Clinical research advances in thoracic surgery and thoracic oncology in 2025: a panoramic narrative review

Abstract

Background and Objective

The year 2025 brought important refinements in thoracic surgery and thoracic oncology, particularly in risk-adapted screening, biomarker-guided perioperative therapy, minimally invasive surgery, and perioperative recovery. This narrative review summarizes clinically relevant evidence reported in 2025 and discusses how these updates may influence multidisciplinary practice.

Methods

We conducted a narrative review of literature and conference reports published or presented between January 1 and December 31, 2025. PubMed, Google Scholar, and the official abstract proceedings/websites of the American Society of Clinical Oncology (ASCO), the World Conference on Lung Cancer (WCLC), and the European Society for Medical Oncology (ESMO) were searched using combinations of disease-, treatment-, and surgery-related keywords. Earlier pivotal trials were cited selectively to provide background context where needed.

Key Content and Findings

The National Comprehensive Cancer Network (NCCN) updated lung cancer screening criteria toward broader risk inclusion. In resectable non-small cell lung cancer (NSCLC), 2025 data further refined perioperative strategies for molecularly selected and driver-negative populations, including long-term outcomes from immunotherapy trials and phase III evidence for neoadjuvant targeted therapy. Robot-assisted thoracic surgery (RATS) continued to show advantages in selected complex settings, especially after induction treatment. Progress in mesothelioma, esophageal cancer, perioperative rehabilitation, and digital symptom monitoring also underscored the increasing importance of whole-patient care. However, several emerging tools, including circulating tumor DNA (ctDNA) and artificial intelligence (AI)-assisted models, still require prospective validation, standardization, and broader accessibility before routine implementation.

Conclusions

Recent advances in thoracic oncology increasingly support a more precise, multidisciplinary, and patient-centered model of care. The most meaningful progress in 2025 lay not only in new treatments, but also in better integration of molecular stratification, surgical decision-making, and perioperative management.

Introduction

Thoracic malignancies remain leading causes of cancer-related mortality worldwide, with lung cancer accounting for approximately 1.8 million deaths annually. The landscape of thoracic oncology has undergone rapid transformation through advances in early detection technologies, molecular diagnostics, immunotherapy, and surgical techniques. The year 2025 witnessed critical developments across the disease continuum, from screening and early diagnosis to curative-intent treatment and perioperative care. Because the greatest volume of practice-changing evidence reported in 2025 was concentrated in lung cancer, the present review is necessarily lung cancer-centered while also summarizing selected developments in other thoracic malignancies, including mesothelioma, esophageal cancer, and thymic tumors.

This review synthesizes key clinical trial results, technological innovations, and practice-changing evidence published in 2025, focusing on non-small cell lung cancer (NSCLC), small cell lung cancer (SCLC), malignant pleural mesothelioma (MPM), esophageal cancer, and thymic malignancies. We examine how these advances translate into improved patient outcomes and discuss remaining challenges in optimizing multidisciplinary care for thoracic cancer patients. We present this article in accordance with the Narrative Review reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0478/rc).

Methods

This narrative review focused on clinically relevant advances in thoracic surgery and thoracic oncology reported in 2025. Searches were conducted in PubMed and Google Scholar, together with manual screening of the official conference proceedings/websites of American Society of Clinical Oncology (ASCO), World Conference on Lung Cancer (WCLC), and European Society for Medical Oncology (ESMO), for material published or presented between January 1 and December 31, 2025. Search terms combined disease-related terms (“NSCLC”, “SCLC”, “mesothelioma”, “esophageal cancer”, “thymic tumor”), management-related terms (“screening”, “ctDNA”, “immunotherapy”, “targeted therapy”, “perioperative”, “adjuvant”, “neoadjuvant”), and surgery-related terms (“thoracic surgery”, “segmentectomy”, “lobectomy”, “robot-assisted thoracic surgery”, “video-assisted thoracic surgery”, “prehabilitation”, and “ERAS”). Boolean operators (AND/OR) were used as appropriate. The full search strategy, including databases, search terms, and selection criteria, is summarized in Table 1.

Table 1. The search strategy summary.

Items	Specification
Date of search	December 31, 2025
Databases and other sources searched	PubMed, Google Scholar, and official abstract proceedings/websites of ASCO, WCLC, and ESMO
Search terms used	Combinations of disease-related, treatment-related, and surgery-related terms (e.g., “NSCLC”, “screening”, “ctDNA”, “immunotherapy”, “targeted therapy”, “RATS”, “VATS”, “ERAS”, “prehabilitation”); Boolean operators AND/OR were used
Timeframe	January 1, 2025 to December 31, 2025; earlier pivotal studies were cited selectively for clinical context
Inclusion and exclusion criteria	Inclusion: guidelines, clinical trials, major follow-up analyses, meta-analyses, and clinically relevant cohort studies/meeting abstracts. Exclusion: animal studies, purely mechanistic preclinical reports, and publications without direct thoracic oncology relevance
Selection process	The study selection was independently conducted by Z.S. and X.L., with discrepancies resolved by discussion

ASCO, American Society of Clinical Oncology; ctDNA, circulating tumor DNA; ERAS, enhanced recovery after surgery; ESMO, European Society for Medical Oncology; RATS, robot-assisted thoracic surgery; VATS, video-assisted thoracic surgery; WCLC, World Conference on Lung Cancer.

