The Therapeutic Value of Volumetric-Modulated Arc Therapy In Advanced Non-Small Cell Lung Cancer Primary Tumors: A Multicenter Retrospective Analysis Based On Propensity Score Matching And Comparison With Intensity-Modulated Radiotherapy

Objective:.To explore the feasibility of volumetric-modulated arc therapy (VMAT) instead of intensity-modulated radiotherapy (IMRT) for primary tumors of advanced non–small-cell lung cancer (A-NSCLC). Methods:.We used propensity score matching (PSM) and multicenter retrospective analysis to study the ecacy and toxicity of VMAT technology in radiotherapy for primary tumors of A-NSCLC.We used the chi-squared test to analyze the response rate(RR), local control (LC), acute radiation injury, and dose-volume parameters; the Kaplan–Meier and log-rank tests were used to determine local-regional progression-free survival (LRPFS) and overall survival (OS). Results: LRPFS was signicantly prolonged in stage III patients treated with IMRT before PSM (P< .05) and cases of grade 1 or 2 acute radiation esophagitis (RE) or radiation pneumonia(RP) in the IMRT than the VMAT(P< .05), but cases of grade 3 or 4 RE and RP were not signicantly different between the groups (P> .05). Before PSM, there was no signicant difference in the LRPFS and OS rates of the whole study group and the RR, LC, and stage IV subgroups, respectively (P> .05). After PSM, there was no signicant difference in the RR, LC, LRPFS, OS, RP, or RE of patients treated with VMAT and IMRT (P> .05), normal whole-lung (V5, V20, MLD), heart (V30, V40, MHD), and equal dose-volume parameters were signicantly lower in the VMAT group (P< .05). Conclusion:.Radiation therapy of A-NSCLC primary tumors using VMAT can achieve a similar ecacy to that of IMRT but with signicantly lower low-dose volume parameters of normal tissues and organs; this is true especially in stage N2 cases, where the reduction in radiation damage may be more favorable.


Introduction
As three-dimensional radiotherapy is widely used in advanced non-small-cell lung cancer (A-NSCLC), prospective studies have shown that, based on chemotherapy, molecular targeted therapy combined with three-dimensional conformal radiotherapy (3D-CRT), intensity-modulated radiotherapy (IMRT), stereotactic body radiotherapy (SBRT), or stereotactic radiation ablation (SABR) can prolong overall survival (OS) [1][2][3][4], especially in patients with oligometastasis. Although two phase II prospective studies con rmed that randomization to 3D-CRT or IMRT for stage IV NSCLC prolonged survival at whole-lung V20 ≤ 32% and maximum spinal dose ≤ 50 Gy, the characteristics of 3D-CRT or IMRT were not analyzed [5,6]. A phase II randomized controlled study of SBRT and drug-only treatment based on four cycles of chemotherapy or three months of molecular targeted therapy bene ting oligometastatic stage IV NSCLC patients recorded a signi cant prolongation of both progression-free survival (PFS) and overall survival (OS), but radiotherapy-only treatment involved irradiation of residual lesions and radiotherapy. The therapy plan was determined by the radiotherapists participating in the study, and there was no choice of radiotherapy technology or divided dose administration [7]. Another course of SABR or hyperfractionated radiotherapy for EGFRnegative stage IV NSCLC could also be used to irradiate the primary tumor or mediastinal metastatic lymph nodes. A randomized controlled phase II prospective study examining the bene t of chemotherapy alone or metastatic lesions also reported a signi cant prolongation of PFS; however, although it described the divided dose, this study did not mention the choice of three-dimensional radiotherapy technology to achieve SABR [8]. A meta-analysis of 21 studies con rmed that primary tumor radiotherapy signi cantly prolonged OS and PFS; among other results, it noted that radical dose radiotherapy may be more conducive to prolonging survival results. The studies included in the analysis used 3D-CRT, IMRT, SBRT, etc., and no radiotherapy was also used. The report included an analysis of technical characteristics [9]. T3-4 and N2-3 of A-NSCLC account for about 64% and 83%, respectively. The primary tumor is of a late stage and large size, with spatial diversity [5], which are also factors that lead to the failure of rst-line chemotherapy and molecular targeted therapy [10,11]. Therefore, the local control rate and survival are obtained to control radiation damage and improve tumor dose. In this study, volumetric modulated arc therapy (VMAT) was used to improve the conformity of radiotherapy for large-volume thoracic esophageal cancer and to reduce the V20 and dose (MU) of the whole lung [12]. In addition, in the external irradiation of cervical cancer, VMAT at an energy dose of 6-MV has the best conformation of two-arc therapy, the lowest bladder and rectal doses considering adjacent organs, and reliable radio-physical quality assurance [13,14]. The clinical value of VMAT in the treatment of primary tumors of A-NSCLC was retrospectively analyzed by pairing multicenter analysis and propensity score matching.

