Clinical Outcomes of Several IMRT Techniques for Patients With Head and Neck Cancer: A Propensity Score–Weighted Analysis

Accepted Manuscript Clinical outcomes of several IMRT techniques for patients with head and neck cancer: A propensity score weighted analysis Jean-Emmanuel Bibault, MD, MSc, Sophie Dussart, MD, Pascal Pommier, MD, PhD, Magali Morelle, MSc, Marius Huguet, MSc, Pierre Boisselier, MD, Bernard CocheDequeant, MD, Marc Alfonsi, MD, Etienne Bardet, MD, Michel Rives, MD, Valentin Calugaru, MD, PhD, Enrique Chajon, MD, Georges Noel, MD, PhD, Hinda Mecellem, MD, Stephanie Servagi Vernat, MD, PhD, Lionel Perrier, PhD, HDR, Philippe Giraud, MD, PhD PII:

S0360-3016(17)33522-8

DOI:

10.1016/j.ijrobp.2017.06.2456

Reference:

ROB 24390

To appear in:

International Journal of Radiation Oncology • Biology • Physics

Received Date: 13 December 2016 Revised Date:

9 March 2017

Accepted Date: 19 June 2017

Please cite this article as: Bibault J-E, Dussart S, Pommier P, Morelle M, Huguet M, Boisselier P, Coche-Dequeant B, Alfonsi M, Bardet E, Rives M, Calugaru V, Chajon E, Noel G, Mecellem H, Servagi Vernat S, Perrier L, Giraud P, Clinical outcomes of several IMRT techniques for patients with head and neck cancer: A propensity score weighted analysis, International Journal of Radiation Oncology • Biology • Physics (2017), doi: 10.1016/j.ijrobp.2017.06.2456. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Clinical outcomes of several IMRT techniques for patients with head and neck cancer: A propensity score weighted analysis

RI PT

Running title: Clinical outcome of IMRT for head and neck cancer

Jean-Emmanuel Bibault, MD, MSc 1, Sophie Dussart MD 2, Pascal Pommier MD, PhD 2,

SC

Magali Morelle MSc 3, Marius Huguet MSc 4, Pierre Boisselier MD 5, Bernard CocheDequeant MD 6, Marc Alfonsi MD 7, Etienne Bardet MD 8, Michel Rives MD 9, Valentin

M AN U

Calugaru MD, PhD 10, Enrique Chajon MD 11, Georges Noel MD, PhD 12, Hinda Mecellem MD 13, Stephanie Servagi Vernat MD, PhD14, Lionel Perrier PhD, HDR 3, Philippe Giraud MD, PhD1

Paris, France

TE D

1 Paris Descartes University, Paris Sorbonne Cité, Hôpital Européen Georges Pompidou,

EP

2 Leon Berard Cancer Centre, Lyon, France,

AC C

3 Univ Lyon, Léon Bérard Cancer Center, GATE L-SE UMR 5824, F-69008 Lyon, France, 4 Univ Lyon, University Lumière Lyon 2, GATE L-SE UMR 5824, F-69130 Ecully, France, 5 Montpellier Cancer Institute, Montpellier, France, 6 Oscar Lambret Cancer Centre, Lille, France, 7 Sainte Catherine Institute, Avignon, France, 8 René Gauducheau Cancer Centre, Saint Herblain, France,

ACCEPTED MANUSCRIPT 9 Claudius Regaud Institute, Toulouse, France, 10 Curie Institute, Paris, France,

12 Paul Strauss Cancer Centre, Strasbourg, France,

RI PT

11 Eugène Marquis Cancer Centre, Rennes, France,

13 Lorraine Institute of Oncology, Vandoeuvre-lès-Nancy, France,

M AN U

SC

14 Jean Godinot Institute, Reims, France

Corresponding author Pr. Philippe Giraud

TE D

Department of Oncology Radiotherapy Georges Pompidou European Hospital

EP

20, rue Leblanc, 75015 Paris, France

AC C

E-mail adress: [email protected]

Role of funding source

This work was supported by the National Institute of Cancer (INCa), 52 Avenue André Morizet, 92513 Boulogne Billancourt Cedex, for the collection of data.

Acknowlegdement

ACCEPTED MANUSCRIPT This study has received funding from the National Institute of Cancer (INCa). The

AC C

EP

TE D

M AN U

SC

RI PT

authors thank the 14 French centers involved in the study.

ACCEPTED MANUSCRIPT

Clinical outcome of several IMRT techniques for patients with head and neck cancer Running title: Clinical outcome of IMRT for head and neck cancer

RI PT

Summary

Helical Tomotherapy (HT) and Volumetric-modulated Arc Therapy RapidArc (RapidArc)

SC

are two different IMRT techniques used to treat patients with head and neck cancer. In this study, we have prospectively compared their efficacy and toxicity. After inverse

M AN U

probability of treatment weighting (IPTW), loco-regional control and cancer specific survival rates at 18 months were significantly better in the Tomotherapy® group. No

AC C

EP

TE D

significant difference was shown on progression-free or overall survival.

