Loco-regional relapse pattern and timing after segmentectomy in patients with c-IA non-small cell lung cancer Authors: Hironobu Wada, Hidemi Suzuki, Takahide Toyoda, Yuki Sata, Terunaga Inage, Kazuhisa Tanaka, Yuichi Sakairi, Yukiko Matsui, Shigetoshi Yoshida, Ichiro Yoshino Published: 2024 ABSTRACT AbstractBackgroundSegmentectomy has been recognized as the standard procedure for small peripheral lung cancer; however, it has been shown that loco-regional relapse is more common with segmentectomy than with lobectomy. This study aims to investigate the long-term outcomes and loco-regional relapse patterns in patients with clinical stage IA (c-IA) non-small cell lung cancer (NSCLC) after segmentectomy and compare them with those after lobectomy.MethodsWe retrospectively compared the long-term outcomes of 115 patients who underwent segmentectomy for c-IA NSCLC with those of 292 patients who underwent lobectomy between January 2008 and December 2015. Segmentectomy was indicated intentionally or chosen in patients who were considered intolerable to lobectomy. New isolated growing lung lesions were defined as relapses if they were not diagnosed with a second primary lung cancer.ResultsThe median observation period was 2,150 days. The 10-year overall survival (OS) rates and relapse-free survival (RFS) rates of the two groups were similar: 79.4% and 68.7% for segmentectomy, and 68.2% and 61.2% for lobectomy. Even after propensity score matching, no significant differences were observed in the OS and RFS rates between the groups. The segmentectomy group had a higher loco-regional relapse rate (14% vs. 8%), including the surgical margin, remnant lobe, ipsilateral lung, mediastinal lymph node, and ipsilateral dissemination; however, no relapse was observed in the ipsilateral hilar lymph node. Loco-regional relapse occurred significantly later after segmentectomy than after lobectomy (median: 1,246 vs. 512 days, P=0.03), especially four years after segmentectomy. Loco-regional relapse occurred even when the tumor diameter was <1.0 cm. Most patients with loco-regional relapse had solid-dominant tumors.ConclusionsSegmentectomy, both intentional and compromised, showed comparable long-term outcomes to lobectomy; however, loco-regional relapse can develop in a later phase than lobectomy, requiring careful follow-up.Keywords: Non-small cell lung cancer (NSCLC), segmentectomy, lobectomy, loco-regional relapse Highlight box Key findings • Following long-term observations, the 10-year overall survival (OS) and relapse-free survival (RFS) after segmentectomy, including intentional and compromised, and lobectomy in patients with clinical stage IA (c-IA) non-small cell lung cancer (NSCLC) were similar: 79.4% and 68.7% for segmentectomy and 68.2% and 61.2% for lobectomy. Even after propensity score matching, the OS and RFS rates were not different between the groups (P=0.88 and P=0.56, respectively). Loco-regional relapse occurred more frequently and later after segmentectomy than after lobectomy, especially more than four years after segmentectomy. What is known and what is new? • Segmentectomy is a standard procedure for peripheral small NSCLC; however, large-scale clinical trials have shown that segmentectomy causes more loco-regional relapse than lobectomy, which is a great concern that must be resolved. • We demonstrated that loco-regional relapse occurred later in the segmentectomy group than in the lobectomy group, especially more than 4 years after surgery. Loco-regional relapse occurred even when the tumor diameter was <1.0 cm, and most of these patients had a solid-dominant tumor with a consolidation to tumor ratio >0.5. What is the implication, and what should change now? • We need to be concerned about loco-regional relapse in the late phase after segmentectomy in patients with solid-dominant c-IA NCSLC, even in smaller tumors (<1.0 cm). Patients with solid-dominant tumors, regardless of the tumor size, may require a long-term follow-up. Introduction In 2022, the results of two multicenter phase 3, randomized, non-inferiority trials [the Japan Clinical Oncology Group (JCOG) 0802/the West Japan Oncology Group (WJOG) 4607L (1) and the Cancer and Leukemia Group B (CALGB)/Alliance 140503 (2)] had a major impact on thoracic surgeons worldwide. Although the standard surgical procedure for non-small cell lung cancer (NSCLC) has been lobectomy with systemic lymph node dissection since a prospective, randomized study conducted by the Lung Cancer Study Group in 1995 (3), two recent clinical trials clearly demonstrated the superiority or non-inferiority of sublobar resection in terms of the overall survival (OS) compared to lobectomy when selecting peripherally located small NSCLC. In the JCOG0201 trial, radiologically non-invasive lung cancer, ≤2 cm in size with a consolidation to tumor ratio (CTR) of ≤0.25, was defined and recognized to have a favorable prognosis (4,5). Starting with this trial, various nationwide clinical trials were conducted in Japan based on the tumor diameter and CTR on preoperative computed tomography (CT). The JCOG0804/WJOG4507L (6) trial, including