Moderately Hypofractionated Intensity Modulated Radiotherapy (IMRT) With a Simultaneous Integrated Boost for Locally Advanced Head and Cancer – Do Modern Techniques Held Their Promise?


 PurposeIntensity-modulated-radiotherapy (IMRT) is still a standard of care for radiotherapy in locally advanced head and neck cancer (LA-HNSCC). Simultaneous-integrated-boost (SIB) and moderately hypofractionation offer an opportunity of individual dose painting and reduction of overall treatment time. We present retrospective data on toxicity and local-regional-control of a patient cohort with LA-HNSCC treated with an IMRT-SIB-concept in comparison to normofractionated 3D-conformal radiotherapy (3D-RT) after a long-term follow-up.MethodsBetween 2012 and 2014, n=67 patients with HNSCC (stages III/IV without distant metastases) were treated with IMRT-SIB either definitive (single/total doses: 2.2/66Gy, 2.08/62.4Gy, 1.8/54Gy in 30 fractions) or in the postoperative setting (2.08/62.4Gy, 1.92/57.6Gy, 1.8/54Gy). These patients' clinical course was matched (for gender, primary, and treatment concept) as part of a matched-pair-analysis with patients treated before mid-2012 with normofractionated 3D-CRT (definitive: 2Gy/50Gy followed by a sequential boost up to 70Gy; postoperative: 2Gy/60-64Gy). Chemotherapy/immunotherapy was given concomitantly in both groups in the definitive situation (postoperative dependent on risk factors). Primary endpoints were acute and late toxicity; secondary endpoint was loco-regional-control (LRC).Results67 patients treated with IMRT-SIB (n = 20 definitive, n = 47 adjuvant) were matched with 67 patients treated with 3D-RT. There were minor imbalances between the groups concerning non-matching-variables like extracapsular extension (ECE) and chemotherapy in IMRT-SIB.Significantly less toxicity was found in favor of IMRT-SIB concerning dysphagia, radiation dermatitis, xerostomia, fibrosis, and lymphoedema. After a median follow-up of 63 months, median LRC was not reached (IMRT-SIB) vs. 69.5m (3D-RT) (p=0.63).ConclusionThis moderately hypofractionated IMRT-SIB-concept showed to be feasible with less toxicity compared to conventional 3D-RT in this long-term follow-up observation.


Introduction
In locally advanced head and neck cancer (LA-HNSCC), radiotherapy (RT) is an essential element in curative treatment strategies, either in the de nitive (in case of inoperability or to avoid mutilating surgery in oropharyngeal cancers) or in the postoperative-adjuvant situation [1].Thereby, concomitant -usually Cisplatin-based -chemotherapy improves the prognosis in the de nitive [2], as well as in the postoperative-adjuvant situation, in particular in case of extracapsular spread of nodes (ECE) and/or microscopically involved resection margins [3].
Moderately hypofractionated RT becomes more common in clinical practice.It is already the standard of care in postoperative-adjuvant RT of breast cancer after breast-conserving surgery [4] and is a guidelinebased alternative in RT for prostate cancer [5,6].Similarly, it has also already been used in HNSCC [7].
Nowadays, intensity modulated radiotherapy (IMRT) is used as the standard of care in RT for HNSCC to lower the risk of high-grade chronic toxicity [8].The implementation of IMRT also offers the possibility of simultaneous integrated boost (SIB) radiation for individual dose painting and -using moderately hypofractionated concepts -for reduction of overall treatment time [9].
Here, we present a cohort of patients with locally advanced HNSCC in the curative setting treated with RT or radiochemotherapy (RCT), either de nitive or postoperative-adjuvant, using an at our institution implemented moderately hypofractionated IMRT-SIB concept.Data on toxicity and loco-regional-control are reported and compared to a historical cohort of patients treated with normofractionated 3Dconformal radiotherapy (3D-RT) before the IMRT-era using a matched-pair analysis.

