Early clinical outcomes by combination of intensity modulated radiation therapy and intensity modulated proton therapy in treating oropharynx cancer patients

Purpose: To report the early clinical outcomes of combining intensity‐modulated radiation therapy (IMRT) and intensity‐modulated proton therapy (IMPT) in comparison with IMRT alone in treating the oropharynx cancer (OPC) patients. Materials and Methods: The medical records of 148 OPC patients were retrospectively reviewed, who underwent definitive radiotherapy (RT) with concurrent systemic therapy, from January 2016 till December 2019 at Samsung Medical Center. During the 5.5 weeks’ RT course, the initial 16 (or 18) fractions were delivered by IMRT in all patients, and the subsequent 12 (or 10) fractions were either by IMRT in 81 patients (IMRT only) or by IMPT in 67 (IMRT/IMPT combination), respectively, based on comparison of adaptive re-plan profiles and availability of equipment. Propensity‐score matching (PSM) was done on 76 patients (38 from each group) for comparative analyses. Results: With the median follow‐up of 24.7 months, there was no significant difference in overall survival and progression free survival between groups, both before and after PSM. Before PSM, IMRT/IMPT combination group experienced grade ≥3 acute toxicities less frequently: mucositis in 37.0% and 13.4% (p<0.001); and analgesic quantification algorithm (AQA) in 37.0% and 19.4% (p=0.019), respectively. The same trends were observed after PSM: mucositis in 39.5% and 15.8% (p=0.021); and AQA in 47.4% and 21.1% (p=0.016), respectively. In multivariate logistic regression, grade ≥3 mucositis was significantly less frequent in IMRT/IMPT combination group, both before and after PSM (p=0.027 and 0.024, respectively). AQA score ≥3 was also less frequent in IMRT/IMPT combination group, both before and after PSM (p=0.085 and 0.018, respectively). Conclusions: In treating the OPC patients, with comparable early oncologic outcomes, more favorable acute toxicity profiles were achieved following IMRT/IMPT


Introduction
Radiation therapy (RT) has the capability of organ preservation, and plays a key role, with or without chemotherapy, in managing the oropharyngeal cancer (OPC) patients with nonmetastatic disease extent [1][2][3]. With the technical advancement, intensity modulated radiation therapy (IMRT), when compared to the traditional 2-or 3-dimensional RT techniques, has enabled high enough radiation dose delivery to the targets at reduced risks of severe acute and delayed side effects. Though IMRT has currently become the most popular and recommended RT technique in treating most head and neck cancer (HNC) types, a significant proportion of HNC patients, however, still suffer from annoying side effects and lowered quality of life during and following high dose RT [4][5][6][7].
Proton beam therapy (PBT), by virtue of physical property of Bragg-Peak phenomenon, can generate more advantageous dose distribution profile than photon-based RT techniques, including IMRT, and has been in clinical use in treating several cancer types including HNC [8][9][10][11]. Nevertheless, more clinical evidences are still in need to confirm whether the physical advantage of PBT genuinely can lead to better therapeutic outcomes in real-world practice setting. Considering the substantial costs and resources needed for installation and operation of PBT facilities, answering this issue seems even more important.
With these theoretic backgrounds, we intended to apply PBT in treating the HNC patients since December of 2015, when our PBT facility began its operation [12]. However, the average waiting time interval before PBT to start after therapeutic decision has been about 4~6 weeks, because of limited PBT resources when compared to the clinical demands. It has been well addressed, however, that long waiting before treatment initiation in treating the HNC patients could result in significantly unfavorable oncologic outcomes [13,14]. In order to overcome this long waiting before treatment initiation, we developed our strategy to begin the RT course by IMRT (helical Tomotherapy, HT) and then to determine whether to continue IMRT or to switch into intensity modulated PBT (IMPT), based on the rival re-plan comparison, which corresponds to our adaptive re-plan policy. We previously reported the early clinical outcomes and acute toxicity profiles following IMRT only and IMRT/IMPT combination in treating the nasopharynx cancer (NPC) patients [15], and would report our experience in treating the OPC patients.

Patients' characteristics
The characteristics of all patients and 76 matched patients based on the propensity scores (38 in each group) were summarized in Table 1. The median age of the whole patients was 60 years (range, 38~76 years), and the majority was male (137 patients, 92.6%). Among all patients, most characteristics were similarly distributed between groups, but the patients in IMRT/IMPT combination group more frequently had lower T stage (p=0.025) and received unilateral neck irradiation (p<0.001), respectively. Among 76 matched patients, however, all characteristics distributed similarly between groups.