radiologically non-invasive lung cancer, demonstrated that sublobar resection of these lesions is extremely favorable even after a 10-year follow up (7). The JCOG1211 trial demonstrated favorable outcomes after segmentectomy for ground-glass opacity (GGO)-dominant lung cancers, including tumors ≤3 cm in size (8). The presence of a GGO component is also being increasingly considered as a reliable prognostic factor, even in radiologically invasive lung cancer with a CTR of ≥0.5 (9,10). The JCOG0802/WJOG4607L trial demonstrated that even in patients deemed to have an unfavorable prognosis due to the absence of GGO components, the OS was better with segmentectomy than with lobectomy (1). In terms of short-term postoperative outcomes, nationwide retrospective studies with a propensity score-matched analysis clarified that segmentectomy has been performed relatively safely compared with lobectomy (11,12). Given this background, the indications for sublobar resection, especially segmentectomy, are undoubtedly expanding and expected to become one of the standard surgical procedures in patients with early-stage NSCLC. However, loco-regional relapse following segmentectomy is a critical issue that needs to be resolved because it might have been avoided if lobectomy had been performed in certain cases. Few studies have investigated the detailed characteristics and risk factors for postoperative loco-regional relapse in patients with early-stage NSCLC who underwent segmentectomy compared to those who underwent lobectomy. Therefore, the present study investigated the long-term prognosis and relapse patterns of segmentectomy for patients with clinical stage IA (c-IA) NSCLC and discussed an appropriate follow-up strategy for detecting post-segmentectomy loco-regional relapse, the rate of which is sure to increase in the near future. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-783/rc). Methods Ethics statement This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the Institutional Review Board of Chiba University Hospital (No. HK202308-08). Owing to the retrospective nature of this study, the need to obtain written informed consent from each patient was waived. Study design and patient selection This was a retrospective study with a propensity score-matched analysis, demonstrating the OS, relapse-free survival (RFS), and proportion of loco-regional relapse after segmentectomy compared with those after lobectomy. Patients with c-IA NSCLC who underwent segmentectomy or lobectomy between January 2008 and December 2015 at the Chiba University Hospital were recruited. Clinical tumor, node, and metastasis (TNM) staging was performed using chest high-resolution CT (HRCT), brain magnetic resonance imaging, and 18F-fluorodeoxyglucose-position emission tomography (FDG-PET), using the seventh edition of the TNM classification by the International Association for the Study of Lung Cancer (13). The maximum tumor diameter was used for the pathological evaluation instead of the solid part diameter because of the retrospective nature of this study. Patients whose tumors showed pure GGO were excluded because they had a good prognosis and were classified as clinical stage 0 in the eighth edition. Patients with tumors >3 cm in diameter or multicentric lung cancers in the ipsilateral lung were excluded. Patients with contralateral isolated simultaneous lung cancers that could be curatively resected after the initial surgery were included. Indication of segmentectomy Tumor location, tumor diameter, solid component diameter, and CTR were assessed using HRCT images. A solid-dominant tumor was defined as a tumor with a radiologically solid component greater than the GGO component (CTR >0.5) and a GGO-dominant tumor was defined as a tumor with a CTR ≤0.5. We performed intentional segmentectomy based on previous studies (1,8,14,15) in patients with solid-dominant tumors ≤2 cm in size or GGO-dominant tumors ≤3 cm in size who could tolerate lobectomy as well as in patients in whom sufficient surgical margin distances (>2 cm or greater than the tumor diameter) could be secured. We also selected segmentectomy for patients who were unable to tolerate lobectomy because of an impaired vital organ function, comorbidities, poor performance status (PS based on the Eastern Cooperative Oncology Group scale; ≥2), or planned contralateral lung resection for simultaneous multiple lung cancers, expressed as compromised segmentectomy. Surgical procedure Almost all surgical procedures are performed via thoracotomy or hybrid video-assisted thoracic surgery. The segments to be resected were preoperatively determined based on the distance between the tumor and intersegmental vein on thin-section CT. Complete anatomical segmentectomy was performed (16) with dissection and division of the segmental bronchus, arteries and veins. Extended segmentectomy (17) and/or multi-segmentectomy was performed when the tumor was located close to the intersegmental plane. Intraoperatively, the inflation-deflation line and intersegmental pulmonary veins were used to identify the intersegmental plane. The