Because this was a narrative review, the aim was not exhaustive study capture but transparent selection of clinically influential evidence. Priority was given to guidelines, phase II/III trials, major follow-up analyses, meta-analyses, and practice-relevant cohort studies reported in 2025. Earlier pivotal studies were included selectively when necessary to provide background context for 2025 developments. Animal studies, purely mechanistic preclinical reports, and publications without direct relevance to thoracic oncology practice were generally excluded. An example detailed PubMed search strategy may be provided in Table S1 to improve transparency.

Advances in early diagnosis and screening

Updates in National Comprehensive Cancer Network (NCCN) screening guidelines

The NCCN clinical practice guidelines for lung cancer screening (v1.2025) updated the recommendation by removing the restriction that individuals who quit smoking more than 15 years ago do not require continued screening (1). This historic revision recognizes that while lung cancer risk decreases with cessation duration, it never returns to baseline. Multiple 2025 cohort studies demonstrated former smokers maintain elevated risk warranting lifelong surveillance, thereby broadening the population potentially considered for screening beyond previous USPSTF (U.S. Preventive Services Task Force) criteria.

Universal screening beyond traditional risk stratification

Wenhua Liang’s team published in JAMA showing low-dose computed tomography (LDCT) achieved 1–2% detection rates in never-smokers and light smokers, comparable to traditional high-risk populations, challenging current risk-stratification models (2). A meta-analysis of 30 studies (n=3,287) showed tumor-informed circulating tumor DNA (ctDNA) strategies achieved 0.97 specificity vs. 0.93 for tumor-agnostic approaches in landmark analysis, while tumor-agnostic methods showed higher sensitivity (0.79 vs. 0.76) in longitudinal monitoring (3). Circulating tumor cell (CTC) meta-analysis (8 studies, n=1,601) demonstrated 75% sensitivity and 89% specificity for lung cancer detection, while a seven-autoantibody classifier achieved >92% specificity in 1,987 patients with pulmonary nodules (4).

Multimodal diagnosis and artificial intelligence (AI) integration

The DECIPHER-NODL study integrated LDCT radiomics with plasma cfDNA fragmentomics in 1,356 subjects, significantly improving malignant nodule classification and invasive adenocarcinoma identification over single-modality imaging (5). Yongde Liao’s team used LightGBM algorithm with computed tomography (CT) features to predict spread through air spaces (STAS) in 1,325 lung adenocarcinoma patients across three centers, achieving area under the curve (AUC) 0.756 in external validation using SHapley Additive exPlanations (SHAP) method for feature contribution quantification (6). Yilong Wu’s ADAURA subgroup analysis in Nature Medicine showed 68% of recurrence events occurred after osimertinib cessation, with 58% within 12 months, confirming minimal residual disease (MRD) monitoring can identify patients needing extended therapy (7). Table 2 summarizes the performance metrics and clinical applications of these innovative diagnostic technologies, illustrating the paradigm shift from image-based judgment to “imaging-molecular-intelligent” collaborative decision-making.

Table 2. Emerging technologies for early diagnosis, recurrence monitoring, and risk stratification in thoracic oncology.

Technology/approach	Study/platform	Population	Key performance metrics	Clinical application
Universal LDCT screening	Liang/He team (JAMA)	Non-risk-stratified general population	Detection rate: 1–2% (comparable to high-risk)	Challenges traditional risk models; supports broader screening
Multimodal AI model	DECIPHER-NODL	Pulmonary nodules (n=1,356)	Superior to single-modality imaging	Malignancy risk stratification; reduce unnecessary biopsies
STAS prediction model	Liao team (Frontiers)	Lung adenocarcinoma (n=1,325)	AUC 0.756 (external validation)	Guide surgical extent (segmentectomy vs. lobectomy)
ctDNA MRD (tumor-informed)	Meta-analysis (30 studies)	Postoperative NSCLC (n=3,287)	Specificity: 0.97; sensitivity: 0.76	Recurrence monitoring; guide adjuvant duration
ctDNA MRD (tumor-agnostic)	Meta-analysis (30 studies)	Postoperative NSCLC (n=3,287)	Specificity: 0.88; sensitivity: 0.79	Broader applicability without tumor tissue
CTC detection	Meta-analysis (8 studies)	Lung cancer screening (n=1,601)	Sensitivity: 75%; specificity: 89%	Complement to LDCT; reduce false positives
Autoantibody panel	Clinical cohort	Pulmonary nodules (n=1,987)	Specificity: >92%; PPV: >70%	Risk stratification of indeterminate nodules

AI, artificial intelligence; AUC, area under the curve; CTC, circulating tumor cell; ctDNA, circulating tumor DNA; LDCT, low-dose computed tomography; MRD, minimal residual disease; NSCLC, non-small cell lung cancer; PPV, positive predictive value; STAS, spread through air spaces.

Despite their clinical promise, each modality faces distinct limitations. ctDNA assays are affected by inter-platform variability, pre-analytical factors, and the lack of standardized detection thresholds (8). cfDNA fragmentomics remains technically complex, with evolving analytical frameworks and limited prospective validation (9,10). AI-assisted imaging and integrative models rely on high-dimensional data and require further large-scale validation, with ongoing challenges in generalizability and clinical integration (11). In addition, high costs and unequal access across healthcare settings further hinder widespread adoption. Prospective validation, harmonized workflows, and improved interoperability will be essential before routine implementation.

