Prognostic Impact of Reactive Oxygen Species Modulator 1 in Surgically Resected Epidermal Growth Factor Receptor-Mutant Lung Adenocarcinomas

Lung cancer is the leading cause of cancer-related mortality globally, with 1.8 million deaths attributed to lung cancer in 2020 worldwide [1]. In Korea, approximately 18,500 lung cancer-related deaths were recorded in 2021 [2, 3]. The high mortality rate of lung cancer is owing to the fact that most patients were diagnosed at an advanced stage and that the recurrence rates are higher than those of other cancers after curative resection even for early stage disease [4]. In a recent real-world study using the SEER-Medicare database, recurrence-free survival rates at 5 years were as low as 29.3% in patients with stage I–IIIA non-small-cell lung cancer (NSCLC) after curative surgery [5]. Although investigations to identify more effective therapeutic approaches have been widely performed to reduce this high mortality, the survival gain of neoadjuvant and adjuvant chemotherapy for early resected NSCLC was only 5% at 5 years, as evident in historical meta-analyses [6, 7].

Epidermal growth factor receptor (EGFR) mutations are one of the major driving genetic alterations in NSCLC, occurring in approximately 15% of European and US patients and up to 50% of Asian patients with lung adenocarcinoma [8, 9]. Owing to the clinical application of EGFR-tyrosine kinase inhibitors (TKIs), the discovery of primary and secondary resistance mechanisms, and the development of newer-generation TKIs, the survival of EGFR-mutant patients has been significantly prolonged over the past decades [10‒12]. For resected early-stage diseases, the presence of EGFR mutations is associated with favorable clinical outcomes compared to those with wild-type EGFR [13, 14]. However, most patients with EGFR-mutant lung cancer eventually experience recurrence [14]. Thus, it is important to predict recurrence and select high-risk patients before surgical resection to improve the clinical outcomes of those groups of patients.

Reactive oxygen species modulator 1 (Romo1) is a mitochondrial membrane protein involved in intracellular reactive oxygen species (ROS) production [15]. Romo1-induced ROS promotes the proliferation of both cancer and normal cells, as well as resistance to anticancer treatment [16, 17]. Since the first study reporting the relationship between Romo1 overexpression and poor prognosis in hepatocellular carcinoma [18], similar results have been reported in other cancers, including gastric, bladder, and colorectal cancers [19‒21]. Interestingly, Romo1 expression is significantly higher in lung cancer tissues than in the surrounding normal cells, and serum Romo1 levels are higher in patients with lung cancer than in healthy individuals [22]. In addition, Romo1 overexpression in tumor tissues was significantly associated with poor survival in surgically resected NSCLC patients and was related to poor response and clinical outcomes in patients with advanced NSCLC treated with platinum-based chemotherapy [23, 24]. These data suggest that Romo1 is a potential predictive and prognostic biomarker for various human malignancies. However, there are limited data on its clinical implications in cancers harboring driving genetic alterations. In this study, we aimed to investigate whether Romo1 expression is associated with clinical outcomes of the patients with EGFR-mutant lung cancer who received surgical resection.

We retrospectively enrolled patients with EGFR-positive lung adenocarcinoma who underwent curative surgery at Kyung Hee University Hospital, a referral hospital in South Korea, between July 2008 and June 2021. Patients who lacked sufficient survival data, those who underwent neoadjuvant chemotherapy or radiotherapy, had an incomplete resection, with a history of other cancers within 5 years before the diagnosis of lung cancer, or harbored other driving genetic alterations, including ROS proto-oncogene 1 and anaplastic lymphoma kinase fusions, were excluded from the study.