Materials And Methods
Criteria for case selection. Eligible patients had pathologically diagnosed and newly treated stage IIIA to IV (American Joint Committee on Cancer, eighth edition) NSCLC; were between 18 and 80 years old; had a Karnofsky performance status of at least 70; underwent primary tumor complete VMAT or IMRT radiotherapy with a radiation dose of at least 40Gy/20f; underwent chemotherapy for more than two cycles; received epidermal growth factor receptor EGFR (M+) rst-line treatment with EGFR tyrosine kinase inhibitors; underwent complete recent results collection and acute radiotoxicity evaluations; and had complete follow-up data available. To determine the metastasis status, we de ned oligometastasis (OM) as one to ve distance metastases [15,16], while cases of at least six metastases were de ned as non-OM (NOM).
Primary tumor radiotherapy. We rst selected a 6-MV X-ray from Elekta In nity linear accelerator then positioned patients in the supine position, performed thermoplastic lm xation, performed computed tomography simulation with a 5-mm-layer thickness enhancement scan to obtain the positioning image and transmit it to Pinnacle 3 from Philips, and completed target acquisition according to the default lung window and mediastinal window of the planning system outline. The primary tumor (gross tumor volume [GTV]) is de ned as the primary lesion and the metastatic lymph nodes in the drainage area. The metastatic lymph node must be a single lymph node with a short diameter of ≥ 1.0 cm or at least three lymph nodes in the same lymph drainage area with a diameter of ≥ 0.5 cm.
For this study, the clinical target area (clinical target volume [CTV]) was de ned as a three-dimensional expansion of the GTV edge by 0.6 cm and combined with anatomical barriers, and the planning target volume (PTV) was de ned as a three-dimensional expansion of the CTV edge by 0.5 to 1 cm. The Pinnacle3 system was used to complete the radiotherapy plan design. IMRT of the primary tumor employed four to eight coplanar or non-coplanar elds, and VMAT required two to four arcs. The plan evaluation required that the prescribed dose cover 100% of the GTV volume; 95% of the prescribed dose for stage III cases included 95% or more of the PTV volume, and 90% of the prescribed dose for phase IV cases included 98% or more of the PTV volume. Considering a normal average dose of whole lung (total lung volume minus GTV) ≤ 20 Gy and whole-lung (V20) dose ≤ 32%,for primary tumor irradiation, we adopted late-course accelerated hyperfractionation ( rst course, 36-40Gy, 2Gy/time, 1 time/1 day, 5 days/week; latter course, 1.5Gy/time, 2 times/day, with an interval of 6 hours, 5 days/week). Drug therapy. Chemotherapy involved a platinum-based two-drug combination regimen [17] of platinum drugs, and the dose regimen was as follows: cisplatin (DDP) 75 mg/m 2 given as an intravenous infusion over two to four days or carboplatin (CBP) 300 to 350 mg/m 2 given on day 2; paclitaxel (TAT) 135 to 175 mg/m 2 given on day 1 or docetaxel (DOC) 60 to 75 mg/m 2 given on day 1, with vinorelbine 30 mg/m 2 also given on day 1; and pemetrexed (PEM) 500 mg/m 2 given on day 1 (for cases of non-squamous cell carcinoma). A total of 21 to 28 days was considered one cycle, with two to six cycles performed in a row.
Molecular targeted therapy employs rst-line drugs based on the results of sensitive mutations in driving genes, with ge tinib 250 mg taken orally every day, icotinib 150 mg taken orally three times a day, osimertinib 80 mg taken orally every day, and crizotinib 250 mg taken orally twice a day, until the disease progresses.
Evaluation of the recent results for primary tumors. According to the RECIST 1.1 standard [18], the e cacy evaluation allocated patients into complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD) groups; for this study, CR+PR was de ned as effective (response rate [RR]), and a result of CR+PR+SD suggested the disease to be controlled (disease control rate [DCR]). The effective rate = (CR+PR)/total number of cases × 100%, while the disease control rate = (CR+PR+SD)/total number of cases × 100%.
Acute radiation injury classi cation. According to the radiation injury classi cation standards of the Radiotherapy Oncology Group (RTOG), acute injuries to the lungs, esophagus, and/or heart were recorded.
Statistical method. Using the Statistical Package for the Social Sciences version 26.0 software program (IBM Corporation, Armonk, NY, USA), characteristics including sex, age, pathological type, T/N/M staging, GTV volume, primary lesion location, clinical stage, metastasis status, targeted therapy, chemotherapy cycle, prescription dose, and other factors for IMRT and PSM of VMAT showed a matching tolerance of 0.02. Survival analysis was tested using the Kaplan-Meier and log-rank methods; local control rate was calculated by the life table method; and recent measurements, radiation injury, and dose-volume parameters were tested using the chi-squared test. A P-value of .05 or less was considered statistically signi cant.