Targeted therapy in early-stage NSCLC

Actionable driver genes including EGFR, ALK, ROS1, BRAF V600E, MET ex14, RET, HER2, and KRAS G12C constitute mature diagnostic pillars, with molecular detection advancing from late-stage salvage to preoperative diagnosis in resectable disease (12). The strategic focus has shifted from late-stage salvage therapy to preoperative diagnosis in resectable disease, establishing a comprehensive framework with precise molecular typing as the starting point and surgical resection as the core anchor. Key perioperative trials discussed in this review are summarized in Table 3.

Table 3. Key clinical trials shaping perioperative systemic therapy and related therapeutic strategies in NSCLC reported or updated in 2025.

Study name	Population	Intervention	Primary endpoint	Key results	Clinical impact
NeoADAURA	Stage II–IIIB EGFR+ NSCLC	Neoadjuvant osimertinib + chemo vs. chemo alone	MPR rate	MPR: 26% vs. 2% (P<0.0001); N2 downstaging: 53% vs. 21%	First phase III evidence for neoadjuvant targeted therapy in EGFR+ population
CheckMate 816 (5-year)	Stage IB–IIIA NSCLC	Neoadjuvant nivolumab + chemo vs. chemo alone	OS	5-year OS: 65% vs. 55%; pCR patients: 95% 5-year OS	Established pCR as reliable surrogate for long-term survival
KEYNOTE-671	Stage II–IIIB NSCLC	Perioperative pembrolizumab + chemo vs. chemo + placebo	EFS, OS	EFS: HR 0.58; OS: HR 0.72; pCR: 18.1% vs. 4.0%	Validated ‘sandwich’ perioperative immunotherapy model
RATIONALE-315	Stage II–IIIA NSCLC (Chinese)	Perioperative tislelizumab + chemo vs. chemo + placebo	EFS, OS	pCR: 40.7% vs. 5.5%; 36-month OS: 79.3% vs. 69.3% (HR 0.62)	Strongest OS benefit in Asian population

Chemo, chemotherapy; EFS, event-free survival; EGFR, epidermal growth factor receptor; HR, hazard ratio; MPR, major pathological response; NSCLC, non-small cell lung cancer; OS, overall survival; pCR, pathological complete response.

Epidermal growth factor receptor (EGFR)-mutant disease: neoadjuvant breakthrough

While osimertinib has already established standard-of-care paradigms for EGFR-mutant NSCLC in both the post-surgical adjuvant setting [ADAURA, demonstrating an 88% 5-year overall survival (OS)] (13) and the unresectable stage III setting [LAURA, achieving an unprecedented 39.1-month median progression-free survival (PFS) following chemoradiotherapy] (14), optimizing pre-surgical downstaging remained a critical unmet need for thoracic surgeons. Filling this crucial perioperative gap, the phase III NeoADAURA trial emerged as a landmark neoadjuvant breakthrough. The study demonstrated that neoadjuvant osimertinib plus chemotherapy achieved a 26% major pathologic response (MPR), and osimertinib monotherapy achieved 25%, both vastly outperforming the 2% MPR seen with chemotherapy alone (P<0.0001) (15). Crucially for surgical planning, in baseline N2 disease, these targeted regimens facilitated a robust 53% lymph node downstaging rate compared to just 21% with chemotherapy (10). The biological rationale for such profound neoadjuvant efficacy is supported by prospective analyses of postoperative specimens, which revealed that patients achieving MPR/pathologic complete response (pCR) exhibit significant CD8⁺ T cell and M1 macrophage enrichment with reversed immunosuppression [although anaplastic lymphoma kinase (ALK) fusion cohorts showed even higher baseline immune activation than EGFR-mutant patients] (16).

Bispecific antibodies and emerging targets

The HARMONI-2 trial demonstrated ivonescimab [programmed cell death protein 1 (PD-1)/vascular endothelial growth factor (VEGF) bispecific antibody] achieved 11.14 months median PFS in programmed death-ligand 1 (PD-L1) positive NSCLC, significantly exceeding pembrolizumab’s 5.8 months [hazard ratio (HR) 0.51], with dual mechanism blocking VEGF-induced vessel normalization to improve drug penetration (17). Neoadjuvant alectinib in ALK-positive stage IB–IIIA NSCLC achieved 67% MPR and 24% pCR with 96% proceeding to surgery and 100% R0 resection, while ALINA trial confirmed adjuvant alectinib achieved 93% 3-year disease-free survival (DFS) vs. 63% with chemotherapy (HR 0.24), reducing central nervous system (CNS) recurrence from 18% to 3%. Although HARMONI-2 was conducted in advanced rather than resectable disease, it highlights the potential of dual-pathway biologics that may inform future perioperative development.

Surgical integration challenges

Long-term targeted therapy induces local tissue scarring and dense mediastinal lymph node fibrosis, significantly increasing anatomical difficulty and thoracotomy conversion likelihood (18,19). For BRAF V600E mutation with extensive mediastinal metastasis, short-term neoadjuvant BRAF+ MEK inhibitor successfully achieved “unresectable” to “R0 resection and pCR” conversion (20). Dynamic monitoring revealed genetic transformation such as EGFR to ALK fusion after recurrence, emphasizing importance of postoperative re-biopsy and ctDNA for molecular spectrum updates (21).