Computed tomography (CT), ¹⁸F-fluorodeoxyglucose positron emission tomography-CT, and brain magnetic resonance imaging were performed for staging workup. Tumor-node-metastasis (TNM) staging was established based on the 8th edition of the American Joint Commission on Cancer TNM staging system for NSCLC [4]. Patients with stage IIA or higher disease and high-risk stage IB patients with poorly differentiated tumors, tumor size >4 cm, vascular invasion, visceral pleural involvement, those who underwent wedge resection, or incomplete lymph node sampling received adjuvant chemotherapy according to the clinical guidelines. Clinical data were obtained retrospectively by reviewing electronic medical records. This study was reviewed and approved by the Institutional Review Board of Kyung Hee University Hospital (KHUH 2022-03-033). Written informed consent was obtained from all participants who were alive. All the studies were conducted in compliance with the principles of the Declaration of Helsinki.

EGFR mutation testing was performed on the tumor samples. Sections with 5-μm-thickness were prepared from formalin-fixed, paraffin-embedded tissue, and genomic DNA was extracted using the High Pure Template Preparation kit (Roche Applied Science, Mannheim, Germany) and was stored at −20°C until analysis. PANA Mutyper™ (PANAGENE Inc., Daejeon, South Korea), a PNA-clamping-based EGFR mutation detection kit, was used to detect the mutations. The primer sets included mutations or deletions in exons 18–21 that encode the tyrosine kinase domain of EGFR. The results were interpreted according to the manufacturer’s instructions.

Romo1 protein expression was evaluated by immunohistochemical staining of tissue specimens. A BOND-MAX Immunoautostainer (Leica Biosystems, Newcastle, UK) was used for the staining. Formalin-fixed paraffin-embedded tumor tissue specimens were sectioned into 4-μm sections and deparaffinized with Bond Dewax Solution (Leica Biosystems). Antigen retrieval was performed by heating slides at 98°C for 20 min using Epitope Retrieval Solution 1 (Leica Biosystems) and cooling for 10 min in 0.01 m citrate buffer (pH 6.0). The slides were washed in distilled water, and endogenous peroxidase activity was blocked using a Bond Polymer Refine Detection Kit (Leica Biosystems, Newcastle, UK) for 5 min. After washing, the slides were placed in Tris-buffered saline and incubated for 30 min with Romo1 monoclonal antibody (OriGene Technologies, Rockville, MD, USA) at 1:200 dilution. Subsequently, sections were developed with 3,3′-diaminobenzidine chromogen solution for 7 min, counterstained with hematoxylin, and dehydrated. Positive and negative controls were defined using human colon adenocarcinoma tissues and the exclusion of the primary antibody, respectively.

Blinded to the clinical information, a pathologist (K Na) assessed Romo1 staining under a light microscope at ×40–200 magnification and interpreted it as positive when cytoplasmic staining was observed. The staining intensity was graded as 0 (no staining), 1 (weak), 2 (distinct), or 3 (strong), and positive staining was quantified based on the percentage of tumor cells with a specific staining intensity. The final histochemical scores (H scores) were calculated by multiplying the percentage of tumor cells by the staining intensity (possible range, 0–300).

The cutoff H score between low and high Romo1 expression levels was defined as the point with the lowest p value, according to the log-rank test for all possible H scores. Comparison of Romo1 expression levels according to clinicopathological parameters was performed using the Wilcoxon rank-sum test, and comparison of the proportions of high and low Romo1 expression within each parameter was performed using the χ² or Fisher’s exact tests, as appropriate. Disease-free survival (DFS) was defined as the period from the day of surgery until recurrence or death. Data from patients without tumor recurrence or death were censored at the last follow-up. The relationship between these parameters and DFS was estimated using univariate and multivariate analyses with the log-rank test, followed by Cox proportional hazards regression analysis. Parameters with p values <0.1 in the univariate analysis were further evaluated using multivariate analysis. Statistical significance was set at p < 0.05. Survival rates were estimated using the Kaplan-Meier method. Statistical analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA) and R (version 4.2.0; R Foundation for Statistical Computing, Vienna, Austria).

During the study period, 915 patients underwent surgical resection for lung cancer. Of these, 125 were EGFR-mutant adenocarcinomas. Ten patients with inadequate-quality tissue samples, 6 patients with insufficient survival data, five with concomitant cancers, four who received neoadjuvant treatment, and two without R0 resection were excluded. Ultimately, 98 patients were included in the analysis.