ACCEPTED MANUSCRIPT

Clinical outcomes of several IMRT techniques for patients with head and neck cancer: A propensity score weighted analysis

RI PT

Running title: Clinical outcomes of IMRT for head and neck cancer

Keywords

AC C

EP

TE D

M AN U

SC

Head and neck cancer; Intensity-modulation Radiation Therapy

1

ACCEPTED MANUSCRIPT Abstract (272/300 words) Purpose: The ART-ORL study (NCTXXXXXX) was performed to prospectively evaluate the clinical and economic aspects of Helical Tomotherapy® and Volumetric Modulated

RI PT

Arc Therapy RapidArc® for patients with head and neck cancer. Materials and methods: Fourteen centers participated in this prospective comparative study. Randomization was not possible due to the availability of equipment. Patients

SC

with epidermoid or undifferentiated nasopharyngeal carcinoma, epidermoid carcinoma of oropharynx and oral cavity (T1-T4, M0, N0-N3) were included between February

M AN U

2010 and February 2012. Only the results of the clinical study are presented here, as the results of the economic assessment have been published previously. Inverse probability of treatment weighting (IPTW) using the propensity score analysis was undertaken in an effort to adjust for potential bias due to non-randomization. Loco-regional control,

TE D

specific and overall survival assessed 18 months after treatment were evaluated, as well as long-term toxicity and salivary function.

EP

Results: The analysis included 166 patients. The following results are given after IPTW adjustment. Loco-regional control rate at 18 months was significantly better in the

AC C

Tomotherapy® group: 83.3% (95% CI: [72.5%; 90.2%]) vs 72.7% (95% CI: [62.1%; 80.8%]) in the RapidArc® group (p=0.025). Cancer specific survival rate was better in the Tomotherapy® group: 97.2% (95% CI: [89.3%; 99.3%]) vs 85.5% (95% CI: [75.8%; 91.5%]) in the RapidArc® group (p=0.014). No significant difference was shown on progression-free or overall survival. Tomotherapy® induced fewer acute salivary disorders (p=0.012). Post-treatment salivary function degradation was worse in the RapidArc® group (p=0.012).

2

ACCEPTED MANUSCRIPT Conclusion: Tomotherapy® provided better loco-regional control and cancer specific survival than RapidArc®, with fewer salivary disorders. No significant difference was shown on progression-free and overall survival. These results should be explored in a

AC C

EP

TE D

M AN U

SC

RI PT

randomized trial.

3

ACCEPTED MANUSCRIPT Introduction Intensity Modulated Radiation Therapy (IMRT) is the standard of care in head and neck cancer patients treated by radiotherapy (1). It has shown superior preservation of the salivary gland function in Upper Aerodigestive Tract (2, 3), when compared with 3D

RI PT

conformal radiotherapy. Several IMRT techniques are available: Helical Tomotherapy® (Accuray) and VMAT (Elekta SmartArc® or Varian RapidArc®). These different modes result in dosimetric differences that were previously evaluated (4, 5). Tomotherapy®

SC

reduced high doses to the planning target volumes (PTVs) compared to RapidArc® and provided a more homogeneous dose distribution with an increased Non Tumoral

M AN U

Integral Dose (NTID). Whether or not these differences translate into the clinic has not been prospectively assessed yet. We performed a prospective multicentric study investigating both clinical outcomes and costs of TomoTherapy® (Accuray) and VMAT (Elekta or Varian RapidArc®) in patients with head and neck cancer under real-life

TE D

conditions. We already published the results of the cost analysis, which showed that, on the time horizon of radiotherapy, Tomotherapy® appeared to be more expensive than RapidArc®, mainly due to the higher price of the accelerator, the higher cost of

EP

maintenance, and the longer duration of session treatment (6). This paper presents the

AC C

clinical outcomes (loco-regional control, disease-free, overall survival, acute and longterm toxicities) of Tomotherapy® versus RapidArc®, techniques in fourteen academic French cancer centers.

Materials and methods Tomotherapy® was first installed in 2007 through the NCI new technology implementation project (7) and in 2011 the ART-ORL trial was created to compare cost 4

ACCEPTED MANUSCRIPT and clinical effectiveness of several IMRT techniques (8). The inclusion criteria were patients ≥ 18 years of age with epidermoid or undifferentiated nasopharyngeal (UCNT) carcinoma, epidermoid carcinoma of the oropharynx and oral cavity (T1-T4, M0, N0N3), without previous surgery, with a WHO performance status ≤ 2, and who were

RI PT

treated exclusively with radiotherapy of the bilateral cervical lymph nodes. Patients received 3D Intensity Modulated Arc Therapy (IMAT) with TomoTherapy® (Accuray), or VMAT (Elekta SmartArc® or Varian RapidArc®). Neoadjuvant chemotherapy or

SC

concomitant chemotherapy, including Cetuximab, was allowed (9). Exclusion criteria were previous invasive cancer, cervicofacial radiotherapy, postoperative radiation

M AN U

therapy, treatment with Amifostine, brachytherapy, and indications for re-irradiation. All patients provided informed consent. This prospective, multicenter, current care study (ClinicalTrials.gov; NCTXXXXXX) conducted in the framework of the French National Cancer Institute (INCa) involved 14 French centers located within academic