Perioperative immunotherapy

Following breakthroughs in metastatic NSCLC, immune checkpoint inhibitors rapidly expanded to early and locally advanced resectable stages, forming coexistence of neoadjuvant, adjuvant, and perioperative strategies (22). Based on CheckMate 816, AEGEAN, and KEYNOTE-671 phase III trials, NCCN, ESMO, and CSCO guidelines listed neoadjuvant/perioperative immunochemotherapy as standard treatment for stage II–III driver gene-negative resectable NSCLC, significantly improving pCR, MPR, and event-free survival (EFS).

CheckMate 816 5-year milestone

CheckMate 816 5-year follow-up published in NEJM showed 65% OS in nivolumab group vs. 55% in chemotherapy, with pCR patients achieving 95% 5-year survival and no lung cancer-related deaths (23). This provided longer-term support for the association neoadjuvant immunotherapy’s deep pathological remission translates into survival benefits, establishing immunoneoadjuvant therapy as standard for stage IIIA NSCLC with negative driver mutations.

Perioperative sandwich mode standardization

KEYNOTE-671 updated analysis showed neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant maintenance achieved EFS HR 0.58 and OS HR 0.72 in stage II–IIIB NSCLC, with 18.1% pCR vs. 4.0% and 3-year EFS of 93% in pCR patients vs. 67% in non-pCR patients (24). These trials support perioperative immunochemotherapy as an evidence-based option for selected patients with driver-negative resectable NSCLC; however, the optimal sequencing, duration, and biomarker-guided tailoring of therapy remain areas of ongoing investigation.

Chinese evidence and strategy comparison

RATIONALE-315 showed tislelizumab achieved 36-month OS rate 79.3% vs. 69.3% (HR 0.62) with 40.7% pCR vs. 5.5%, becoming first Chinese phase III trial confirming perioperative immunochemotherapy OS benefits (25). Neotorch study showed toripalimab achieved overall HR 0.40 for EFS, particularly prominent in stage III subgroup (HR 0.40) (26). JNCI network meta-analysis of 11 randomized trials showed neoadjuvant/perioperative immunochemotherapy significantly improved EFS (HR ~0.58) and OS (HR ~0.65) vs. chemotherapy alone, while adjuvant-only immunotherapy showed limited benefits especially in PD-L1 tumor cell score (TC) <50% population (27). Table 3 provides a comprehensive comparison of these landmark trials, highlighting their distinct contributions to the perioperative treatment landscape.

Biomarkers and special populations

Tumor mutational burden (TMB) ≥10 mutations/megabase predicted enhanced benefit with 72% 3-year EFS vs. 54% in TMB-low patients (28). EGFR/ALK-positive resectable NSCLC showed only 12.6% MPR vs. 38.9% (P<0.001), with current consensus preferring neoadjuvant targeted therapy (29). Miyakoshi et al. identified “low TPS/high IC” phenotype [tumor proportion score (TPS) <50%/immune cell score (IC) >1] where dual immunotherapy containing cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) inhibitor achieved 12.4 months median PFS vs. pembrolizumab + chemotherapy’s 6.6 months [restricted mean survival time (RMST) difference 1.5, P=0.049] (30). Multicenter real-world study of initially unresectable stage III NSCLC showed 41.3% imaging reduction rate after neoadjuvant immunochemotherapy, with 28.7% achieving R0 resection and 2-year OS 81.6% vs. 49.2% in non-operation group (31). NEO-PIONEER trial using “radiotherapy (24 Gy/3) + tislelizumab + anlotinib” achieved 32.4% pCR and 54.1% MPR in stage II–IIIA resectable NSCLC (32). Table 4 stratifies these biomarker-guided efficacy patterns, providing evidence-based recommendations for treatment selection across different molecular subgroups.

Table 4. Biomarker-stratified patterns of benefit from perioperative immunotherapy in resectable NSCLC.

Biomarker	Subgroup	Representative EFS/clinical benefit measure	Clinical recommendation	Evidence level
PD-L1 TPS	≥50%	0.41–0.52	Strongly recommended; greatest benefit	High (phase III RCTs)
	1–49%	0.58–0.69	Recommended; moderate benefit	High (phase III RCTs)
	<1%	0.72–0.88 (some NS)	Consider; limited but potential benefit	Moderate (subgroup analyses)
TMB	High (≥10 mut/Mb)	3-year EFS: 72%	Recommended; enhanced benefit	Moderate (exploratory analyses)
	Low (<10 mut/Mb)	3-year EFS: 54%	Standard recommendation	Moderate
Driver mutations	EGFR/ALK positive	0.89–1.02 (NS)	Not recommended; prefer targeted therapy	High (pooled analyses)
	EGFR/ALK negative	0.55–0.68	Strongly recommended; standard of care	High (phase III RCTs)
ctDNA clearance	Post-neoadjuvant clearance	2-year EFS: 86% vs. 54% (non-clearers)	Favorable prognostic indicator; guide adjuvant decisions	Moderate (post-hoc analyses)

Statistical terms: HR <1 indicates a reduced risk of an event in the treatment group compared with the control group. The third column summarizes representative efficacy data reported in the cited studies, including hazard ratios and time-point EFS/survival comparisons where available. ALK, anaplastic lymphoma kinase; ctDNA, circulating tumor DNA; EFS, event-free survival; EGFR, epidermal growth factor receptor; HR, hazard ratio; NS, not significant; NSCLC, non-small cell lung cancer; PD-L1, programmed death-ligand 1; RCT, randomized controlled trial; TMB, tumor mutational burden; TPS, tumor proportion score.

