Reasonable Timing of Radiotherapy for Stage IV Non-Small-Cell Lung Cancer During Targeted Therapy Based on Tumour Volume Change

Purpose The aim of this study was to investigate the reasonable timing of radiotherapy for stage IV non-small-cell lung cancer (NSCLC) with EGFR-positive mutations during targeted therapy based on tumour volume change (TVC). Patients and Methods Simulation Computed Tomography Scan (SCTS) measurements were taken to test TVC in patients with stage IV NSCLC during targeted therapy at intervals of 10 days. The SCTS measurement was terminated when the tumour volume shrinkage rate in the latter simulation compared with the previous simulation was ≤5% or when the time after treatment was 90 days. Then, primary tumour radiotherapy was performed. Related parameters of the radiotherapy plan were compared between the implementation and simulation plans. Results Twenty-seven patients were enrolled in the analysis. After treatment, shrinkage of the primary tumour was observed in all patients, but the rate and speed were inconsistent. The average tumour volume decreased obviously within 40 days and was significantly different every 10 days (P ≤ 0.001). The average volume decreased slowly and tended to be stable (P>0.05) after 40 days. After the termination of SCTSs, 21 patients accepted primary tumour radiotherapy. No patients experienced grade 3+ acute radiation toxicity. The implementation radiotherapy plan was significantly better than that before treatment (all P<0.05) but not better than that on the 40th day after treatment (all P>0.05). Conclusions To obtain a high radiation dose and control radiation toxicity, the 40th day after targeted therapy may be a reasonable time to start radiotherapy for stage IV NSCLC with EGFR-positive mutations. Clinical Trial Registration https://www.clinicaltrials.gov/ct2/show/NCT03258671, identifier, NCT03258671.


BACKGROUND
First-line treatment of stage IV non-small-cell lung cancer (NSCLC) has evolved from chemotherapy alone to chemotherapy, targeted therapy, and immunotherapy (1), and targeted therapy for patients with positive mutations in driver Genes, such as human epidermal growth factor receptor (EGFR), anaplastic lymphoma kinase (ALK)/ C-ros oncogene 1 receptor tyrosine kinase (ROS1) and T790M, has been shown to significantly prolong progression-free survival (PFS) (2)(3)(4)(5). Higginson DS (6) et al. analysed stage III/IV NSCLC patients who received only systemic chemotherapy and found that the state of the primary tumour (large central tumour, pulmonary symptoms, and bronchial or vascular compression) was associated with poor OS. More importantly, recent studies have shown that targeted therapy, chemotherapy, and immunotherapy combined with threedimensional radiotherapy of primary tumours and metastatic lesions can significantly improve overall survival (OS) (7)(8)(9) and significantly reduce the treatment failure rate of primary tumours from 80%-90% to less than 30% (10). A meta-analysis suggested that primary tumour radiotherapy, especially with radical doses, might further prolong survival (11). The local failure (12) was 82% for stage IV NSCLC treated with only EGFR-TKI. Previous studies (13,14) showed that targeted therapy can increase the sensitivity of radiotherapy, and the combination therapy has the best inhibitory effect on cancer cell proliferation compared with radiotherapy alone or targeted therapy alone. OS benefits may be derived from the synergistic combination of radiotherapy and targeted therapy (15)(16)(17)(18). However, the tumour burden of stage IV NSCLC is relatively large, with T 3 -4 accounting for 60%-70% and N 2-3 accounting for 70%-80%, and the median volume of the primary tumour is approximately 134 cm 3 (7,19). The large tumour volume results in a low local control rate (LCR) due to the low radiation dose to reduce the rates of severe radiation toxicities and can also lead to an increase in radiation-induced toxicities due to an increased radiation dose. Therefore, we designed a prospective clinical trial to reduce the tumour to a certain size and maintain a relatively stable state by using EGFR-tyrosine kinase inhibitors (EGFR-TKIs), which have an objective response rate (ORR) of more than 70%, to realize the reasonable timing of radiotherapy to reduce normal tissue toxicity and increase the radiation dose, and to provide a reference for further randomized controlled studies on the reasonable timing of radiotherapy.
Tumour volume measurement process: (1) A Simulation Computed Tomography Scan (SCTS) was planned within 1 week before targeted therapy and every 10 days after the first day of treatment that patients underwent one simulation scan in sequence for a maximum of 90 days; (2) the SCTS within 1 week before targeted therapy was defined as C 0 ; after the start of treatment, the SCTSs every 10 days were defined as C 10 -C 90 ; the primary tumour volume before treatment (V P ), volume of metastatic lymph nodes in the drainage area (V N ) and gross tumour volume (GTV) were defined as V P0 , V N0 and GTV 0 , respectively; and the volumes measured on the SCTSs were V P10 -V P90 , V N10 -V N90 and GTV 10 -GTV 90 , respectively; (3) termination criteria for the SCTS were a tumour volume shrinkage (TVS) rate ≤5% in the latter simulation compared with the previous simulation or when the time after treatment was 90 days.
Delineation and calculation of tumour volume: Intensitymodulated radiotherapy (IMRT) was given via 6 MV X-ray. The patient was positioned in the supine position with thermoplastic film fixation, and the 5-mm-thick enhanced Computed Tomography (CT) scans were transferred to the Pinnacle3 planning system. V P was outlined with a lung window (W: 1,600, L: -300), and V N was outlined with a mediastinal window (W: 400, L: 800). Tumour volume was calculated, and the GTV compromised V P and V N . The GTV was outlined on the last simulation CT image. The clinical target volume (CTV) was defined as the GTV plus a margin of 0.6 cm, and the planning target volume (PTV) was defined as the CTV plus another margin of 0.5 to 1.0 cm. The TVS rate of C N was calculated as follows: TVS rate = (pre-treatment volumesimulation volume of C N )/pre-treatment volume × 100%.
Implementation radiotherapy plans and simulation plans: (1) IMRT was given via 6 MV X-ray. The implementation radiotherapy plans were created with the last simulation CT image. The radiotherapy dose was given to patients according to the tolerability of normal tissue and was maintained at ≤76 Gy. For all individual treatment plans, the percentage of the total lung volume receiving ≥20 Gy (V 20 ) was maintained at ≤32% (≤25% in crizotinib-treated patients), V 5 at ≤ 70%, mean lung dose (MLD) at ≤20 Gy, mean heart dose (MHD) at ≤26 Gy and maximum point dose to the spinal cord (SC-MPD) at ≤50 Gy. Radiotherapy plans were evaluated as 100% of the prescription dose line including 100% of the GTV and 95% of the prescription dose including 95% of the PTV or 90% of the prescription dose including 98% of the PTV. Patients received late-course accelerated hyperfractionated radiotherapy (LCAHRT) (20)(21)(22)(23) to the primary tumour. The first course of radiotherapy was given in 1.8-Gy fractions, 5 days per week, to a total dose of 36-40 Gy/18-20 f. LCAHRT was then delivered in twice-daily fractions of 1.5 Gy each, separated by 6 to 8 hours per day, to a total dose of 21-30 Gy/14-20 f.
Simulation plans were created with the pre-treatment simulation (C 0 ) and 40 days post-treatment (C 40 ) simulation images. Implementation radiotherapy plans were adjusted according to the same dose or the same radiation toxicity control criteria for each patient, and the dose-volume histogram (DVH) was recorded.