TE D

institutions, as previously detailed (6). Randomization between treatments with TomoTherapy®, SmartArc® or RapidArc® was not possible as participating centers

EP

offered only one of the competing arc therapy techniques. All eligible patients seen in the radiotherapy departments participating in the study were included. Initially set at 12

AC C

months, the inclusion period was extended by one more year, due to a slower recruitment rate than expected in accordance with the French National Institute of Cancer which supported the work. The expected number of patients to be included in the study was 120 for TomoTherapy®, 120 for RapidArc® and 60 for SmartArc®. Target volumes were treated with three different dose levels (simultaneous integrated boost) according to the risk of relapse. High, medium and low risk volumes received respectively 70 Gy (2 Gy per fraction), 63 Gy (1.8 Gy per fraction), 56 Gy (1.6 Gy per fraction) in 35 fractions delivered 5 days per week over a 7 week time period. The 5

ACCEPTED MANUSCRIPT clinical endpoints of this study were the loco-regional control rate, progression free survival (PFS), specific and overall survival 18 months post treatment. In the cost analysis study, the endpoint on toxicity was restricted to the number of acute adverse events occurring to the treatment time horizon (day 1 to the last day of irradiation) in

RI PT

accordance with the resource consumption collection. In this study, we analyzed early (<3 months) and long-term (>6 months) toxicities scored using CTCAE v3. Salivary function was evaluated at 3, 6, 12, 18 months after treatment with different scales: a

SC

modified scale toxicity grading from the Radiation Therapy Oncology Group (RTOG) (10), a Visual Analogic Scale (VAS) and with a measure of salivary weight. Patients were

M AN U

asked to chew on a 5x5cm plastic paraffin film (Parafilm®) for two minutes and swallow their saliva. The saliva produced during five minutes after this stimulation was then weighted. The study was conducted in accordance with the ethical principles for medical research involving human subjects developed in the Declaration of Helsinki by the

TE D

World Medical Association (WMA). The study received approval in France from the National Ethics Committee (N°909226) and the National Committee for Protection of

EP

Personal Data (N°09- 203).

AC C

Statistical methodology

The analyses were performed using Stata 13.0 (StataCorp, College Station, TX) and SAS software (version 9.4; SAS Institute, Cary, NC). Loco-regional recurrence was defined as an infield or a borderline relapse (tumor and cervical lymph nodes). Cancer specific survival was defined as the time from inclusion to death from neoplastic diseases. Progression Free Survival (PFS) was defined as the time elapsed between inclusion and disease progression (loco-regional or metastatic) or death from any cause. Overall

6

ACCEPTED MANUSCRIPT survival (OS) was defined as the time from inclusion to death from any cause or censored at the last follow-up. Crude analysis

score

analysis

on

the

estimation

of

the

effect

RI PT

Because it is important that the reader be able to estimate the impact of the propensity of

the

treatment,

a

preliminary analysis of the original dataset was performed as followed: OS and PFS

SC

functions were calculated using the Kaplan-Meier method (11). Survival distributions were compared between the 2 arms using a log –rank test (12). Median follow-up time

M AN U

was calculated using a reverse Kaplan-Meier estimate (13). Because there were possible competitive events to loco-regional relapse (death from any cause without loco-regional relapse and distant relapse) and to death from neoplastic diseases (death from second cancer, toxicity, and other cause), the competing risk approach was used to estimate

TE D

loco-regional relapse-free survival and cancer specific survival from inclusion (14). The cumulative incidence function developed by Kalbfleisch and Prentice (15) and nonparametric Gray’s test (16) were used to estimate and compare cumulative

EP

incidence function between the 2 arms. Event-free survival probabilities were reported

AC C

as (1 - [cumulative incidence probability]). Results of standard multivariate analyses were also provided since the regression adjustment and propensity score are complementary methods to control the selection bias in observational studies and including both approaches can improve the robustness of the results (17, 18). The adjustment of the treatment effect in regressions analyses was performed on the same set of co-variables included in the propensity score analysis (see below). Propensity score analysis by the Inverse Probability of Treatment Weighting (IPTW)

7

ACCEPTED MANUSCRIPT In a second phase, we implemented the IPTW method using the propensity score in order to control the potential selection bias associated with non-randomization. The propensity score is the conditional probability for a patient to be treated with TomoTherapy®, conditionally to observables characteristics. We determined the

RI PT

probability of receiving TomoTherapy® by fitting a logit model of an indicator variable denoting TomoTherapy® versus RapidArc® on age, gender, performance status, histology, tumor stage, nodal status, initial localization, specific comorbidities, and use of

SC

induction chemotherapy. As most participating centers offered only one of the competing techniques, it was not feasible to take account in the propensity score model

M AN U

the clustered structure of the data (patients nested in hospitals) (19). The IPTW method balances the covariate of the two groups by weighting all patients of the database by the inverse of their propensity score (1/(ps)) in the TomoTherapy® group and 1/(1-ps) in the RapidArc® group). The stabilized weighted of the IPW proposed by Robins (20) was

TE D

used in order to reduce the volatility of the weights and preserving the sample size in pseudo-datasets (21). The standardized differences approach allows to assess the

EP

balance between treatment groups before and after weighting. A standardized difference (d) ≥ 20% for a given covariate indicated a strong imbalance. The adjusted

AC C

Kaplan Meier Estimator (AKME), proposed by Xie and Liu was applied in order to compare survival distributions (22). In order to test for a significant difference in survival curves for the two groups, we used the Cox test of equality instead of a log-rank test, to take into account the fact that patients treated with TomoTherapy® versus RapidArc® are no longer independent after weighting using the IPTW (23). Heterogeneity of treatment effects was explored by subgroups analysis when appropriate.

