Safety and Efficacy of Single-Fraction Carbon-Ion Radiotherapy for Early-Stage Lung Cancer with Interstitial Pneumonia

Simple Summary Patients with lung cancer complicated by interstitial pneumonia (IP) often lose treatment options early because of concerns about acute exacerbation of IP. Carbon-ion radiotherapy (CIRT) is expected to provide both superior tumor control and low toxicity owing to its superior dose concentration; however, it is not yet a well-established therapy. In this study, we confirmed that 50 Gy single-fraction CIRT can be performed even in IP-complicated lung cancer with acceptable efficacy and tolerability. Lung dose is a significant predictor of overall survival, indicating the need for further efforts to reduce lung dose. Abstract Patients with lung cancer complicated by interstitial pneumonia (IP) often lose treatment options early owing to acute exacerbation of IP concerns. Carbon-ion radiotherapy (CIRT) can provide superior tumor control and low toxicity at high dose concentrations. We conducted a retrospective analysis of the efficacy and tolerability of a single-fraction CIRT using 50 Gy for IP-complicated lung cancer. The study included 50 consecutive patients treated between April 2013 and September 2022, whose clinical stage of lung cancer (UICC 7th edition) was 1A:1B:2A:2B = 32:13:4:1. Of these, 32 (64%) showed usual interstitial pneumonia patterns. With a median follow-up of 23.5 months, the 3-year overall survival (OS), cause-specific survival, and local control rates were 45.0, 75.4, and 77.8%, respectively. The median lung V5 and V20 were 10.0 and 5.2%, respectively (mean lung dose, 2.6 Gy). The lung dose, especially lung V20, showed a strong association with OS (p = 0.0012). Grade ≥ 2 pneumonia was present in six patients (13%), including two (4%) with suspected grade 5. CIRT can provide a relatively safe and curative treatment for patients with IP-complicated lung cancer. However, IP increases the risk of severe radiation pneumonitis, and further studies are required to assess the appropriate indications.


Introduction
Lung cancer (LC) is often associated with interstitial pneumonia (IP).Up to 20% of patients with idiopathic pulmonary fibrosis (IPF), the most common form of IP, are reported to develop LC, with an estimated five-fold incidence [1][2][3].However, treatment options for IP-complicated LC are limited, even in the earliest stages, because of concerns that all treatment modalities, such as surgery, chemoimmunotherapy, and radiation therapy (RT), can induce acute exacerbations of IPF [4][5][6][7][8][9][10][11][12].RT for early-stage LC, particularly stereotactic body RT (SBRT) for non-small cell LC (NSCLC), has proven to be an effective and well-tolerated treatment for patients who are medically inoperable.However, most clinical trials have excluded cases of pre-existing IP, making reaching a consensus on the safety of SBRT in such patients challenging.The risk factors and their borderlines remain unclear, and treatment for such patients has not improved.
Conversely, particle therapy has a favorable dose concentration among radiation modalities [13,14], and when treating lung tumors, the dose for normal lungs could be minimized.Therefore, even for IP-complicated LC, which is difficult to treat with other modalities, curative treatment can be provided with relative safety [15,16].In particular, carbon-ion radiation therapy (CIRT) exhibits biological effects because of its high linear energy transfer radiation [17].The National Institute of Radiological Sciences, currently the QST Hospital, initiated CIRT for NSCLC in 1994.Based on the results of dose-escalation studies, a single irradiation protocol of 50 Gy (described as the relative biological effect [RBE]-weighted dose based on the modified microdosimetric kinetic model) has been employed for early-stage LC [18] and has achieved extremely high safety and efficacy in cases without IP [19][20][21].However, even for CIRT, limited clinical data on its safety are available for patients with IP complications.In this study, we evaluated the toxicity and efficacy of single-irradiation CIRT for NSCLC complicated by IP, especially analyzing the risk factors for acute exacerbation of IP.