intersegmental plane was divided using electrocautery and stapling. In some patients who underwent intentional segmentectomy, a pathological assessment of the hilar lymph nodes was performed intraoperatively, and if positive, the surgical mode was converted to lobectomy. The need for intraoperative pathological assessment was waived for patients who underwent compromised segmentectomy. When the surgical margins were determined to be insufficient intraoperatively, conversion to lobectomy or additional resection of the lung parenchyma was performed to secure sufficient surgical margins, based on the surgeon’s decision, if the patient was able to tolerate additional lung resection. Another option was that stapler cartridges used for pulmonary excision were lavaged intraoperatively to confirm the absence of cancer cells at the surgical margins. Follow-up Chest CT was performed at least every 6 months during the first 2 years and at least every 12 months from 2 years after surgery. Relapse was determined on the basis of radiographic features or histological evidence, and not all relapses were histologically proven. Local relapse was defined as the occurrence of relapse at the surgical margins of the lung, bronchus, or in a preserved lobe. Regional relapse was defined as an occurrence in an ipsilateral lobe other than the preserved lobe, hilar and mediastinal lymph nodes, including the contralateral side, and ipsilateral pleural space as pleural dissemination or malignant effusion. Other types of relapses were defined as distant relapses. Newly developed lung lesions in the lung parenchyma were considered relapses unless they showed histology or subtypes different from those of the primary lung carcinoma. Regarding metastatic diseases, in the absence of histological proof, such cases were considered relapse, not heterochronous multiple carcinomas. OS was defined as the time between surgery (segmentectomy or lobectomy) and death due to any cause. RFS was defined as the time between surgery for lung cancer and disease relapse or death from any cause, whichever occurred first. These outcomes were censored on the last day on which the patient survived. Statistical analysis Continuous and categorical variables were compared using the Mann-Whitney U and chi-squared tests, respectively. OS and RFS were estimated using the Kaplan-Meier method and compared using the log-rank test. Propensity scores were estimated using a logistic regression model that included age, sex, smoking index, forced vital capacity (FVC), forced expiratory volume in one second (FEV1), FEV1/FVC, tumor diameter, solid component diameter, CTR, and pathological staging. These propensity scores were then used to create a 1:1 matched cohort of segmentectomy and lobectomy patients with a caliper width equal to 0.2 of a standard deviation. The cumulative incidence of loco-regional relapse (CILR) was estimated using a cumulative incidence function that defined distant relapse and mortality without relapse before loco-regional relapse as competing events. The cumulative incidence of distant relapse (CIDR) was estimated similarly, loco-regional relapse and mortality without relapse before distant relapses as competing events. Patients with both loco-regional and distant relapses simultaneously were classified as having the distant relapse. In the CILR and CIDR analyses, patients were censored if they had survived without relapse at the last follow-up. The Fine and Gray method was used to compare the CILR and CIDR between groups. Two-sided p-values were calculated for all analyses and a P value <0.05 was considered statistically significant. CILR and CIDR analyses were performed using the EZR software program (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a modified version of R commander designed to add statistical functions frequently used in biostatistics (18). All OS and RFS curves were created using the EZR software program. Other statistical analyses were performed using the JMP software program, version 15 (SAS Institute Inc., Cary, NC, USA). Results Patients’ characteristics This study cohort included 115 patients who underwent segmentectomy and 292 who underwent lobectomy for c-IA NSCLC (Figure 1). The average age of this cohort was 67.2±9.3 years old. Of the 407 patients, 244 (60.0%) were men, 181 (44.5%) were aged ≥70 years, and 328 (80.6%) had adenocarcinomas. The median tumor and solid tumor diameters were 2.0 (range, 0.6–3.0) cm and 1.6 (range, 0.2–3.0) cm, respectively. The median CTR was 1.0 (range, 0.12–1.0), and 238 (58.5%) patients had a CTR of 1.0. In terms of pathological nodal status, 359 (88.2%) patients had pathological nodal factor (pN) 0, 29 (7.1%) had pN1 disease, and 19 (4.7%) had pN2 disease. The flow chart of the patients included in this study. c-IA, clinical stage IA; NSCLC, non-small cell lung cancer. Detailed segmentectomy procedures are shown in Table 1. Left upper division segmentectomy and superior segmentectomy (S6) on both sides were relatively common, accounting for 50.4% of cases. Complex segmentectomy was selected for 41.7% of patients, and a variety of