Comparative effectiveness of simultaneous integrated boost vs sequential intensity-modulated radiotherapy for oropharyngeal or hypopharyngeal cancer patients: A population-based propensity score-matched analysis.

There were 2 common radiotherapy dose fractionation strategies in head-and-neck cancer patients (such as oropharyngeal cancer [OPC] or hypopharyngeal cancer [HPC]) treated with radiotherapy: intensity-modulated radiotherapy using simultaneous integrated boost (IMRT-SIB) and sequential IMRT (IMRT-SEQ). There is a lack of high-level clinical evidence to compare IMRT-SIB vs IMRT-SEQ specifically for OPC or HPC patients. The present study investigated the survival outcomes of OPC or HPC patients receiving definite concurrent chemoradiotherapy (CCRT) with either IMRT-SIB or IMRT-SEQ via a population-based propensity score (PS)-based analysis.The localized stage OPC or HPC patients diagnosed between 2011 and 2015 were identified based on the Health and Welfare Data Science Center database in Taiwan. These patients received definitive CCRT with either IMRT-SIB or IMRT-SEQ. We constructed a PS-matched cohort (1:1 for IMRT-SIB vs IMRT-SEQ) to balance observable potential confounders. We compared the hazard ratio (HR) of death between IMRT-SIB and IMRT-SEQ during the entire follow-up period. We also evaluated other disease outcome or subgroups.Our study population constituted 200 patients with well balance in observed covariables. The HR of death when IMRT-SIB was compared to IMRT-SEQ was 1.23 (95% confidence interval 0.84-1.80, P = .29). The results were similar for other disease outcome or subgroups.We found the survival outcome might be comparable for those treated with IMRT-SIB vs those treated with IMRT-SEQ.

Introduction

Radiotherapy is an important treatment modality of head-and-neck (HN) cancer. Many radiotherapeutic factors were well studied in the past, including the radiation field, the total dose, and overall treatment time. A systematic review published in 2014 had shown that using intensity-modulated radiotherapy (IMRT) reduced the incidence of grade 2 to 4 xerostomia in patients with HN cancer without compromising locoregional control and overall survival (OS), in comparison with conventional technique or three-dimensional radiotherapy.[ Chemotherapy (especially concurrent chemoradiotherapy [CCRT]) further improves survival in patients with locally advanced HN squamous cell carcinoma (SqCC).[

However, there are several different dose-fractionation strategies to perform HN IMRT.[ Two common strategies are described in this study. The first technique, sequential IMRT (IMRT-SEQ) via conventional dose fractionation, uses usually 180 to 200 cGy to treat all risk regions, and then boost the high-risk region with smaller field thereafter. The other is IMRT with simultaneous integrated boost (IMRT-SIB) via nonconventional dose fractionation, which treats the patients with different dose levels to 2 to 3 planning target volumes at the same time. But the field of radiotherapy is always identical during the whole treatment course.

To compare these 2 techniques, there were several dosimetric studies to compare the dose to the organ at risks when lesions are treated with same coverage.[ However, there is still controversy about which one is better. For other malignancies, the debate over the 2 techniques is also under investigation now.[

A recently (2019) published meta-analysis reported that for HN cancer patients, there was no significant difference in OS when IMRT-SIB was compared to IMRT-SEQ (hazard ratio [HR] 0.94; P = .71).[ Among the several articles included in this meta-analysis, 3 were randomized controlled trials (RCT) and all 3 focused on nasopharyngeal cancer (NPC). All 3 RCTs reported similar outcome between IMRT-SEQ and IMRT-SIB. Therefore, there is high-level evidence that there was no statistical difference in OS when IMRT-SIB was compared to IMRT-SEQ. Similar results were also seen in the low-level evidence in this systematic review. For example, Tao et al[ analyzed 107 NPC patients from a single institute and reported no significant difference in 5-year OS (80.9% vs 80.5%, P = .568).