Table 1 summarizes the clinical characteristics of the study population. All subjects were Korean, and their median age was 63 years (range, 42–87 years). Thirty-five (35.8%) were aged ≥70 years, 71 (72.4%) were female, and 77 (78.6%) had never smoked. Seventy-seven (78.6%) patients had an Eastern Cooperative Oncology Group performance status of 0 or 1. Eighty-four (85.7%) patients had stage I or II disease and 14 (14.3%) had stage IIIA disease. Eighty (81.6%) patients had tumors with well-to-moderate differentiation, and 85 (86.7%) had lymphovascular invasion in their tumors. Forty-nine (50.0%) patients had a deletion mutation in exon 19 (19del), 45 (45.9%) had an L858R point mutation in exon 21, and 4 (4.1%) had uncommon mutations, including S768i, g719A, and L861q. Seventy-three (74.5%) patients underwent lobectomy, and 30 (30.7%) received adjuvant chemotherapy using a platinum doublet: 6 for stage IB, 10 for stage II, and 14 for stage IIIA disease.

Representative sections with different H scores are shown in Figure 1. Because Romo1 is located in the mitochondrial membrane, it was primarily localized in the cytoplasm of cancer cells, as expected. The median H score of the study population was 140 (range, 10–300, Fig. 1d).

To investigate which clinicopathological parameters were associated with the level of Romo1 expression, we compared the median Romo1 H score between the groups for each parameter. The median H score was significantly higher in smokers, patients with more advanced stages, and patients who had tumors with lymphovascular invasion (all p < 0.05, online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000540521). We compared the distribution of patients according to the Romo1 group. The optimal cutoff H score for low and high Romo1 expression was determined at 200 using the log-rank test. Using this cutoff, 73 (74.5%) and 25 (25.5%) patients were classified into the low and high Romo1 groups, respectively. As shown in Table 1, Romo1 expression was significantly associated with advanced-stage disease (p = 0.0103) and lymphovascular invasion (p = 0.0147).

The median follow-up period for our study subjects was 71.5 months (range, 10.1–96.0 months). Survival analysis results with respect to clinicopathological parameters are summarized in Table 2.

The median DFS for all study subjects was 72.1 months (range, 36.9–104.8 months). In the univariate analysis, poor differentiation, lymphovascular invasion, advanced stage, and adjuvant chemotherapy showed significantly shorter DFS (all p < 0.05). In addition, high Romo1 expression was significantly associated with a shorter DFS (36.0 vs. 73.1 months, p = 0.0376). Multivariate analysis showed that advanced stage (hazard ratio [HR] = 2.75, 95% confidence interval [CI]: 1.04–4.37, p = 0.0247), adjuvant chemotherapy (HR = 1.95, 95% CI: 1.12–4.64, p = 0.0413), and high Romo1 expression (HR = 2.18, 95% CI: 1.10–4.32, p = 0.0261) were independently associated with poor DFS. Kaplan-Meier survival curves showed that patients with Romo1 overexpression were likely to have poor survival in terms of DFS (Fig. 2a).

We exploratively investigated whether the association between Romo1 expression and DFS could apply to patients with different mutational subtypes. For the patients with 19del, poor differentiation, lymphovascular invasion, advanced stage, and high Romo1 expression showed significantly shorter DFS in univariate analysis (all p < 0.05). Multivariate analysis showed that advanced stage (HR = 2.24, 95% CI: 1.11–4.18, p = 0.0123) and high Romo1 expression (HR = 2.09, 95% CI: 1.01–3.74, p = 0.0287) were independently associated with poor DFS. For the patients with L858R, the advanced-stage disease (HR = 2.10, 95% CI: 1.34–3.96, p = 0.0367) was the only factor that was significantly associated with DFS in the multivariate analysis. Kaplan-Meier survival curves showed that patients with Romo1 overexpression had poor DFS in those with 19del (Fig. 2b) but not in those with L858R (Fig. 2c).