Oncologic outcomes
The tumor response was excellent and the rates of overall and complete response, evaluated at There were no significant differences of overall survival (OS) and progression-free survival (PFS) profiles between groups, both before and after propensity score matching (PSM) ( Figure   1 (Figure 2A). Among the matched patients, the same trends were observed: the frequencies of grade ≥3 mucositis were 39.5 % and 15.8% (p=0.021); and those of AQA score ≥3 were 47.4 % and 21.1% (p=0.016), respectively ( Figure 2B).  Table 2. Acute toxicity profiles.

Discussion
Radiation oral mucositis is very common and unavoidable acute side effect affecting most HNC patients who receive high dose RT. Oral mucositis typically causes oral soreness, swallowing difficulty, decreased oral intake, and subsequent weight loss. Severe oral pain usually necessitates taking pain-killers, sometimes narcotics, and the patients may become prone to various adverse effects of the medication. It was reported that RT-related complications, such as oral mucositis, can increase the treatment cost by up to 17,000 USD per patient, and its severity is proportionally associated with the incremental healthcare cost [16,17]. Moreover, the modification and/or interruption of the planned RT schedule occasionally is necessary, in order not to compromise the precision of RT and to compensate for the body contour change incurred by these sequence of events in addition to tumor shrinkage itself.
Considering these respects, IMPT, by virtue of Bragg-Peak phenomenon, can provide more effective sparing of the oral mucosa from moderate to high dose radiation exposure, especially when treating well-lateralized target lesions, and subsequently reduce the severity of oral mucositis, when compared to IMRT. Likewise, saving of anteriorly located oral cavity mucosal lining, if the target is posteriorly located as in most OPC patients, could be achieved more effectively by using IMPT.
Xiao et al. demonstrated that the detrimental effects by increased TTI was mediated by tumor progression during the waiting time [18], which was demonstrable by comparing the initial clinical stages and the surgical stages, and found that even 1 week's delay could be detrimental and suggested timely intervention within the first 4 weeks. Although induction chemotherapy before definitive local treatment modality could also be considered to bridge the gap, which, however, might increase the treatment-related morbidity risk and the care cost without significant clinical benefit [19]. Based on these backgrounds, we decided to begin upfront RT by starting by IMRT, instead of IMPT, in all OPC patients.
Adaptive re-plan during the RT course has been highly recommended in order to accommodate the body contour changes in treating most HNC patients [20,21]. The body contour change is usually more significant during the early CCRT course than during the later course. We previously measured the mean tumor volume reduction rates by the time of re-plan, which were 40.7% in the OPC patients and 41.9% in the NPC patients, respectively [22,23]. These clinically relevant data by the authors strongly endorse the adaptive re-plan strategy, which has long been our institutional policy. For the adaptive RT, we generated two rival plans, one by IMRT and the other by IMPT, and then to determine the subsequent RT modality (whether to continue IMRT or to switch into IMPT), based on the dosimetric profiles in addition to the availability of IMPT. By following these strategies, we could shorten the waiting from 4~6 weeks to a few days, avoid the break during the RT course due to significant and abrupt body contour changes, and determine optimal RT modality on the individual basis.
We previously reported the early clinical outcomes and acute toxicity profiles following IMRT only and IMRT/IMPT combination in treating the NPC patients [15], and found that combination of IMRT and IMPT was more advantageous in weight loss, analgesic use, with the equivalent oncologic outcomes. To the best of our knowledge, there have been only few retrospective studies on the OPC patients, which evaluated the causal relationship between of IMPT's dosimetric advantage and RT-related toxicities. Blanchard et al. performed a casematched analysis comparing IMRT and IMPT for OPC patients, and reported reduced rates of gastrostomy tube dependency and severe weight loss (defined as >20% weight loss from the baseline) in IMPT group [24]. Sio et al., based on 81 OPC patients, demonstrated that patientreported symptom burden was lower following IMPT than IMRT [25]. These two studies neither did thorough multivariate analyses, nor did quantitative measurement of the toxicities, including mucositis and analgesic usage. The current study intended to investigate whether similar effects as in the NPC patients could be obtained in the OPC patients by combining IMRT and IMPT. As described above, the oncologic outcomes of OS and PFS were not different between groups, while IMRT/IMPT combination, compared with IMRT only, resulted in more favorable acute toxicity profiles in terms of grade ≥3 mucositis and AQA score ≥3 through the quantitative measurement and multivariate analyses. Our study could have complemented the limitations of aforementioned studies and, at the same time, have supported the consistent finding of improved acute toxicity profiles in treating the OPC patients.
The current study has a weak point of uneven distribution of several characteristics between the treatment groups, mainly by virtue of the retrospective nature. We did propensity-score matching and multivariate logistic regression to mitigate this weakness. In addition to the main observations described above, our IMRT/IMPT combination regimen could reduce the direct RT cost up to 28% according to the Korean National Health Insurance plan, when compared with 30 fractions' IMPT only throughout the RT course. manual, which mainly depended on the anatomic disease extent but not on the HPV status [26].