operative procedures were performed. No single resection procedure was performed for segments 4, 5, and 7. Of 115 patients who underwent segmentectomy, 69 (60.0%) underwent intentional segmentectomy. Table 1Resected segments in patients who underwent segmentectomySurgical type of segmentectomySegmentNumber (%)Simple segmentectomy67 (58.3) RightS610 (8.7)S7/S8/S9/S102 (1.7) LeftS1+2/S339 (33.9)S4/S56 (5.2)S69 (7.8)S8/S9/S101 (0.9)Complex segmentectomy48 (41.7) RightS13 (2.6)S29 (7.8)S34 (3.5)S2/S31 (0.9)S2b/S3a1 (0.9)S83 (2.6)S91 (0.9)S6/S81 (0.9)S8/S92 (1.7)S9/S101 (0.9) LeftS1+25 (4.3)S36 (5.2)S1+2c/S4/S51 (0.9)S84 (3.5)S103 (2.6)S8/S91 (0.9)S9/S102 (1.7) Resected segments in patients who underwent segmentectomy A comparison of patient characteristics between segmentectomy and lobectomy is shown in Table 2. Age, sex, smoking index, and body mass index were similar between the groups. With respect to spirometry, the segmentectomy group had a significantly lower FEV1/FVC than the lobectomy group. The radiological parameters showed significant differences, demonstrating that segmentectomy was indicated for smaller tumors with a lower CTR. The distribution of primary lung tumor sites was significantly different between the two groups, as left upper division segmentectomy and right upper lobectomy were common in the segmentectomy and lobectomy groups, respectively. The pathological results showed that the lobectomy group contained significantly more advanced-stage patients with lymph node metastasis than did the segmentectomy group. Table 2Patient characteristics comparing segmentectomy with lobectomyCharacteristicsSegmentectomy (n=115)Lobectomy (n=292)P valueAge (years)68.1±9.466.8±9.30.21Sex, male67 (58.3)177 (60.6)0.66BMI (kg/m2)22.9±3.122.7±3.60.19Smoking index (pack-year)24.1±27.428.9±31.60.25preFVC (L)2.97±0.693.06±0.760.27preFEV1 (L)2.19±0.562.30±0.610.08FEV1/FVC (%)73.8±10.475.4±9.10.01Tumor size on CT (mm)17.9±4.920.6±5.3<0.001Solid size on CT (mm)12.9±6.117.3±6.7<0.001CTR (%)72.7±28.584.4±23.3<0.001c-T factor<0.001 T1mi16 (13.9)9 (3.1) T1a29 (25.2)46 (15.8) T1b54 (47.0)136 (46.6) T1c16 (13.9)101 (34.6)Location of primary tumor<0.001 Right upper lobe18 (15.7)90 (30.8) Right middle lobe027 (9.3) Right lower lobe20 (17.4)61 (20.9) Left upper lobe57 (49.6)62 (21.2) Left lower lobe20 (17.4)52 (17.8)Pathology0.51 Adenocarcinoma96 (83.5)232 (79.5) Squamous cell carcinoma12 (10.4)43 (14.7) Others7 (6.1)17 (5.8)p-N factor0.006 N0110 (95.7)249 (85.3) N13 (2.6)26 (8.9) N22 (1.7)17 (5.8)p-Stage0.01 IA94 (81.7)202 (69.2) IB15 (13.0)42 (14.4) II4 (3.5)30 (10.3) III–2 (1.7)18 (6.2)Data were expressed as mean ± standard deviation or n (%). BMI, body mass index; FVC, forced vital capacity; FEV1, forced expiratory volume in one second; CT, computed tomography; CTR, consolidation to tumor ratio; c-T, clinical T; p-N, pathological N; p-Stage, pathological stage. Patient characteristics comparing segmentectomy with lobectomy Data were expressed as mean ± standard deviation or n (%). BMI, body mass index; FVC, forced vital capacity; FEV1, forced expiratory volume in one second; CT, computed tomography; CTR, consolidation to tumor ratio; c-T, clinical T; p-N, pathological N; p-Stage, pathological stage. Survival analyses After a median follow-up of 5.9 years (2,150 days; range, 30–4,603 days), segmentectomy was not inferior to lobectomy in terms of both OS and RFS. The 5- and 10-year OS rates were 91.8% and 79.4%, respectively, after segmentectomy, and 89.9% and 68.2%, respectively, after lobectomy. The 5- and 10-year RFS rates were 79.8% and 68.7%, respectively, after segmentectomy, and 77.2% and 61.2%, respectively, after lobectomy (Figure 2A,2B). No 30-day mortalities were observed. When segmentectomy patients were further divided into two groups, intentional and compromised segmentectomy, although the OS of intentional segmentectomy appeared to be better, there was no significant difference among the three groups. In contrast, the RFS showed a significant difference among the three groups: the intentional segmentectomy group showed the most favorable RFS, and the compromised segmentectomy group showed the worst RFS among the three groups (Figure 2C,2D). The long-term outcomes after segmentectomy and lobectomy. The overall (A) and relapse-free (B) survival after segmentectomy and lobectomy. The overall (C) and relapse-free (D) survival among three groups, including intentional, compromised segmentectomy, and lobectomy. OS, overall survival; RFS, relapse-free survival; Seg, segmentectomy; Lob, lobectomy. Propensity score matching between the segmentectomy and the lobectomy groups was conducted to eliminate selection bias caused by confounding factors, and 99 patients in each group were matched. The characteristics of the matched patients are presented in Table S1. There were no statistically significant differences between the two groups in terms of radiological and pathological factors. Of the 99 patients in the segmentectomy group, 42 (42.4%) underwent compromised segmentectomy. Even after propensity score matching, OS and RFS after segmentectomy were similar to