 mm to create respective PTVs.

The consensus guidelines, which serve to standardize contouring for head and neck tumors, result in the treatment of subclinical disease to 70 Gy through the 5 mm expansion of the initial GTV_70Gy. As subclinical disease is treated to 60-66 Gy in the post-operative setting^30,31, it is possible that expansion of the GTV_70Gy may not be necessary to sterilize subclinical disease in the definitive setting. We recently reported our clinical outcomes of patients with p16-positive oropharyngeal squamous cell carcinoma (OPSCC) treated with a direct GTV_70Gy to PTV_70Gy expansion with an intermediate dose CTV_60Gy covering the gross disease with a concentric expansion of 10 mm³². In our analysis, local and regional control was approximately 90% with nearly all failures occurring within the initially defined GTV_70Gy. There is data to suggest that microscopic disease in this patient population that is located outside the 70 Gy dose but within the 60 Gy intermediate region is effectively sterilized³³. However, the applicability of this approach to non-p16-positive OPSCC is unknown. We therefore analyzed patients with LSCC undergoing comprehensive head and neck irradiation with or without systemic therapy treated as described above and report patterns of failure, dosimetric implications of this approach compared with consensus contouring guidelines, and toxicities.

Material and Methods

Approval for this study was obtain from the University of Wisconsin-Madison Institutional Review Board. We identified 73 patients with LSCC treated with definitive comprehensive head and neck IMRT with or without systemic therapy between 2003 and 2018. AJCC 8^th edition staging system was used to report patient data.

Treatment

Patients were immobilized and simulated using a thermoplastic head and neck mask and intravenous contrast unless medically contraindicated. The 70 Gy high-dose gross tumor volume (HD-GTV_70Gy) included the primary tumor and pathologic lymph nodes as determined by the treating Radiation Oncologist using data gathered by physical examination and cross-sectional imaging. A 60 Gy intermediate-dose CTV (ID-CTV_60Gy) consisted of a 10 mm expansion of the HD-GTV_70Gy - trimmed from bone and air – plus inclusion of involved nodal stations and stations thought to be at significant risk of harboring subclinical disease. A 54-56 Gy low-dose CTV (LD-CTV_54-56Gy) was used for prophylactic coverage of low-risk uninvolved nodal stations. All CTVs were volumetrically expanded by 2-3 mm to create respective PTVs. Concurrent systemic therapy consisted of either weekly cisplatin at 30-40 mg/m², Q3 weekly cisplatin at 100 mg/m², or cetuximab with a 400 mg/m² loading dose followed by weekly doses of 250 mg/m² weekly.

Patterns of failure determination

Patients with local, regional, and distant failures were identified. Imaging at the time of failure was co-registered, using deformable techniques, with the treatment planning CT and 95% isodose lines using MIM software as previously described (MIM Software Inc, Cleveland, OH)³⁴.

Comparison of dose to DARS using consensus contouring guidelines (cPTV_70Gy) versus direct GTV_70Gy to PTV_70Gy (dPTV_70Gy).

Patients with clinical T3-4N0 LSCC were contoured per recent consensus guidelines²⁸ and using a direct GTV_70Gy to PTV_70Gy approach as described above. Nodal basins II-IV were contoured per Radiation Therapy Oncology Group guidelines and treated to 56 Gy. cPTV_70Gy and dPTV_70Gy contours were used for planning with a TomoTherapy-based IMRT approach. DARS including the larynx, residual larynx (larynx minus gross disease), superior, middle, and lower constrictors, and cricopharyngeus were contoured as previously described³⁵.

Statistics

Overall survival, local and regional control, and gastrostomy and tracheostomy rates were analyzed by the Kaplan-Meier method³⁶. Cox regression analyses were performed to determine associations between covariates with overall survival and local control. The Fisher exact test was used to assess the relationship between type of systemic therapy and incidence of severe toxicities. Differences in doses to DARS were compared using the t-test with Bonferroni correction to account for multiple comparisons.