Radiotherapy to Metastatic Lesions
For oligometastatic NSCLC, all metastatic lesions were treated with radiotherapy. For non-oligometastatic NSCLC, radiotherapy to metastatic lesions was determined by clinical necessity, such as, brain metastasis, bone metastasis with cancer pain or risk of fracture.
Study endpoints and statistical methods: The primary endpoints were the change patterns of the V P , V N and GTV before and during treatment, and the secondary endpoints were acute radiation pneumonitis (RP) (within 3 months after the end of radiotherapy), oesophagitis (RE) (NCICTC 3.0 criteria) and DVH parameters. Statistical analysis was performed using SPSS software (version 23.0). Measurement data are expressed as the mean ± standard deviation (SD) and were analysed with t-tests or Mann-Whitney U-tests. P<0.05 was considered a statistically significant difference.

Patient Characteristics
Thirty patients met the inclusion criteria, and 27 patients were eligible for analysis (refusal in 1 patient and SCTS not as planned in 2 patients). The ratio of males to females was 1.25, and the median patient age was 60 years. The most common site of metastatic disease at diagnosis was the bone, brain and lung ( Table 1). The V P0 , V N0 and GTV 0 were 6.23~470.00, 0~362.97 and 28.86~470.00 cm 3 , respectively. The median and average GTV 0 were 149.42 cm 3 and 189.23 ± 127.03 cm 3 , respectively (the rest are shown in Table 1). Twenty-seven patients completed the SCTS and volumetric measurements according to the termination criteria. Twentythree patients harboured EGFR-positive mutations: an exon 19 deletion mutation (19del) was observed in 14 patients, and an exon 21 deletion mutation (L858R) was observed in 9 patients. Four patients harboured an ALK rearrangement. Targeted therapy involved gefitinib in 16 patients, icotinib in 7 patients and crizotinib in 4 patients ( Figure 1).