those after lobectomy. The 5- and 10-year OS rates were 93.7% and 79.6%, respectively, after segmentectomy, and 93.6% and 69.9%, respectively, after lobectomy. The 5- and 10-year RFS rates were 81.0% and 70.0%, respectively, after segmentectomy, and 83.4% and 66.2%, respectively, after lobectomy (Figure 3A,3B). The overall survival (A) and relapse-free survival (B) after segmentectomy and lobectomy using propensity score-matched analysis. OS, overall survival; RFS, relapse-free survival; Seg, segmentectomy; Lob, lobectomy. Relapse patterns The loco-regional and distant sites of first relapse are listed in Table 3. Loco-regional relapses were nearly 2-fold more frequent after segmentectomy than after lobectomy (13.9% vs. 7.9%). In contrast, distant relapses were confirmed more often after lobectomy than after segmentectomy (5.2% vs. 11.6%), thus leading to similar relapse rates in both the groups (19.1% after segmentectomy and 19.5% after lobectomy). With regard to local relapse, there were two patients with relapse on the intersegmental staple lines after compromised segmentectomy and two with isolated pulmonary nodules within the preserved lobes after intentional segmentectomy. There was no ipsilateral hilar lymph node relapse, and three patients (2.6%) had ipsilateral mediastinal lymph node relapse after segmentectomy, while 21 (7.2%) patients developed ipsilateral hilar or mediastinal lymph node relapse after lobectomy. Ipsilateral dissemination or pleural effusion occurred at a similar frequency in both the groups. All patients with first relapse at multiple sites were included in the lobectomy group. Table 3The first relapse sites after lung resection in patients with c-IA non-small cell lung cancerRelapse patternsSegmentectomy (n=115)Lobectomy (n=292)Relapse22 (19.1)57 (19.5) Local4 (3.5)1 (0.3) Loco-regional16 (13.9)23 (7.9) Distant6 (5.2)27 (9.2) Both*07 (2.4)Local relapse site Surgical margin2 (1.7)0 Bronchial stump01 (0.3) Remnant lobe2 (1.7) [2]#–Regional relapse site Ipsilateral lung3 (2.6)1 (0.3) Ipsilateral hilar LN04 (1.4) Ipsilateral mediastinal LN3 (2.6) [1]#17 (5.8) Contralateral hilar LN01 (0.3) Contralateral mediastinal LN2 (1.7)6 (2.1) Ipsilateral dissemination4 (3.5) [2]#4 (1.4) Ipsilateral pleural effusion05 (1.7)Distant relapse site Contralateral lung4 (3.5)13 (4.5) Bone2 (1.7)12 (4.1) Brain09 (3.1) Abdominal organs09 (3.1) Others01 (0.3)Data were expressed as n (%). *, both included simultaneous loco-regional and distant relapses. #, the numbers represent the number of patients with postoperative relapse after the intentional segmentectomy. Some patients showed multiple sites at the first relapse. Six patients with regional relapse after lobectomy had two different regional relapse sites. Of the seven with both regional and distant relapses, three showed two different regional relapse sites in addition to distant sites. Of the 34 patients with distant relapse after lobectomy, 7 showed multiple distant relapse organs. LN, lymph node; c-IA, clinical stage IA. The first relapse sites after lung resection in patients with c-IA non-small cell lung cancer Data were expressed as n (%). *, both included simultaneous loco-regional and distant relapses. #, the numbers represent the number of patients with postoperative relapse after the intentional segmentectomy. Some patients showed multiple sites at the first relapse. Six patients with regional relapse after lobectomy had two different regional relapse sites. Of the seven with both regional and distant relapses, three showed two different regional relapse sites in addition to distant sites. Of the 34 patients with distant relapse after lobectomy, 7 showed multiple distant relapse organs. LN, lymph node; c-IA, clinical stage IA. Time to first relapse The CILR and CIDR of the segmentectomy and lobectomy groups are shown in Figure 4. The CILR of segmentectomy overlapped with that of lobectomy up to 4 years postoperatively; however, after 4 years, the CILR of segmentectomy increased up to 17%, while that of lobectomy plateaued at 8%, although there were no significant differences between the groups. Of the 16 patients who developed loco-regional relapse after segmentectomy, loco-regional relapse was observed within 3 years in 8 patients (50%), and after 5 years in 3 patients (19%). In contrast, of the 23 patients who underwent lobectomy, 19 (83%) developed loco-regional relapse within 3 years, while no patients developed loco-regional relapse after 5 years. The median time to loco-regional relapse after segmentectomy was 1,245.5 days, which was significantly longer than that at 512 days after lobectomy (P=0.03). The cumulative incidence of loco-regional relapse (A) and distant relapse (B) after segmentectomy and lobectomy. CILR, cumulative incidence of loco-regional relapse; CI, confidence interval; Seg, segmentectomy; Lob, lobectomy; CIDR, cumulative incidence of distant relapse. Regarding the CIDR, the median times until the development of distant metastases after segmentectomy and lobectomy were 707.5 and 724 days, respectively, without a significant difference. Distant relapse