Treatment scheme
All patients underwent contrast enhanced CT-based simulation with the open mouth. According to our institutional "selective neck irradiation" policy, three target volumes were delineated: gross tumor volume (GTV); high-risk clinical target volume (HR-CTV) which included the tissue and lymphatics adjacent to GTV; and low-risk clinical target volume (LR-CTV), respectively [27]. LR-CTV was individually determined and did not include the clinically uninvolved lymphatics that were two stations away from GTV. The same target delineation policy was applied to all patients regardless of the actual RT modality assigned and/or the HPV status. The dose schedules varied along with the study period, which mainly reflected the resource allocation limitation at the authors' institute. The typical dose schedules to the GTV, HR-CTV, and LR-CTV were 66~68.4 Gy, 60 Gy, and 36 Gy over 30 fractions in 97 patients until January of 2018, while those were 67.2 Gy, 56 Gy, and 32~36 Gy over 28 fractions in 51 since after February of 2018, respectively. The differential dose delivery was possible by combining the simultaneous integrated boost and the adaptive re-plan and shrinking field concept, which eliminated LR-CTV during the later RT course.  (Figure 3). We intended to assign RT modality during the later RT course based on the rival plan comparison. The actual RT modality assignment, however, did not solely depend on dosimetric superiority, but had to be allocated considering the practical resource limitation and availability. The patients who showed equivalence or superiority by IMRT plan were allocated to IMRT. Meanwhile, those who showed dosimetric superiority by IMPT but should have had RT break longer than a week for subsequent IMPT were allocated to IMRT in order to avoid the undesirable treatment interruption. Consequently, 81 patients (54.7%) continued to receive IMRT (IMRT only group) and 67 (45.3%) received IMRT + IMPT (IMRT/IMPT combination group), respectively (Table 1). in nine (6,1%), and oral cetuximab (400 mg/m 2 loading dose followed by 5 weekly dose of 250 mg/m 2 ) in 12 (8.1%), respectively ( Table 1). The vast majority of patients (130, 93.5%) were able to complete the planned chemotherapy cycles, while nine did not because of toxicity.
Seven patients among 118 (5.9%) in whom 2 cycles of triweekly cisplatin was planned received only 1 cycle, and two among nine (22.2%) in whom weekly cisplatin were planned received <6 cycles, and all 12 in whom cetuximab was planned were able to complete the intended dose schedule, respectively.

Propensity score matching
In order to adjust the differences in the baseline characteristics in groups, a PSM method was used. After building a multivariate logistic regression model including the variables with a pvalue < 0.1 on the Chi-square test or Fisher's exact, two variables were considered significant (clinical T stage and bilateral neck irradiation). In order to guarantee the homogeneity, only the patients who receive cisplatin-based chemotherapy and the primary tumor site of tonsil or base of tongue were included at time of matching. Based on the calculated propensity score, the matching ratio was 1:1 with the caliper set at 0.2.

Assessment of acute side effects and response, and follow-up schedule after treatment
The acute toxicity profiles during RT were evaluated at least once a week on each patient by the radiation oncologist in charge: the Common Terminology Criteria for Adverse Events (CTCAE) ver. 4.03 [28] to monitor radiation dermatitis, oral mucositis, and weight loss; and the AQA scoring system to quantify the analgesic usage (Supplementary table 1) [29].
Response evaluation was done by neck CT taken 1 month of RT completion and PET-CT taken in 3 months thereafter, respectively. PET response criteria in solid tumors (PERCIST) was used to assess the early tumor response [30]. Subsequent follow-up evaluations, including neck CT, were regularly scheduled: at every 3-4 months' interval during the first 2 years; and at every 6 months' interval thereafter. Locoregional failure was defined as any development of new lesion or progression of preexisting lesion, either within or near the initial disease sites.

Statistical analysis
All statistical analyses were performed using the SPSS software version 24.0 (IBM Corporation, Armonk, NY) and R version 4.0.0 (R Development Core Team, Vienna, Austria, http://www.rproject.org). The OS and PFS rates of the two groups were calculated using the Kaplan-Meier estimate and compared by log-rank tests. To compare the patient characteristics and acute toxicity profiles between the two treatment groups, the Chi-square test or Fisher's exact test was used for categorical variables while the Student's t-test was used for continuous variables.
Furthermore, multivariate logistic regression was performed in order to identify factors that are independently associated with acute toxicity. Factors with a p-value < 0.1 on the univariate analysis or factors considered clinically significant were included in the multivariate analysis, after exclusion of the possible confounding factors.

Conclusion
In summary, our strategy of combining IMRT and IMPT could avoid undesirable long waiting before treatment initiation, and lead to favorable acute toxicity profiles, at similar oncologic outcomes in treating the OPC patients. Further analyses with larger sample size and longerterm observation including the delayed side effect profiles would be needed.

Conflict of interest
None.