However, the literatures were less clear and inconclusive regarding oropharyngeal cancer (OPC) or hypopharyngeal cancer (HPC). Among the non-RCTs in the abovementioned systematic review,[ none focused specifically on OPC or HPC, whereas 2 single-institution retrospective studies on HN cancer (roughly half of the patients had OPC or HPC) reported similar disease outcome.[ However, one reported lower rate of acute side effect for IMRT-SEQ,[ whereas the opposite result was reported in the other study.[

Since the clinical benefit of IMRT-SIB vs IMRT-SEQ in definitive CCRT for OPC or HPC is unclear and there is a lack of population-based study, we sought to compare IMRT-SIB vs IMRT-SEQ for locally advanced OPC or HPC patients treated with definitive CCRT via a population-based propensity score (PS)-based analysis.

Materials and methods

Data source

The data were obtained from the Health and Welfare Data Science Center (HWDC) database, including the Taiwan cancer registry (TCR), death registry, and reimbursement data for the whole Taiwan population provided by the Bureau of National Health Insurance (NHI). The TCR was established in 1979. The central cancer registry was reformed in 2002 to include details regarding the stage at diagnosis and the first course of treatment. Additional risk factors such as use of alcohol, betel nuts, or smoking were also available since 2011. All recordings in TCR are made by the professional cancer registrar(s) in each hospital in collaboration with relevant physicians, and are further reviewed by the TCR. The excellent quality (97% completeness) of the TCR data has been confirmed.[ The NHI is a single-payer, compulsory social insurance program that provides insurance coverage to almost all citizens in Taiwan. All of these data were included in the HWDC with personal identifiers removed. This study was approved by the Central Regional Research Ethics Committee, China Medical University, Taichung, Taiwan (CRREC-108-080).

Study population and design

The study flow chart is depicted in Figure 1, as suggested in the STROBE statement.[ The study population consisted of locally advanced (nonmetastatic stage III-IV) OPC or HPC patients (SqCC) who were diagnosed between 2011 and 2015 and received definitive CCRT with external beam radiotherapy. CCRT was defined as concurrent systemic therapy with radiotherapy per recording in TCR. The date of diagnosis in the cancer registry was adopted as the index date and the explanatory variable of interest was treatment with IMRT-SIB or IMRT-SEQ. We classified patients as IMRT-SIB or IMRT-SEQ based on the dose range recommended in practice guideline[ (see footnote in Fig. 1 for definition). Data regarding covariables were also collected to adjust for potential nonrandomized treatment selections (see section 2.3). The survival status of each patient was obtained from the death registry (until December 31, 2017).

Figure 1

. STROBE study flowchart and the number of individuals at each stage of the study. 1We only included those treated (class 1–2) by any single institution without synchronous or metachronous cancer to ensure data consistency. 2Clinical stage III-IV and cM0; by seventh American Joint Committee on Cancer staging. 3International classification of diseases oncology, third edition, site codes C090-C091, C098-C104, C108-C109 for OPC and C129-C132, C138-C139 for HPC. 4Simultaneously integrated boost was defined as highest dose 70 Gy with 2 Gy/fraction (Fx) and lower dose 54 to 63 Gy with 1.6 to 1.8 Gy/Fx as well as same Fx for both highest and lower dose, whereas sequential was defined as highest dose 70 Gy with 2 Gy/Fx and lower dose 44 to 50 Gy with 2 Gy/Fx. 5Without missing information in Taiwan cancer registry and death registry.

Other explanatory covariables

In this study, our covariables included patient demographics (age, gender, residency region]), patient characteristics (comorbidity, smoking, betel nut chewing, drinking status, socioeconomic status [SES]), disease characteristics (OPC vs HPC, clinical T-stage and N-stage, stage), and treatment characteristics (radiotherapy delivery method [image-guided radiotherapy [IGRT] or non-IGRT], radiotherapy break, and neoadjuvant or adjuvant systemic therapy). The selection and definition of these factors were based on our experiences in clinical care and modified from our previous related studies.[ The covariables were defined as follows. Patient residency region was classified as northern Taiwan or elsewhere. Comorbidity was classified as yes or no via Charlson comorbidity score (excluding existing cancer). SES was classified as high (income greater than minimal wage) or not. T-stage was classified as T1-T2 or T3-T4. N-stage was classified as N0-N1 or N2-N3. Overall stage was classified as III or IV. radiotherapy break was defined as more than 1 week or not. IGRT, neoadjuvant or adjuvant systemic therapy, smoking, betel nut chewing, and drinking status were classified as yes or no.