that developed after 5 years included two patients (33%; isolated bone metastasis and isolated contralateral lung nodule in one each) who underwent segmentectomy and three patients (9%; multiple bone metastases with mediastinal lymph node metastases, isolated contralateral lung nodule, and multiple pulmonary metastases in one each) who underwent lobectomy. Loco-regional relapse after segmentectomy Detailed information on the patients with loco-regional relapse after segmentectomy is shown in Table 4. Three patients had a solid diameter of <1.0 cm, and only one patient was GGO-dominant. In the segmentectomy group, the CTR was significantly higher in patients with loco-regional relapse than in those without any relapse (median, 100% vs. 66.7%, P=0.01), while there was no marked difference in the tumor diameter (median, 18.5 vs. 18.0 mm, P=0.79) or solid diameter (median, 15.0 vs. 12.0 mm, P=0.07) (Tables S2-S4). Therefore, larger tumors did not necessarily develop more loco-regional relapses after segmentectomy. Of the 16 patients listed in Table 4, 7 (patients 1–7) might have avoided developing loco-regional relapses if lobectomy with systemic lymph node dissection had been performed. The median tumor diameter, solid diameter, and CTR were 1.8 cm, 1.6 cm, and 100%, respectively. Most tumors were pure-solid tumors >1.0 cm in diameter; however, a pure-solid tumor with a diameter of 0.9 cm developed surgical margin relapse. Of the four patients with pleural dissemination, three showed pathological pleural invasion beyond the elastic layer. Table 4Characteristics of patients who developed loco-regional relapse after segmentectomyNo.Age, years/sexRadiological evaluationOperation (intentional or compromised)PathologyRelapse site (time, years)Outcome (time, years)Tumor size, cmSolid size, cmCTR, %p-NplHistology172/F2.52.5100Rt S8 (compromised)N0pl0AdSurgical margin (3.9)Alive (11.2)272/M0.90.9100Rt S3 (compromised)N0pl0SqSurgical margin (5.7)Alive (9.1)350/M1.81.8100Rt S3 (intentional)N0pl0AdRemnant lobe (5.8)Alive (9.7)470/M1.91.263.2Lt S6 (intentional)N0pl1AdRemnant lobe (6.0)Dead (7.5)578/M1.61.6100Lt S8 (compromised)N1pl1AdIpsilateral MLN (3.0)Alive (5.9)669/M1.41.4100Rt S2 (intentional)N0pl0AdIpsilateral MLN (2.3)Dead (5.9)778/M2.12.1100Rt S7-10 (compromised)N0pl0SqIpsilateral MLN (0.7)Alive (2.2)877M1.21.2100Lt S6 (compromised)N0pl0SqIpsilateral lung (4.2)Dead (7.9)972/M1.31.3100Lt S1+2+3 (compromised)N0pl0AdIpsilateral lung (2.4)Dead* (2.8)1066/M2.11.780.9Lt S6 (compromised)N0pl0AdIpsilateral lung (4.5)Alive (5.5)1167/F2.52.5100Rt S2 (compromised)N0pl2AdContralateral MLN (1.2)Dead (8.1)1265/F1.20.866.7Lt S1+2 (compromised)N0pl0AdContralateral MLN (2.1)Alive (5.0)1369/F1.91.9100Lt S4+5 (intentional)N1pl1AdDissemination (3.8)Dead (6.9)1454/F1.60.637.5Rt S8+9 (intentional)N0pl0AdDissemination (0.8)Dead (2.0)1578/M2.01.155.0Lt S1+2 (compromised)N0pl1AdDissemination (4.9)Alive (6.6)1671/F1.91.9100Lt S1+2+3 (compromised)N0pl1AdDissemination (1.4)Alive (2.0)*, dead due to diseases other than lung cancer. CTR, consolidation tumor ratio; p-N, pathological N; pl, pleural invasion; M, male; F, female; Rt, right; Ad, adenocarcinoma; Sq, squamous cell carcinoma; Lt, left; MLN, mediastinal lymph node. Characteristics of patients who developed loco-regional relapse after segmentectomy *, dead due to diseases other than lung cancer. CTR, consolidation tumor ratio; p-N, pathological N; pl, pleural invasion; M, male; F, female; Rt, right; Ad, adenocarcinoma; Sq, squamous cell carcinoma; Lt, left; MLN, mediastinal lymph node. Discussion This study demonstrated equivalent OS and RFS of segmentectomy and lobectomy in patients with c-IA (≤3 cm) NSCLC, even in cohorts that included patients who underwent compromised segmentectomy. These results are based on long-term follow-up and may offer the potential for segmentectomy to be considered as an effective surgical procedure for c-IA NSCLC. In contrast, the high frequency of loco-regional relapse after segmentectomy was similar to that reported in previous studies (1-3). Most patients with loco-regional relapse had a solid-dominant tumor. The time to loco-regional relapse was longer after segmentectomy than that after lobectomy. Therefore, loco-regional relapse after segmentectomy in the late phase, especially after 4 years, should be considered when the tumor is solid-dominant, even if it is small (<1.0 cm). These findings may contribute to the development of an appropriate follow-up strategy for segmentectomy. The classification of c-IA (≤3 cm) NSCLC originating in Japan, based on tumor diameter and CTR, has been helpful in determining the treatment strategy for lung cancer (1,6-8). However, the efficacy of segmentectomy for solid-dominant tumors >2 cm in size remains unclear. Some retrospective studies have demonstrated comparable outcomes of segmentectomy compared to lobectomy among 2- to 3-cm solid-dominant lung cancers with (19,20) or without (21,22) GGO components. In our study, OS after compromised segmentectomy, which included many 2- to 3-cm solid-dominant lung cancers, overlapped with that after lobectomy. These