Study Design and Patient Selection
Consecutive patients treated with CIRT for NSCLC complicated by IP between April 2013 and September 2022 were identified and retrospectively analyzed.The treatment protocols and procedures were approved by our institution's ethics committee, and written informed consent was obtained from all patients who participated in the study.The inclusion criteria for the study were as follows: (1) histologically or pathologically confirmed NSCLC treated with single-fraction CIRT with 50 Gy, (2) clinical stage 1-2 (UICC7th) with no nodal or distant metastasis, (3) medically inoperable or refused surgery, (4) Eastern Cooperative Oncology Group Performance Status (PS) 0-2, and (5) patients who met the above criteria and were clinically or pathologically diagnosed with IP.Central tumors were not included in this analysis, as they were treated using a different protocol with 12 fractions due to safety concerns.The availability of treatment with other modalities (surgery, photon SBRT), acceptability of CIRT, and treatment strategies were determined by a multidisciplinary cancer board, which included thoracic surgeons and physicians.Screening and diagnosis of IP were conducted at referral institutions prior to or in parallel with the diagnosis of LC and were also reviewed at our institution based on clinical and laboratory findings.The usual interstitial pneumonia (UIP) pattern was diagnosed using high-resolution computed tomography (CT) images, according to the new international guidelines for IPF diagnosis [22].Two radiation oncologists and one radiologist made the final diagnosis.The GAP model (and the revised GAP model used in Japan; rGAP model) for predicting the prognosis of IPF [23,24] was also used to evaluate the patients.

Carbon-Ion Radiotherapy Procedure
The details of the CIRT planning and delivery at our institution have been described previously [18,19,[25][26][27][28].The carbon ion dose was calculated by multiplying the physical dose by the RBE and was expressed in Gy.A fixed dose of 50 Gy was delivered in a single fraction through a series of 2-4 fixed ports.Treatment planning for CIRT was conducted using four-dimensional (4D) CT at 1-2-mm intervals.The gross tumor volume (GTV) was delineated, including the lung tumor at the lung window.The clinical target volume (CTV) was created by adding a margin of 0.5-1.0cm to the GTV.In addition, we defined the beam field-specific target volume (FTV) by extending the 3D treatment planning technique to 4D [29] and a setup margin of 2-3 mm to create the planning target volume (PTV).The total dose was applied to the isocenter and tuned to cover the PTV with a 95% isodose line of the prescribed dose.For organs at risk (OARs), dose constraints were set based on our previous clinical trials [18] and strictly adhered to the following criteria: spine (Dmax) < 10 Gy, esophagus (D0.2cc) < 10 Gy, and mainstem bronchus (D2cc) < 30 Gy, taking priority over the target coverage.No dose constraints on the lung doses were defined.

Follow-Up
Post-CIRT follow-up consisted of physical examination, blood tests, chest radiography, and contrast-enhanced CT performed every 3 months for the first 2 years and thereafter, at least every 6 months.In cases where continuous examinations at our hospital were difficult, the latest medical reports and diagnostic images were sent to us.Local recurrence was defined as a progressive abnormality on CT.In cases of suspected local recurrence, 18F-fluorodeoxyglucose positron emission tomography was performed, and when feasible, a biopsy was performed.Acute and late radiation pneumonitis (RP) was assessed using the Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0 [29].Acute AEs were defined as AEs occurring within 3 months and late AEs as those occurring later.Many patients had respiratory symptoms before CIRT; therefore, their symptoms were graded before and after CIRT, and the onset or progression of symptoms after treatment was evaluated.Mild progression (addition or change of medication due to exacerbation) was classified as grade 2, and severe exacerbation (need for inpatient treatment and introduction of home oxygen therapy [HOT]) was classified as grade 3. Grades 4 and 5 were the same as those used in the common evaluation.