Results

We identified 73 patients with squamous cell carcinoma of the glottic or supraglottic larynx who received comprehensive RT-based larynx preservation. Patient, disease, and treatment characteristics are detailed in Table 1. Median follow-up was 58.8 months, 78.1% of patients presented with supraglottic primary tumors, 61.6% were clinically node positive, and 75.3% received systemic therapy. Five-year cause-specific, laryngectomy-free, and overall survival rates were 77.1% (95% CI: 66.5%-89.4%), 55.0% (95% CI: 43.6%-69.5%), and 57.8% (95% CI: 46.0%-72.5%), respectively (Figure 1A–C). Rates of local and regional control at 5-years were 79.8% (95% CI: 70.4%-90.4%) and 88.2% (95% CI: 80.3%-96.9%), respectively (Figure 1D). There was no statistically significant difference in locoregional control between patients with Stage I and II disease versus Stages III-IV. Further, we did not identify a significant difference in locoregional control for patients with Stage III-IV disease treated with radiotherapy and either concurrent cisplatin or cetuximab.

Table 1.

Patient and treatment characteristics

	Number	Percent
Age
Median	63
< 55	42	57.5
≥ 65	31	42.5
Sex
Female	15	20.5
Male	58	79.5
Tumor site
Glottic larynx	16	21.9
Supraglottic larynx	57	78.1
Tobacco use
Never	0	0
Former smoker	27	37.0
Current smoker	46	63.0
Pack years
Median	47
Alcohol
Never drinker	6	8.2
0-21 drinks per week	32	43.8
>21 drinks per week	31	43.5
T stage
T1	6	8.2
T2	12	16.4
T3	44	60.3
T4	11	15.1
N stage
N0	28	38.4
N1	9	12.3
N2	35	47.9
N3	1	1.4
Stage
I	3	4.1
II	5	6.8
III	23	31.5
IVA	42	57.3
Chemotherapy
Concurrent platinum	39	53.4
Neoadjuvant chemotherapy	1	1.4
Concurrent cetuximab	13	19.2
Lapatinib	1	1.4
None	18	24.7

^*Patients with T1N0 or T2N0 tumors were all supraglottic primaries treated with comprehensive head and neck irradiation.

Figure 1.

Five-year (A) laryngectomy-free survival 55.0% (95% CI: 43.6%-69.5%), (B) cause-specific survival 77.1% (95% CI: 66.5%-89.4%), (C) overall survival 57.8% (95% CI: 46.0%-72.5%), (D) local control 79.8% (95% CI: 70.4%-90.4%) and regional control 88.2% (95% CI: 80.3%-96.9%) for patients (n=73) with glottic or supraglottic squamous cell carcinoma undergoing RT-based larynx preservation.

On univariate analysis, patients who reported > 50 pack-years (median reported pack-years of cohort) (HR 2.41; 95% CI: 1.11-5.25) exhibited an association with worse overall survival while patients presenting with node-negative disease demonstrated a trend towards improved overall survival (HR 0.42; 95% CI: 0.17-1.03). The association between smoking and overall survival was maintained in multivariate analysis (HR 2.5; 95% CI: 1.07-5.88). We did not identify patient, disease, or treatment factors associated with local control (Table 2).

Table 2.

Univariate and multivariate analysis

Univariate	Local Control		Overall Survival
	HR	95% CI	HR	95% CI
Over 21 drinks per week	1.53	0.48-4.87	1.22	0.51-2.88
Smoking > 50 pack-years	2.02	0.71-5.77	2.41	1.11-5.25
Age > 65 years	1.41	0.49-4.09	1.43	0.66-3.09
T4	2.68	0.84-8.58	1.40	0.53-3.75
N0	0.85	0.28-2.53	0.42	0.17-1.03
Chemotherapy	1.84	0.61-5.49	1.07	0.43-2.66
Multivariate
Smoking > 50 pack years			2.5	1.07-5.88

There were 19 failures in the entire cohort. Of these, 17 experienced a local, regional, or locoregional recurrence while 2 developed metastatic-only failure. Fifteen of these seventeen patients had treatment plans available for detailed dosimetric review. Of these, 80% experienced failures that were 95% contained within the high-dose treatment volume. One patient experienced a marginal failure of the primary tumor in the intermediate risk primary tumor region. One patient failed in the intermediate risk nodal region, and one patient failed in the low risk nodal region (Figure 3).

