BACKGROUND:The standard treatment for resectable non-small cell lung cancer (NSCLC) has traditionally been upfront surgery, followed by adjuvant therapy according to pathological stage. Although recent trials have demonstrated the superiority of surgery after neoadjuvant chemoimmunotherapy over chemotherapy alone, studies directly comparing it with upfront surgery remain lacking. This trial aims to determine the optimal perioperative treatment strategy for resectable NSCLC. METHODS:This is a multicenter, randomized, phase III trial designed to assess the superiority of neoadjuvant chemoimmunotherapy over upfront surgery followed by adjuvant therapy in terms of overall survival in patients with resectable clinical stage II-III NSCLC. Patients are randomized in a 1:1 ratio to receive either (1) upfront surgery followed by cisplatin-based chemotherapy (every 3 weeks for 4 cycles) and immune checkpoint inhibitor monotherapy (up to 1 year for patients with PD-L1 TPS≥1%) or (2) neoadjuvant nivolumab plus platinum-based chemotherapy (every 3 weeks for 3 cycles) followed by surgery. The primary endpoint is the overall survival. Secondary endpoints include progression-free survival, time to distant metastasis, objective response rate to neoadjuvant therapy, and adverse events. A total of 330 patients will be enrolled over 5 years. The trial was initiated in March 2025 and registered in the Japan Registry of Clinical Trials (study number: jRCTs031240747). CONCLUSIONS:As immunotherapy has become an integral component of NSCLC management in the perioperative setting, this trial is expected to provide critical evidence to refine optimal perioperative treatment strategies for patients with resectable stage II-III NSCLC.

We have previously reported radiological patterns of precursor lesions of lung squamous cell carcinoma (SqCCs) using computed tomography. SqCCs were found to develop from five unique precursor patterns. We histologically analyzed 61 cases from the previous cohort to evaluate whether these precursor patterns reflect possible developmental pathways. Of the precursor patterns, the endobronchial group demonstrated involvement of large bronchi in the hilar region. In the subsolid nodule pattern group, which exhibited ground-glass opacities radiologically, the tumors showed intra-alveolar wall infiltration, nuclear inclusion/vacuoles, low mitotic count, and absent atypical mitosis. In two of the seven cases in this group, lepidic growth, corresponding to ground-glass opacity, occurred at the tumor periphery. In the cyst wall thickening group, the tumors were located subpleurally, characterized by papillary growth structures. In the multiple-nodule group, all tumors displayed elastofibrous collapse and intra-alveolar wall infiltration. Overall and recurrence-free survival varied across the progression patterns. The radiological progression patterns were reflected in the histological features, highlighting the unique characteristics of subsolid and multiple nodule groups corresponding to peripheral-type SqCC. This study represented a new approach to the assessment of tumor evolutional pathways with histological features in combination with temporal radiographic changes.

INTRODUCTION:We aimed to clarify the prognostic impact of adjuvant chemotherapy intensity in patients with high-grade neuroendocrine carcinoma (HGNEC) of the lung using a post hoc analysis of the phase III JCOG1205/1206 trial comparing irinotecan plus cisplatin (IP) with etoposide plus cisplatin (EP). MATERIAL AND METHODS:We included 217 patients (109 and 108 received EP and IP, respectively) with p-stage I-IIIA HGNEC. Prognosis was assessed based on treatment completion (four cycles) and relative dose intensity (RDI) of adjuvant chemotherapy, calculated for each regimen and agent, using Cox proportional hazards models. RESULTS:Median regimen RDI was similar (87.2% vs. 89.6%), whereas treatment completion rate was higher in EP (89.0% vs. 74.1%; P = 0.0046). Five-year relapse-free survival (RFS) among patients who completed versus did not complete treatment was 66.0% versus 73.3% in EP arm (hazard ratio [HR], 1.944; 95% CI, 0.590-6.404) and 67.4% versus 64.1% in IP arm (HR, 0.874; 95% CI, 0.437-1.745). RDI ≥90% was associated with worse RFS in EP arm (HR, 2.024; 95% CI, 1.081-3.787) but not in IP arm (HR, 1.123; 95% CI, 0.608-2.072). In EP arm, cisplatin and etoposide RDI ≥90% were associated with worse RFS (HR, 1.926; 95% CI, 1.029-3.605 and HR, 2.024; 95% CI, 1.081-3.787, respectively), whereas no significant associations were observed for cisplatin or irinotecan in the IP arm. CONCLUSION:In completely resected HGNEC, high-intensity adjuvant chemotherapy was not clearly associated with improved prognosis. Higher RDI in the EP regimen was associated with poorer RFS, suggesting that the pursuit of high RDI may not be necessary in adjuvant HGNEC treatment.

