Andrew D. Marten

Background: The lung fosters a diverse microbiome that may have a complex interplay with underlying diseases, such as lung cancer. However, existing literature has revealed diverse results, likely owing to small sample sizes and diverse sampling methods. In this context, we aim to characterize the lung microbiome and its association with malignancy via multiple sampling methods during lung resection. Methods: Patients undergoing lung resection for confirmed malignancy were prospectively included in the study. During the operation, a bronchial swab and a parenchymal biopsy of lung tissue were collected from the resected specimen and sent for analysis via 16S rRNA gene amplicon (V4) sequencing. An established bioinformatics pipeline was used to determine taxonomic identification and compare these results to clinical and tumor data. Results: There were 83 participants included, with 75 (90.4%) patients having primary lung tumors and 8 (9.6%) having metastatic lesions. Resections included lobectomy (69.9%), wedge (25.3%), and segmentectomy (4.8%). Lung microbiome composition differed by primary malignancy site (lung vs metastatic; p=0.014) and tumor mutations (TP53, p=0.040; KRAS, p=0.054). For primary lung cancer, there was no difference in microbiome composition by tumor histology. A near-significant association was noted for patients with prolonged air leaks and alpha diversity of the lung microbiome (p=0.070). Conclusions: In this study, we demonstrate associations between the lung microbiome and underlying malignancy and tumor characteristics. Validation of these results may assist with diagnostic, prognostic, and therapeutic interventions in the operative and peri-operative management of lung cancer.

BACKGROUND Machine learning (ML)-based medical image analysis is increasingly prominent in surgical research. However, implementation often requires specialized expertise in artificial intelligence and software development, a substantial barrier for surgeon-scientists wishing to implement them for clinical use. Simple solutions that allow clinicians to perform rapid image analysis within a unified framework are lacking. OBJECTIVE In this context, we developed the RAPID software platform to facilitate image processing analysis with reduced technical burden while maintaining methodological rigor. METHODS RAPID is divided into two components. First is RAPID Core which contains routines to convert DICOM images into useable NIFTI format, automate segmentation of target anatomy using the open-source MONAI library, preprocess images with region of interest cropping and pixel preservation from segmentations, extract radiomics from preprocessed images using PyRadiomics, and train and test machine learning models. Second is RAPID GUI, a graphical user interface that organizes and parameterizes RAPID Core components to form a pipeline of tasks. RAPID GUI generates ready-to-use Python scripts for the RAPID pipeline that can be run on any computer containing RAPID Core. RESULTS We tested the RAPID software platform through two tasks. The first identified radiomic features(n=112) predictive of aortic aneurysm in a balanced cohort(n=1,114). The second identified features(n=225) of psoas muscles that correlate with sarcopenia as a surrogate for frailty and predictor of mortality for patients(n=116) undergoing thoracoabdominal aortic aneurysm repair. Pipelines of RAPID Core processes were designed for each of these tasks with RAPID GUI. Radiomics-based Extreme Gradient Boosting (XGBoost) models produced AUROC=0.956 and 0.929 for these tasks, respectively. CONCLUSIONS RAPID is an accessible platform that shortens time from design to implementation. It enables surgeon-scientists with minimal computer programming experience to perform advanced radiomics and ML analysis, facilitating broader adoption of medical image analytics that enhances surgical research. CLINICALTRIAL n/a

Background: Although the lung is of low biomass, it harbors a diverse and dynamic microbiome that may influence disease and healing. Existing studies have used diverse sampling methods with high propensities for contamination and sampling error, leading to diverse and unclear results. Here, we characterized the lung microbiome via airway and parenchymal samples to determine variation across patients and sampling methods. Methods: We recruited adult patients undergoing lung resection for suspected or confirmed malignancy. After resection and under sterile conditions, a 1 cm cubic piece of non-cancerous lung parenchyma and a swab from the specimen’s bronchus were collected and sent for microbiome analysis via 16S rRNA gene amplicon (V4) sequencing on an Illumina platform. An established bioinformatics pipeline was used to determine taxonomic identification. Baseline clinical and demographic data were compared to microbiome composition. Results: A total of 86 patients were included in the study. Beta diversity (microbial composition) varied significantly by sampling method (biopsy of lung parenchyma versus airway swabs), so all further results were analyzed within sample types. Further analyses revealed significant differences in beta diversity by lobe of the lung, indicating a different microbial composition by anatomic location. Analyses of patient demographics revealed significant differences by age and comorbidities, including chronic obstructive pulmonary disease and atrial fibrillation. Conclusions: The lung harbors a diverse microbiome that differs by anatomic location and patient characteristics. This study provides a framework for more accurate future lung microbiome sampling and characterization.

Prophylactic antibiotics are an effective intervention in reducing fracture related infection (FRI) in open fractures after orthopaedic trauma. Despite routine antibiotic use, the rate of FRI in Gustilo Anderson (GA) type III open fractures remains high; henceforth, we investigated the rate of open tibia fractures and trends in prophylactic antibiotic usage at our institution. This was a retrospective study of 133 type III open tibia fractures treated at an academic tertiary level I trauma center over a 15-year period (2007-2021). We collected information on administered prophylactic antibiotics, local antibiotics, soft tissue coverage, and FRI. Chi-squared tests analyzed trends in antibiotic utilization and FRI. Multivariable logistic regression was performed to evaluate association between GA type, prophylactic antibiotics, local antibiotics, and flap coverage with FRI outcomes. Analysis revealed a higher percentage of 1st/2nd generation cephalosporin use among type IIIA compared to IIIB/IIIC fractures (p < 0.001), an increase in utilization of 3rd generation cephalosporins over the 15-year period (p = 0.018), and decline in 1st/2nd generation cephalosporins. FRIs occurred in 24.8% (n = 26) of type III open tibia fractures. On multivariable logistic regression analysis, when controlling for GA type, prophylactic antibiotic regimen, and flap coverage: local antibiotic use was the only independent predictor of FRI (OR 9.3, 95% CI = 3.0-28.7, p < 0.001). When evaluating patients with adequate follow-up, local antibiotics remained the only independent predictor of FRI (OR 9.4, 95% CI 2.9-30.1, p < 0.001). Institutional protocols for orthopaedic care and hospital-wide education may be beneficial to improve recognition and appropriate prophylactic treatment of high-risk open fractures.