Based on the National Lung Screening Trial data, the adoption of low-dose computed tomography (CT) as a screening tool for lung cancer has increased worldwide [1]. As a result, the detection of early-stage lung cancer with ground-glass opacity (GGO) has increased significantly, up to 63% [2]. The TALENT study also demonstrated the effectiveness of low-dose CT screening in a predefined high-risk group of non-smokers and reported a detection rate of T0 lung cancer of up to 2.6% [3].

Most ground-glass nodules (GGN) or subsolid nodules found on chest CT are transient or have benign findings such as inflammation, fibrosis, or hemorrhage, but their malignant potential increases if they persist or increase in size. To identify tumors and adopt treatment at the optimal time, several academic societies have recently suggested guidelines for GGN or subsolid nodules, but no established statement has been published yet [4,5]. Most guidelines suggest determining the timing of routine follow-up according to the nodule size and recommend surgery if the total size and consolidation increase above certain criteria, but the underlying level of evidence is weak. Therefore, in the real world, when determining surgical indications, the experience of the surgeon or institution is often relied upon.

Meanwhile, promising results of stereotactic body radiotherapy (SBRT) have been revealed through pooled analyses of the STARS and ROSEL trials and the revised STARS trial for patients with early-stage lung cancer [6,7]. As a result, SBRT is being used as an alternative option for patients with early-stage lung cancer who are difficult to treat surgically [8]. However, limited reports have been published with GGO-predominant lesions.

Therefore, we performed a single-institution retrospective study to investigate the clinical outcome after SBRT in patients with early-stage lung cancer with GGO-predominant lesions.

Eighty-nine patients were found to be eligible for the analysis, and their characteristics are shown in Table 1. The median age was 72 years (range, 45 to 90 years), and 55% of the patients were men. Approximately 15.7% of patients had an Eastern Cooperative Oncology Group performance status score of 2 or 3, and 14.6% had an underlying lung disease such as chronic obstructive pulmonary disease, asthma, or interstitial lung disease. Fifty-one patients (57.3%) were previously diagnosed with lung cancer, and most of these patients underwent surgery, although some patients were treated with SBRT or a combination of SBRT and surgery. Seventy-two patients (80.9%) had single primary lung cancer, whereas 17 patients (19.1%) had multiple primary lung cancer, with up to four tumors identified in one patient. In the case of multifocal lesions, some were treated with multiple sessions of SBRT, and some were treated with a combination of surgery and SBRT; a few small, slow-growing lesions were observed. The characteristics of each lesion are shown in Table 2. In all cases, the necessity and possibility of biopsy and surgery were discussed in a multidisciplinary team meeting. Among them, 44 lesions (44.4%) were pathologically diagnosable, and 55 (55.6%) were clinically diagnosed as lung cancer through imaging tests such as serial chest CT and PET-CT. All lesions identified using histological examination were adenocarcinomas. Pure GGN was observed in 43 cases (43.4%), and lesions with subsolid nodule was 56 cases (56.6%). Median value of CTR was 0.3, ranging from 0.0 to 0.5.

The overall follow-up period for the study was median 33.0 months (range, 9.9 to 65.9 months). The follow-up period for tumor response was 23.2 months (range, 1.8 to 65.1 months), during which there was no local progression, and 100% LC was achieved. Any type of recurrence occurred in a total of five patients, and the characteristics of each patient are shown in Table 3. As for the initial pattern of failure, three patients (3.4%) had a regional recurrence, and two patients (2.2%) had a distant recurrence as the first event. It took an average of 12.6 months for the first recurrence to occur and be confirmed on CT and PET-CT images in all cases, and no pathologic confirmation was made. A distant recurrence of abdominal LNs occurred in one of the five patients after a regional recurrence, and all but one patient received systemic therapy as a salvage treatment. A total of eight patients expired, of whom five died of an unknown cause. Of the three patients whose causes of death could be determined, one died of pneumonia and sepsis, and the other two died of cardiac problems unrelated to cancer.

Three-year FFRR, DMFS, DFS, and OS rates were 96.3%, 93.0%, 92.6%, and 91.6%, respectively (Fig. 2A–D). A univariate analysis was performed, and the results are shown in Table 4. None of the risk factors associated with DFS were identified. Patients with a younger age or a higher level of diffusing capacity of the lung for carbon monoxide (DLCO) showed significantly improved OS. Female patients, patients with good performance status, and patients with high levels of forced vital capacity also seemed to have better survival outcomes, but there was no statistical significance. The results of the log-rank test with the Kaplan-Meier method were similar to those of the univariate analysis, showing low 3-year OS rates in the older population (age ≥ 75 years vs. < 75 years, 84.3% vs. 96.8%; p=0.007) and in those with low DLCO (DLCO ≥ 60% vs. < 60%, 96.4% vs. 62.5%; p=0.001) (Fig. 3A and B). Patients with subsolid nodules and pathologically confirmed nodules seemed to have worse survival rates, but this was not significant (Fig. 3C and D).

The toxicity results are shown in Table 5. No patients experienced grade ≥ 3 serious adverse effects for both acute and late toxicities. Five patients with grade 2 radiation pneumonitis (RP) were identified, but they recovered with oral steroid administration.