Results
Clinical characteristics of the case data. From September 2008 to March 2020, 637 cases met the case selection criteria. In the study population, the male to female ratio was 2.5:1; the median age was 58 years (range, 22-79 years); and there were 51, 91, 68, and 427 cases of stages IIIA, IIIB, IIIC, and IV cancer, respectively. A total of 102 cases of driver gene-sensitive mutations were detected before treatment due to economic and medical insurance factors (including 83 cases of EGFR mutations and 19 cases of ALK mutations), and only 10.20% of patients received molecular targeted therapy (including ge tinib in 28 cases, icotinib in 22 cases, ositinib in two cases, and crizotinib in 13 cases), while chemotherapy was used in 89.80% of cases. The proportions of N2-3 and IV stages among IMRT patients before PSM was higher than those among VMAT (P < .05), and the prescribed dose and GTV volume were similar (P > .05). After PSM, there were 308 patients (154 pairs) with a median age of 58 years. The clinical baseline conditions of the two groups were similar (P > .05). See Table 1 for details.  LRPFS of different techniques for radiotherapy of primary tumors before and after PSM. Before PSM, the one-year and two-year LRPFS rates following treatment with IMRT or VMAT were 91.1% and 91.9% and 76.0% and 86.1%, respectively, and there was no statistically signi cant difference (χ 2 = 0.246; P = .620). The strati ed analysis showed that the LRPFS of stage III at one and two years were 97.6% and 90.6% and 91.9% and 83.0%, respectively, and the LRPFS of IMRT treatment was prolonged (χ 2 = 4.183; P = .041). The one-year and two-year LRPFS of stage IV were 91.0% and 93.2% and 67.8% and 88.7%, respectively, with no statistically signi cant difference (χ 2 = 2.097; P = .148). See Figure 1 (a-c)  respectively (χ 2 = 3.132; P = .077), and there was no statistically signi cant difference. See Figure 2 (a-c).
The one-and two-year OS of radiotherapy with different techniques after PSM. The one-year and two-year OS rates after IMRT or VMAT treatment were 70.1% and 69.9% and 31.3% and 50.1%, respectively (χ 2 = 1.504; P = .220), and there was no statistically signi cant difference. Strati ed analysis showed that the one-year and two-year OS rates of stage III patients were 81.8% and 71.6% and 44.7% and 67.1%, respectively (χ 2 = 0.022; P = .881), and there was no statistically signi cant difference. Finally, the one-year and two-year OS rates of stage IV patients were 67.3% and 61.3% and 21.4% and 37.6%, respectively (χ 2 = 2.207; P = .137), and there was no statistically signi cant difference. See Figure 3 (a-c) for details.
RP and RE of different radiotherapy techniques before and after PSM. There was no grade VI injury in the whole study group. The incidence rates of RE and RP in the IMRT group before PSM were 73.70% and 36.02%, respectively, which were signi cantly higher than those of the VMAT group (P < .05), The incidence rates of grade III or IV RE and RP were 8.28% and 7.14% and 7.03% and 7.79% for IMRT and VMAT, respectively, with no statistically signi cant difference (χ 2 = 2.77; P = .096). There was also no signi cant difference after PSM. See Table 3 for details. IMRT and VMAT dose-volume parameters after PSM. N2 stage IMRT whole-lung V5, V20, MLD, heart V30, V40, MHD, and esophagus V50 were signi cantly higher than in the VMAT group (P < .05); further, N3 stage IMRT whole-lung V5 was higher than that in the VMAT group (P < .05). There was no signi cant difference in esophageal V50 of the whole group of patients (P > .05). See Table 4 for details. .046 .545