Chart Review
The doses to the target and OARs were evaluated using dose-volume histogram (DVH) analysis.In this study, as indicators of the dose to normal lungs (lungs-GTV), the mean total lung dose (MLD total) and percentage of lung volume irradiated above 5/10/20 Gy (lungs V5/V10/V20) were retrieved.Overall survival (OS), cancer-specific survival (CSS), local control, and progression-free survival (PFS) were analyzed using the Kaplan-Meier method, and subgroups were compared using log-rank statistics.Cox proportional hazards regression methods were used for risk factor analysis.Fisher's exact test was used to examine the independence of events and risk factors.Each outcome was calculated from the date of CIRT initiation to the date of the event or the last follow-up date.Statistical analyses were performed using R software version 4.2.3 (https://www.r-project.org/,accessed on 23 December 2023), and statistical significance was set at p < 0.05.Each indicator of respiratory function and serum marker was evaluated based on borderline values.

Patient Characteristics
We identified patients with NSCLC who received single-fraction CIRT at 50 Gy between April 2013 and September 2022.Among these, 50 cases complicated by IP were included in this study.Patient characteristics are summarized in Table 1.The median age of the patients was 76 (interquartile range [IQR] 71-81) years, and the male-to-female ratio was 45:5.The clinical stage was 1A:1B:2A:2B = 32:13:4:1.All participants were clinically or pathologically diagnosed with IP.Of these, 32 (64%) showed UIP patterns and were diagnosed with IPF, and 9 (18%) were suspected to be associated with autoimmune diseases (mainly rheumatoid arthritis, other dermatomyositis, and anti-neutrophil cytoplasmic antibody: ANCA-related diseases).Eleven (22%) and five (10%) patients were on corticosteroids and HOT, respectively, and seven (14%) had a history of acute exacerbations.The proportions of patients with GAP stage 1:2:3 and rGAP stage 1:2:3 were 26:19:5 and 22:22:6, respectively.

Treatment Outcomes
The median follow-up time was 23.5 (IQR 10-47.5)months for all patients and 28 (IQR 9-47) months for survivors.At the final follow-up, 35 patients (70%) had died, with 12 of those deaths attributed to other causes.Of the 23 deaths from other causes, 2 were related to AEs, 13 were respiratory-related (pneumonia, respiratory failure) deaths excluding those related to AEs, and 8 were due to other causes (2 cardiac diseases, 2 other cancers, 1 gastric ulcer, and 3 unknown).
DVH data were evaluated for all patients.Table 2 presents the treatment characteristics of the patients.The lung dose was overestimated in four patients because part of the lung was out of range for the planning CT.The median GTV volume was 7.15 (IQR 3.80-15.08)cm 3 .In all cases, the GTV dose was within the prescribed range of ±5%.The doses delivered to the main OARs are listed in Table 2.For the lungs, the lung V5, V10, and V20 were 10.0 (IQR 6.49-12.12)%,7.59 (IQR 5.23-10.23)%and 5.18 (IQR 3.72-6.92)%,respectively.The MLD was 2.63 (IQR 1.87-3.18)Gy.

Prognostic Factors for OS
Table 3 summarizes the results of the univariate analysis of the factors associated with OS.Factors known to be associated with OS in early-stage LC, including age, PS, and cancer stage, were not significantly predictive of OS in this study.In addition, the parameters reported to be useful in predicting the prognosis of interstitial pneumonia, such as causative diseases, serum markers, and respiratory function, were not significantly associated with OS.In contrast, regarding dosimetric variables, V5, V20, and MLD showed significant associations in univariate analysis, and multivariate analysis indicated that V20 remained a significant risk factor.(p = 0.0012).The association between GAP/rGAP stage and OS was not significant (p = 0.28/0.58).