INTRODUCTION:Segmentectomy has become a standard treatment for small-sized peripheral NSCLC. However, its advantages compared with lobectomy in NSCLC harboring pathologically invasive features remain unclear. METHODS:We conducted a post hoc supplementary analysis of randomized controlled phase 3 trial JCOG0802/WJOG4607L in patients with clinical stage IA NSCLC (≤2 cm) harboring one or more of the following pathologically invasive features: lymphatic, vascular, or pleural invasion and lymph node metastasis. The primary outcome was overall survival (OS), and secondary outcomes were relapse-free survival, cumulative incidence of locoregional relapse, lung cancer-specific death, and death from other causes. RESULTS:Of the 1106 participants randomly assigned to the lobectomy or segmentectomy arm, 298 (27%) were included in this analysis: 164 patients were in the lobectomy arm and 134 in the segmentectomy arm. At a median follow-up of 7.1 (interquartile range: 5.9-8.4) years, segmentectomy demonstrated a trend toward better OS than lobectomy, with hazard ratio (HR) of 0.657 (95% confidence interval [CI]: 0.401-1.077; p = 0.0936). The cumulative incidence of death from other causes was significantly reduced in the segmentectomy arm compared with the lobectomy arm (HR = 0.359 [95% CI: 0.143-0.900]). Relapse-free survival and cumulative incidence of lung cancer-specific death were not significantly different between the two arms, whereas the cumulative incidence of locoregional relapse was significantly higher in the segmentectomy than in the lobectomy arm (HR = 2.234 [95% CI: 1.334-3.741]). CONCLUSIONS:Segmentectomy demonstrated better OS compared with lobectomy even for NSCLC harboring pathologically invasive features. However, the risk of locoregional relapse after segmentectomy should be considered.

8009 Background: JCOG0802/WJOG4607L confirmed the superiority of segmentectomy on overall survival (OS) over lobectomy for small-sized peripheral non-small cell lung cancer (NSCLC). However, it remains unclear whether technically demanding complex segmentectomy offers therapeutic efficacy compared to lobectomy, similar to simple segmentectomy. This study therefore aimed to evaluate the oncological outcomes of complex and simple segmentectomy compared to those of lobectomy using final analysis data from JCOG0802/WJOG4607L. Methods: Simple segmentectomy was defined as resection of the bilateral superior division (S 6 ), left lingular (S 4+5 ), or superior division (S 1-3 ); all other segmentectomies were defined as complex. The primary outcome was OS, and secondary outcomes were respiratory function, relapse-free survival (RFS), cumulative incidence of locoregional relapse, and lung cancer-specific death. To ensure comparability, a location-adjusted analysis was performed, separately comparing lobectomy to simple segmentectomy in patients with tumors in the bilateral S 6 , left S 1-3 , or S 4+5 , and to complex segmentectomy in patients with tumors in the right S 7-10 , S 1-3 , or left S 8-10 , S 1-3 . Results: 1106 patients were assigned to either lobectomy (n = 554) or segmentectomy (n = 552) including 318 complex and 234 simple segmentectomies. At a median follow-up of 10.5 years, the 10-year OS was 83.5% (95% CI: 78.8–87.3%), 83.5% (95% CI: 77.7–87.9%) and 79.8% (95% CI: 76.1–83.0%) for complex, simple segmentectomy, and lobectomy, respectively. In the location-adjusted analysis, the 10-year OS were 83.6% (95% CI: 78.9–87.3%) for complex segmentectomy and 79.2% (95% CI: 74.7–82.9%) for lobectomy (HR 0.839 [95% CI 0.608–1.158]). Similarly, 10-year OS were 82.9% (95% CI: 76.9–87.4%) for simple segmentectomy and 78.0% (95% CI: 72.5–82.6%) for lobectomy (HR 0.791 [95% CI 0.540–1.158]). The reduction of median forced expiratory volume in 1 second at 1 year was less after complex (−7.9%) and simple segmentectomy (−9.0%) than lobectomy (−12.0%; p < 0.001 and p < 0.001, respectively). RFS and lung cancer-specific death were not significantly different among the arms, while the cumulative incidence of locoregional relapse was higher after complex and simple segmentectomy than lobectomy (HR 2.124 [95% CI 1.327–3.339] and HR 1.817 [95% CI 1.071–3.083], respectively). Surgical margins were shorter in complex (median 2.2 cm, p < 0.001) and simple (2.5 cm, p < 0.001) segmentectomy than lobectomy (4.0 cm). Conclusions: Complex segmentectomy showed a trend toward improved OS compared to lobectomy in small-sized peripheral NSCLC similar to simple segmentectomy. However, due to a higher risk of locoregional relapse, complex segmentectomy should be performed with meticulous attention to securing adequate surgical margin. Clinical trial information: UMIN000002317.