8

ACCEPTED MANUSCRIPT Results Patients and treatment characteristics in the unweighted population From February 2010 to February 2012, 180 patients were accrued instead of the

RI PT

planned 300 due to recruitment difficulties. Six patients included in the initial cohort were not treated on the study, as detailed in the previous publication (6). Seventy-four patients were treated with TomoTherapy®, 92 with RapidArc® and 8 with Elekta. The

SC

small sample of patients treated with SmartArc® was not generalizable and therefore excluded from the analysis. Patient and tumor characteristics in the unweighted

M AN U

population are presented in Table 1. Histology, ECOG performance status and nodal status appeared strongly unbalanced between treatment groups (d ≥20%). Percentages of patients who received concomitant chemotherapy were not significantly different between groups (73.6% with TomoTherapy® and 70.3% with RapidArc®, d=-13%.).

TE D

Comparative characteristics of concomitant chemotherapy and dosimetric data are provided in the supplementary tables. In the crude analysis prior to IPTW adjustment, the median follow-up was 20.6 months (range, 0.9-41.1 months). There were 8

EP

premature discontinuations of treatment, 7 patients treated with TomoTherapy® and 1

AC C

patient treated with RapidArc®.

Recurrence and survival Unweighted analysis. A loco-regional relapse was observed in 15 (20.3%) of the patients treated with TomoTherapy® and in 29 (31.5%) of the patients treated with RapidArc®. A distant relapse occurred in 8 (10.8%) and 11 (12.0%) of the patient respectively. Fourteen patients (18.9%) died in the TomoTherapy® group (of which 4 9

ACCEPTED MANUSCRIPT (5.2%) from their disease) and 16 patients (17.4%) in the RapidArc® group (of which 14 patients (15.5%) from their disease). Loco-regional control rate and cancer specific survival rate were significantly better in the TomoTherapy® group than in the RapidArc® group (p=0.033 and p=0.029, respectively). PFS rates and overall survival

RI PT

(OS) rate were not significantly different between the TomoTherapy® group and RapidArc® group (p=0.813 and p=0.809, respectively). Multivariate regression analysis indicates a significant treatment effect on loco-regional relapse (p=0.018), with no effect

SC

of neo-adjuvant chemotherapy (p=0.897) (Table 2). No effect of TomoTherapy® or neoadjuvant chemotherapy were found between treatment groups for specific (p=0.0546

M AN U

and p=0.558, respectively), PFS (p=0.435 and p=0.073, respectively) or overall survival (p=0.970 and p=0.147, respectively). Weighted analysis. After IPTW adjustment, all standardized difference of weighted comparisons were <5%, which indicated that the distribution of baseline patients and tumor characteristics was similar between

TE D

treatment group (Table 1). Loco-regional control rate at 18 months were 83.3% (95% CI: [72.5%; 90.2%]) in the TomoTherapy® group and 72.7% (95% CI: ([62.1%; 80.8%])

EP

in the RapidArc® group, respectively (p=0.024) (Fig. 1a). A subgroup analysis suggested a treatment effect in terms of loco-regional control larger for T3 and N2 or higher

AC C

tumors (Fig 2) (HR=0.37 95% CI: [0.16; 0.89]). Cancer specific survival rate were 97.2% (95% CI: [89.3%; 99.3%]) and 85.5% (95% CI: [75.8%; 91.5%], respectively (p=0.014) (Fig. 1b). PFS rates were 68.0% (95%CI: (54.9%; 78.0%) and 65.7% (95%CI: (53.8%; 75.3%) respectively (p=0.4835) (Fig. 1c). Overall survival (OS) rates were 85.5% (95% CI: (74.0%; 92.2%) and 84.3% (95% CI: (73.5%; 90.9%) respectively (p=0.742) (Fig. 1d).

10

ACCEPTED MANUSCRIPT Acute toxicity Unweighted analysis. In the unweighted population, eighty eight patients (53.0%) reported acute toxicities (grade ≥ 3), 46 (62.2%) in the TomoTherapy® group and 42

RI PT

(45.6%) in the RapidArc® group (p=0.034). Cutaneous toxicities were most frequent in TomoTherapy® (p=0.001). Nineteen patients had a grade 3 and above xerostomia (5 in the TomoTherapy® group and 14 in the RapidArc® group (p=0.089). Weighted

SC

analysis After IPTW adjustment, mycosis and salivary disorders reached the 5% significant level. Mycosis were significantly most frequent in TomoTherapy® group

M AN U

(p=0.035) and salivary disorders were significantly most frequent in RapidArc® group (p=0.012). Detailed results are reported in table 3.