Adverse Events
AEs were discussed in 49 of the 50 patients, excluding one patient who died of the primary disease within 2 months.Of the 49 patients in the present study, RP of grades 2, 3, and 5 were encountered in two patients (4.1%) each, totaling six patients (13%) (Table 4).Figure 2 shows the pretreatment thin-slice CT and dose distribution of the two patients with grade 5 pneumonia.The median duration of onset was 3 months (range, 1-6 months).Two patients who developed grade 5 RP presented with symptoms within 1 month of starting CIRT, indicating that early onset should be carefully monitored as a sign of severe pneumonia.Both patients died of respiratory failure 5 months after treatment and multiple steroid pulse therapies.No tumor recurrence was observed in either of the patients.Five of 32 patients (15.6%) with a UIP pattern developed grade 2 or higher RP, whereas only one (5.9%) of 17 patients without a UIP pattern.Statistical evaluation failed to show any significant correlation between RP and any of the patient backgrounds, serum markers, respiratory function, or lung irradiation doses previously reported to be associated with RP.When patients were classified using the rGAP model, RP tended to be present in patients with a higher rGAP stage (Table 5).multiple steroid pulse therapies.No tumor recurrence was observed in either of the pa tients.Five of 32 patients (15.6%) with a UIP pattern developed grade 2 or higher RP whereas only one (5.9%) of 17 patients without a UIP pattern.Statistical evaluation failed to show any significant correlation between RP and any of the patient backgrounds, serum markers, respiratory function, or lung irradiation doses previously reported to be associ ated with RP.When patients were classified using the rGAP model, RP tended to be pre sent in patients with a higher rGAP stage (Table 5).4) are shown.Both patients had a UIP pattern and clinical stage 1B (tumor diameter 39 and 31 mm, respectively).Case 5 had a mild IP, rGAP-stage 1, but a relatively higher lung dose (lungV5/V20 of 19.18%/8.07%,respectively).Case 6, on the other hand, was a pa tient with severe IP in rGAP-stage 3 and on HOT but with a lower lung dose (lungV5/V20 8.64%/5.48%,respectively).The red, orange, magenta, green, blue, and purple lines show dose level of 95, 90, 70, 50, 30, and 10%, respectively; CTV is drawn in yellow.4) are shown.Both patients had a UIP pattern and clinical stage 1B (tumor diameter 39 and 31 mm, respectively).Case 5 had a mild IP, rGAP-stage 1, but a relatively higher lung dose (lungV5/V20 of 19.18%/8.07%,respectively).Case 6, on the other hand, was a patient with severe IP in rGAP-stage 3 and on HOT but with a lower lung dose (lungV5/V20 8.64%/5.48%,respectively).The red, orange, magenta, green, blue, and purple lines show dose levels of 95, 90, 70, 50, 30, and 10%, respectively; CTV is drawn in yellow.

Radiation Therapy for IP-Complicated LC
This retrospective study evaluated the efficacy and safety of a single CIRT in 50 patients with IP-complicated localized LC.This is one of the largest reports on radical irradiation in patients with IP and, to our knowledge, the first report of single-carbon-beam irradiation.
In photon radiotherapy, ILD complications have long been identified as risk factors for severe RP [30][31][32][33][34][35].In particular, patients diagnosed with IP are at a high risk of fatal RP and are considered unsuitable for irradiation, resulting in limited reports [10,[36][37][38].Yamashita et al. reported that seven of nine patients with IP (78%) who underwent SBRT with 48 Gy in four fractions developed grade 4-5 pneumonia [10].Tsurugai et al. reported that SBRT for 42 idiopathic interstitial pneumonias complicated by LCs resulted in grade ≥ 3 RP in 12% of patients [36].These results indicate that the risk of fatal RP is considerably higher in patients with IP complications, even with SBRT for stage I NSCLC.
In contrast, particle therapy, that is, proton beam therapy (PBT) and CIRT, has been increasingly reported to have a relatively low risk of irradiation in cases with pulmonary interstitial changes [26,[39][40][41][42][43].Among them, studies focusing on IP are still scarce; however, Hashimoto et al. reported in 2019 that 29 patients with IP-complicated LC (including 10 patients with UIP pattern) underwent PBT, resulting in four patients (13.8%) with grade 2-3 RP [40].Noh JM et al. reported in 2020 the results of radical PBT in 54 patients with IP-complicated LC, resulting in seven patients (13.0%) of grade ≥ 3 pneumonia [41].The present study found grade ≥ 2 RP in 6 of 50 patients (12%).This is comparable to previous reports on particle therapy for IP-complicated LC, suggesting that our single CIRT is not inferior in safety (Table 6) [10,36,40,41].In contrast, in our previous report on Cancers 2024, 16, 562 9 of 14 the same regimen for early-stage LC without IP, none of the 57 patients had G2 or higher pneumonitis [19].