Objective:Surgery remains the standard treatment for clinical stage I non-small cell lung cancer (NSCLC). Conventional prognostic factors are often subjective and variable, highlighting the need for objective prediction. We developed and validated an annotation-free artificial intelligence (AI) model using computed tomography (CT) images and clinical data to predict prognosis in stage I NSCLC. Methods:In step 1, an AI algorithm was developed to predict pathological classifications from CT and clinical data collected in 3 prospective multi-institutional trials. In step 2, the model was refined and validated using a cohort from the National Cancer Center Hospital East. Models were trained to predict 5-year disease-free survival and overall survival. Performance was evaluated by sensitivity, specificity, predictive values, and receiver operating characteristic area under curve (AUC). Results:We analyzed 1217 patients in step 1 and 1338 in step 2. Models integrating CT imaging and clinical data outperformed models using either dataset alone. The pathology prediction model achieved an AUC of 0.787. The highest performance for 5-year disease-free survival (AUC = 0.757) and overall survival (AUC = 0.756) was obtained by combining preoperative clinical information, physician CT assessments, and an AI-based CT model. Adding AI outputs to clinical factors improved risk stratification, better separating high-risk from low- to intermediate-risk groups. Conclusions:Annotation-free AI models that integrate CT imaging with clinical data provide accurate, objective prediction of recurrence and survival in stage I NSCLC, complement conventional diagnostics, and support personalized multidisciplinary treatment planning.

PURPOSE:To clarify the prognostic impact and associated complications of mediastinal lymph node dissection (MLND) in older adults with radiologically pure solid clinical stage I non-small cell lung cancer (NSCLC), using data from a nationwide surgical database in Japan (n = 18,973). PATIENTS AND METHODS:Of 2075 older adults (≥75 years of age), who underwent complete anatomical resection for clinical stage I NSCLC and registered in the database, data from 1313 with pure-solid tumours on preoperative computed tomography were included. To compare survival and the incidence of postoperative complications, patients were classified into two groups according to the extent of lymph node dissection: MLND; and non-MLND. An inverse probability of treatment weighting method with propensity scoring was used to reduce selection bias. After matching, 1162 patients were analysed for overall survival (OS), and 1153 for disease-free survival (DFS) and cumulative incidence of recurrence (CIR). RESULTS:Before matching, the MLND group exhibited significantly better DFS and OS than the non-MLND group (P = .030 and P = .013, respectively). Multivariable analysis revealed that MLND did not affect the incidence of postoperative complications (≥ grade III). After matching, there were no significant differences between the two groups in terms of DFS, OS, or CIR (P = .536, P = .409, and P = .933, respectively). CONCLUSION:MLND did not yield a survival benefit in older adults (≥75 years of age) even with radiologically pure-solid clinical stage I NSCLC. Anatomical pulmonary resection without MLND may be an alternative treatment option for these patients.

Objective In the Japan Clinical Oncology Group 0802 trial, lung cancer-specific mortality was comparable between lobectomy and segmentectomy despite a higher rate of locoregional relapse after segmentectomy. We evaluated relapse patterns and postrelapse survival following lobectomy or sublobar resection. Methods This single-institution retrospective study included patients who underwent surgery for primary non–small cell lung cancer (NSCLC) between 2003 and 2022 and subsequently developed relapse. Overall survival was defined as the period from surgery (OS1) or from the date of relapse diagnosis (OS2) to death from any cause. Results Among 362 patients with clinical stage IA NSCLC who developed relapse, isolated locoregional relapse accounted for 110 of 297 relapsed patients (37.0%) after lobectomy and 36 of 65 relapsed patients (55.4%) after sublobar resection. Among 59 patients treated with local consolidative therapy for isolated locoregional relapse, the 4-year OS2 rate was 76.4% (95% CI, 64.9%-89.8%). After inverse probability of treatment weighting, the 7-year OS1 rates were 49.3% (95% CI, 42.2%-57.6%) after lobectomy and 64.9% (95% CI, 52.4%-80.3%) after sublobar resection, and the 4-year OS2 rates were 49.6% (95% CI, 42.4%-57.9%) and 62.7% (95% CI, 49.5%-79.3%), respectively. Conclusions In this exploratory cohort of patients who developed postoperative relapse, relapse patterns differed between surgical groups, with isolated locoregional relapse representing a larger proportion of relapse in the sublobar resection group. Long-term survival was observed in selected patients treated with modern postrelapse therapies, including local consolidative therapy. These findings should be interpreted cautiously as hypothesis-generating.