The results of the limited studies on the clinical outcomes of SBRT for GGO-predominant lung cancer lesions are shown in S1 Table [10–12]. Except for the study by Nagata et al. [11], all showed LC of 100%. In studies using photon beams, various doses were administered depending on the location and size of the tumor, but generally, approximately 50 Gy was prescribed. In the study by Onishi et al. [12], patients were recruited by defining GGO-predominant lung tumors with a CTR of ≤ 0.5 using the same criteria as ours, reporting an LC of 100% and a 3-year OS of 94.6%. Although not all reported rates of regional or distant recurrences are the same, in our case, the recurrence rate appears to be slightly higher than others. This may be because the proportion of patients with pathologically confirmed cancer was quite high at 44.4%, and patients with a history of lung cancer accounted for more than half at 57.3%. According to the subgroup analysis of the study by Onishi et al. [12], the 2-year distant failure rate for patients with a history of lung cancer was 25.0%. Additionally, in a surgical study published in Taiwan, only patients with stage I lung cancer were included, and patients with a higher pathologic T category had a higher chance of any recurrence, DM, and lower OS rates, and patients with a pathology of non-squamous cell carcinoma had a more distant recurrence [13]. Even though recurrences occurred in some patients in our study, all survived well with appropriate salvage therapy and are well managed. The PACIFIC-4 trial, which is currently in progress, is a study on the administration of durvalumab after definitive SBRT in patients with clinical stage I or II node-negative non–small cell lung cancer (NSCLC) [14]. If the results of the PACIFIC-4 trial are promising, they will provide a solid foundation for administering immunotherapeutic agents to patients such as those included in our study. Usually, after SBRT in early-stage lung cancer, RP grade ≥ 2 is known to occur in < 10% of cases, but as many as 29% cases have been reported [15,16]. Symptomatic RP was reported in 5% of participants in the present study, which may be due to a small target volume and advances in planning and targeting techniques.

The 5-year survival rate of patients with clinical stage I NSCLC is 68%–92% [17]. Regarding the long-term outcome for patients with GGO-predominant lung cancer lesions reported by Samsung Medical Center 2015, the 5-year survival rate after wedge resection was 98.6%, and recurrences occurred in 5.1% of patients, which was superior to the prognosis of typical patients with early-stage lung cancer [18]. Tomita et al. [19] published a study comparing the outcomes of SBRT and surgery in patients with stage I NSCLC, including some patients with GGN. There was no significant difference between the two groups in any of the oncologic outcomes, including local recurrence-free survival, DFS, cause-specific survival, or OS. Based on the results of a study of early-stage NSCLC, SBRT and surgery seem to yield comparable results for patients with GGN [6].

The greatest advantage of SBRT is that it can be a useful alternative as a non-invasive local therapy. Patients with insufficient remaining lung volume after multiple lung surgeries or with poor underlying lung function may benefit the most. However, SBRT also has some limitations. First, lobectomy with mediastinal LN sampling or dissection is still suggested as the standard of care. It has the advantage of being able to search for occult LNs surgically. However, occult LN metastasis is rare in GGO-predominant lung cancer lesions and almost always absent in pure nodules [20,21]. Second, many patients receive RT without pathologic confirmation. If salvage treatment is required to recurrence later, it may be difficult to use appropriate drugs without histological information. However, on the other hand, it takes quite a while for recurrence to occur, and since molecular or immunological status changes may occur during that time, it may be better to perform a biopsy again if necessary [22]. Third, evaluating the response of the treated lesion with certainty is often difficult. Owing to various presumed mechanisms, the density and size often increase in the post-SBRT CT image, and the treated nodule may be obscured because of fibrosis [23]. Similar results were obtained in a study of CT findings after SBRT for GGN, and a careful evaluation of the initial response was recommended [10].

The strength of this study is that we recruited a sufficient number of participants with GGO-predominant lung cancer lesions and had a consistent treatment policy as a single institution study. Furthermore, all cases were discussed at the tumor board, and decisions were made on the basis of multi-departmental consensus. As for the radiation dose, a sufficient dose was administered unless the tumor was adjacent to a major organ, such as a central tumor, and 88.9% of patients received an ablative biologically effective dose > 130 Gy₁₀. Several recent studies have shown that survival is improved with an ablative dose for lung cancer [24,25]. Therefore, the results of our institution’s study have a unique advantage in that they better reflect the latest opinion on total dose than those of other institutions (S1 Table).

There are some limitations to our study. First, despite the recommendation of some guidelines to postpone treatment for small lesions, even these lesions were irradiated [4,5]. However, studies have shown that lesions with increased size or density on serial chest CT have a high possibility of malignancy; considering this, we attempted to start treatment as soon as possible [26,27]. Nevertheless, even if some of them are classified as a high-risk group in imaging tests, they may have an indolent characteristic with slow growth and are routinely examined mainly by a department of pulmonary medicine. Unfortunately, when designing this study, we collected the data of patients who had received radiotherapy, so there was no pool of patients who had regular follow-up only without treatment, so direct comparison was not possible. Therefore, it would be useful to carry out comparative analysis with these patients in the future. The second limitation, as mentioned above, is the ambiguity in response evaluations due to changes in the lung parenchyma, such as fibrosis. Accordingly, there are efforts to assess the response using various other imaging techniques other than chest CT [9,28]. Finally, considering that lung cancer with an adenocarcinoma has a longer latency period to recurrence than squamous cell carcinoma, our follow-up period may seem short. This observation period can be explained by the fact that we included only patients treated within the last five years and that there were many elderlies who expired of causes unrelated to lung cancer. However, the RTOG 0618 study also had a follow-up period of about 48 months (range, 15.4 to 73.7 months) for patients with early-stage lung cancer who were operable, and compared to this, it was thought that our study period was not that insufficient [29]. If the next study is conducted over a longer period of time with more patients, it will be of great help in the future.

In conclusion, for patients with GGO-predominant lung cancer lesions, the application of SBRT has been shown to yield high LC and appears to be a safe treatment with low toxicity. SBRT is likely to be one of the alternatives to surgery in patients who have multiple lesions, a surgical history, or who are medically inoperable owing to age or comorbidities.