Long-term toxicity

TE D

Unweighted analysis. Twenty-nine patients (17.5%) reported at least one long-term grade 2 or above toxicity, 13 (17.6%) in the TomoTherapy® group and 16 (17.4%) in

EP

the RapidArc® group (p=0.976) and the occurrence of the various long-term toxicities were no significantly different between groups. The most common toxicity was salivary

AC C

disorders (15 patients). Weighted analysis. Results on long-term toxicities were unchanged compared to unweighted analysis. Detailed results are shown in table 4.

Post-treatment salivary function evaluation Unweighted analysis. Salivary function degradation after treatment, as graded with the modified RTOG scale, was worse in the RapidArc® group (p=0.013). Mouth dryness assessed with a Visual Analogic Scale (VAS) also found that patients from the RapidArc® 11

ACCEPTED MANUSCRIPT group had a worse salivary function degradation (median maximal VAS=6, min=0; max=10) compared to patients in the TomoTherapy® group (median maximal VAS=4, min=0; max=9; p=0.0142). Minimum salivary weight after treatment was 440mg/min in median in the TomoTherapy® group and 200 mg/min in median in the RapidArc®

RI PT

group, which was not significantly different (p=0.1892). Weighted analysis. After IPTW adjustment, the conclusions were similar compared to the unweighted analysis (p=0.012

SC

with the RTOG scale, p<0.001 with VAS, and p=0.602 for salivary weight).

M AN U

Discussion

Comparing different radiation delivery techniques can be challenging. Many radiation oncologists argue that IMRT has not proven its clinical superiority on conventional radiotherapy in most of the indications we use it for today (24). Fortunately, head and

TE D

neck cancers are not one of them. Since the publication of the phase III PARSPORT trial, comparing IMRT to conventional radiotherapy, it is now demonstrated that sparing the

EP

parotid glands with IMRT significantly reduced the incidence of xerostomia, leading to recovery of saliva secretion and better quality of life. But IMRT can be delivered with

AC C

different techniques, Helical Tomotherapy® (Accuray) and VMAT (Elekta SmartArc® or Varian RapidArc®) being the most common. We have previously shown that, beyond different treatment planning requirements, these techniques had different costs (6, 25, 26) and that their dosimetric output was also significantly different (4, 5). What remained to assess was whether these differences actually translated into any meaningful clinical impact. To this day, no trial has ever prospectively compared the outcome of Tomotherapy and VMAT in any treatment localization.

12

ACCEPTED MANUSCRIPT The ART-ORL trial was designed as a multicenter prospective trial assessing the cost and efficacy of three IMRT techniques. Patients received Tomotherapy®, RapidArc® or SmartArc®, depending on the technique available in the center treating them. Only 8 patients were accrued in the SmartArc® group, which was removed from the analysis.

RI PT

Since randomization was not performed, we controlled the potential selection bias using the Inverse Probability Treatment Weighting (IPTW) method. We showed that locoregional control rate at 18 months was significantly better in the Tomotherapy® group.

SC

This result was confirmed by standard multivariate regression analysis. This difference could be explained by the better dosimetric homogeneity of Tomotherapy®, as we

M AN U

showed in the dosimetric analysis of this trial (5). However, no statistical difference was observed for progression-free (PFS) and overall survival (OS). But since 90% of patients with solid tumors are actually dying from metastatic disease (27), the correlation of loco-regional control and overall survival is still controversial (28), meaning that this

TE D

results were not unexpected. Tomotherapy® and RapidArc® provided the same low long-term toxicity, as no difference was shown for toxicities of grade 2 and above

EP

between the two groups. However, a significant difference was found in occurrence of acute salivary disorders grade≥3 in disfavor of RapidArc®. Moreover, Tomotherapy®

AC C

spared post-treatment salivary function more efficiently than RapidArc®, according to the evaluations performed with the RTOG scale. Mouth dryness assessed by the patients with a Visual Analogic Scale was also worse in the RapidArc® group. In the short-time horizon of the cost analysis (restricted to the radiotherapy course itself, due to budget constraints), the number of adverse events grade ≥3 occurring during RT was not significantly different between strategies. However, the Tomotherapy® appeared to be more expensive than RapidArc® (6). At this stage, a long-term analysis that includes

13

ACCEPTED MANUSCRIPT modeled data is needed to provide policy makers with full information to determine which strategy is cost-effective over a patient’s life time. We acknowledge that the main limitation of this trial was that it was not randomized.

RI PT

This study is the first to our knowledge to demonstrate the relative importance of the radiation delivery methods we use for the patients in terms of dosimetry, cost and clinical analysis (5, 6), even within the field of IMRT. High quality evidence for many of

SC

our treatment techniques are still very much lacking and with new techniques constantly becoming available, it should not be considered as trivial to create trials

M AN U

comparing novel delivery techniques, for example in the field of stereotactic body radiation therapy. Creating such randomized phase III trials will probably prove difficult, or even impossible. More now than ever, new ways to generate high level evidence are needed: the creation of detailed, high-granularity national or even

Conclusion

EP

beyond its current state.