The present study demonstrated that Romo1 overexpression is significantly associated with early recurrence in EGFR-mutated lung adenocarcinomas treated with curative resection. To our knowledge, this is the first study to demonstrate the association between Romo1 expression and clinical outcomes in early-stage EGFR-mutant NSCLC. The current data highlight that Romo1 could be a potential prognostic biomarker for this group of patients and that the Romo1-high population may be a distinct subgroup warranting active surveillance and a tailored therapeutic approach.

Romo1 is a novel protein that was first identified in a patient with head and neck cancer who was resistant to chemotherapy after recurrence [25]. Studies have revealed that Romo1 is a major regulator of intracellular ROS production and that Romo1-induced ROS play a critical role in cell proliferation and chemoresistance [16, 26]. Interestingly, clinical studies have consistently demonstrated that Romo1 overexpression is associated with poor clinical outcomes in a variety of human malignancies, including lung cancer. High Romo1 expression in tumor tissue is significantly associated with poor DFS in patients who underwent surgical resection as well as poor progression-free survival and overall survival (OS) in patients who received chemotherapy for early and advanced NSCLC, respectively [23, 24]. Lee et al. recently reported that Romo1 overexpression was significantly associated with poor progression-free survival and OS in patients with EGFR-mutant lung adenocarcinoma treated with frontline targeted therapy and was associated with the less frequent development of secondary T790M mutations after TKI failure [27]. Our data are in accordance with previous results, suggesting that the Romo1 expression can be prognostic in both early and advanced EGFR-mutant NSCLC.

The exact mechanism by which Romo1 is involved in the early recurrence of surgically resected EGFR-mutant tumors is unclear; however, it can be partly attributed to the invasiveness and aggressiveness of cancer cells which is caused by Romo1-induced intracellular ROS production [18]. Previous data have suggested that Romo1 overexpression is associated with vascular invasion or lymph node metastasis in patients with hepatocellular carcinoma and colorectal cancer [18, 19]. In addition, our data showed that Romo1 expression was associated with smoking and lymphovascular invasion, which are well-known factors associated with poor prognosis in lung cancer. Upregulated ROS level is one of the characteristics of cancer cells, and increased ROS levels are critically involved in carcinogenesis, metastasis, and chemoresistance [28, 29]. Evidence suggests that ROS play a critical role in the resistance to targeted therapy. Oxidative stress induces TKI resistance of EGFR-mutant cells by activating phosphorylation and impairing the dimerization of EGFR [30, 31]. In addition, ROS play a role in the process of epithelial-mesenchymal transition and the activation of Src via the Raf/ERK signaling pathway [32, 33], all of which are critical mechanisms of TKI resistance. TKI can also affect mitochondrial function; morphological changes in mitochondria and enhanced intracellular oxidative stress have been observed after long-term exposure to EGFR-TKIs [27, 34]. Moreover, TKI resistance can be alleviated by antioxidants that inhibit the ROS-induced aberrant phosphorylation of EGFR and AKT [27]. Taken together, enhanced ROS production in the Romo1-upregulated tumor can be one of the reasons for the invasiveness and aggressiveness of EGFR-mutant tumors.