Figure 3.

Dosimetric comparison of dysphagia/aspiration related structures treated with cPTV_70Gy or dPTV_70Gy. Mean dose to the (A) larynx (p < 0.0005), (B) residual larynx (p < 0.03), (C) inferior constrictors (p < 0.002). V70Gy to the (D) larynx (p < 0.001), (E) residual larynx (p < 0.002), (F) inferior constrictors (p < 0.01), n=5.

Among 54 recurrence-free patients only two (3.7%) retained a gastrostomy tube at last follow-up. One of these patient was lost to follow-up at 9 months and the second passed away following complications of bacterial tracheitis. A total of 10 tracheostomy tubes were placed in non-recurrent patients. Two patients had tracheostomy tubes placed prior to treatment secondary to tumor-related symptoms. During and within 1 month of treatment, 4 tracheostomy tubes were placed with 3 being from treatment related edema and 1 given the need for airway access given a critical illness and difficulty with intubation. Following 1 month of treatment, 4 tracheostomy tubes were placed for persistent edema (n=1), laryngeal dysfunction (n=2), and chondronecrosis (n=1), all of which were likely treatment related. One of 2 tracheostomy tubes placed prior to radiation therapy was removed, while 3 of the 4 placed during or within 1 month of radiation therapy were able to be removed. Three of 4 tracheostomy tubes placed one month after completion of radiation remained in place at last follow-up. Thus, only 4 (7.4%) tracheostomy tubes placed during or after radiation therapy initiation were present at last follow-up with 1 patient ultimately undergoing total laryngectomy. Aspiration pneumonia occurred in 12 non-recurrent patients with a median time to occurrence of 16 months. Finally, hypopharyngeal/esophageal stricture dilation occurred in 8 patients with a median time of 20 months. There were no statistically significant relationships identified between the type of systemic therapy, cisplatin versus cetuximab, delivered and the incidence of the above evaluated toxicities. Finally, the associations between supraglottic and glottic larynx subsite with occurrence of tracheostomy, aspiration, or stricture were not statistically significant.

Patients (n=5) with node-negative LSCC who received comprehensive head and neck radiotherapy were contoured according to consensus guidelines²⁸ and with direct GTV_70Gy to PTV_70Gy expansions with both receiving 56 Gy to bilateral nodal basins II-IV. Dosimetric comparisons of DARS (total larynx, residual larynx, superior/middle/inferior constrictors, and cricopharyngeus) was performed. There was a significant difference in the mean PTV_70Gy volume between cPTV_70Gy and dPTV_70Gy (mean 35.3 mL: range 19.6 mL – 40.9 mL versus mean 16.4 mL: range 7.6 mL – 23.5 mL; p < 0.001). There was a significant reduction in the mean and V70 for total larynx, residual larynx, and inferior constrictor (Figure 4A-C). There was no difference in dose to the superior and middle constrictor and cricopharyngeus. There was no difference in the maximum dose delivered to cPTV_70Gy and dPTV_70Gy (mean 74.5 Gy: range 72.0 Gy – 75.9 Gy versus mean 74.9 Gy: range 73.9 Gy – 75.3 Gy; p = 0.60).

Discussion

RT-based larynx preservation is a well-established treatment modality for appropriately selected patients with LSCC^1–6,37. Excluding patient with Stage I and Stage II glottic SCC primaries, patients with larynx cancer without extension through the thyroid cartilage or extending more than 1 cm into the base of tongue, rates of larynx preservation and laryngectomy-free survival is approximately 65.8-83.6% and 34.0%-47.0%, respectively, at 5-years when managed with either radiation alone or radiation combined with systemic therapy⁷. We demonstrate that using reduced high-dose target volumes excellent local and regional control is similarly achievable as our 5-year primary tumor control was 80% and laryngectomy-free survival 55%.