Patients And Methods
From 2012 to 2014, the analyzed 67 patients with locally advanced (stage III / IV without distant metastases; mostly squamous cell carcinomas) cancers were treated at our department with RT in curative intent with the following moderately hypofractionated IMRT-SIB concepts.For de nitive RCT we used a slightly modi ed fractionation as proposed by RTOG 0022: 66 Gy (daily dose 2.2 Gy) for gross tumor volumes, 62.4 Gy (2.08 Gy) for elective cervical nodes considered to be at exceptionally high risk for subclinical disease, and 54 Gy (1.8 Gy) for elective cervical nodes.In case of postoperative-adjuvant treatment, 62.4 Gy (daily dose 2.08 Gy) for the primary tumor region and cervical regions with involved nodes with ECE, 57.6 Gy (1.92 Gy) for cervical regions with involved nodes without ECE, and 54 Gy (1.8 Gy) for elective cervical nodes were given.All patients were immobilized with a thermoplastic mask, including the head and the neck and shoulder regions.A planning CT scan with a minimum slice thickness of 3mm was obtained in all patients.On each CT slice, the gross tumor volumes (GTV) were delineated by the treating physician, as well as the areas at especially high risk of potential microscopic disease and other potentially affected regions including lymph nodes (CTV).The margins to compensate for set up variability and organ motions were mostly 5mm.Furthermore, organs at risk (OAR) like the parotids, spinal cord, brachial plexus, and brainstem were drawn.All calculations were done with Monaco (Electa) by experienced physicists.A phantom measurement with the PTW OCTAVIUS® (4D) Phantom (PTW Freiburg GmbH -Germany) and the corresponding PTW VeriSoft ® in the latest available version was performed to verify each plan.Planning objectives like dose prescriptions and normal tissue constraints had to be realized according to ICRU-83 [10] and to QUANTEC-data (spinal cord Dmax < 45 Gy, brachial plexus Dmax < 54 Gy, contralateral parotid gland Dmean < 23 Gy) [11].To reach these objectives/constraints, the PTV coverage was modi ed if necessary.For RT, linear accelerators with 6MV photon energy were used.
For comparison, another cohort was additionally analyzed using a matched-pair analysis.The 67 patients of this cohort were treated before the IMRT-era at our institution from 2008 to 2012 with normofractionated 3D-RT as follows: Primary and involved as well as elective cervical nodes up to a dose of 50 Gy followed by a sequential boost to the primary and involved nodes (or regions with ECE in case of adjuvant therapy) to a total dose of 70 (de nitive) or 60 to 64 Gy (postoperative-adjuvant).Planning was done by multi-eld 3D conformal forward planning using 6, 10, and 15MV photon beams and a "shrinking eld approach".Dose prescriptions had to be realized according to ICRU-50.To avoid long-term toxicity, the supraclavicular lymph nodes were mostly spared at 46 to 54 Gy and the spinal cord at 45 Gy.
In both cohorts, concomitant chemotherapy was given regularly in the de nitive situation, in the postoperative-adjuvant setting in case of ECE or microscopically involved resection margins.
All patients were clinically assessed weekly during RT and three months after that by experienced staff to evaluate and grade acute toxicity (oral mucositis, dysphagia, radiation dermatitis, and nadir of hemoglobin levels, white blood cell count, and platelet count) according to CTCAE v4.03 [12].Afterward, patients were asked to show up for yearly follow-up visits to score late toxicities according to LENT/SOMA (xerostomia, taste alteration, brosis, lymphedema, hoarseness, stula, necrosis of mandible, and trismus) [13] in order to generate a long-term follow-up observation, since late toxicity partly occurs some years later.

Statistics
For this retrospective matched-pair analysis, about 200 consecutive patients who received and completed 3D-RT in curative intent for HNSCC between 01.2008 and 05.2012 were screened as controls.Furthermore, 67 consecutive patients who received hypofractionated IMRT-SIB between 06.2012 and 04.2014 were documented.To select 3D patients as controls, three variables had to match between both groups: gender, site of the primary tumor, and treatment concept (de nitive vs. postoperative).
Retrospectively demographical, histopathological, clinical, and toxicity data were collected from the charts.Staging was done according to the 7th version of the TNM-manual.

After data collection, for comparison between the patients' characteristics in both cohorts the McNemar-
Test for binary characteristics or Cohens Kappa for characteristics with more levels were used.Locoregional-control (LRC) was analyzed using the Kaplan-Meier-method [14].Differences in toxicity were tested for statistical signi cance with the Wilcoxon signed-rank test.We used SPSS Version 26 to do the statistical analyses.Signi cance was de ned as p < 0.05.

Results
Overall, in each group 67 patients were analyzed (20 de nitive, 47 postoperative in each cohort).Median follow-up for all patients was 63 months.The essential patients' characteristics are summarized in Table 1.They were su ciently balanced between the groups.However, there were signi cantly more patients with ECE (27% vs. 8%) and with concomitant chemotherapy (70% vs. 49%) treated in the IMRT-SIB cohort, as well as six vs. three patients were suffering from cN2c disease.
Acute and chronic toxicity are shown in table 2. There were no signi cant differences in acute oral mucositis incidences, although there were slightly more patients with at least grade 3 mucositis with 3D-RT (48% vs. 40%).However, a statistically signi cant difference was documented with lower toxicity in the IMRT-SIB group for overall dysphagia (p = 0.044) and radiation dermatitis (p = 0.002). Concerning