Surgical practice and adjuvant follow-up

Real-world cohort showed hilar and mediastinal lymph node fibrosis incidence significantly increased after neoadjuvant immunochemotherapy (31% vs. 12%), with robot-assisted thoracic surgery (RATS) conversion rate significantly lower than video-assisted thoracic surgery (VATS) (3.1% vs. 8.9%) and superior lymph node dissection quality (33). IMpower010 5-year data showed atezolizumab achieved 68.3% 5-year DFS vs. 51.2% (HR 0.48) in PD-L1 TC >50% stage II–IIIA population, representing longest follow-up evidence in adjuvant immunotherapy field (34).

Surgical advances

Sublobar resection precision

A study of 284 pure ground-glass opacity (GGO) adenocarcinoma patients showed no significant 5-year RFS difference between full-margin (97.9%) and insufficient-margin groups (98.2%, P=0.530) by NCCN criteria, but “margin >5 mm and ratio >0.25” joint standard was independent risk factor (HR ~7.80, P=0.025) (35). Multicenter cohort (n=999) showed segmentectomy reduced “death from other causes” (2.5% vs. 5.9%), including secondary malignancies (1.2% vs. 2.6%) and cardiovascular deaths (1.2% vs. 3.3%), with higher proportion receiving guideline-consistent treatment for secondary tumors due to retained lung function (36). Fudan cohort (n=888 T1c solid-dominant) showed segmentectomy achieved non-inferior 5-year RFS (89.0% vs. 81.4%) and OS (91.7% vs. 88.1%) after matching, with low-risk patients without visceral pleural invasion (VPI)/lymphovascular invasion (LVI) showing superior RFS (HR ~0.12–0.13), while VPI/LVI presence favored lobectomy for local control (37). Table 5 delineates these precision criteria for sub lobar resection, enabling individualized surgical decision-making based on lesion-specific characteristics.

Table 5. Precision criteria for sublobar resection in early-stage NSCLC.

Lesion characteristics	Recommended procedure	Margin criteria	Key evidence	5-year outcomes
Pure GGO (any size)	Segmentectomy preferred	Combined: ≥5 mm AND ratio ≥0.25	Retrospective (n=284); HR 7.80 if criteria not met	RFS: 97.9% (full-margin) vs. 98.2% (insufficient-margin); very low recurrence (1.8%)
Part-solid GGO (CTR <0.5)	Segmentectomy	Standard anatomic margins	Multiple cohorts	Equivalent to lobectomy
T1c solid-dominant (2–3 cm) without VPI/LVI	Segmentectomy	Standard anatomic margins	Propensity-matched (n=888); HR ~0.12–0.13	Superior RFS vs. lobectomy in low-risk subgroup
T1c solid-dominant (2–3 cm) with VPI/LVI	Lobectomy	N/A	Propensity-matched (n=888)	Superior locoregional control with lobectomy
T1a-b solid (≤2 cm)	Segmentectomy	Standard anatomic margins	JCOG0802, CALGB140503	Non-inferior to lobectomy

CTR, consolidation-to-tumor ratio; GGO, ground-glass opacity; HR, hazard ratio; LVI, lymph vascular invasion; N/A, not applicable; NSCLC, non-small cell lung cancer; RFS, recurrence-free survival; VPI, visceral pleural invasion.

Robot-assisted surgery evidence

Journal of Robotic Surgery systematic review confirmed RATS lymph node dissection and R0 resection rates equivalent to VATS with medium-to-high evidence quality for short-term benefits (38). Meta-analysis of randomized trials showed RATS and VATS similar in hospital stay, complications, and 30-day mortality, but RATS showed improvement trends in N2 lymph node dissection and thoracotomy conversion rates (39). National Cancer Database (NCDB) analysis (n=301,123, 2010–2021) showed RATS exceeded VATS in 2019 (9,579 vs. 9,454 cases) and reached 65.4% of minimally invasive lobectomies by 2021, with academic centers adopting faster than community hospitals (40). Multicenter cohort (n>600) showed RATS and VATS generally similar in blood loss, chest tube duration, hospital stay, and 30-day complications, with RATS showing lower thoracotomy conversion trend (41). BRAVO trial cost analysis (n=76) showed no significant total cost difference: RATS R$ 35,590 vs. VATS R$ 41,067 (P=0.564) (42). In the BRAVO cost analysis, similar total costs appeared to reflect a trade-off between higher robotic material expenses and lower downstream resource use, including fewer readmissions and lower follow-up-related costs within 90 days.

In post-neoadjuvant therapy patients, RATS significantly reduced the incidence of thoracotomy conversion (3.1% vs. 8.9%) and improved blood loss, operative time, air-leakage incidence, and lymph node dissection quality (33). Japanese propensity-matched study (n=55 pairs) showed no significant difference in 1-year post-lobectomy lung function decline [%forced vital capacity (%FVC), %forced expiratory volume in one second (%FEV1), %peak expiratory flow (%PEF), %diffusing capacity of the lungs for carbon monoxide (%DLCO)] between RATS and VATS (43). Table 6 quantifies the comparative performance of RATS vs. VATS across different clinical scenarios, highlighting where robotic platforms provide measurable advantages.

Table 6. Comparative performance: RATS vs. VATS in complex scenarios.

Clinical scenario	Outcome measure	RATS	VATS	P value	Clinical significance
Post-neoadjuvant therapy	Conversion to thoracotomy	3.1%	8.9%	<0.05	RATS maintains minimally invasive approach
Post-neoadjuvant therapy	Operative time (min)	Shorter	Longer	<0.05	RATS more efficient in fibrotic tissue
Post-neoadjuvant therapy	Total lymph nodes harvested	Higher	Lower	<0.05	RATS superior nodal dissection
Post-neoadjuvant therapy	N2 station dissection rate	Higher	Lower	<0.05	RATS more thorough staging
Standard lobectomy	Hospital length of stay	Similar	Similar	NS	Equivalent recovery
Standard lobectomy	30-day complications	Similar	Similar	NS	Equivalent safety
Standard lobectomy	Total hospital cost	R$35,590	R$41,067	NS (P=0.56)	RATS cost-neutral in experienced centers
Sleeve resection	Operative time	Shorter	Longer	<0.05	RATS technical advantage
1-year post-op	Lung function (%FEV1, %FVC, %DLCO)	Equivalent	Equivalent	NS	No long-term functional difference

DLCO, diffusing capacity of the lung for carbon monoxide; FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity; NS, not significant; RATS, robotic-assisted thoracic surgery; VATS, video-assisted thoracic surgery.

ERAS integration and technology frontier

Prospective cohort (n=82) implementing ERAS with RATS showed most patients reached high QoR40 levels within 24 hours, with only 20% reporting moderate pain at 3 weeks, identifying preoperative anxiety, intraoperative bleeding, and analgesia strategies as recovery quality factors (44). Regional hospital randomized controlled trial (RCT) confirmed enhanced recovery after surgery (ERAS) significantly shortened hospitalization, reduced pulmonary complications, and improved recovery quality scores (45). Shurui single-port robot prospective study showed successful completion of predetermined single-port path in vast majority of operations with perioperative risk control similar to multi-port RATS and core indicators equivalent to traditional VATS (46). Fast track pathway study integrated ion robotic bronchoscopy with robotic resection, achieving 90% diagnostic rate, 100% localization success, and median 6-day hospitalization in 10 patients with small nodules (47).

Other thoracic malignancies

MPM

Review article noted nivolumab + ipilimumab became systemic treatment cornerstone in unresectable stage, with observational studies showing the lung-sparing pleurectomy/decortication (PD) improved functional outcomes vs. extrapleural pneumonectomy (EPP) with similar survival (48). Cancers retrospective analysis using ninth edition tumor-node-metastasis (TNM) staging found that strictly limiting surgical indications to epithelioid histology, lymph node-negative status, and complete gross resection showed P/D combined with systemic treatment had greater survival benefit than non-surgical group (49). Phase II study in Nature Medicine showed neoadjuvant nivolumab + ipilimumab achieved 28.6 months median OS vs. 18 months historical control, with >80% completing surgery within planned window and controllable toxicity (50). Hypersensitive whole-genome ctDNA detection showed patients with >95% clearance/reduction after neoadjuvant immunotherapy had significantly longer survival, while continuous positivity indicated early progression risk. EORTC-SPECTRA Arcagen platform identified actionable mutations in ~30% of rare thoracic tumors including MPM, adjusting treatment strategies accordingly (51).

Esophageal cancer and antibody-drug conjugates (ADCs)

Phase III SCIENCE trial showed sintilimab combined with concurrent chemoradiotherapy achieved 60% pCR vs. 47.3% with chemoradiotherapy alone or 13% with immunochemotherapy, confirming synergistic effect with all patients achieving R0 resection (52). Real-world study (n=99) confirmed similar pathological remission rates without increased major complications, with baseline neutrophil/lymphocyte ratio associated with tumor regression degree (53). KEYNOTE-590 5-year follow-up showed pembrolizumab plus chemotherapy increased 5-year survival from 3.0% to 10.6% in advanced esophageal cancer, more than tripling long-term survival chances, with 1-year PFS increasing from 0% to 5.5% (54). For metastatic esophageal squamous cell carcinoma (ESCC) after immunotherapy progression, EGFR/human epidermal growth factor receptor 3 (HER3) bispecific ADC BLB01D1 phase 1b study (n=82) showed 32.9% objective response rate (ORR) and 71.2% disease control rate (DCR), with 2.5 mg/kg dose achieving 39.6% ORR, 79.2% DCR, 5.4 months median PFS, and 11.5 months median OS (55). Another practice-informing development in 2025 was the phase III SANO trial, in which active surveillance after a clinical complete response to neoadjuvant chemoradiotherapy was non-inferior to standard surgery for 2-year OS in the modified intention-to-treat analysis. These findings suggest that response-adapted organ-preservation strategies may be feasible in carefully selected patients, although longer follow-up and rigorous surveillance protocols remain essential (56).

Thymic tumors

EORTC-SPECTRA Arcagen cohort included 46 thymic epithelial tumors (TETs) cases (23 thymic cancer, 23 thymoma), identifying actionable mutations in ~30% through comprehensive genomic analysis to support clinical trial entry or off-label targeted drug use (51). This provides a realistic path for precise treatment: surgery obtains high-quality tissue for molecular typing in resectable/borderline resectable cases, while recurrent/progressive cases can rely on molecular tumor board (MTB) consultation to find early clinical trials. Table 7 summarizes these advances in rare thoracic malignancies, demonstrating how precision medicine principles are extending beyond NSCLC.

Table 7. Advances in rare thoracic malignancies.

Disease	Study/intervention	Design	Key results	Breakthrough significance
Malignant pleural mesothelioma	Perioperative nivolumab + ipilimumab	Phase II	Median OS: 28.6 months (vs. ~18 months historical); ctDNA clearance predicts survival	First perioperative immunotherapy benefit in MPM; validates ctDNA monitoring
Esophageal squamous cell carcinoma	SCIENCE trial: Neoadjuvant sintilimab + chemoRT	Phase III	pCR: 60% vs. 47.3% (chemoRT alone)	Establishes synergy of immunotherapy + chemoradiotherapy
Esophageal cancer (advanced)	KEYNOTE-590 (5-year follow-up)	Phase III	5-year OS: 10.6% vs. 3.0% (chemo alone)	>3-fold increase in long-term survival; paradigm shift to “clinical cure”
Esophageal SCC (post-immunoresistance)	BLB01D1 (EGFR/HER3 bispecific ADC)	Phase Ib	ORR: 32.9% (39.6% at 2.5 mg/kg); mPFS: 5.4 months	Opens “second front” after immunotherapy failure
BRAF V600E+ NSCLC (unresectable)	Neoadjuvant BRAF + MEK inhibitor	Case report/series	Transformation to R0 resection + pCR	Proof-of-concept for rare mutation-targeted neoadjuvant therapy

ADC, antibody-drug conjugate; chemoRT, chemoradiotherapy; ctDNA, circulating tumor DNA; EGFR, epidermal growth factor receptor; HER3, human epidermal growth factor receptor 3; mPFS, median progression-free survival; MPM, malignant pleural mesothelioma; NSCLC, non-small cell lung cancer; ORR, objective response rate; OS, overall survival; pCR, pathological complete response; SCC, squamous cell carcinoma.

Prehabilitation and pulmonary rehabilitation

UK prospective cohort showed standardized pulmonary rehabilitation (PR) program (respiratory training, aerobic exercise, health education) reduced postoperative pulmonary complication (PPC) risk by ~60% [relative risk (RR) ~0.55] and significantly improved quality of life at 6 weeks and 6 months, with physical function scores ~6.6% higher than non-participants (57). Updated review emphasized multimodal prehabilitation for high-risk groups (advanced age, chronic lung disease, limited exercise tolerance) over 2–4 weeks significantly improved 6-minute walking distance and respiratory muscle strength, associated with downward PPC trends and shortened hospital stays (58).

Airway management and inhalation therapy

Multidisciplinary international expert consensus issued in 2025 systematically integrated airway management experience through improved Delphi method, reaching >70% agreement on preoperative evaluation, anesthesia strategies, non-intubated minimally invasive surgery, and postoperative management (59,60). Consensus clearly recommended tubeless minimally invasive thoracic surgery in appropriate patients to reduce airway mechanical stimulation and intubation-related complications. Update significantly strengthened inhalation therapy [inhaled corticosteroid (ICS) and long-acting bronchodilators] advancement to perioperative standard path, emphasizing individualized scheme adjustment according to airway inflammation load and bronchial hyperresponsiveness, with hierarchical management strategy for chronic obstructive pulmonary disease (COPD) and asthma patients.

Postoperative cough and lung function protection

Single-center RCT showed budesonide/glycopyrronium/formoterol (BGF) triple inhaler significantly reduced incidence of cough lasting ≥14 days post-lobectomy (13.7% vs. 40.4%), with significantly better cough-related quality of life scores at days 14 and 30, and only mild adverse reactions such as transient palpitations (61). A retrospective cohort study using portable budesonide-formoterol dry powder inhaler showed Leicester Cough Questionnaire in Mandarin-Chinese (LCQ-MC) cough score increased by ~2 points at 1 month post-thoracoscopic pneumonectomy, exceeding minimum clinically important difference threshold, with FEV1 and PEF decreasing less than control group (62).

ERAS pathway deepening and analgesia

Multiple 2025 studies reinforced ERAS protocol value in shortening hospital stay, reducing complications, and improving quality of life, incorporating “protocol adherence” and “analgesia quality” as key performance indicators, though significant heterogeneity was noted across countries, healthcare systems, and institutions (63). Multicenter practice study showed high adherence to key components (early chest tube removal, early mobilization, nutritional support) significantly associated with reduced hospital stay, decreased pulmonary complications, and lower costs (64). Meta-analysis of RCTs (n=1,524) showed regional blockade techniques (paraspinal, serratus anterior plane, intercostal nerve) significantly reduced Visual Analog Scale (VAS) pain scores and decreased patient-controlled analgesia use without increasing serious adverse events (65). Real-world pain score observation found combination of thoracoscopic/robotic surgery, regional blockade, and standardized ERAS achieved “mild” pain control in most patients by postoperative days 3–5, accelerating daily activity recovery (44).