TE D

international databases could help bolster the value of radiation therapy (29, 30)

AC C

In this current care prospective multicenter trial, head and neck patients treated with Tomotherapy® appear to display better loco-regional control, cancer specific survival and post-treatment salivary function sparing compared to patients treated with RapidArc®. A subgroup analysis suggested a greater treatment effect with Tomotherapy® in terms of loco-regional control larger for T3 and N2 or higher tumors. No difference in progression-free or overall survival was found. The results of this study should be explored in a randomized trial before further conclusion can be provided.

14

ACCEPTED MANUSCRIPT References 1. Marta GN, Silva V, de Andrade Carvalho H, et al. Intensity-modulated radiation therapy for head and neck cancer: systematic review and meta-analysis. Radiother. Oncol. J. Eur.

RI PT

Soc. Ther. Radiol. Oncol. 2014;110:9–15. 2. Chao KS, Majhail N, Huang CJ, et al. Intensity-modulated radiation therapy reduces late salivary toxicity without compromising tumor control in patients with oropharyngeal

M AN U

Ther. Radiol. Oncol. 2001;61:275–280.

SC

carcinoma: a comparison with conventional techniques. Radiother. Oncol. J. Eur. Soc.

3. Nutting CM, Morden JP, Harrington KJ, et al. Parotid-sparing intensity modulated versus conventional radiotherapy in head and neck cancer (PARSPORT): a phase 3 multicentre randomised controlled trial. Lancet Oncol. 2011;12:127–136.

TE D

4. Servagi Vernat S, Ali D, Puyraveau M, et al. Is IMAT the ultimate evolution of conformal radiotherapy? Dosimetric comparison of helical tomotherapy and volumetric modulated arc therapy for oropharyngeal cancer in a planning study. Phys. Medica PM Int. J. Devoted

EP

Appl. Phys. Med. Biol. Off. J. Ital. Assoc. Biomed. Phys. AIFB. 2014;30:280–285.

AC C

5. Vernat SS, Ali D, Messina C, et al. Intensity modulated arc therapy in bilaterally irradiated head and neck cancer: a comparative and prospective multicenter planning study. Cancer Invest. 2014;32:159–167. 6. Perrier L, Morelle M, Pommier, et al. Cost analysis of complex radiation therapy for patients with head and neck cancer. Int J of Radiation Oncol Biol Phys. 2016; 95: 654662. 7. Giraud P, Kantor G, Yassa M, et al. Two-year clinical experience with tomotherapy: the

15

ACCEPTED MANUSCRIPT French national cancer institute project on implementing new technology. Cancer Invest. 2011;29:557–563. 8. Anon. Protocole Advanced RadioTherapy ORL - Full Text View - ClinicalTrials.gov.

RI PT

9. Bibault J-E, Morelle M, Perrier L, et al. Toxicity and efficacy of cetuximab associated with several modalities of IMRT for locally advanced head and neck cancer. Cancer Radiother. J. Soc. Francaise Radiother. Oncol. 2016;20:357–361.

SC

10. Cox JD, Stetz J, Pajak TF. Toxicity criteria of the radiation therapy oncology group

M AN U

(RTOG) and the European organization for research and treatment of cancer (EORTC). Int. J. Radiat. Oncol. Biol. Phys. 1995;31:1341–1346.

11. Kaplan EL, Meier P. Nonparametric Estimation from Incomplete Observations. J. Am. Stat. Assoc. 1958;53:457–481.

TE D

12. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. analysis and examples. Br. J. Cancer.

EP

1977;35:1–39.

13. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time.

AC C

Control. Clin. Trials. 1996;17:343–346. 14. Gooley TA, Leisenring W, Crowley J, et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat. Med. 1999;18:695–706. 15. Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. John Wiley & Sons; 2011. 16. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a 16

ACCEPTED MANUSCRIPT competing risk. Ann. Stat. 1988:1141–1154. 17. Stuart E. Matching methods for causal inference: A review and a look forward. Stat Sci: 2010. 1-21.

RI PT

18. Stürmer, Joshi M, Glynn R, Avorn J, Rothman K, Schneeweiss S. A review of the application of propensity score methods yielded increasing use, advantages in specific settings, but not substantially different estimates compared with conventional

SC

multivariate methods. Journal of Clinical Epidemiology (2006) 437-447.

Stat. Med. 30: 32 (19): 3373-3387.

M AN U

19. Li F, Zaslavsky AM, Landrum MB. Propensity score weighting with multilevel data.

20. Robins JM. 1997 Proceedings of the American Statistical Association, Section on Bayesian Statistical Science. 1998.

TE D

21. Xu S, Ross C, Raebel MA, et al. Use of stabilized inverse propensity scores as weights to directly estimate relative risk and its confidence intervals. Value Health J. Int. Soc.