The Pattern of Tumour Volume Change
SCTSs from C 0 to C 30 were performed on 27 patients, and C 40 , C 50 , C 60 , C 70 , C 80 and C 90 were performed on 24, 17, 11, 9, 3, and 1 patients, respectively. The GTV of all patients had different degrees of change from C 0 to the last SCTS and showed a trend of gradual shrinkage, in which the largest volume shrinkage rate was 78.1% (gefitinib, thick red solid line) and the smallest was 18.8% (icotinib, thick black solid line) ( Figure 1). According to the graph of mean GTV, V P , and V N (C 0-70 ) values, tumour volume decreased gradually and significantly within the 40th day after treatment and then tended to stabilize (3 patients in C 80 and 1 patient in C 90, not analysed). The mean tumour volume continued to shrink or tended to stabilize after slightly increasing at 50 days ( Figure 2).

GTV Changes in Two Adjacent SCTSs
GTV changes in two adjacent SCTSs showed that the tumour volume shrinkage rate was inconsistent before the C 40 SCTS every 10-day interval, and the tumour volume shrinkage rate was <5% on the C 40 and C 50 SCTSs and every 10-day interval thereafter ( Table 2).

Volume and Shrinkage Rates of Tumours at Different Times After Treatment
The change patterns in V P and V N were similar to that of the GTV after treatment, with the most significant shrinkage rate in the first 10 days (C 10 ). The shrinkage rates of the GTV 10-40 were 22.21%, 14.64%, 5.54% and 4.37%, respectively. In every interval from C 40 to C 90 , 3, 7, 6, 2, 6 and 2 patients met the termination criteria due to having a shrinkage rate (adjacent comparison) <5%. The average shrinkage rate from C 40 to C 70 was 2.67%. Only 1 patient continued to have a >5% shrinkage rate at C 90 (Table 3).

Acute Radiotherapy Toxicity
Twenty-one patients (6 of whom refused radiotherapy after the termination of simulation) were treated with primary tumour radiotherapy according to the last CT simulation image and were followed up until 90 days after the end of radiotherapy. There were 5 (23.8%), 2 (9.5%), 1 (4.8%) and 5 (23.8%), 3 (14.3%), 0 (0%) cases of grade I, II and III acute RP and RE, respectively.

Comparison of DVH Parameters Between the Implementation Plan and the Corresponding Simulation Plan
The comparison between the implementation plan and the corresponding simulation plan with the same primary tumour dose revealed the following. The lung V 20 , MLD, and MHD of the C 0 plan were significantly higher than those of the implementation plan and the C 40 plan. The lung V 5 , SC-MPD and oesophageal V 50 also tended to increase. The C 40 plan was similar to the implementation plan ( Table 4). The comparison between the implementation plan and the simulation plan with the same radiation damage control criteria revealed the following. The C 40 plan increased the radiotherapy dose from 63 ± 7 Gy at C 0 to 66 ± 7 Gy (P<0.001), and the implementation plan increased the radiotherapy dose to 68 ± 7 Gy (P<0.001). The radiotherapy dose of the C 40 plan was similar to that of the implementation plan (P=0.110).

