Tumor stage, squamous histology and male sex predict positive resection margins in non-small cell lung cancer patients undergoing anatomical resection—a retrospective study from a high-volume thoracic surgery unit Authors: Mircea Gabriel Stoleriu, Konrad Steinestel, Matteo Grimmer, Lena-Maria Löwer-Kiem, Moritz Witzenhausen, Jan Sölter, Sebastian Brill, Annette Arndt Published: 2026 ABSTRACT BackgroundThe aim of the study is to identify perioperative predictors of positive resection margins in patients with non-small cell lung cancer (NSCLC) undergoing anatomical resections to improve perioperative risk stratification and ultimately, patient care.MethodsAll patients with primary NSCLC admitted at the Bundeswehrkrankenhaus (Armed Forces Hospital) Ulm for anatomical resections between 01.01.2019 and 31.12.2024 were retrospectively included into the study. Based on resection margins, patients were categorized in two groups: Group 1 (with negative resection margins, R0) and Group 2 (with tumor involvement at the level of resection margins/ positive resection margins at bronchial level, or adjacent structures, R+). A comparative analysis of patients’ demographics, topographical, pathological tumor characteristics and surgical approach was performed by Mann-Whitney U test, Chi-squared test, and Fisher test. A logistic regression model was performed to assess the independent predictive value of the selected variables.ResultsOf 232 NSCLC patients [median age 69.00 (63.00, 75.00) years, 22% aged >75 years, and 2.2% with an Eastern Cooperative Oncology Group (ECOG) >2] who underwent anatomical resections, 107 (46.1%) female patients were included. While R0 resections were observed in 214 (92.2%) patients, 18 (7.8%) patients had positive resection margins on histopathological evaluation. R+ was more frequently reported in male patients (83.3% vs. 16.7%, P=0.009), large tumors (>50 mm/> pT2, P<0.001), as well as in resection specimens with lymph node metastasis (pN1–2, P<0.001), lymphangiosis (L1, P=0.006), vascular invasion (V1, P=0.008) and squamous histology (P<0.001). No significant differences were observed between groups in terms of lobar distribution, tumor mutational status, regression grade upon neoadjuvant therapy, as well as surgical approach (open/minimally invasive). Multivariable analysis revealed male sex (P=0.042), squamous histology (P=0.02), and pT >2 (P<0.001) as independent predictors for R+. These predictors increased the risk of resection margin positivity by 4.1-, 3.7-, and 8.5-fold, respectively.ConclusionsT stage (pT >2), squamous histology and male sex predict positive resection margins in lung cancer patients undergoing anatomical resections. Consideration of these parameters could help in patients’ stratification and care and thus, should be further evaluated in larger patients’ cohorts. Highlight box Key findings • Tumor stage (pT >2), squamous histology and male sex predict positive resection margins in lung cancer patients undergoing anatomical resections. What is known and what is new? • Achieving R0 upon lung cancer surgery significantly reduces local relapse rates, and consecutively leads to an improved outcome and survival. Despite substantial clinical evidence, the relationship between tumor histology including growth pattern and mutational status and resection margin completeness in lung cancer patients is still insufficiently characterized. • We specifically addressed the relationship between perioperative clinical demographics, tumor topography, pathological tumor-node-metastasis descriptors, as well as histological subtypes, morpho-molecular pattern of the resected tumor, and likelihood of positive resection margins as predictors of tumor recurrence and survival. What is the implication, and what should change now? • Assessment of tumor stage (pT >2), squamous histology and male sex may guide risk stratification, refine perioperative care, and enhance patient outcomes. • A frequent use of frozen section examination by suspected adjacent tumor involvement could possibly lead to an extended resection and decrease the R+ rate. Introduction Background Complete surgical resection is essential in oncological surgery, in particular in the treatment of lung cancer (1,2), the leading cause of cancer-related mortality worldwide (3). A microscopically margin-negative resection (R0) in patients with non-small cell lung cancer (NSCLC) is a prerequisite for an accurate treatment reflecting surgical quality and expertise (2,4-6). Clinical evidence supporting the multimodal treatment of lung cancer indicates that achieving R0 resection significantly reduces local relapse rates, and consecutively leads to an improved outcome and survival (7-9). On the other side, microscopically (R1) and macroscopically (R2) incomplete resections are associated with a higher risk of recurrence [5-year disease-free survival (DFS) rates 43.2% for R0 vs. 21.1% for R1/2] (10) and significantly decreased survival [5-year overall survival (OS) ranging between 10% and 42% for R0 patients vs. 14% in R1 patients] (11-13). Given the established association between margin positivity and lymph node involvement, recent studies have increasingly examined the risk of uncertain resection attributable to incomplete tumor removal and insufficient lymphadenectomy (10,14-16). Therefore, the International Association for the Study of Lung Cancer (IASLC) proposed a new definition for uncertain surgical resection (Run) and extensively discussed on its potential clinical impact, particularly affecting treatment planning and decision-making based on tumor-node-metastasis (TNM) classification (1,17,18). Rationale and knowledge gap To improve treatment quality and prognosis, many studies already addressed a variety of perioperative factors that could be associated with the presence of residual tumor or questionable lymphadenectomy. Specifically, older age (>75 years), black race, higher Charlson comorbidity index (CCI >3), clinical TNM (cTNM) descriptors (e.g., tumor size, localization, lymph node involvement), as well as surgical extent (e.g., pneumonectomy) were found to be associated with a higher rate of R1 or R2 resections (19-21). These studies provide substantial clinical evidence on TNM descriptors and surgical approach; however, few reports sufficiently address the complete morpho-molecular patterns of lung cancer, leaving further relevant aspects such as growth pattern and mutational status insufficiently characterized. Objectives The aim of the present study is to perform a comprehensive analysis of perioperative clinical and histopathological parameters that could influence the certainty of the resection margins and ultimately influence the curative intent of the treatment. We specifically addressed the potential relationship between perioperative clinical demographics, tumor topography, pathological TNM (pTNM) descriptors, as well as histological subtypes, growth pattern, mutational status and likelihood of positive resection margins as negative predictors of tumor recurrence and survival. These findings may guide risk stratification, refine perioperative care, and enhance patient outcomes. The present work was designed as a retrospective study rigorously conducted in a high-volume Thoracic Surgery Unit. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0511/rc). Methods Study population This monocentric retrospective cohort study was performed after approval by the Ethics Committee of University of Ulm, Germany (No. 112/23 and No. 293/24) and individual consent for this retrospective analysis was waived. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All patients with primary NSCLC admitted for anatomical resections of primary tumors (segmentectomy, lobectomy, bilobectomy, or pneumonectomy) at the Bundeswehrkrankenhaus (Armed Forces Hospital) Ulm, Germany between 01.01.2019 and 31.12.2024 were included into the study. Patients who underwent anatomical resection for lung metastases (n=20) and with indeterminate resection margins (Rx, n=7) of the histological specimen or who were lost to follow-up (n=1) were excluded (Figure 1). Study flow chart illustrating patient enrollment at study entry. Of 260 patients with primary and secondary malignancies of the lung undergoing anatomical resections between 01.01.2019 and 31.12.2024, 20 (7.7%) patients experiencing lung metastases of non-pulmonary tumors, 1 (0.4%) patient lost to follow-up and 7 (2.9%) patients with indeterminate resection status (Rx) were excluded from the study. Accordingly, 240 patients (92.3%) with primary NSCLC were included into the study. Based on tumor involvement of the resection margins, patients were categorized into two groups: Group 1 with negative resection margins (R0, 214 patients, 92.2%) and Group 2 including patients with positive resection margins (R+, 18 patients, 7.8%). NSCLC, non-small cell lung cancer. Data assessments/sources Clinical follow-up (data during in hospital stay following thoracic surgery) was obtained from hospital medical reports included patients’ demographics [age, sex, Eastern Cooperative Oncology Group (ECOG) performance status], lung function tests [forced expiratory volume in 1 second (FEV1) and diffusing capacity of the lung for carbon monoxide (DLCO)], and histological parameters [TNM and World Health Organization (WHO) classification of histological subtypes]. Primary lung tumors were analyzed using the 8th edition of the TNM staging system [2017] (22,23). Histopathological specimens were based on the WHO classification of lung tumors [2015] (24,25). For lung adenocarcinoma or the adenocarcinoma component of adenosquamous carcinoma, the following growth patterns were discerned: low grade/well-differentiated (G1): lepidic with less than 30% high-grade pattern; intermediate grade/moderately-differentiated (G2): acinar or papillary with less than 30% high-grade pattern; and high-grade/poorly-differentiated (G3): solid or micropapillary pattern (26,27). For mixed adeno(squamous) carcinoma including more than two growth patterns, the frequency of each individual and dominant component was separately assessed. In addition, a comprehensive next-generation sequencing approach including assessment of mutations in epidermal growth factor receptor (EGFR), Kirsten rat sarcoma viral oncogene homolog (KRAS), v-Raf murine sarcoma viral oncogene homolog B1 (BRAF), serine/threonine kinase 11 (STK11), and Kelch-like ECH-associated protein 1 (KEAP1) was performed (Illumina TSO500 NGS panel on an Illumina 550 NextSeq, San Diego, CA, USA). Alterations/translocations in anaplastic lymphoma kinase (ALK), ROS proto-oncogene 1 receptor tyrosine kinase (ROS1), and RET/rearranged during transfection were assessed by immunohistochemistry (IHC) and/or fluorescence in situ hybridization (FISH) following routine protocols (26,27). In patients undergoing neoadjuvant or perioperative chemoimmunotherapy, a tumor regression grade was applied on the resection specimen according to the IASLC including major pathological response (MPR; ≤10% vital tumor cells) or complete pathological response (cPR; no vital tumor cells) (28,29). Data on patients’ treatment included surgical approach [open or minimally invasive by video-assisted thoracic surgery (VATS) or robotic-assisted thoracic surgery (RATS)] and extent of anatomical resection (segmentectomy, lobectomy, bilobectomy, and pneumonectomy). Outcomes and endpoints The primary outcome of the study was the assessment of the resection margins status upon surgery. Margins completeness refers to microscopically proven negative margins (R0), while incomplete resections (R+) discriminate a microscopic (R1) and a macroscopic (R2) involvement. Following the aim of the study, R+ was defined in accordance to the IASLC proposal as tumor involvement at the margin of the resected bronchus, vessels (artery, vein), central or peripheral parenchyma, parietal pleura/chest wall, as well as lymph node involvement at the highest mediastinal station (15). Accordingly, patients were categorized in two groups: Group 1 (with negative resection margins, R0) and Group 2 (with tumor involvement at the level of resection margins/positive resection margins, R+). The primary endpoint was defined as R+. The correlation between abovementioned clinical parameters and the resection margin status (R0 vs. R+) was performed in order to identify potential risk factors for positive resection margins. Statistical analysis Continuous variables are assessed as median and quartiles (1st quartile, 3rd quartile). Comparisons between groups were performed using the Mann-Whitney U test for continuous variables [age, pack years (PY), lung function parameters, tumor size] and by Chi-squared test or Fisher’s exact test (if appropriate) for categorical variables (sex, smoking history, ECOG, tumor localization, TNM, Union for International Cancer Control (UICC), histological growth pattern, mutational status, surgical approach). Multivariable analysis was performed by binary logistic regression analysis with three selection methods [enter, forward likelihood ratio (LR) and backward LR]. To minimize potential confounding bias, the discrimination ability of the logistic regression model was investigated by receiver operating characteristic (ROC) curves and decision tree analysis. An algorithm for clinical setting was developed using the independent predictors identified in the multivariable analysis, using decision tree analysis [classification and regression trees (CRT) algorithm]. The internal validation of the cohort was performed using a 5- and 10-fold cross-validation. Odds ratios (ORs) with 95% confidence intervals (CIs) were assessed to show the independent value of the selected parameters in predicting resection margins positivity. Missing data were accordingly reported, while any records with essential missing variables were excluded from the analysis (Tables 1-6). All analyses were performed using the software SPSS (version 26, IBM, Armonk, NY, USA). P values <0.05 were considered as statistically significant. Table 1Demographic characteristics of study populationPatient demographics at study entryGroup 1 (R0, n=214)Group 2 (R+, n=18)P valueAge (years)69.00 (62.00, 75.00)68.00 (63.75, 75.25)0.95 >7547/214 (22.0)4/18 (22.2)0.98Sex0.009* Female104/214 (48.6)3/18 (16.7) Male110/214 (51.4)15/18 (83.3)Smoking history Never smokers41/213 (19.2)2/18 (11.1)0.39 Active smokers65/213 (30.5)6/18 (33.3)0.80 Ex-smokers >1 year107/213 (50.2)10/18 (55.6)0.67PY35.00 (20.00, 50.00)50.00 (31.25, 67.50)0.02* ≥30109/167 (65.3)15/16 (93.8)0.02*ECOG ECOG 0111/192 (57.8)5/16 (31.3)0.040* ECOG 165/192 (33.9)10/16 (62.5)0.02* ECOG 211/192 (5.7)1/16 (6.3)0.93 ECOG 35/192 (2.6)0/16 (0.0)>0.99† ECOG >116/192 (8.3)1/16 (6.3)0.77Lung function parameters FEV1 (% predicted)82.00 (69.00, 97.00)63.50 (51.00, 76.50)0.001* DLCO (% predicted)77.00 (66.00, 92.00)65.00 (53.50, 81.50)0.10Demographic characteristics of lung cancer patients undergoing anatomical resections, stratified by resection margins. Median and quartiles (1st, 3rd) or absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Mann-Whitney U test for continuous variables (e.g., age, PY, FEV1, DLCO) and Chi-squared or Fisher’s exact test (if appropriate) for categorical variables. Missing data were found when assessing smoking history (one patient in R0 group) and ECOG status (n=22 patients in R0 group and n=2 patients in R1 group). †, Fisher’s exact test. *, P<0.05. DLCO, diffusing capacity of the lung for carbon monoxide; ECOG, Eastern Cooperative Oncology Group; FEV1, forced expiratory volume in 1 second; PY, pack years. Demographic characteristics of study population Demographic characteristics of lung cancer patients undergoing anatomical resections, stratified by resection margins. Median and quartiles (1st, 3rd) or absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Mann-Whitney U test for continuous variables (e.g., age, PY, FEV1, DLCO) and Chi-squared or Fisher’s exact test (if appropriate) for categorical variables. Missing data were found when assessing smoking history (one patient in R0 group) and ECOG status (n=22 patients in R0 group and n=2 patients in R1 group). †, Fisher’s exact test. *, P<0.05. DLCO, diffusing capacity of the lung for carbon monoxide; ECOG, Eastern Cooperative Oncology Group; FEV1, forced expiratory volume in 1 second; PY, pack years. Table 2Tumor characteristics of study populationTumor characteristicsGroup 1 (R0, n=214)Group 2 (R+, n=18)P valueTumor side0.65 Left95/214 (44.4)7/18 (38.9) Right119/214 (55.6)11/18 (61.1)Tumor localization Left upper lobe42/214 (19.6)2/18 (11.1)0.38 Left lower lobe45/214 (21.0)3/18 (16.7)0.66 Left combined/centrally located8/214 (3.7)2/18 (11.1)0.14 Right upper lobe57/214 (26.6)6/18 (33.3)0.54 Right middle lobe11/214 (5.1)–– Right lower lobe41/214 (19.2)3/18 (16.7)0.80 Right combined/centrally located10/214 (4.7)2/18 (11.1)0.24Tumor diameter (mm)22.00 (14.00, 38.00)66.00 (40.50, 92.25)<0.001* >3072/214 (33.6)16/18 (88.9)<0.001* >5030/214 (14.0)12/18 (66.7)<0.001*TNM7 classification pTis2/214 (0.9)0/18 (0.0)>0.99† pT010/214 (4.7)0/18 (0.0)0.35 pT1105/214 (49.1)1/18 (5.6)<0.001* pT252/214 (24.3)4/18 (22.2)0.84 pT328/214 (13.1)4/18 (22.2)0.28 pT417/214 (7.9)9/18 (50.0)<0.001* > pT197/214 (45.3)17/18 (94.4)<0.001* > pT245/214 (21.0)13/18 (72.2)<0.001*Lymph node involvement pN0167/213 (78.4)6/18 (33.3)<0.001* pN118/213 (8.5)6/18 (33.3)<0.001* pN228/213 (13.1)5/18 (27.8)0.09 pN30/213 (0.0)1/18 (5.6)0.08†Lymphangiosis0.006* L0120/176 (68.2)5/15 (33.3) L156/176 (31.8)10/15 (66.7)Vascular invasion0.008* V0147/167 (88.0)8/13 (61.5) V120/167 (12.0)5/13 (38.5)UICC stage 02/214 (0.9)0/18 (0.0)>0.99† I111/214 (51.9)1/18 (5.6)<0.001* II41/214 (19.2)2/18 (11.1)0.40 III47/214 (22.0)9/18 (50.0)0.008* IV13/214 (6.1)6/18 (33.3)<0.001* UICC >1101/214 (47.2)17/18 (94.4)<0.001* UICC >260/214 (28.0)15/18 (83.3)<0.001*Pathological characteristics of resected lung cancer specimens, stratified by resection margins positivity (R0 vs. R+). Median and quartiles (1st, 3rd) or absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Mann-Whitney U-test for continuous variables (e.g. tumor diameter) and Chi-squared or Fisher’s exact test (if appropriate) for categorical variables. Missing data were found when assessing pN status (one patient in R0 group), lymphangiosis (n=38 patients in R0 group and n=3 patients in R1 group), and vascular invasion (n=47 patients in R0 group and n=5 patients in R1 group). †, Fisher’s exact test. *, P<0.05. pTNM: TNM8 staging system applied for intraoperative histopathological specimens (19,22); UICC: tumor staging according to the UICC (22,23). pTNM, pathological tumor-node-metastasis; TNM, tumor-node-metastasis; UICC, Union for International Cancer Control. Tumor characteristics of study population Pathological characteristics of resected lung cancer specimens, stratified by resection margins positivity (R0 vs. R+). Median and quartiles (1st, 3rd) or absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Mann-Whitney U-test for continuous variables (e.g. tumor diameter) and Chi-squared or Fisher’s exact test (if appropriate) for categorical variables. Missing data were found when assessing pN status (one patient in R0 group), lymphangiosis (n=38 patients in R0 group and n=3 patients in R1 group), and vascular invasion (n=47 patients in R0 group and n=5 patients in R1 group). †, Fisher’s exact test. *, P<0.05. pTNM: TNM8 staging system applied for intraoperative histopathological specimens (19,22); UICC: tumor staging according to the UICC (22,23). pTNM, pathological tumor-node-metastasis; TNM, tumor-node-metastasis; UICC, Union for International Cancer Control. Table 3Histological characteristics of resected tumorsHistological characteristics of resected tumorsGroup 1 (R0, n=214)Group 2 (R+, n=18)P valueHistology of the primary tumor (WHO) Adenocarcinoma145/214 (67.8)6/18 (33.3)0.003* Squamous cell carcinoma41/214 (19.2)11/18 (61.1)<0.001* Carcinoid 18/214 (8.4)0/18 (0.0)0.20 Adenosquamous carcinoma6/214 (2.8)1/18 (5.6)0.51 Large cell carcinoma5/214 (2.3)0/18 (0.0)>0.99†Dominant growth pattern in adenocarcinoma Lepidic36/151 (23.8)0/7 (0.0)0.14 Acinar62/151 (41.1)4/7 (57.1)0.45† Papillary13/151 (8.6)2/7 (28.6)0.08 Micropapillary2/151 (1.3)1/7 (14.3)0.13† Solid34/151 (22.5)0/7 (0.0)0.16 Mucinous4/151 (2.6)0/7 (0.0)>0.99†Mutational status None117/214 (54.7)14/18 (77.8)0.058 EGFR 26/214 (12.1)2/18 (11.1)0.90 KRAS G12C28/214 (13.1)0/18 (0.0)0.10 KRAS (other)26/214 (12.1)1/18 (5.6)0.40 BRAF4/214 (1.9)0/18 (0.0)>0.99† ALK3/214 (1.4)0/18 (0.0)>0.99† ROS2/214 (0.9)1/18 (5.6)0.22† STK119/214 (4.2)0/18 (0.0)0.38 KEAP13/214 (1.4)0/18 (0.0)>0.99† RET2/214 (0.9)0/21 (0.0)>0.99†Histological characteristics of resected tumors in lung cancer patients undergoing anatomical resections, stratified by resection margins positivity (R0 vs. R+). For all adenocarcinoma specimens, the frequency of the dominant component in tumors with mixed growth pattern was assessed. Absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Chi-squared or Fisher’s exact test (if appropriate). †, Fisher’s exact test. *, P<0.05. ALK, anaplastic lymphoma kinase; BRAF, v-Raf murine sarcoma viral oncogene homolog B1; EGFR, epidermal growth factor receptor; KEAP1, Kelch-like ECH-associated