High doses of radiotherapy to DARS was significantly lower using a dPTV_70Gy approach compared to recommended guidelines²⁸. The endpoint of larynx preservation has to be evaluated in the context of the potential long-term sequelae from RT-based approaches such as dysphagia, aspiration, and aspiration-induced pneumonia. Acutely, rates of dysphagia, as defined by the ability to eat only soft foods or take liquids, at one year was reported to be 23% for patients undergoing CRT in a large randomized trial². Further, in other series rates of feeding tube presence at 1-year has been shown to be as high as 20%³⁸. Dysphagia in the long-term setting (> 1 year) has been reported to be as high 59%-83%^16,39. The impact on quality of life from long-term dysphagia⁴⁰ is compounded by the related complication of aspiration and aspiration-induced pneumonia, which has been demonstrated to negatively impact overall survival^41–44. Given the severity of these potential life-altering events, investigators have evaluated dosimetric parameters associated with their occurrence and have proposed recommendations to reduce their severity and incidence through various RT planning approaches.

O’hare et al. evaluated patients with larynx cancer treated with RT or CRT and found that a mean dose of 43.6 Gy (95% CI: 30.5 – 56.7; p = 0.023) and maximum dose of 58.3 Gy (95% CI: 47.9 – 68.7; p = 0.06) to the cricopharyngeus were associated with a risk of death from aspiration-induced pneumonia¹⁴. Retrospective analysis of long-term survivors from RTOG 91-11, 97-03, and 99-14 that evaluated the use of CRT for patients with locally advanced HNSCC demonstrated an association of dose received to the inferior hypopharynx (OR 1.023 per Gy; p = 0.016) with the development of chronic grade 3-4 pharyngeal/laryngeal toxicity and/or requirement for a feeding tube ≥ 2 years after registration and/or potential treatment-related death such as pneumonia within 3 years⁴⁵. Evaluation of dosimetric parameters in conjunction with late swallowing dysfunction and dysphagia following RT for oropharyngeal cancers identified significant associations between these late toxicities with increasing supraglottic mean and median dose and supraglottic volume receiving > 60 Gy¹⁶.

Eisbruch et al. identified on post-therapy videofluoroscopy common abnormalities that contributed to aspiration (e.g. reduced elevation of the larynx, reduced epiglottic inversion, and prolonged pharyngeal transit time) and that radiotherapy dose to the pharyngeal constrictors, supraglottic and glottic larynx was likely responsible for these toxicities¹³. Dosimetric studies were performed on patients with either larynx (n = 6), hypopharynx (n = 2), tonsil (n = 3) and base of tongue (n = 9) using 3D-RT, standard IMRT (stIMRT), and dysphagia-optimized IMRT (doIMRT), which was similar to standard IMRT expect for minimizing the volume of dysphagia/aspiration-related structures (DARS) receiving ≥ 50 Gy. In their analysis, stIMRT was able to significantly spare the V₅₀ to the pharyngeal constrictors by an average of 10% (range, 0-36%, p < 0.001) and the larynx by an average of 7% (range, 0-56%, p = 0.054) compared to 3D-RT. However, doIMRT demonstrated even greater sparing of pharyngeal constrictors by increasing the dose reduction by 10% (range, 0-38%, p < 0.001) compared to stIMRT and for the larynx and additional 11% (range 0-41%, p = 0.002). As there was overlap with planning target volumes and DARS, the maximum dose was not different between the treatment modalities.

Using a dPTV_70Gy approach we were able to demonstrate a significant reduction in dose to the larynx, residual larynx, and inferior constrictors, which as described above have been associated with long-term sequelae of RT-based larynx preservation. We were able to accomplish this dose reduction without reducing disease control in this patient cohort.

The target contouring approach described in this report provides a > 50% reduction in the volume of normal tissue receiving 70 Gy while maintaining excellent local tumor control rates. Similarly, there was a significant reduction in mean dose and V70 dose for total larynx, residual larynx, and inferior constrictor muscles. Of the 73 patients reported in this series, only 2 developed distant metastases as their only site of disease progression. This finding suggests that future studies might emphasize approaches to improve local and regional tumor control with maintenance of quality organ function rather than high emphasis on distant disease. Given the reduction of radiotherapy dose to surrounding normal tissue structures (including DARS) using a dPTV_70Gy, we hypothesis that this technique may permit future therapy intensification potentially through modulation of radiation fractionation or the use of radiosensitizing agents.

Figure 2.

Patterns of failure.