A subaortic anomalous left brachiocephalic (innominate) vein is a rare systemic venous anomaly. Although usually asymptomatic, when it traverses the aortopulmonary window it may obscure the operative field and increase the risk of vascular injury during left upper mediastinal procedures, including station 4 L lymph node dissection and exposure or control of the left pulmonary artery. A man in his seventies presented with a 4.2-cm solid mass in the left upper lobe (S3b) and an enlarged station 4L lymph node. Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) of station 4L demonstrated carcinoma consistent with non-small cell lung cancer (NSCLC). Given resectable single-station N2 disease and pre-existing interstitial lung disease with a usual interstitial pneumonia (UIP) pattern, a surgery-first strategy was selected. Preoperative contrast-enhanced computed tomography (CT) with three-dimensional reconstruction demonstrated an anomalous left brachiocephalic vein coursing along the left lateral side of the aortic arch and then beneath the aortic arch, superior to the left pulmonary artery, and crossing the aortopulmonary window before draining into the superior vena cava (Takada pattern b/c). Video-assisted thoracoscopic left upper lobectomy with systematic mediastinal lymph node dissection was performed. Early identification of the left vagus nerve enabled safe proximal tracing to the left recurrent laryngeal nerve. En bloc dissection of the station 4L and station 5 lymph node regions and left pulmonary artery management were completed without neural or vascular injury. The ductus arteriosus ligament (DL) could not be identified intraoperatively, suggesting a retroductal course (Takada pattern c). Preoperative identification of a subaortic anomalous left brachiocephalic vein on contrast-enhanced computed tomography (CT) with three-dimensional (3D) reconstruction, together with early identification of the vagus and recurrent laryngeal nerves and careful pulmonary artery management after establishing a safe dissection plane, may facilitate safe dissection in the aortopulmonary window and help prevent neural or vascular injury.

Figure S4. List of genes of which the frequencies of mutation were different between recurrence group and control group. Green, red, and black indicate missense mutation, nonsense mutation, and multi-hits, respectively.

Figure S7. Concordance rates (A) between identified 6 genes and p-values (B).

Figure S3. Mutational signature analysis in Recurrence group (A) and in Control group (B).

Objective:Deaths not associated with primary lung cancer (other deaths) and secondary malignancies affect long-term survival after surgery for early-stage non-small cell lung cancer. The JCOG0802 trial suggested that segmentectomy may lower the risk of such deaths. This study examined the relationship between the incidence of other deaths and surgical procedures. Methods:This retrospective, single-center study included patients with clinical stage 0 or IA non-small cell lung cancer who underwent segmentectomy or lobectomy between 2013 and 2016. Other deaths were defined as deaths from causes unrelated to the original lung cancer. Guideline-concordant treatment was assessed based on the National Comprehensive Cancer Network Guidelines. Results:The 5-year overall survival was 90.7%, with a median follow-up of 5.6 years. Among 122 total deaths, 51 (41.8%) were other deaths. Segmentectomy (n = 241) was associated with a lower rate of other deaths (2.5% vs 5.9%) than lobectomy (n = 758), including deaths from other malignancies (1.2% vs 2.6%) and nonmalignancy-related deaths (1.2% vs 3.3%). These findings remained consistent after adjusting for patient characteristic imbalances between the groups. Further, the rate of guideline-concordant treatment for secondary malignancies was higher in the segmentectomy group than in the lobectomy group (71.4% vs 63.8%), with a 9.8% higher rate for second primary lung cancer. Conclusions:Segmentectomy was associated with a lower incidence of deaths not associated with primary lung cancer than lobectomy. The higher rate of guideline-concordant treatment for secondary malignancies in the segmentectomy group may partially explain the reduced incidence of deaths from secondary malignancies.

Clinicopathologic characteristics of patients divided by OLFM1/BMP6 status in cohort 2a (N = 125) and cohort E (patients with EGFR mutation; N = 108).

Table S2. The list of differently expressed genes in Cohort 1

Figure S11. Recurrence-free (A) and overall survivals (B) based on BMP6 and OLFM1 expression status in the entire Cohort 2.

Recurrence-free survival of patients in Cohort 2a in the high expression group and low expression group for each gene (A, BMP6; B, KCNK3; C, NFASC; D, OLFM1; E, PEG3; and F, TNXB).