EP

Pharmacoeconomics Outcomes Res. 2010;13:273–277. 22. Xie J, Liu C. Adjusted Kaplan-Meier estimator and log-rank test with inverse

AC C

probability of treatment weighting for survival data. Stat. Med. 2005;24:3089–3110. 23. Cole SR, Hernan MA. Adjusted survival curves with the inverse probability weights. Comput Methods Programs Biomed. 2004; 75:45-49 24. Intensity Modulated Radiation Therapy Collaborative Working Group. Intensitymodulated radiotherapy: current status and issues of interest. Int. J. Radiat. Oncol. Biol. Phys. 2001;51:880–914. 25. Perrier L, Morelle M, Dussart S, et al. In Reply to Fodor and Di Muzio. Int. J. Radiat. 17

ACCEPTED MANUSCRIPT Oncol. Biol. Phys. 2016;96:1124–1126. 26. Perrier L, Morelle M, Huguet, Marius, et al. In Reply to Escande et al. Int. J. Radiat. Oncol. Biol. Phys. forthcoming.

RI PT

27. Mehlen P, Puisieux A. Metastasis: a question of life or death. Nat. Rev. Cancer. 2006;6:449–458.

28. Chargari C, Soria J-C, Deutsch E. Controversies and challenges regarding the impact

SC

of radiation therapy on survival. Ann. Oncol. Off. J. Eur. Soc. Med. Oncol. 2013;24:38–46.

M AN U

29. Chen RC, Gabriel PE, Kavanagh BD, et al. How will big data impact clinical decision making and precision medicine in radiation therapy. Int. J. Radiat. Oncol. 2015. 30. Bibault J-E, Giraud P, Burgun A. Big Data and machine learning in radiation oncology:

AC C

EP

TE D

State of the art and future prospects. Cancer Lett. 2016.

18

ACCEPTED MANUSCRIPT Figures

Figure 1. Figure 1. Loco-regional relapse-free (a), cancer specific (b), survival progression-free (c) and overall survival (d) after weighting with the IPTW. - Note: Event-free survival probabilities were reported as (1-[cumulative incidence probability]).

RI PT

Figure 2. Subgroup analysis for T3 or N2 or larger tumors (HR=0.37 95% CI: [0.16;

AC C

EP

TE D

M AN U

SC

0.89]).

19

ACCEPTED MANUSCRIPT Table 1. Patients characteristics

Before weighting by IPTW TomoTherap RapidArc® d y® (n=74) (n=92)

After weighting by IPTW TomoTherap RapidArc d y® (pseudo ® data) (pseudo data)

AC C

EP

TE D

M AN U

SC

RI PT

Age ≤55 years
 
 36 (48.6%) 40 (43.5%) 10% 48.57% 48.12% 0,9% >55 years 38 (51.4%) 52 (56.5%) 51.43% 51.88% Sex Female 15 (20.3%) 18 (19.6%) 2% 20.56% 21.48% -2,3% Male 59 (79.7%) 74 (80.4%) 79.44% 78.52% Performance Status 0 40 (54.8%) 69 (75.0%) -43% 64.24% 65.06% -1,7% Performance Status 1, 2 33 (45.2%) 23 (25.0%) 35.76% 34.94% Histology Epidermoid 77.66% 77.80% 0,3% carcinoma 51 (68.9%) 79 (85.9%) 42% Other 23 (31.1%) 13 (14.1%) 22.34% 22.20% Tumor stage T1, T2 24 (32.4%) 38 (41.3%) -19% 38.49% 38.24% 0,5% T3, T4 50 (67.6%) 54 (58.7%) 61.51% 61.76% Nodal status N0, N1 28 (37.8%) 44 (47,8%) -20% 42.92% 44.85% -3,9% N2, N3 46 (62.2%) 48 (52.2%) 57.08% 55.15% Localization Nasopharynx 19 (25.7%) 19 (20.7%) 12% 26.42% 26.77% -0,8% Oropharynx or oral 55 (74.3%) 73 (79.3%) 73.58% 73.23% cavity Specific 7 (9.5%) 13 (14.1%) 14% 10.53% 11.97% 4,6% comorbidities No comorbidities 67 (90.5% 79 (85.9%) 89.47% 88.03% Use of induction chemotherapy Yes 29 (39.2%) 42 (45.7%) -13% 41.98% 43.30% -2,7% No 45 (60.8%) 50 (54.3%) 58.02% 56.70% d= absolute standardized difference. The sizes of the pseudo-datasets are 90.37 in the Tomotherapy® group and 92.57 in the RapidArc® group.

ACCEPTED MANUSCRIPT Table 2. Multivariate analysis of loco-regional relapse-free survival

M AN U TE D EP AC C

P 0.592 0.741 0.003 0.044 0.561 0.491 0.125 0.154 0.897 0.018

RI PT

Hazard ratio 0.828 1.142 2.914 3.424 1.263 1.259 2.298 1.817 0.959 0.428

SC

Factor Age >55 years (ref : ≤ 55 years) Sex Male (ref : female) Performance status 1 or 2 (ref : 0) Histology Epidermoid carcinoma (ref : other) Tumor stage T3, T4 (ref : T1, T2) Nodal status N2, N3 (ref : No,N1) Localization Oropharynx or oral cavity (ref : Nasopharynx) Specific comorbidities (ref: No) Use of induction chemotherapy (ref : No) Use of TomoTherapy® (ref : RapidArc®)