protein 1; KRAS, Kirsten rat sarcoma viral oncogene homolog; RET, rearranged during transfection; ROS, ROS proto-oncogene 1 receptor tyrosine kinase; STK11, serine/threonine kinase 11; WHO, World Health Organization. Histological characteristics of resected tumors Histological characteristics of resected tumors in lung cancer patients undergoing anatomical resections, stratified by resection margins positivity (R0 vs. R+). For all adenocarcinoma specimens, the frequency of the dominant component in tumors with mixed growth pattern was assessed. Absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Chi-squared or Fisher’s exact test (if appropriate). †, Fisher’s exact test. *, P<0.05. ALK, anaplastic lymphoma kinase; BRAF, v-Raf murine sarcoma viral oncogene homolog B1; EGFR, epidermal growth factor receptor; KEAP1, Kelch-like ECH-associated protein 1; KRAS, Kirsten rat sarcoma viral oncogene homolog; RET, rearranged during transfection; ROS, ROS proto-oncogene 1 receptor tyrosine kinase; STK11, serine/threonine kinase 11; WHO, World Health Organization. Table 4Technical aspects of the surgical approachFeatures of the surgical approachGroup 1 (R0, n=214)Group 2 (R+, n=18)P valueSurgical approach Open (thoracotomy)126/214 (58.9)14/18 (77.8)0.12 VATS29/214 (13.6)1/18 (5.6)0.33 RATS31/214 (14.5)3/18 (16.7)0.80 VATS conversion to open22/214 (10.3)0/18 (0.0)0.15 RATS conversion to open6/214 (2.8)0/18 (0.0)0.47Resection extent Segmentectomy24/214 (11.2)2/18 (11.1)0.99 Lobectomy or bilobectomy176/214 (82.2)12/18 (66.7)0.11 Pneumonectomy14/214 (6.5)4/18 (22.2)0.02* Lymphadenectomy211/214 (98.6)16/18 (88.9)0.050†Neoadjuvant treatment0.52 No192/214 (89.7)17/18 (94.4) Yes22/214 (10.3)1/18 (5.6)Tumor regression grade No vital tumor cells (CPR, IASLC)10/22 (45.5)0/1 (0.0)>0.99† Vital tumor cells ≤10% (MPR, IASLC)3/22 (13.6)0/1 (0.0)>0.99† Vital tumor cells >10%9/22 (40.9)1/1 (100.0)0.44†Technical aspects of the surgical approach in lung cancer patients undergoing anatomical resections, stratified by resection margins positivity (R0 vs. R+). Absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Chi-squared or Fisher’s exact test (if appropriate). †, Fisher’s exact test. *, P<0.05. CPR, complete pathological response; IASLC, International Association for the Study of Lung Cancer; MPR, moderate pathological response; RATS, robotic-assisted thoracic surgery; VATS, video-assisted thoracic surgery. Technical aspects of the surgical approach Technical aspects of the surgical approach in lung cancer patients undergoing anatomical resections, stratified by resection margins positivity (R0 vs. R+). Absolute values with relative frequency (n/total, %) were presented. Group comparisons (R0 vs. R+) were performed using the Chi-squared or Fisher’s exact test (if appropriate). †, Fisher’s exact test. *, P<0.05. CPR, complete pathological response; IASLC, International Association for the Study of Lung Cancer; MPR, moderate pathological response; RATS, robotic-assisted thoracic surgery; VATS, video-assisted thoracic surgery. Table 5Clinical description of patients with R+ following lung cancer resectionPatient No.R statusPreop. demographicsPreop. lung functionHistology (WHO), pTNM8, morpho-molecular patternMutationsSurgical approachSurgery time (min)Localization R+Age (years)SexECOGFEV1 (%)FEV1 (L)HistologyLUAD diff.pTpNcMLVAnatomical extentOpen approachSideUpper lobeMiddle lobeLower lobeCentral/combinedPatient 1Par78Male2461.32LUADPap, sol, ac, leppT4pN0cM0L0V0 EGFR SegmYesLeftYesNoNoNo238Patient 2Par60Male310.96LUADAcpT1pN0cM0L0V00SegmNoRightYesNoNoNo76Patient 3Par64Male1582.33SQCA–pT4pN2cM0L1V00PneumoYesRightNoNoNoYes179Patient 4Par, Bro76Male0732.31SQCA–pT4pN3cM0L1V10LobectYesRightNoNoYesNo180Patient 5Par, Ves65Male1421.24SQCA–pT4pN2cM0L1V10PneumoYesLeftNoNoNoYes260Patient 6Chest87Female1881.5SQCA–pT3pN1cM0L1V00LobectYesLeftNoNoYesNo126Patient 7Ves65Male0762.47SQCA–pT2pN2cM0L1V10LobectYesRightYesNoNoNo198Patient 8Bro69Male1671.98SQCA–pT2pN1cM0L0V00LobectYesRightYesNoNoNo141Patient 9Chest71Female1781.72SQCA–pT4pN0cM0 0LobectYesRightNoNoYesNo133Patient 10Par65Male1481.43SQCA–pT4pN2cM0L1V00LobectNoRightYesNoNoNo249Patient 11Par, Ves67Male1742.37LUADPappT4pN0cM0L0V0KRAS (other)LobectNoLeftNoNoYesNo346Patient 12Ves, Bro59Male0914LUADAc, pap, solpT2pN2cM0L1V1 EGFR PneumoYesRightNoNoNoYes197Patient 13Bro63Male1591.92SQCA–pT4pN1cM0L1V10PneumoYesLeftNoNoNoYes154Patient 14LK Hilus69Male1521.77SQCA–pT2pN1cM0 0LobectYesRightYesNoNoNo186Patient 15Par, Bro70Male0621.85LUADMpap, acpT4pN0cM0L0V0 ROS LobectYesRightNoNoYesNo214Patient 16Chest81Male1651.8SQCA–pT3pN1cM1L1 0LobectNoRightYesNoNoNo279Patient 17Chest75Female01132.35LUADAc, pap, leppT3pN0cM0 0LobectYesLeftNoNoYesNo99Patient 18Par, Bro55Male–551.87LUADAc, pappT3pN1cM0L1 0LobectYesLeftYesNoNoNo216Detailed description of patients with positive resection margins (R+) following lung cancer resection with focus on the tumor and R+ localization, as well as clinical, topographical and morpho-molecular profile of the resection margins. Ves: artery/vein; Chest: pleura/chest; LK Hilus: highest mediastinal positive lymph node station; pTNM: TNM8 staging system applied for intraoperative histopathological specimens (19,22). Ac, acinar; Bro, bronchus; ECOG, Eastern Cooperative Oncology Group; EGFR, epidermal growth factor receptor; FEV1, forced expiratory volume in 1 second; KRAS, Kirsten rat sarcoma viral oncogene homolog; L, lymphangiosis; lep, lepidic; lobect, lobectomy or bilobectomy; LUAD, lung adenocarcinoma; mpap, micropapillary; pap, papillary; Par, parenchyma; pneumo, pneumonectomy; pTNM, pathological tumor-node-metastasis; ROS, ROS proto-oncogene 1 receptor tyrosine kinase; segm, segmentectomy; sol, solid; SQCA, squamous cell carcinoma; TNM, tumor-node-metastasis; V, vascular involvement; Ves, vessels; WHO, World Health Organization. Clinical description of patients with R+ following lung cancer resection Detailed description of patients with positive resection margins (R+) following lung cancer resection with focus on the tumor and R+ localization, as well as clinical, topographical and morpho-molecular profile of the resection margins. Ves: artery/vein; Chest: pleura/chest; LK Hilus: highest mediastinal positive lymph node station; pTNM: TNM8 staging system applied for intraoperative histopathological specimens (19,22). Ac, acinar; Bro, bronchus; ECOG, Eastern Cooperative Oncology Group; EGFR, epidermal growth factor receptor; FEV1, forced expiratory volume in 1 second; KRAS, Kirsten rat sarcoma viral oncogene homolog; L, lymphangiosis; lep, lepidic; lobect, lobectomy or bilobectomy; LUAD, lung adenocarcinoma; mpap, micropapillary; pap, papillary; Par, parenchyma; pneumo, pneumonectomy; pTNM, pathological tumor-node-metastasis; ROS, ROS proto-oncogene 1 receptor tyrosine kinase; segm, segmentectomy; sol, solid; SQCA, squamous cell carcinoma; TNM, tumor-node-metastasis; V, vascular involvement; Ves, vessels; WHO, World Health Organization. Table 6Binary logistic regression model including three independent predictorsIndependent predictors of positive resection marginsExp(B) (95% CI)P valueMale sex4.08 (1.05–15.87)0.042Squamous cell carcinoma histology3.65 (1.21–11.04)0.02pT >28.47 (2.71–26.48)<0.001Binary logistic regression model assessing the independent predictors of positive resection margins in patients with NSCLC undergoing anatomical resections, stratified by resection margins positivity (R0 vs. R+). The multivariable regression analysis was performed by using the enter method and validated by forward LR and backward LR selection methods. pT: tumor stage according to the TNM8 staging system [2017] (19,22); Exp(B): odds ratio; CI, lower bound–upper bound. CI, confidence interval; LR, likelihood ratio; NSCLC, non-small cell lung cancer; TNM, tumor-node-metastasis. Binary logistic regression model including three independent predictors Binary logistic regression model assessing the independent predictors of positive resection margins in patients with NSCLC undergoing anatomical resections, stratified by resection margins positivity (R0 vs. R+). The multivariable regression analysis was performed by using the enter method and validated by forward LR and backward LR selection methods. pT: tumor stage according to the TNM8 staging system [2017] (19,22); Exp(B): odds ratio; CI, lower bound–upper bound. CI, confidence interval; LR, likelihood ratio; NSCLC, non-small cell lung cancer; TNM, tumor-node-metastasis. Results Study population Of 260 patients with primary NSCLC or pulmonary metastases admitted for anatomical resections, 20 patients with pulmonary metastases, 1 patient who was lost to follow-up and 7 patients with indeterminate resection status (Rx) were excluded from the study. Accordingly, 232 patients were further characterized based on clinical demographics, tumor localization, TNM, and WHO histological classification of primary tumor, resection margins, mutational status and surgical approach. The patient selection process is depicted in Figure 1. Following the primary outcome of the study, 214 (92.2%) patients had negative resection margins (Group 1), while 18 (7.8%) patients had positive resection margins on histopathological evaluation (Group 2). Univariate analysis Patient demographics are presented in Table 1. Missing data were found when assessing smoking history (one patient in R0 group) and ECOG status (22 patients in R0 group and 2 patients in R1 group). The median age of the whole patient’s cohort was 69.00 (63.00, 75.00) years with 22.0% of patients being aged 75 years or older. No significant age differences were observed between groups. Of 232 patients, 107 (46.1%) patients were female. Positive resection margins were more frequent in male patients (R0 51.4% vs. R+ 83.3%, P=0.009). While smoking history (active, never, ex) was not significantly correlated with R+, smoking intensity (quantified in PY) was significantly associated with the resection margins completeness. In particular, a smoking intensity >30 PY was significantly associated with R+. When considering sex distribution, more never smokers among female patients [27/106 (25.5%) vs. 16/125 (12.8%), P=0.01], and more active smokers among male patients [25/106 (23.6%) vs. 46/125 (36.8%), P=0.03] were observed. In addition, male patients smoked significantly more than female patients [20.00 (0.00, 40.00) vs. 40.00 (20.00, 50.00) PY, P<0.001]. Patients with reduced preoperative lung function parameters were more likely to present with positive resection margins [FEV1; R0 82.00 (69.00, 97.00) vs. R+ 63.50 (51.00, 76.50) % predicted, P=0.001; Table 1]. One hundred and ninety-one (82.3%) patients were admitted with low ECOG grades (0 and 1), while only 17 (7.3%) patients presented with an ECOG >1 with no significant differences with respect to resection status (R0 8.3% vs. R+ 6.3%, P=0.77; Table 1). Tumors were homogenously distributed in both lungs and no specific localization was particularly associated with R+ status. Tumors with pT >1 (pT2–4, 45.3% vs. 94.4%, P<0.001), lymph node metastasis (pN+, 21.6% vs. 66.7%, P<0.001), lymphangiosis (L1, 31.8% vs. 66.7%, P=0.006) and vascular invasion (V1, 12.0% vs. 38.5%, P=0.008) were significantly more frequent in the R+ group. Accordingly, an UICC stage >1 was significantly associated with R+ status (47.2% vs. 94.4%, P<0.001). The tumor characteristics including localization, size, TNM and UICC status are summarized in Table 2. Adenocarcinoma histology was more frequent among R0 resections (67.8% vs. 33.3%, P=0.003), while squamous histology was more frequently reported in the R+ group (19.2% vs. 61.1%, P<0.001). Among patients with adenocarcinoma histology more never smokers were found in comparison to the patients presenting squamous carcinoma histology [34/151 (22.5%) vs. 1/51 (2.0%), P=0.001]. In contrast, patients with squamous histology were more frequently active smokers [43/151 (28.5%) vs. 21/51 (41.2%), P=0.09] and presented a significantly increased nicotine consumption, when compared to adenocarcinoma patients [30.00 (1.25, 45.00) vs. 42.50 (29.25, 60.00) PY, P<0.001]. When analyzing each individual (not the dominant) component of the growth pattern in mixed adeno(squamous) carcinoma, papillary tumors were more frequent in the R+ group (25.2% vs. 71.4%, P=0.007). When analyzing the dominant component of the growth pattern in mixed adeno(squamous) carcinoma, no significant differences were reported between groups (Table 3). While no specific driver mutation was associated with R status, the absence of such genetic alterations showed borderline significant association with positive resection margins (54.7% vs. 77.8%, P=0.058). The histopathological and molecular characteristics of resected tumors are summarized in Table 3. Open surgery was more frequent in patients group experiencing incomplete resections (58.9% vs. 77.8%, P=0.12; Table 4). No significant differences were reported between patients undergoing thoracoscopic (VATS, 13.6% vs. 5.6%, P=0.33) or robotic (RATS, 14.5% vs. 16.7%, P=0.80) minimally invasive procedures. In patients with large tumors, who underwent pneumonectomies, a significantly higher rate of incomplete resections was reported (6.5% vs. 22.2%, P=0.02; Table 4). Specifically, patients undergoing pneumonectomies experienced more frequently tumors >5 cm (> pT2, 61.1% vs. 23.0%, P<0.001 or UICC >2, 88.9% vs. 28.5%, P<0.001) and presented more frequently a lymph node involvement (pN+ 66.7% vs. 22.7%, P<0.001). In addition, squamous cell carcinoma histology appeared significantly more frequent in specimens upon pneumonectomy than other anatomical resections (55.6% vs. 20.3%, P=0.01). Of 23 patients undergoing neoadjuvant treatment, one patient experienced an incomplete R+ resection and present >10% viable tumor cells in the histological specimen. Technical aspects of the surgical approach are summarized in Table 4. Combining all parameters presented above, a detailed description including clinical demographics, lung function parameters, morpho-molecular pattern of the tumor, as well as surgical approach was assessed in the R+ group (n=18 patients, Table 5). Of note, a descriptive analysis did not reveal a significant difference in the distribution of bronchovascular and mediastinal positive margins between segmentectomy, lobectomy and pneumonectomy cases (Tables S1,S2). Accordingly, 9 of 18 patients (50%), 4 of 18 patients (22.2%), and 4 of 18 (22.2%), patients experienced positive resections margins at the level of parenchyma, vessels (artery/vein) and pleura/chest wall, respectively. One patient (5.6%) presented a lymph node involvement at the highest mediastinal station, while 6 of 18 patients (33.3%) presented tumor cells at the level of the resected bronchus. In 8 of 18 (44.4%), a tumor involvement at the level of the resected parenchyma (R1, Par) at the final histological analysis was reported. In these patients, no intraoperative frozen section examination was performed. Patients with R1, were in majority male sex (15 of 18, 83.3%) and had a median age of 68.00 (63.75, 75.25) years, a reduced FEV1 [63.50 (51.00, 76.50) % predicted] and large tumors (pT3: 4 of 18, 22% and pT4: 9 of 18, 50%). The dominant and the most frequent pattern in adeno(squamous)carcinoma was acinar pattern (4 of 7, 57.1% and 6 of 7, 85.7%, respectively). Patients with incomplete resections underwent frequently an open approach (14 of 18, 77.8%) and extended resections (at least lobectomy, 16 of 18, 88.9%). Of note, none of the R+ patients experienced a bronchial stump insufficiency following surgical resection. This complication was reported in only one female patient (72-year-old, ex-smoker, 30 PY) with a R0 resected squamous cell carcinoma localized in the left lower lobe. Multivariable analysis The clinically relevant and statistically significant variables at univariate analysis were incorporated into a binary logistic regression model. Accordingly, male sex (P=0.042), squamous histology (P=0.02), and pT >2 (P<0.001) were identified as independent predictors for R+. These predictors increased the risk of resection margins positivity by 4.1-, 3.7-, and 8.5-fold, respectively (Table 6). The independent predictiveness of these parameters was separately confirmed by using three selection methods for the logistic regression analysis (enter, forward LR, and backward LR). The discrimination ability of the logistic regression model was investigated by ROC curves (Figure S1). An algorithm for the clinical setting using the abovementioned independent predictors was performed by decision tree analysis (CRT algorithm) using a 5- and 10-fold cross-validation as internal validation method (Figure S2). Given the various R+ localizations and patterns upon surgery (Table 5, Tables S1,S2), the surgical extent (categorized as segmentectomies, lobectomies, bilobectomies, and pneumonectomies) was further analyzed in relation to the abovementioned predictors. When incorporating the surgical extent as covariate in the final logistic regression model, the analysis turned out unchanged results, while surgical extent was not found to independently predict resection margin positivity (Table S3). Discussion Complete surgical resection is a key aspect in the multimodal treatment of lung cancer that directly influences the curative intent of the treatment and ultimately, the short- and long-term prognosis of patients (30). Despite existing evidence on the role of TNM descriptors and surgical approach on the patients’ prognosis, the description of the morpho-molecular patterns of lung cancer in relation to the resection margins patterns remains still insufficient. This is in part due to the fact that tumor genetic analyses are mostly performed in non-resectable/advanced cases. To address this knowledge gap, our study focuses on the relationship between perioperative clinical, histopathological, and molecular parameters and the resection margin status. The present study addressed all NSCLC patients admitted for anatomical surgical resections at the Bundeswehrkrankenhaus (Armed Forces Hospital) Ulm, Germany between January 1st, 2019 and December 31st, 2024. Our cohort had a median age of 69 years, and included 46.1% female patients. These clinical demographics align with previous patients cohorts (31-34). Following the aim of the study, the R+ rate in the histopathological specimen and potential associations with perioperative clinical and morpho-molecular parameters were further evaluated. The definition of R+ includes the tumor involvement at the margin of the resected bronchus, vessels (artery, vein), parenchyma, pleura/chest wall, as well as lymph node involvement at the highest mediastinal station. This definition is in line with the publication by Edwards et al. (IASLC Staging Project database, n=14,293 patients) (15). The rate of R+ found in our study was 7.8%. It was comparable with the rate reported by Rasing et al. (R1: 7.1%) in a Dutch national database including 7156 patients (21), but higher than the rate published by Edwards et al. (R+: 2–4%) in the IASLC database (15). Importantly, Edwards et al. reported a stage dependent R1 rate (pT1/pT2/pT3/pT4: 0.5%/1.7%/4.3%/7.3%) and R2 rate (pT1/pT2/pT3/pT4: 0.4%/0.9%/1.8%/5.7%). Thus, our results align with the R1 rate reported in pT4 by Edwards et al. (15). One potential explanation for the lower R1 rate in the IASLC database might be the higher percentage of patients with pT1/2 in the IASLC database, in the context of a systematic screening. In contrast, our study included more patients with pT4 (12.1%) than the IASLC database (pT4: 5.7%), potentially reflecting more challenging cases and explaining the higher R+ rate. Beside the registry evidence described by Rasing et al. (21) and Edwards et al. (15) a wider R+ range between 4% and 20% was reported in previous studies (19,30,35,36). Multifactorial explanations include differences in R+ definition, individual clinical parameters, demographics, tumor topography, time of diagnosis, and pTNM descriptors in different countries and databases. In detail, the highest R+ rates were reported in large (>5 cm, pT3–4) NSCLC with radiologically confirmed invasion of neighboring structures including mediastinal vessels, trachea, esophagus, nerves or chest wall (19). Importantly, R+ was also influenced by patient-independent factors such as individual surgical experience, and intraoperative management of resection margins in hospitals with varying surgical volume. These aspects underscore the importance of complete surgical resections in order to achieve a curative intent and ultimately, best outcomes in NSCLC patients. Our patient cohort revealed significantly reduced preoperative lung function parameters in patients experiencing R+ resections, which aligns with a higher rate of larger tumors reported in the R+ group. This result underscores the relationship between lung function and tumor stage, as previously validated in other patients cohorts by us or others (33,37). Notably, R+ was frequent in patients with large tumors (> pT2, UICC >2), lymph node metastases (pN+), lymphangiosis (L1), and vascular invasion (V1) in line with previous studies (19,21,38). Correspondingly, complete resections were more frequent in adenocarcinoma, while we observed more incomplete resections in squamous cell carcinoma (19). This aspect could be potentially explained by the topography of the resected tumors, with adenocarcinoma occurring more frequently in the lung periphery, while squamous cell carcinoma arises in close neighborhood of the bronchial tree and mediastinal structures (39-41). Based on these considerations, it could be assumed that peripheral adenocarcinomas could be resected more easily in comparison to centrally located squamous cell carcinomas. Even though some studies showed more frequent open approaches in squamous cell carcinomas, and minimally invasive approaches in adenocarcinomas (31,42,43), these findings cannot be generalized due to the advances and steadily increasing expertise on minimally invasive approaches in high volume centers (44-47). Reflecting this, our study showed an almost similar rate of open procedures in both adenocarcinoma (59.6%) and squamous cell carcinoma (63.5%). On the other hand, recent studies showed that larger and centrally located lung tumors (locally advanced lung tumors) can also be resected by minimally invasive approaches (32,34). Importantly, when considering only patients undergoing minimally invasive approaches, 1 patient (3.3%) undergoing VATS and 3 patients (8.8%) undergoing RATS experienced incomplete resections. As expected, open approaches were commonly preferred in larger and centrally located tumors, where infiltration of the bronchial tree and mediastinal structures might impact the completeness of resection. Accordingly, the R+ rate was higher in patients undergoing pneumonectomies. Potential explanations are the presence of large tumors >5 cm (> pT2, 61.1% vs. 23.0%, P<0.001 or UICC >2, 88.9% vs. 28.5%, P<0.001), lymph node involvement (pN+ 66.7% vs. 22.7%, P<0.001), and squamous cell carcinoma histology (55.6% vs. 20.3%, P=0.01). Thus, it seems plausible to consider that centrally located, large tumors with positive lymph node involvement resected by pneumonectomy are more frequently associated with positive resection margins than those tumors approached by segmentectomy or lobectomy. All these correlations are in accordance with previous studies (21,48-51). In the subgroup of patients with neoadjuvant treatment (n=23), only one patient had a R+ resection, with >10% vital tumor cells in the histological specimen. Even though this subgroup includes only a small number of patients, the presence of residual vital tumor cells in the resected specimen could point toward the aggressiveness of the tumor growth and potentially explain incomplete resection. Further potential explanations could be a slower growth pattern of adenocarcinomas (especially in low-grade/lepidic predominant subtypes) and more aggressive biological behavior of squamous cell carcinoma that tends to spread centrally very close to the bronchial tree (39,40,52-55). Taking etiological factors into account, squamous histology is more frequently associated with smoking and chronic obstructive pulmonary disease (COPD) (56-58), while a subgroup of patients experiencing adenocarcinomas are nonsmokers (59,60). Consequently, history of smoking/COPD associated with parenchymal hyperinflation might additionally impact the extent of the surgical resections, in particular in patients with squamous histology (61,62). Furthermore, postoperative complications such as pneumonia, air leak, and prolonged in-hospital stay are more frequent in these high-risk patients (62). While some studies demonstrate a decreased survival in COPD patients who underwent lung cancer resection (63-65), other authors show no causal relationship with the long-term survival (62). In alignment with previous considerations, histological analysis of the resected specimens showed significantly more squamous cell carcinomas in patients experiencing incomplete resections. On the other side, adenocarcinoma histology was significantly more frequent in patients with complete resections. When considering the frequency of the dominant growth pattern in adenocarcinoma specimens, no significant differences were reported between groups and no pattern was associated with R+. The lack of statistical significance between high grade (G3, micropapillary or solid) tumors and R+ could be justified by the low R+ rate (n=7 of 158 patients, 4.4%) in adeno(squamous) carcinoma. Existing evidence shows however, that micropapillary and solid subtypes are independent predictors for curative resection for stage I lung cancer when complete lymphadenectomy is additionally performed (66). It has also to be noted that histological grading is based on a fraction estimate of high-grade patterns, and these patterns are not necessarily observed in the areas of positive resection margins, but may also be observed in the tumor center. Taking these considerations into account, we still consider reporting the growth pattern in histological specimens as a key aspect with prognostic consequences, that has to be systematically assessed and thoroughly characterized in further high-volume studies. One might discuss that in cases with positive resection margins, the respective growth pattern along the positive margin might be reported, comparable to the recommendations in prostatic adenocarcinoma. This approach might allow for a more granular correlation between actual growth patterns and biological aggressiveness in NSCLC (67). One novel aspect addressed in our study was a comprehensive description of the mutational status of the examined histological specimens in relation to the resection status. More than 50% of the patients included in our study [54.7% (R0) and 77.8% (R+)] did not harbor any oncogenic driver mutations. These rates consistently vary in different studies taking the ethnic, and demographic variability of the patients into account (68-70). The low rate of mutations found in our patient cohort and, in particular, in patients with incomplete resections could be correlated with the fraction of squamous histology. Of note, from nine distinct genes analyzed in the whole patient cohort, only two patients with EGFR mutations, one patient with KRAS (other) mutation and ROS alteration had an incomplete surgical resection. There was no clustering of STK11 and/or KEAP1 mutations which have been previously associated with tumor aggressiveness in the R+ resection group (71). Our results confirm previous data that assume that the histologic subtype and localization of the primary tumor seem to remain the most important factors associated with complete resections, however a relationship with the mutational status of the tumor cannot be excluded, as reported in previous studies (72-74). Mechanistically, early somatic events driving progenitor clonality and secondary tumor growth are considered responsible for the abovementioned association (75,76). Beside tumor stage, and histology of the primary tumor, male sex was identified in our multivariable model, ROC and decision tree analysis as independent predictor of incomplete tumor resections. There is robust confirmatory evidence that male sex is associated with adverse postoperative outcomes and poor survival upon lung cancer surgery, when compared to the female patients. However, to the best of our knowledge, male sex was not identified until now at multivariable analysis as independent predictor of incomplete resections in large scale patients’ cohorts. While the predictive role of squamous cell carcinoma and tumor size (pT >2) was previously confirmed (19,21), we consider the significant correlation between male sex and R+ a novel aspect, to date insufficiently discussed. Specifically, among all patients with incomplete resections, 83.3% of the patients were of male sex. The association between male sex and R+ could be linked with the pTNM descriptors, histology of the primary tumor, surgical approach and nicotine history and intensity. Accordingly, 52.8% of the male patients had a tumor stage > pT1, while 25% showed lymph node metastasis, along with higher fractions of lymphangiosis (33%), vascular invasion (16.8%), and squamous histology (29.6%) in the resected specimen. Notably, 62.4% of the male patients underwent open surgery and 12% underwent multilobar approaches. In addition, the higher nicotine consumption among male patients who are active smokers, along with the higher prevalence of squamous histology in the R+ group, may represent further explanations for this association. When considering only the R+ subgroup in males, 60% of the patients had squamous histology, 93.3% presented with tumor stage > pT1, 73.3% had lymph node metastasis, 64.3% had lymphangiosis, 41.7% had vascular involvement, while 73.3% of tumors were approached as open procedures and 26.7% requiring multilobar resections. Interestingly, no specific mutational profile could be found in the majority of the male patients with incomplete resections [73.3% no oncogenic driver mutation, 13.3% EGFR mutation, 6.7% KRAS (other) mutation, and 6.7% ROS alteration]. Based on these results, we assume that TNM descriptors, squamous histology, and surgical approach could explain an R+ status to some extent. Accordingly, these findings align with existing evidence that advanced TNM descriptors are unfavorable predictors and thus significantly linked with a worse survival in male patients (9,77). On the contrary, a survival advantage upon lung cancer treatment was reported in female patients in several cohorts with varying ethnic and demographic characteristics (77-80). Furthermore, one Swedish high-volume patient cohort found this survival advantage in female patients independent of age, comorbidities, socioeconomic status, physical performance, type and extent of surgery, tumor stage and characteristics (80). In summary, patients with R1, were in majority male sex, and presented a reduced FEV1 [63.50 (51.00, 76.50) % predicted] and large tumors (pT4 9 of 18 patients, 50%) approached in the majority of cases by open procedures (14 of 18 patients, 77.8%) with extended resections (at least lobectomy, 16 of 18 patients, 88.9%). When addressing the individual patients’ characteristics that potentially led to R+ resection, further aspects could be emphasized. The impossibility of surgical extension (and thus no further intraoperative frozen sections examination) might be justified in certain situations by very low FEV1 values <50% (in 4 of 18 patients, 22.2%), vascular or pleura /chest wall involvement (in 8 of 18 patients, 44.4%), and large (pT4) and centrally located tumors (in 9 of 18 patients, 50%). Of note, some patients have a combination of abovementioned factors (e.g., parenchymal/chest/vessel involvement, low FEV1, pT4 tumor) thus justifying the R+ outcome. These parameters reflect the heterogeneity of NSCLC patients admitted for surgical resection. In clinical practice, managing such diverse patient groups requires individualized decisions regarding the extent of resection. Consequently, no predefined surgical approach (e.g., segmentectomy or lobectomy or pneumonectomy) can be applied in advance for all patients. Thus, we deliberately pooled all patients and analyzed the entire cohort irrespective of the extent of surgery, in order to present real-word data from our surgical unit. One important aspect that could be improved in the surgical routine is the surgical management of the resected margins. Basically, frozen section examination of the resection margins is not mandatory in macroscopically non-suspicious cases. However, in 8 of 18 (44.4%) histological specimens where parenchyma resection margins were found positive (R1) at the final histological analysis, no intraoperative frozen section examination was performed. As a point of criticism for our study, we consider that a more frequent use of frozen section examination by suspected adjacent tumor involvement could possibly lead to an extended peripheral or central resection, which in turn, could decrease the R+ rate. The present study has several limitations. Due to the retrospective nature of the study, the assessment of clinical parameters (demographics, comorbidities, postoperative course and follow-up) was not always possible. Thus, the analysis, focused on descriptive distributions could not be adjusted for potential confounders. Due to the various R+ localizations and patterns upon surgery, pooling all surgical procedures into a single analytical framework, could introduce potential heterogeneity and confounding effects. In addition, TNM descriptors and histology may partly reflect procedure-selection bias in preoperative decision-making regarding the extent of resection. To minimize methodological bias, our study design aligned with large registry databases (e.g., Dutch national database with n=7,156 patients, assessed by Rasing et al.) that clearly identified TNM descriptors and squamous histology as independent predictors for surgical margin completeness (21). In addition, we focused on the morpho-molecular profile of the primary tumors and R+ status, characteristics that have been consistently investigated in our analysis, with robust results aligning with previous studies. One additional limitation represents the relatively low sample size in comparison to large scale database registries. Consecutively, the relatively low number of events limits the robustness of the multivariable model, potentially leading to misleading interpretations. A final limitation is the monocentric design of the study, without an external validation. In order to demonstrate their generalizability, the presented findings should be further evaluated in large prospective multicenter studies. Conclusions Taken together, our study addressing the relationship between topographical, morpho-functional and mutational profile of the primary tumor with the resection margin completeness, identified three clinical parameters that influence resection status: tumor stage, squamous histology, and male sex. A careful assessment of baseline clinical demographics, TNM descriptors, and morpho-molecular status of the tumor can guide risk stratification, refine perioperative care, and enhance patient outcomes. Supplementary The article’s supplementary files as Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0511/prf