results indicate the efficacy of segmentectomy for these tumors. Currently, a multicenter, randomized, controlled phase III trial is underway in Japan to confirm the non-inferiority of segmentectomy to lobectomy in patients with clinical stage IA3 pure-solid peripheral NSCLC (23). Further evidence for segmentectomy will be established for the treatment of more challenging lung cancers. Several previous studies have reported late-onset loco-regional relapse beyond five years after segmentectomy (24-26). Nakao et al. reported that 4 of 26 patients (15%) with lepidic growth-dominant adenocarcinomas treated with wedge resection developed surgical margin relapses >5 years after surgery (27). Nomori et al. also reported that all five surgical margins relapses among 179 patients (3%) who underwent segmentectomy developed beyond 5 years after the procedure, and four of them were seen in adenocarcinomas with a lepidic pattern (24). They cautioned that local relapses were observed even in GGO-dominant lung cancer and that long-term (>5 years) monitoring is required. In our study, we observed more loco-regional relapses four or more years after segmentectomy than after lobectomy, which we believe is due to the slow-growing nature of the tumor treated with segmentectomy compared to that treated with lobectomy. In the 16 patients with loco-regional relapse after segmentectomy, the solid diameter (median, 15 vs. 20 mm, P=0.041) was significantly smaller than that after lobectomy in 23 patients. The key to preventing loco-regional relapses is to secure sufficient surgical margins, which requires appropriate selection of the surgical mode based on careful preoperative simulation and accurate intraoperative identification of the intersegmental lines. Three-dimensional CT is widely used for preoperative simulations. It can aid thoracic surgeons in identifying affected veins that run between segments and perceiving the precise distance between a tumor and intersegmental margins (28,29). If the surgical margin of the tumor is insufficient in the simulation, then lobectomy should be performed instead of segmentectomy. Otherwise, extended segmentectomy (17), or additional segmentectomy or subsegmentectomy should be considered to secure sufficient margins. The precise identification of intersegmental veins is crucial in the area of the hilum, and various techniques are used in the peripheral region, including the inflation-deflation method with selected aeration to the affected bronchus (30), indocyanine green endobronchial (31,32) and intravenous injection (33) with near-infrared imaging. Recently, virtual-assisted lung mapping (34) and radiofrequency identification (35) demonstrated their efficacy in localizing the tumor and guaranteeing the surgical margin. The optimal surgical margin, considered to be >2 cm or the maximum tumor diameter (36-38), was incorporated into the inclusion criteria of large-scale prospective clinical trials (1,8). Sawabata et al. demonstrated that malignant positive margins were seldom found when the margin distance was greater than the maximum tumor diameter on the cross-section, without staple removal. The results also showed that no cancer cells were observed at the surgical margins when the margin distance was >2 cm (36,37). Morimoto et al. reported that all free tumor clusters, which were considered risk factors for loco-regional relapse, were observed within the diameter of the primary tumor from the edge in patients with micropapillary adenocarcinoma and concluded that adequate surgical margins should be equal to or exceed the diameter of the primary tumor for limited resection (39). Soh et al. demonstrated that cT1c lung cancer with a CTR of 1 showed no significant difference in OS and disease-free survival (DFS) rates between segmentectomy and lobectomy, whereas cT1c with a CTR of 0.5 to 1 showed a significantly lower OS and DFS after segmentectomy than after lobectomy (21). This may indicate that cT1c tumors with a CTR of 0.5 to 1, which can be >3cm in size, may not obtain a sufficient surgical margin after segmentectomy. In contrast, the optimal surgical margin distance for GGO-dominant lung cancer is likely to differ from that for solid-dominant lung cancer. In the JCOG0804/WJOG4507L study, which enrolled lung cancer with a diameter ≤2 cm and CTR ≤0.25, the macroscopic surgical margin mandated ≥5 mm from the tumor to the parenchymal staple or cut line (6,7). The results showed excellent local control with only one patient (0.3%) showing local relapse at the resection line after 10 years of follow up (7). This suggests that radiologically non-invasive lung cancer might not require >2 cm or the tumor diameter to obtain sufficient surgical margins. In the present study, surgical margin relapse developed in two compromised patients. In patient 1, the tumor was pure-solid, 2.5 cm in size, and was located in the middle part of S8. After staple removal, pathology revealed only a 1.5 mm between the intersegmental resection margin and tumor edge. The more centrally localized the tumor, the more difficult it was to secure surgical margins. Right