Of note, DFS was significantly shorter in patients who received adjuvant treatment, suggesting a high recurrence rate in those groups of patients. This was an unexpected finding, however, and is consistent with a previous report [35]. Lee et al. investigated the prognostic factors using tissues from 131 patients who had complete resection of stages I–IIIA EGFR-mutated lung adenocarcinomas [35]. In that study, adjuvant chemotherapy is one of the independent factors to predict early recurrence as shown in the present study. The reasons for the association between adjuvant treatment and poor prognosis are unclear; however, it can be attributed to the more frequent administration of adjuvant treatment in patients with advanced-stage disease, and more importantly, adjuvant chemotherapy was not effective in preventing recurrence. Thus, the previous data and ours suggest a more effective treatment strategy for reducing relapse and improving clinical outcomes in patients with EGFR-mutant lung cancer. ADJUVANT-CTONG1104 phase III trial showed that adjuvant gefitinib treatment resulted in significantly longer DFS than chemotherapy in patients with completely resected stages II–IIIA EGFR-mutant NSCLC [36]. However, in the final analysis, this DFS advantage did not translate to a significant difference in OS [37]. In a recent ADAURA trial, adjuvant osimertinib for the same patients group showed a significant reduction of recurrence and also proved OS benefit in the overall population (HR = 0.49; 95.03% CI: 0.34–0.70; p < 0.001) and patients with stages II to IIIA (HR = 0.49; 95.03% CI: 0.33–0.73; p < 0.001) [38]. However, considering the less prominent DFS benefit in some subgroups (smokers, patients with stage IB disease, and patients with L858R mutation) and the cost-effectiveness of long-term treatment duration (3 years) of adjuvant osimertinib [14], a more precise and appropriate selection of high-risk patients using biomarkers is essential for the application of adjuvant EGFR-TKIs.

The present data suggest that high Romo1 expression confers a distinct aggressive phenotype that may require different treatment strategies. Based on Romo1 expression, more active application of adjuvant TKI even in stage IB disease and other therapeutic strategies, such as adjuvant TKI plus chemotherapy combination in high-risk patients, might be feasible, as evidenced by the FLAURA2 trial, in which osimertinib plus platinum doublet showed better clinical outcomes than osimertinib alone in advanced EGFR-mutant NSCLC [39]. Large-scale prospective studies are required to determine the optimal treatment strategy for patients with different Romo1 expression levels.

Our study had several limitations. First, it was a small, retrospective study conducted at a single center. To reduce possible selection bias, we included the most eligible patients over approximately 12 years in our institution, collected several pathological parameters, and provided long-term follow-up data (>5 years). Second, as the postoperative use of osimertinib was approved in early 2021 in Korea, the clinical impact of adjuvant TKI could not be evaluated. Third, we did not simultaneously collect other specimens, including serum. Previous studies have demonstrated that serum Romo1 levels correlate well with tissue Romo1 expression and are associated with poor clinical outcomes in resected early-stage NSCLC [22, 40]. The baseline levels and dynamics of serum Romo1 will be an interesting area for future studies. Finally, we did not explore other factors that could be associated with recurrence, including the tissue immune phenotype or coexisting genomic profiles. To address these issues, we conducted a comprehensive prospective study to identify prognostic factors, including serum Romo1, immune signatures, and concurrent genomic alterations, in patients with EGFR-mutant NSCLC.

In conclusion, the present data indicate that Romo1 overexpression is associated with the early recurrence of surgically resected EGFR-mutant lung adenocarcinoma. Our findings suggest that Romo1 overexpression may confer a distinctive aggressive phenotype to EGFR-mutant tumors. In addition, future studies should focus on whether these findings can be applied to other malignancies harboring driving genetic alterations and optimal treatment strategies for Romo1-overexpressed, EGFR-mutant tumors. Large prospective studies validating our data may facilitate the clinical use of Romo1 expression for risk stratification and prognostic prediction in these patient populations.

This study was conducted in accordance with the principles of the Declaration of Helsinki and the Ethical Guidelines for Clinical Studies in Korea, with the approval of the Institutional Review Board of Kyung Hee University Hospital (KHUH 2022-03-033). Written informed consent was obtained from all surviving patients, and the requirement for deceased patients was waived.

The authors declare no conflict of interest.

This study was supported by the Basic Science Research Program of the National Research Foundation of Korea and funded by the Ministry of Science, ICT, and Future Planning (No. 2023R1A2C1003763).

Seung Hyeun Lee contributed to the conception and design of the work. Tae-Woo Kim and Seung Hyeun Lee drafted the manuscript. All authors were involved in the acquisition, analysis, and interpretation of data. All authors reviewed and approved the final version of the manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.