Discussion
The results of this study show that the male:female ratio is 2.5:1, the median age is 58 years, and that adenocarcinoma and squamous cell carcinoma cases total about 94.3% of the study population, which is similar to the epidemiological characteristics of the incidence of NSCLC [19]. Also, T3-4 and N2-3 cases totaled 67% and 88.9%, respectively. The average volume of GTV was about 200 cm 3 . There was no signi cant difference in the distribution of primary lesions between patients who received VMAT and IMRT, respectively. The disease stage was late, and mediastinal lymph node metastasis was obvious [20,5]. The proportions of N2-3 and IV patients, respectively, in the pre-PSM IMRT group were relatively high, regardless of stage III or oligometastasis of more than 50% among stage IV NSCLC patients. Both results show that the primary tumor is large in size [21], and it is di cult to obtain the local control rate by increasing the dose under the premise of controlling the damage. It is necessary to use the physical characteristics of different three-dimensional radiotherapy techniques to explore the clinical application value.
The difference between VMAT and IMRT as treatment is that, during the accelerator irradiation process, the rotating irradiation is realized by continuous changes of the gantry speed, collimator angle, and dose rate, and it has the characteristic of a short irradiation time. For elderly patients older than 80 years old with irregular target areas and more protected organs at risk, studies on head and neck tumors have shown that VMAT can ensure target dose and improve conformability while signi cantly reducing the exposure to endangered organs so as to obtain the same tumor-free survival rate as younger patients. Considering the tumor-free survival rate [22], a small A-NSCLC radiation therapy plan dose-volume parameter study showed that VMAT increased the V95% and conformity of the planned target area and reduced the average dose to the lungs, esophagus, and heart [23] compared to IMRT. Therefore, in this study, our multicenter and retrospective analysis of VMAT treatment results through the trend score showed that the RR, DCR, and LCR of PSM before and after IMRT treatment were similar (P > .05), suggesting that the volume and primary tumor dose are similar. Under this premise, application of VMAT in A-NSCLC patients is effective in the treatment of primary tumors, and the recent results were not less than those of IMRT. Although it was shown before PSM that the LRPFS of stage III NSCLC cases treated with IMRT was longer than that of VMAT, there was no signi cant difference in LRPFS in the whole group nor during strati ed analysis under the premise of the same baseline conditions after PSM, indicating that VMAT combined with drugs can be used in the rst-line treatment of A-NSCLC with a long-term e cacy similar to that of IMRT. Moreover, the one-and two-year OS rates of VMAT combined with chemotherapy in the rst-line treatment of stage IV NSCLC patients were 61.3% and 37.6%, respectively, which may be much higher than the one-and two-year OS rates of 35% and 10%, respectively, associated with platinumcontaining two-drug regimen rst-line chemotherapy [24], suggesting a prolonged survival role [4]. In summary, the effect of VMAT in the treatment of A-NSCLC is similar to that of IMRT. It improves recent results and improves OS.
The proportion of primary tumor volume shrinkage is also negatively correlated with the risk of progression failure, and OS is prolonged through radiotherapy, as well as an increased dose and local control rate [5,25].
It is well known that acute radiation injury is the dose-limiting toxicity of radiotherapy, RE and RP of grade 3 or higher are unfavorable factors for the prognosis of NSCLC [26], and the volume of normal lung low-dose radiation is related to the occurrence of RP. VMAT irradiation needs to be performed through a rotating arc of the gantry, but the low dose volume of normal tissues does not necessarily increase. This study showed that the incidence rates of RP and RE of primary A-NSCLC tumors with a median dose of 64Gy and an average volume of more than 200 cm 3 after VMAT treatment were lower than that of IMRT before PSM. There was no signi cant difference in grade 3 and 4 acute radiation injury between groups, and both results were determined after PSM. The absence of a signi cant difference suggests that the rotating irradiation mode of VMAT did not increase the acute radiation injury caused by the low dose volume of normal tissue in clinical practice. In addition, this study analyzed the MHD isometric volume parameters of the heart in the dose-volume histogram as well as the important indicators for predicting the occurrence of RE and RP; esophagus V50; and whole-lung V5, V20, and MLD [27][28][29][30][31]. After PSM, the cardiac dosevolume parameters of patients treated with VMAT and the low dose volume of normal lungs were signi cantly lower than those of the IMRT group. In particular, the strati ed analysis showed that the dose parameters, such as MHD, esophagus V50, and normal whole-lung V5 of N2 staging cases were signi cantly lower than those of the IMRT. The IMRT group was signi cantly reduced (P < .05), while N3 patients with only normal whole-lung V5 had parameters signi cantly lower than those of the IMRT group (P = .000), further suggesting that the VMAT rotary irradiation method has a more reasonable dose distribution and maintains the low dose volume to reduce radiation damage. The VMAT rotary irradiation method has the advantages of a more reasonable dose distribution and better control of low dose volume to reduce radiation damage, especially in patients with mediastinal lymph node metastasis only on the same side. More importantly, the reduction in dose volume may reduce the late damage of normal tissues, especially considering the reduction of heart damage that is positively correlated with the radiation dose and clinical dysfunction that occurs for several years or longer after irradiation [32][33].

Conclusions
The selection of VMAT for primary tumor radiotherapy in A-NSCLC patients can lead to the same recent and longterm e cacy results as IMRT, and it may be more advantageous to use this approach to reduce acute radiation injury, especially for late cardiac injury.

Availability of data and materials
The data that support the ndings of this study are available on request from the corresponding author.
Code availability.
Not applicable.

Ethics approval and consent to participate
This retrospective analysis was approved by the ethics committee of the ethics committee of the A liated Hospital of Guizhou Medical University Ethics approval document, (approval number:2021-275). Informed consent was obtained from all individual participants included in the study.

Consent for publication
Informed consent was obtained from all individual participants included in the study.