S9 segmentectomy may have been necessary in this case. In patient 2, although the tumor was located 8 mm from the intersegmental vein (V2c), the vein was preserved when right S3 segmentectomy was performed, which might have caused insufficient surgical margins and local relapse. When the tumor is located close to the intersegmental vein, then the vein should be sacrificed to secure sufficient surgical margins, as has long been advocated as extended segmentectomy (17). Because both patients who developed surgical margin relapse underwent compromised segmentectomy, re-resection was not performed to obtain satisfactory pathological margins after surgery. Interestingly, segmentectomy did not increase ipsilateral hilar lymph node relapse in patients, even though intrapulmonary lymph nodes in the remnant lung segment could not be resected during segmentectomy. We previously reported that intrapulmonary nodal metastasis to the outside of the tumor-bearing segments is rarely observed when the tumor is free of extrapulmonary nodal metastasis and it is associated with non-pure-solid, or peripheral small (≤2 cm) lesions (40). This was encouraged by the recently published supplemental analysis of JCOG0802/WJOG4607L, which argued that mediastinal lymph node dissection could be avoided in patients with non-pure-solid tumors (41). Recent study also indicated that lymph node dissection could be omitted in cases with small (≤3 cm) GGO-dominant tumors (42). In contrast, as many as 10% of pure-solid tumors have lymph node metastases, including nonadjacent interlobar lymph node metastases (41). Because part-solid and pure-solid tumors have different malignant behaviors, the extent of lymph node dissection should be determined depending on tumor characteristics. There is no doubt that appropriate intraoperative inspection of extrapulmonary lymph nodes is crucial for patients with a solid-dominant tumor. Several limitations of the present study warrant mention. First, since this was a single-institutional retrospective study, preoperative selection bias could not be eliminated even after conducting a propensity score-matched analysis. Clinical factors that reflect the malignant grade of tumors, such as standardized uptake values on FDG-PET, and those related to local relapses, such as the location of tumors, whether peripheral or central, were not investigated. Second, the extent of lymph node dissection varied from case to case regardless of surgical mode, which might have affected pathological staging and postoperative lymph nodal relapses. It does not follow that mediastinal lymph node metastases could have been avoided if lobectomy had been performed instead of segmentectomy. Third, neither the existence of lymphovascular and pleural invasion, spread through air space, nor the final surgical margin distance on the pathological specimen were examined in all patients, which might have influenced the outcomes. Furthermore, the patients received a variety of postoperative adjuvant therapies based on the pathological results. Thus, the survival analysis as well as the relapse pattern analysis might have been affected by differences in postoperative treatments. Finally, since not all loco-regional relapses were confirmed by a biopsy, we could not completely exclude false-positive diagnoses, including second primary lung cancer and lymph node metastases. Overestimation of the postoperative relapse rate could not be ruled out. Conclusions Segmentectomy, both intentional and compromised, showed comparable long-term outcomes to lobectomy; however, loco-regional relapse developed later than lobectomy, especially four or more years after segmentectomy, in patients with a solid-dominant tumor. The intensity and length of follow-up may be determined according to the CTR. Patients with solid-dominant tumors, regardless of tumor size, may require a long-term careful follow-up. Supplementary The article’s supplementary files as Acknowledgments We would like to thank Japan Medical Communication (www.japan-mc.co.jp) for editing this manuscript. Funding: None. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the Institutional Review Board of Chiba University Hospital (No. HK202308-08). Owing to the retrospective nature of this study, the need to obtain written informed consent from each patient was waived. Footnotes Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-783/rc Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-783/coif). I.Y. reported consulting fees from Medicaroid, Covidien, Johnson and Johnson, and payment or honoraria for lectures, presentations from Covidien, Johnson and Johnson, Intuitive Surgical, Astra Zeneca, Chugai Pharmaceutical, Taiho Pharmaceutical, and MSD. The other authors have no conflicts of interest to declare. Data Sharing Statement Available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-783/dss References Available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-783/prf Supplementary Materials The article’s supplementary files as Data Availability Statement Available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-783/dss