Effectiveness assessment

The survival status at the end of the follow-up period was obtained from TCR and the registry of deaths. This information was used to compare the OS as our primary outcome of interest. We also evaluated incidence of pharyngeal cancer mortality (IPCM) according to TCR and death registry.

Statistical and sensitivity analyses

Statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC) and R (version 3.5.1, R Development Core Team, R Foundation for Statistical Computing, Vienna, Austria). To overcome the inherent limitations of retrospective studies in the potential imbalance of covariables between treatment groups, several analytic approaches were available in the literature,[ including the traditional regression method that relied on specific assumptions about the relationship between the covariables and outcome.[ PS approach was increasingly used in the recent years with modest improvement when compared to traditional Cox regression,[ so we adopted PS approach in the present study. We did not try the other advocated approach (instrumental variable) due to the difficulty in finding a valid instrument in our study setting.[ We used PS-matched method as advocated in the literature.[ We modeled the use of IMRT-SIB (vs IMRT-SEQ) as the dependent variable and the abovementioned covariables as independent variables, and used logistic regression to model the probability of receiving IMRT-SIB. We then used the logit of the probability as the PS, as commonly used in the literature.[ Tabulation and standardized differences were used to assess the balance of the covariables.[ The HR of death between IMRT-SIB vs IMRT-SEQ during the entire follow-up period was compared using Cox model via a robust variance estimator.[ We further used the E-factor as suggested in the recent (2019) literature to evaluate the impact of potential unmeasured confounder(s) on OS.[ The sub-distribution HR from the clustered Fine–Gray model was used to evaluate IPCM.[ We also performed 2 subgroup analyses (SA) for OPC and HPC patients, respectively.

Results

Study population

As shown in Figure 1, 200 eligible patients who received definitive CCRT with external beam radiotherapy between 2011 and 2015 were identified as our primary study population. They were divided into 2 groups: those who received IMRT-SIB (n = 100) and those who received IMRT-SEQ (n = 100). These 2 groups were balanced (standardized differences < 0.25) after matching, although some imbalance existed before matching (Table 1). The median (range) of lower radiotherapy dose in IMRT-SIB was 63 (56–63) Gy (all by 35 fractions), whereas the median (range) of lower radiotherapy dose and fractions in IMRT-SEQ was 50 (44–50) Gy and 25 (22–25) fractions.

Table 1

Patient characteristics of the study population in the primary analysis.

Primary analysis

After a median follow-up period of 33 months (range 3–84) (median 45 and range 24–84 for the survivors), 90 deaths were observed (47 for IMRT-SIB vs 43 for IMRT-SEQ). There was no statistical significance when IMRT-SIB was compared to IMRT-SEQ (HR for death 1.23, 95% confidence interval [CI] 0.84–1.80, P = .29]. The observed HR of 1.23 for OS could be explained by an unmeasured confounder that was associated with both selections of SIB/SEQ and live/death by a risk ratio of 1.58 (E-factor) fold each, but weaker confounding could not do so. The OS curve is shown in Figure 2. The 5-year OS rates were 47% in IMRT-SIB vs 54% in IMRT-SEQ. There were no statistical differences for IPCM (HR 1.40, 95% CI 0.89–2.21, P = .14, Fig. 3).

Figure 2

. Overall survival in the primary analysis.

Figure 3

. Incidence of pharyngeal cancer mortality in the primary analysis. IMRT-SEQ = sequential intensity-modulated radiotherapy, IMRT-SIB = intensity-modulated radiotherapy with simultaneous integrated boost.

Subgroup analyses

In the first SA for OPC, well-balanced covariables were seen after PS-matching (Table 2). The HR for death when IMRT-SIB was compared with IMRT-SEQ was 1.39 (95% CI 0.87–2.20; P = .17). In the second SA for HPC, well-balanced covariables were seen after PS matching (Table 3). The HR for death when IMRT-SIB was compared with IMRT-SEQ was 1.11 (95% CI 0.65–1.89; P = .71).

Table 2

Characteristics of the OPC patient in the first subgroup analysis.

Table 3

Characteristics of the HPC patient in the second subgroup analysis.