Discussion
With the advent of IMRT in RT for HNSCC as a standard of care, diverse individual concepts have been implemented in different radiotherapy departments.Thereby, the application of a SIB and moderately hypofractionation are often applied to ensure individual dose painting and reduction of overall treatment time, which is crucial in RT for HNSCC [15].
To our knowledge, there is only little prospectively randomized evidence to evaluate the e cacy of IMRT in comparison to 3D-RT in RT of HNSCC.One study (PASPORT) randomized n = 94 patients between IMRT and 3D-RT (with parallel opposed lateral elds) [16].This study was focused on avoiding xerostomia.At 12 and 24 months, xerostomia at least grade 2 was signi cantly less prevalent after IMRT.
Other late toxicities and loco-regional-control, or overall survival did not differ between both groups.Comparable results were obtained in a small randomized trial (n = 60 patients) from India [17].
In nasopharyngeal cancer, this advanced technique has demonstrated a higher oncological e cacy in n = 616 patients compared to outdated 2D planning techniques: OS and progression-free survival were signi cantly improved.At the same time, high-grade chronic toxicity was reduced [18].
The recently published GORTEC 2004-01 randomized phase III trial showed again that the IMRT technique can even reach a dose escalation with markedly decreased late xerostomia, but without a signi cant improvement of local tumor control [31].The authors used a slightly different irradiation concept with a sequential moderate hypofractionated boost to 75 Gy overall dose (25Gy/10F boost dose) in a total of 35 fractions.The frequency of grade ≥ 2 xerostomia was around 2/3 lower after 1 und 3 years in the IMRT group.
At last, Gupta et al. showed repeatedly in a prospective randomized trial with a very long follow-up and enough power and sample size that the moderate hypofractionation with 66 Gy in 30 fraction and IMRT technique leads to a meaningful reduction in severe xerostomia and brosis with comparable locoregional control and overall survivial in the 3D control group [32].
Other observational studies support the assessment of lower acute and chronic toxicity by IMRT in RT for HNSCC in comparison to 3D planning.In this context, Jirkovska et al. demonstrated that acute toxicity and xerostomia were signi cantly reduced in HNSCC treated by IMRT [19].Modesto et al. showed similar data, especially for severe late toxicities like xerostomia, dysphagia, or feeding-tube dependency [20].Our data con rm these ndings showing lower toxicity in the IMRT-SIB group for dysphagia, dermatitis, xerostomia, brosis, and edema.
Other retrospective studies also showed an advantage for IMRT concerning prognosis in LRC [21] or even OS [22].In contrast, in our patients LRC was equal between both groups.However, OS differed to the disadvantage of IMRT-SIB.This nding is most likely explained by more aggressive tumors in IMRT-SIB, as more patients had ECE and had to receive concomitant chemotherapy.Furthermore, despite the matched pair analysis, other biases due to the study's retrospective nature might play a role.However, such a nding is not totally in con ict with the literature.A recent meta-analysis on IMRT versus 3D-CRT in RT for HNSCC con rmed the superiority of IMRT in terms of toxicity (mainly xerostomia) but did not nd improved oncological outcomes.The authors concluded that a positive impact of IMRT on tumor control and survival mains to be proven [8].
As concomitant chemotherapy in RT for locally advanced HNSCC is crucial for prognosis in the de nitive and speci c postoperative-adjuvant situations, RT approaches have to be designed so that the dosefractionation concepts do not compromise the use of concurrent systemic therapy and vice versa.Therefore, we chose a chemotherapy protocol with weekly low dose cisplatin (40 mg/m² BSA up to at least cumulative ≥ 200 mg/² BSA, see above) as a continuous radiosensitizer and decided against a higher hypofractionated RT schedule > 2.2 Gy in the volume.Such a low dose weekly cisplatin application is an established regimen [23] besides likewise often used high dose application -for example 100 mg/m² BSA twice or thrice during radiotherapy [3].A cumulative cisplatin dose of approximately 200 mg/m 2 BSA, independent of the schedule, might be su cient to yield a bene cial antitumor effect [24].However, prospective studies in adequately sized phase III trials on this subject are still pending [25].We saw good tolerance and feasibility of our approach with moderately hypofractionation without compromising one part of the combined treatment.Other studies actually show that higher hypofractionation (with single doses in SIB volumes up to 2.25 or 2.4 Gy) combined with chemotherapy seems to be possible [26,27].
IMRT-SIB RT concepts will further be modi ed according to human papillomavirus (HPV) status in locally advanced HNSCC.On the one hand, HPV positive tumors have a better prognosis and are possible candidates for dose reduction, which is the subject of several ongoing clinical trials [28].On the other hand, other studies examine the feasibility of dose-escalated hypofractionated chemoradiation in HPV-negative cancer [29].Unfortunately, in our retrospective patient cohorts, HPV-status was not available for most of the tumors.
In summary, the presented moderately hypofractionated IMRT-SIB-concept showed to be feasible with an acceptable loco-regional-control and less toxicity compared to conventional 3D-CRT.IMRT is the standard of care in RT for locally advanced HNSCC.The optimal dose-/fractionation concept concerning moderate hypofractionation still has to be de ned.

Declarations
Ethics approval and consent to participate The study protocol was approved by the local ethics committee (number 1795-2013) and all patients declared their written informed consent.The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments.

Table 1 Patients
** Negative Cohens Kappa cannot be interpreted, there for there is no statistical signi cance † The current treatment was due to a relaps. *