Patient Background Related to Severe RP
The UIP pattern on high-resolution computed tomography (HRCT) is a well-known risk factor for AEs in the natural history of IPF [44,45] and an important predictor of pneumonitis due to cancer treatment, including radiotherapy [38,46,47].In this study, the incidence of grade ≥ 2 pneumonia was approximately 2.7 times higher in the group with UIP pattern than in those without it (16% in UIP vs. 6% in non-UIP).This supports previous reports on the association between the UIP pattern and IP severity.Furthermore, the only statistically significant risk factor for pneumonia in this study was pretreatment transcutaneous oxygen saturation (SpO 2 < 95%).This is a brief indicator of the arterial partial pressure of oxygen (PaO 2 < 80%) used in the Japanese IPF severity classification [48].This is consistent with previous reports and the clinical sense that higher IPF severity is associated with a higher risk of severe pneumonia [35,[49][50][51].
Other reported parameters include patient background, such as age and sex, collagen disease, IP history (corticosteroids, HOT, and AEs), and several respiratory function indicators [33,52,53].Serum markers have been studied, particularly in the field of radiotherapy, with KL-6, SP-D, and CRP levels being the most common [10,27,54,55].These indicators were not significantly associated with the severity of pneumonia in the present study.This could be partly due to bias and the small sample size; however, it may also be a result of the complexity of the multiple factors involved in the development of pneumonia.Therefore, attempts have been made to score multiple risk factors for each treatment modality.The GAP/rGAP model evaluated in this study tended to be related to OS.With a larger number of cases, estimating the risk of radiotherapy by modifying or combining existing scores for other treatment modalities is possible.

Importance of Lung Doses
Previous reports have often identified lung dose (V2-25, MLD) as a risk factor for severe pneumonia [31,[55][56][57].Furthermore, the two references on PBT cited in the previous section reported inferior OS in patients with a UIP pattern/IPF diagnosis.In this study, lung doses were not observed to be significantly associated with severe pneumonia; however, lungV20 (<5%) was strongly associated with OS, and other lung dose parameters tended to be similar.One reason for this may be that the lung dose is a comprehensive indicator of tumor size, shape, and location.At the same time, radiation-induced lung fibrosis and loss of respiratory function potentially affect the long-term prognosis of patients with IP, even if they do not lead to severe pneumonia or AEs.In other words, in radiotherapy, especially in cases with IP complications, the reduction of lung doses is an important issue not only for controlling AEs but also for the prognosis and quality of life.Particle therapy, especially CIRT, has excellent depth-dose characteristics known as Bragg peaks.It allows the creation of a region called the spread-out Bragg peak (SOBP) adapted to the depth and shape of the treatment target with only several beam angles, minimizing dose scattering to the nearby OAR and increasing the dose to the target [13].As a result, CIRT enables reduced lung doses, particularly in the low-dose area, which is probably related to lower toxicity in CIRT for IP-complicated LCs [10,[32][33][34]37].Clearly, IP complications are associated with severe pneumonia, even in CIRT, compared with the results using the same regimen for non-IP cases [19].In addition, two cases of grade 5 RP occurred in the present study, but the underlying risk remains to be clarified.In contrast, 21 patients who met both the diagnosis of rGAP-stage 1 and treatment with lung doses lower than the set criteria (V5 < 10%, V20 < 5%, and MLD < 3 Gy), including those showing a UIP pattern, experienced no grade ≥ 2 pneumonitis.To determine the indications for irradiation in IP cases, we should consider the background factors and dose constraints from both sides or in combination.
Although not directly applicable to photon SBRT owing to in biological effects and the number of fractions, experience with CIRT could be applied to radiotherapy in general by indicating relatively safe subjects and dose criteria for IP-complicated LC.