Discussion

Synopsis of key findings

To the best of our knowledge, we provided the first population-based study specifically for OPC or HPC patients treated with definitive CCRT using either IMRT-SIB or IMRT-SEQ. We found the survival outcome might be comparable for those treated with IMRT-SIB vs those treated with IMRT-SEQ.

Clinical applicability of the study

IMRT-SIB technique was developed after the clinical implementation of IMRT. The dose/fractionation of IMRT-SIB was assumed to be at least as effective as the conventional fractionation (ie, IMRT-SEQ) based on radiobiological modeling. However, clinical evidence was needed to confirm this assumption. Our study provides real-world rationale for clinical implementation of IMRT-SIB, although it utilizes nonconventional dose fractionation. However, we should be aware that our observation was interesting but inconclusive with limitations mentioned in section 4.4. Furthermore, our cases were staged by the seventh American Joint Committee on Cancer (AJCC) and its implication for current patients (staged by eighth AJCC) is unsure.

Comparisons with other studies (Table 4)

Presently, all RCTs for NPC reported similar OS as mentioned in the systematic review and meta-analysis,[ similar to what we found for OPC or HPC. There were no fully published journal articles specifically for OPC or HPC to our knowledge, although similar OS had also been reported in 2 single-institution studies for HN cancer patients,[ and similar OS had also been reported in a single-institution study for OPC in another conference paper.[ In additional, IMRT-SEQ was reported to be associated with better OS when used in adjuvant radiotherapy for oral cavity cancer patients in another conference paper.[

Strengths and weaknesses of the study

There were some limitations to our study. First, our study is an observational study, and there might exist some unidentified (due to data limitation in HWDC) confounding factors (such as human papillomavirus status, or systematic or radiation therapy detail) for patient's survival. The treatment groups between prospective randomized studies are homogenous except the treatment but may not be so in retrospective studies (such as our study). Although we used the PS method to diminish the potential imbalance in observable covariables, the 2 groups in our study may still not be similar regarding unmeasured potential confounders. Therefore, one prospective trial with enough follow-up period is warranted to resolve this issue. However, when we searched for phase 3 trials via “Simultaneous Integrated Boost Phase 3” at https://clinicaltrials.gov in September 2019, we found that SIB vs conventional fractionation had been investigated in various tumors such as glioma (NCT01507506), breast cancer (NCT02474641, NCT01322854, NCT02440191), and rectal cancer (NCT01224392), but there is no new ongoing relevant trial for OPC or HPC. Second, there lacks the comprehensive dosimetric comparison between IMRT-SIB and IMRT-SEQ among these patients. Third, treatment-related toxicity such as osteoradionecrosis[ was a potential differentiating factor between SEQ and SIB and maybe the more clinically relevant question when comparing these 2 techniques. Unfortunately, our population-based analysis was unable to explore these toxicity differences due to data limitation.

In conclusion, we provided the first clinical evidence regarding IMRT-SIB vs IMRT-SEQ specifically for OPC or HPC patients treated with definitive CCRT. We found the survival outcome might be comparable for those treated with IMRT-SIB vs those treated with IMRT-SEQ. However, this result should be interpreted with caution given the nonrandomized study design. A prospective study would be needed for further evaluation.

Acknowledgments

The data analyzed in this study were provided by the HWDC, Ministry of Health and Welfare, Executive Yuan, Taiwan. We are grateful to Health Data Science Center, China Medical University Hospital for providing administrative, technical, and funding support. We thank Department of Medical Research China Medical University Hsinchu Hospital for their administrative support.

Author contributions

Conceptualization: Yao-Hung Kuo, Ji-An Liang, Tang-Chuan Wang, Chun-Jung Juan, Chia-Chin Li, Chun-Ru Chien.

Formal analysis: Chia-Chin Li, Chun-Ru Chien.

Investigation: Yao-Hung Kuo, Ji-An Liang, Tang-Chuan Wang, Chun-Jung Juan, Chun-Ru Chien.

Writing – original draft: Yao-Hung Kuo, Ji-An Liang, Tang-Chuan Wang, Chun-Jung Juan, Chun-Ru Chien.

Writing – review and editing: Chia-Chin Li, Chun-Ru Chien.

Chun-Ru Chien orcid: 0000-0002-2365-7641.

Table 4

Studies to compare IMRT-SIB and IMRT-SEQ in head-and-neck cancer.