DISCUSSION
The median survival time (MST) of patients with stage IV NSCLC who received three-dimensional radiotherapy to the   primary tumour combined with chemotherapy was prolonged to 16 months, and radiotherapy may play a very important role in prolonging OS based on the benefits of systemic therapy (21). Stereotactic ablative radiotherapy and stereotactic body radiotherapy to the primary tumour or metastases combined with EGFR-TKIs or first-line chemotherapy (for patients without EGFR mutations) significantly prolonged PFS and OS in patients with oligometastatic NSCLC (16,(24)(25)(26)(27)(28). Increasing the radiotherapy dose to the primary tumour was strongly associated with improved OS, and a radical radiation dose may be more beneficial for OS, especially in patients with oligometastases (10,21,24,26). Radiotherapy has become an important treatment for prolonging OS by reducing the failure rate of the primary tumour in patients with stage IV NSCLC (21). However, it is well known that the radical radiation dose can improve the LCR. However, the volume of the irradiated target area is an important factor that affects an increase in the tumour dose and controls radiation injury to normal tissues (27). The large irradiated volume leads to the phenomenon that radiation injury is aggravated by high doses to improve the LCR, or the LCR is reduced by low doses for fear of radiation injury. Therefore, reducing the volume of the irradiated tumour is the key to both increasing the dose and LCR and decreasing the incidence of radiation injury. However, the primary tumour is large in volume and scattered, and patients mainly have T 3-4 /N 2-3 disease according to research data (11,21,24). In some patients, when radiotherapy and EGFR-TKIs are started simultaneously, the purposes of both increasing the dose to the primary tumour and protecting normal tissues from radiation injury cannot be achieved because of the primary tumour volume. Therefore, this study was designed to take advantage of the ORR of EGFR-TKI treatment (>70%), disease control rate (>90%), and PFS (9-11 months) (29-31) based on the dosimetric property that a ≥15% shrinkage rate in the primary tumour volume can significantly reduce the low-dose volume to the whole lung and reduce radiation injury (27). Patients underwent SCTSs before EGFR-TKI treatment and every 10 days after treatment. The SCTS measurement was terminated, and then primary tumour radiotherapy began when the TVS in the latter simulation compared with the previous simulation was ≤5% or when the  time after treatment was 90 days. The aim was to investigate the timing of administering radiotherapy to the primary tumour to both increase the dose and LCR and reduce the probability of radiation injury. This study showed that although each patient had positive mutations in driver Genes, the rate and degree of tumour volume shrinkage after EGFR-TKI treatment were not consistent. Until the last SCTS, the maximum and minimum shrinkage rates were 78.1% and 18.8%, respectively. The most significant change in the average volume was within 40 days after the start of treatment. Thereafter, the average volume shrinkage rate slowed and was relatively stable at every 10-day interval. The total and average shrinkage rates from C 40 to C 70 were 8% and 2.67%, respectively. On day 50, the shrinkage rate increased slightly (by 3%) and continued to decrease thereafter. The regularity of TVC after EGFR-TKI treatment is that the volume shrinkage gradually slows the volume continues to shrink after increasing in some cases, and tumour shrinkage varies due to the different sensitivities of EGFR-TKI treatment in different patients (32). Therefore, it may be most beneficial to start radiotherapy at the time when the tumour volume continues to shrink to a low level after treatment and stabilizes without waiting until the disease progresses. In this study, the primary tumour volume was measured and compared separately at each 10-day interval. The results showed that the tumour volume shrinkage rates were significant and different within 40 days after the start of treatment. The tumour volumes from days C 40 to C 70 were similar and slow, and the tumour volume increased slightly on day 50 in 1 patient, which suggests that the speed of tumour volume shrinkage is different in each individual. For patients who receive EGFR-TKI treatment, a certain regularity of tumour volume shrinkage may be deduced, or 60 days may be the time to carry out radiotherapy by means of mathematical modelling (33), but an individualized analysis was not performed, and the actual pattern of TVC was not examined. Therefore, the current study shows that TVS was significant within 40 days after EGFR-TKI treatment and entered the stable phase after 40 days in most patients. The 40th day after EGFR-TKI treatment may be a reasonable time to administer radiotherapy to reach the goals of controlling tumours and reducing injury.
The simulated radiotherapy plan and its parameters represent the dose likely to control the primary tumour and the threshold to protect normal tissues from radiation injury (34,35), while the implementation radiotherapy plan and its parameters validate and summarize the efficacy for each individual tumour and the probability of radiation injury for normal tissues after a given dose of radiotherapy (36). Grade 2 and 3 acute RP and RE were observed in only a small number of patients treated with radiotherapy after the termination of SCTSs in this study, which suggests that the safety and efficacy of radiotherapy are acceptable under the premise of injury control criteria. The simulation radiotherapy plans for the primary tumour at C 0 and C 40 were designed at the same dose as the implementation plans of the corresponding patients. The DVH parameters of the 3 plans were compared. The results showed that compared to the C 0 simulation plan, the implementation plan and the C 40 plan significantly reduced the lung V 20 , MLD, and MHD. There was a trend of a significant reduction in the lung V5 and SC-MPD. The reduction in the lung V 20 may significantly reduce the occurrence of RP (28). There was a trend of a significant reduction in the oesophageal DVH of the implementation plans that may reduce the incidence of RE. Under the premise of the same control criteria of radiation injury, the radiation doses were compared among the implementation plan and the C 0 and C 40 simulation plans. The results showed that the implementation plan and C 40 simulation plan could significantly increase the tumour dose compared with the C 0 simulation plan and achieved a radical dose of more than 60 Gy, which not only improved the LCR but also did not increase radiation injury (34)(35)(36)(37). The implementation plan was similar to the C 40 simulation plan in both the tumour dose and DVH parameter regarding radiation injury protection. Therefore, it was further validated that it is a reasonable time to start primary tumour radiotherapy at 40 days after EGFR-TKI treatment in patients with EGFR-positive mutations.
In summary, the tumour volume shrinkage rate after EGFR-TKI treatment in patients with stage IV NSCLC with driver Gene-positive mutations gradually slowed over time and varied in each individual. The shrinkage rate was significant within 40 days after treatment and then entered the stable stage, and it may be the best time to start radiotherapy after 40 days of the initial treatment.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of Guizhou Cancer Hospital, GuiYang, China. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
QL, NL, XZ, YZ, WO, SS, ZM, YH, YG, and XC collected the data. BL conceived the study and participated in its design and coordination. BL and QL performed the statistical analysis and drafted the manuscript. All authors contributed to the article and approved the submitted version.