ACCEPTED MANUSCRIPT

Table 3. Acute toxicities (grade ≥ 3)

TomoThe rapy® N=74 10 (13.5%) 20 (27%)

Mucositis Mycosis Hematologic disorders General disorders

9 (9.8%)

.891

30.1%

26.9%

.637

3 (4.1%) 4 (5.4%)

24 (26.1%) 0 (0%) 6 (6.5%)

.051 .764

4.7% 5. 8%

0% 5.2%

.035 .867

0 (0%)

3 (3.3%)

.117

0%

3.1%

.094

0 (0%)

2 (2.2%)

.503

0%

2.2%

.202

4 (5.4%) 0 (0%)

4 (4.3%) 2 (2.2%)

1.000 .503

5.5% 0%

4.2% 2.0%

.666 .173

0 (0%) 19 (25.7%) 5 (6.8%)

1 (1.1%) 7 (7.6%)

1.000 .001

0% 27.9%

0.9% 8.6%

.365 .007

TE D

Neurological disorders Digestive disorders Infection, infectious syndrome Renal toxicity Cutaneous disorder Salivary disorders

M AN U

(a)

SC

Pain

RapidArc ®N=92

After weighting using the IPTW(c) p-value TomoT RapidArc p-value herapy ® ® .453 14.7% 10.6% .396

RI PT

Before weighting

AC C

EP

14 .089 7.5% 20.4% .012 (15.2%) 0 (0%) 0 (0%) NA 0% 0% NA Fibrosis 0 (0%) 0 (0%) NA 0% 0% NA Cervical edema (b) 1 (1.4%) 2 (2.2%) .692 0.8% 2.1% .484 Other 46 42 .034 62.9% 50.0% .078 Overall (62.2%) (45.6%) (a) asthenia, anorexia ; (b) septic shock, trismus, bleeding of the gastrointestinal tract. (c) the sizes of the pseudo-datasets are 90.37 in the Tomotherapy® group and 92.57 in the RapidArc® group. Abbreviations: NA: Non-applicable

ACCEPTED MANUSCRIPT Table 4. Long-term toxicities (>grade 2) Before IPTW TomoTherapy® RapidArc® N=74 N=92

pvalue*

After IPTW (f) TomoTherapy® RapidArc® p-value

AC C

EP

TE D

M AN U

SC

RI PT

Pain(a) 1 (1.4%) 0 (0%) .263 (1.3%) (0%) .267 Mycosis(b) 1 (1.4%) 0 (0%) .263 (1.8%) (0%) .201 (c) General disorders 1 (1.4%) 1 (1.1%) .877 (0.7%) (0.9%) .861 Neurological 0 (0%) 1 (1.1%) .368 (0%) (0.7%) .423 disorders (d) Cervical edema 0 (0%) 3 (3.3%) .117 (0%) (2.8%) .108 Infection, infectious 0 (0%) 1 (1.1%) .368 (0%) (1.2%) .290 syndrome Fibrosis 0 (0%) 2 (2.2%) .202 (0%) (1.9%) .168 Cutaneous disorder 1 (1.4%) 1 (1.1%)0 .877 (1.8%) (0.8%) .536 Salivary disorders 7 (9.5%) 8 (8.7%) .865 (11.3%) (7.6%) .389 (e) Other 2 (2.7%) 5 (5.4%) .384 (2.3%) (4.0%) .518 Overall 13 (17.6%) 16 (17.4%) .976 (19.3%) (15.3%) .479 (a) dysphagia ; (b) oral candidiasis ; (c) weight loss; (d) hearing impairment ; (e) alopecia, trismus, hyperthyroidism, cramp, groin pain. (f) the sizes of the pseudo-datasets are 90.37 in the Tomotherapy group and 92.57 in the RapidArc group.

ACCEPTED MANUSCRIPT

1,0 0,9

RI PT

0,8

SC M AN U

0,6

TE D

0,5 0,4

RapidArc

0,2

Tomotherapy

0,1

EP

0,3

AC C

Loco-regional free relapse

0,7

0,0 0

6

12

18

24 Months

30

36

42

48

ACCEPTED MANUSCRIPT

1,0 0,9

RI PT SC

0,7

M AN U

0,6 0,5

TE D

0,4

EP

0,3 0,2

AC C

Cancer specific survival

0,8

RapidArc Tomotherapy

0,1 0,0 0

6

12

18

24 Months

30

36

42

48

ACCEPTED MANUSCRIPT

1,0

RI PT

0,8

SC

0,7

M AN U

0,6 0,5

TE D

0,4

EP

0,3 0,2 0,1

RapidArc

AC C

Progression-free survival

0,9

0,0 0

6

12

18

24 Months

30

36

42

48

ACCEPTED MANUSCRIPT

1,0 0,9

RI PT SC

0,7

M AN U

0,6 0,5

TE D

0,4

EP

0,3 0,2 RapidArc

0,1

AC C

Overall survival

0,8

Tomotherapy

0,0 0

6

12

18

24 Months

30

36

42

48

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT