Artificial intelligence (AI) is increasingly becoming integral to medical practice, potentially enhancing outcomes in thoracic surgery. AI-driven models have shown significant accuracy in diagnosing non-small-cell lung cancer (NSCLC), predicting lymph node metastasis, and aiding in the efficient extraction of electronic medical record (EMR) data. Moreover, AI applications in robotic-assisted thoracic surgery (RATS) and perioperative management reveal the potential to improve surgical precision, patient safety, and overall care efficiency. Despite these advancements, challenges such as data privacy, biases, and ethical concerns remain. This manuscript explores AI applications, particularly machine learning (ML) and natural language processing (NLP), in thoracic surgery, emphasizing their role in diagnosis and perioperative management. It also provides a comprehensive overview of the current state, benefits, and limitations of AI in thoracic surgery, highlighting future directions in the field.

Background: Nuclear protein of the testis (NUT) carcinoma (NC) is a rare and exceedingly aggressive neoplasm, typically presenting as midline tumors such as mediastinal, thoracic, or sinonasal tumors, but can also be seen in other locations like the kidneys, liver, and pancreas. NC is frequently underdiagnosed or misdiagnosed due to its rarity, nonspecific clinical and histopathological presentation, lack of clinically trialed standard-of-care therapy, and, ultimately, a lack of awareness about the condition. Methods: We retrospectively reviewed 45 cases of NC from the Surveillance, Epidemiology, and End Results (SEER) database. Results: The patient pool was predominantly male, with the most common tumor sites being the lungs and bronchus (51.1%), followed by the nasal cavity and paranasal sinuses (28.9%).73.3% of patients received systemic therapy, 62.2% received radiotherapy, and only one-quarter of patients underwent surgery. The median survival time was 8 months (95% CI: 4.09–11.91). Systemic therapy (aHR = 0.268, P = 0.009) and radiotherapy (aHR = 0.335, P = 0.02) were the only independent prognostic factors in this cohort. Conclusion: Our study reiterates the rarity of NC in the United States. We also reveal that when faced with the rare NC, systemic therapy and radiotherapy have been found to relatively prolong NC patient survival. Future research is needed to further guide NC treatment and improve the poor survival associated with the diagnosis.

Pericardial effusion is the pathological accumulation of fluid in the pericardial cavity and can lead to life-threatening complications such as cardiac tamponade. While pericardiocentesis and surgical pericardial window creation remain the mainstay of treatment, uniportal video-assisted thoracoscopic surgery (UVATS) has emerged as a minimally invasive alternative with potential advantages over conventional approaches. This study comprehensively reviewed the existing literature to evaluate the safety, efficacy, and clinical outcomes of UVATS in managing pericardial effusion. A comprehensive search across the PubMed, Cochrane, and Google Scholar databases identified studies published from 2000 onward, focusing on the application of UVATS in thoracic surgery. The findings indicate that UVATS is associated with shorter hospital stays, reduced postoperative pain, and lower recurrence rates compared to multiportal VATS and the subxiphoid approach. Additionally, since UVATS is a single-incision technique, this method minimizes intercostal trauma, instead enhancing postoperative recovery. However, challenges remain, including a steep learning curve, technical limitations, and the requirement for specialized instrumentation. Despite these concerns, UVATS continues to evolve with advancements in instrumentation, imaging, and robotic assistance, further improving its feasibility and outcomes. As the adoption of UVATS grows, future research should focus on long-term efficacy, standardization of techniques, and expanded indications to optimize its role in thoracic surgery.

Artificial Intelligence (AI) is increasingly becoming integral to medical practice, potentially enhancing outcomes in thoracic surgery. AI-driven models have shown significant accuracy in diagnosing non-small cell lung cancer (NSCLC), predicting lymph node metastasis, and aiding in the efficient extraction of electronic medical records (EMR) data. Moreover, AI applications in robotic-assisted thoracic surgery (RATS) and perioperative management reveal potential in improving surgical precision, patient safety, and overall care efficiency. Despite these advancements, challenges such as data privacy, biases, and ethical concerns remain. This manuscript explores AI applications, particularly machine learning (ML) and natural language processing (NLP), in thoracic surgery, emphasizing their role in diagnosis and peri-operative management. It also provides a comprehensive overview of the current state, benefits, and limitations of AI in thoracic surgery, highlighting future directions in the field.

Surgical research is beset by a distinct set of challenges that differ meaningfully from those encountered in other medical disciplines. Despite surgeons' historic contributions to medical advancement, the proportion of surgeon-authored publications has declined relative to total medical literature output over the latter half of the twentieth century. This narrative review aims to comprehensively identify, categorise, and discuss the specific barriers surgeons face in conducting and publishing surgical research, with implications for surgical training, research ethics, and academic surgery. A narrative review was conducted through a systematic search of PubMed/MEDLINE, Cochrane Library, and Google Scholar databases. Search terms included combinations of "surgical research," "ethical considerations," "patient confidentiality," "publication barriers," "surgical methodology," and "academic surgery." No restriction was placed on year of publication. Only English-language, peer-reviewed articles were considered. Four authors independently screened retrieved records by title and abstract, and reference lists of included studies were hand-searched for additional relevant publications. Priority was given to narrative reviews, original research articles, and discussion papers addressing the identified themes. The challenges identified fell into two broad domains. In the conduct of surgical research, major barriers include extreme time constraints arising from the structural demands of surgical practice, ethical complexities inherent to the surgeon's dual role as caregiver and investigator, surgeon- and technique-specific variability that complicates standardization and reproducibility, and patient-related obstacles including reluctance to enrol in trials and difficulty maintaining long-term follow-up. In the publication of surgical research, key barriers include communicating technically complex procedures to a broad readership, a worsening peer review crisis driven by reviewer scarcity and burnout within a relatively small surgical community, ethical deficiencies in reporting, including inadequate documentation of ethical approval and informed consent, and challenges in producing and ethically managing high-quality visual and multimedia content. Surgeons face a multilayered set of structural, ethical, and methodological barriers throughout the research and publication process. Addressing these challenges will require targeted interventions at the levels of surgical training, institutional support, and journal infrastructure. Greater awareness of these barriers is an essential first step toward strengthening the evidence base that underpins surgical practice.