Treatment Effect
In radiotherapy for ILD/IP-complicated LC, IP itself has a significant impact on prognosis [10,[32][33][34]38,58].In the present study, mortality from other causes was higher than that from cancer.The 3-year OS was 45%, which was significantly inferior to our results with the same protocol for non-IP cases [19].This is a common trend in reports of radiotherapy for IP-complicated LC.Among Japanese multicenter retrospective studies, the study on SBRT reported a 3-year OS of 42.6% and that on CIRT 48.2% [33,39].Other studies have similarly reported relatively poor OS after treatment of IP-complicated LC, regardless of the radiation modality [38,40,41].Meanwhile, it is clear from several reports that local control of radiotherapy is also reduced in IP-complicated cases [26,40,42,43].The underlying mechanisms remain unclear, but it has been reported that chronic inflammation and fibrosis could be related to poor tumor response [59].Also, the difficulty in distinguishing the tumor from the surrounding interstitial shadows could have led to inaccurate targeting and post-treatment assessment, or efforts to reduce lung irradiation might have been negatively reflected.In patients with IP complications, many factors should be fully considered, including the prognosis of IP, the negative effects of the treatment, and the patient's preferences; then, a careful assessment should be made of the feasibility of the treatment intervention.

Limitation
This study had some limitations.First, the study was conducted at a single institution and had a retrospective design.Therefore, some bias may have been inherent.In addition, CIRT is not an accessible treatment due to its ubiquity and high treatment costs.Second, patients with IP are at a higher risk of undergoing biopsy for LC diagnosis, and in the majority of cases, clinical rather than pathological diagnosis is used.Interstitial pneumonia was also based on clinical diagnosis and was incompletely classified.Therefore, all cases were discussed with our cancer board regarding treatment indications and methods before proceeding with the treatment.Third, the small number of patients may have limited the statistical reliability of the results.Fourth, the safety of single irradiation, which we use in our hospital, has not been established for other modalities at present, making pure inter-modality comparisons difficult.Also, methods to compare the biological effects of CIRT and photon SBRT have not yet been established.Further studies with large sample sizes and basic research are needed to clarify the risk factors for severe RP and the role of CIRT in IP-complicated LCs.

Conclusions
We report our experience with a 50 Gy single CIRT for localized IP-complicated LC.CIRT has the potential to provide a relatively safe and curative treatment for patients who were previously considered difficult to treat with RT.However, IP complications increase the risk of serious RP, as reported in the present study, and are certainly high-risk even

Figure 2 .
Figure 2. Pretreatment CT images and CIRT dose distribution of two patients with grade 5 radiation pneumonitis (cases 5 and 6 in Table4) are shown.Both patients had a UIP pattern and clinical stage 1B (tumor diameter 39 and 31 mm, respectively).Case 5 had a mild IP, rGAP-stage 1, but a relatively higher lung dose (lungV5/V20 of 19.18%/8.07%,respectively).Case 6, on the other hand, was a pa tient with severe IP in rGAP-stage 3 and on HOT but with a lower lung dose (lungV5/V20 8.64%/5.48%,respectively).The red, orange, magenta, green, blue, and purple lines show dose level of 95, 90, 70, 50, 30, and 10%, respectively; CTV is drawn in yellow.

Figure 2 .
Figure 2. Pretreatment CT images and CIRT dose distribution of two patients with grade 5 radiation pneumonitis (cases 5 and 6 in Table4) are shown.Both patients had a UIP pattern and clinical stage 1B (tumor diameter 39 and 31 mm, respectively).Case 5 had a mild IP, rGAP-stage 1, but a relatively higher lung dose (lungV5/V20 of 19.18%/8.07%,respectively).Case 6, on the other hand, was a patient with severe IP in rGAP-stage 3 and on HOT but with a lower lung dose (lungV5/V20 8.64%/5.48%,respectively).The red, orange, magenta, green, blue, and purple lines show dose levels of 95, 90, 70, 50, 30, and 10%, respectively; CTV is drawn in yellow.

Table 3 .
Patients and treatment factors associated with OS.

Table 4 .
Pre-irradiation conditions and characteristics of six patients with G2-5 RP.

Table 4 .
Pre-irradiation conditions and characteristics of six patients with G2-5 RP.

Table 6 .
Previous studies on RP in patients with IP.