Digital patient-reported outcome (PRO) and long-term function

A thoracoscopic surgery prospective cohort (n=550 unilateral multiple pulmonary nodules) evaluated single-port VATS resection by longitudinal PRO, showing multi-lobectomy group had significantly worse pain, dyspnea, sleep disorder, and fatigue scores within 4 weeks, with more obvious walking and daily activity limitations (66). Cross-sectional study (n=441) identified two PRO patterns through latent class analysis: “high symptom burden” and “low symptom burden,” with high burden group showing significantly heavier pain, dyspnea, anxiety, and sleep disorders, significantly decreased Karnofsky scores, related to operation time, kinesiophobia score, and analgesic pump use (67). TD-WELLBEING multicenter RCT (n=355) showed electronic patient-reported outcome (ePRO) intervention group had significantly lower comprehensive symptom scores at 12 weeks, with significantly higher EORTC QLQ-C30 emotional function and overall health scores persisting beyond 6 months (66). Prospective cohort (n=60) using ePRO compared non-intubated thoracic surgery (NITS) with traditional double-lumen anesthesia VATS, showing NITS group significantly better in hospitalization time, chest tube duration and drainage volume, with lower multi-dimensional ePRO scores (pain, dyspnea, cough, general discomfort) at days 7 and 30 (59). Japanese prospective study analyzed health-related quality of life (HRQOL) and employment status changes after anatomical resection with 1-year follow-up, showing most HRQOL dimensions restored to near baseline within 3–6 months, but ~30% did not fully return to work or maintain original work intensity at 12 months, with influencing factors including age, preoperative physical activity level, and severe fatigue/anxiety PRO scores (68). Systematic review and meta-analysis (18 studies, n=88,413) showed obese patients had lower postoperative mortality [odds ratio (OR) 0.73] without significant increase in overall complications or pulmonary/cardiovascular complications, though operation time slightly prolonged, demonstrating “obesity paradox” (69). Table 8 consolidates these evidence-based perioperative interventions, quantifying their clinical impact and providing actionable guidance for implementation.

Table 8. Evidence-based perioperative management interventions.

Intervention category	Specific intervention	Study design	Primary outcome	Effect size	NNT/clinical impact
Prehabilitation	Standardized pulmonary rehabilitation	Prospective cohort	PPC	RR 0.55 (~60% reduction)	NNT ~5–6 to prevent 1 PPC
Postoperative cough	Triple inhaler (ICS/LABA/LAMA)	RCT	Persistent cough ≥14 days	13.7% vs. 40.4% (P<0.001)	NNT ~4 to prevent 1 case of persistent cough
Postoperative cough	ICS/LABA combination	RCT	LCQ-MC score at 1 month	+2 points (exceeds MCID)	Clinically meaningful improvement
Lung function protection	ICS/LABA post-op	RCT	FEV1, PEF decline at 1 month	Less decline vs. control	Preserves functional reserve
Analgesia	Regional blockade (PVB, SAPB)	Meta-analysis (RCTs, n=1,524)	VAS pain scores, rescue analgesia	Significant reduction; no increase in AEs	Multimodal analgesia standard
ERAS protocol	High adherence to ERAS	Multicenter cohort	Hospital LOS, complications, costs	Significant reduction in all three	Adherence more important than individual elements
Digital follow-up	ePRO collaborative platform	RCT (TD-WELLBEING, n=355)	Symptom burden at 12 weeks & 6 months	Significantly lower scores; higher QoL	Sustainable long-term benefit

AE, adverse event; ePRO, electronic patient-reported outcome; ERAS, enhanced recovery after surgery; FEV1, forced expiratory volume in 1 second; ICS, inhaled corticosteroid; LABA, long-acting beta-agonist; LAMA, long-acting muscarinic antagonist; LCQ-MC, Mandarin Chinese version of the Leicester Cough Questionnaire; LOS, length of stay; MCID, minimal clinically important difference; NNT, number needed to treat; PEF, peak expiratory flow; PPC, postoperative pulmonary complication; PVB, paravertebral block; QoL, quality of life; RCT, randomized controlled trial; RR, relative risk; SAPB, serratus anterior plane block; VAS, Visual Analog Scale.

Conclusions

Several priorities are likely to shape the next phase of thoracic oncology. First, biomarker-guided perioperative treatment will need to move from broad subgroup analysis to adaptive treatment algorithms integrating molecular subtype, pathological response, and ctDNA dynamics. Prospective trials should clarify how to escalate, de-escalate, or sequence therapy on the basis of these markers.

Second, multimodal diagnostic platforms that combine imaging, cell-free DNA (cfDNA)/ctDNA, and AI require prospective multicenter validation with attention to cost, interoperability, and deployment beyond tertiary centers. The key challenge is no longer technical feasibility alone, but demonstration of real-world clinical utility and accessibility.

Third, rare thoracic malignancies will likely benefit from innovative trial designs, including platform studies, basket approaches, and registry-based collaborations. In parallel, implementation science and pragmatic studies of prehabilitation, ERAS, and digital symptom monitoring are needed to translate evidence from expert centers into broader routine care.

In summary, the most important advances reported in 2025 were not limited to any single drug or platform. Rather, they collectively strengthened a model of care in which screening eligibility, perioperative therapy, surgical extent, and recovery pathways are increasingly individualized. Lung cancer remained the dominant driver of new evidence, but mesothelioma, esophageal cancer, and thymic tumors also showed meaningful progress toward more tailored management.

At the same time, several promising strategies—particularly ctDNA-guided decision-making, AI-assisted risk models, and response-adapted treatment approaches—remain in transition from research tools to routine practice. Future progress will depend on prospective validation, assay harmonization, equitable access, and implementation across diverse care settings. Overall, 2025 can be viewed as a year in which precision oncology in thoracic disease became more clinically actionable, but also more dependent on careful multidisciplinary integration.

Supplementary

The article’s supplementary files as

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.