Handan Tanyildizi-Kokkulunk

Objectives: This study aimed to apply machine learning (ML) models to enhance lung cancer (LC) classification, distinguishing malignant from benign tumors, using data from 73 patients. Methods: The dataset included PET/CT biomarkers and demographic factors, with SMOTE applied to address class imbalance. Three models were evaluated using 10-fold cross-validation to compare Random Forest (RF), Decision Tree (DT), Extra Trees Classifier (ETC), and XGBoost algorithms based on accuracy and AUC. Results: ETC performed best in Model 1 (86 % accuracy, AUC 0.95), RF in Model 2 (94 % accuracy, AUC 0.97), and XGBoost in Model 3 (94 % accuracy, AUC 0.98). XGBoost consistently outperformed others, particularly in Model 3, which included age and smoking. Feature importance analysis highlighted SUVmax as the most predictive variable, with smoking having a moderate influence and gender being minimal. Conclusions: Integrating clinical and lifestyle data with PET/CT parameters significantly improved LC classification. XGBoost emerged as the most effective model, demonstrating that comprehensive models enhance diagnostic accuracy beyond traditional metrics.

Background/Objectives: Klebsiella pneumoniae is an opportunistic pathogen increasingly resistant to carbapenems and broad-spectrum antibiotics, complicating timely infection management. In critical cases like septic shock, where initiating effective antibiotics within 3 h improves survival, culture-based resistance testing is often too slow. This study evaluates machine learning (ML) algorithms for faster antimicrobial resistance prediction than conventional methods. Methods: In this retrospective study, antibiogram results of 607 Klebsiella pneumoniae isolates collected between 2017 and 2024 were combined with demographic and clinical information of the patients from whom the isolates were obtained. Four different ML algorithms, namely Decision Tree (DT), Support Vector Classifier (SVC), K-Nearest Neighbors (KNN) and Random Forest (RF), were applied to classify the resistance status for 22 antibiotics. Model performances were evaluated using accuracy, precision, recall, F-score, AUC and feature importance metrics. Results: The RF model showed the highest overall performance in accurately predicting resistance to 22 antibiotics, achieving an average AUC value of 0.96. In particular, it predicted resistance to treatment-critical antibiotics such as Ertapenem (100%), Imipenem (93%) and Meropenem (95%) with high accuracy. Conclusions: ML models, especially RF, offer a powerful tool for rapid antibiotic resistance prediction, supporting accurate empirical treatment decisions and antimicrobial stewardship.

Accurate differentiation between non-cancerous, benign, and malignant lung cancer remains a diagnostic challenge due to overlapping clinical and imaging characteristics. This study proposes a multimodal machine learning (ML) framework integrating positron emission tomography/computed tomography (PET/CT) anatomic-metabolic parameters, sarcopenia markers, and inflammatory biomarkers to enhance classification performance in lung cancer. A retrospective dataset of 222 patients was analyzed, including demographic variables, functional and morphometric sarcopenia indices, hematological inflammation markers, and PET/CT derived parameters such as maximum and mean standardized uptake value (SUVmax, SUVmean), metabolic tumor volume (MTV), total lesion glycolysis (TLG). Five ML algorithms—Logistic Regression, Multi-Layer Perceptron, Support Vector Machine, Extreme Gradient Boosting, and Random Forest—were evaluated using standardized performance metrics. Synthetic Minority Oversampling Technique was applied to balance class distributions. Feature importance analysis was conducted using the optimal model, and classification was repeated using the top 15 features. Among the models, Random Forest demonstrated superior predictive performance with a test accuracy of 96%, precision, recall, and F1-score of 0.96, and an average AUC of 0.99. Feature importance analysis revealed SUVmax, SUVmean, total lesion glycolysis, and skeletal muscle index as leading predictors. A secondary classification using only the top 15 features yielded even higher test accuracy (97%). These findings underscore the potential of integrating metabolic imaging, physical function, and biochemical inflammation markers in a non-invasive ML-based diagnostic pipeline. The proposed framework demonstrates high accuracy and generalizability and may serve as an effective clinical decision support tool in early lung cancer diagnosis and risk stratification.

Background/Objectives: Escherichia coli is one of the leading causes of surgical site infections (SSIs) and poses a growing public health concern due to its increasing antimicrobial resistance. High rates of extended-spectrum beta-lactamase (ESBL) production among E. coli strains complicate treatment outcomes and emphasize the need for effective surveillance and control strategies. Methods: A total of 691 E. coli isolates from general surgery clinics (2020–2025) were identified using MALDI-TOF MS. Antibiotic susceptibility data and patient variables were cleaned, encoded, and used to predict resistance using the Random Forest, CatBoost, and Naive Bayes algorithms. SMOTE addressed class imbalance, and model performance was assessed through various validation methods. Results: Among the three machine learning models tested, Random Forest (RF) showed the best performance in predicting antibiotic resistance of E. coli, achieving median accuracy, precision, recall, and F1-scores of 0.90 and AUC values up to 0.99 for key antibiotics. CatBoost performed similarly but was less stable with imbalanced data, while Naive Bayes showed lower accuracy. Feature importance analysis highlighted strong inter-antibiotic resistance links, especially among β-lactams, and some influence of demographic factors. Conclusions: This study highlights the potential of simple, high-performing models using structured clinical data to predict antimicrobial resistance, especially in resource-limited clinical settings. By incorporating machine learning into antimicrobial resistance (AMR) surveillance systems, our goal is to support the advancement of rapid diagnostics and targeted antimicrobial stewardship approaches, which are essential in addressing the growing challenge of multidrug resistance.

Objectives: In this work, it was aimed to employ machine learning (ML) algorithms to accurately forecast the radiation doses for phantoms while accounting for the most popular CT protocols. Methods: A cloud-based software was utilized to calculate the effective doses from different CT protocols. To simulate a range of adult patients with different weights, eight entire body mesh-based computational phantom sets were used. The head, neck, and chest-abdomen-pelvis CT scan characteristics were combined to create a dataset with 33 rows for each phantom and 792 rows total. At the ML stage, linear (LR), random forest (RF) and support vector regression (SVR) were used. Mean absolute error, mean squared error and accuracy were used to evaluate the performances. Results: The female phantoms received higher doses (7.8 %) than males. Furthermore, an average of 11 % more dose was taken to the normal weight phantom than to the overweight, the overweight in comparison to the obese I, and the obese I in comparison to the obese II. Among the ML algorithms, the LR showed 0 error rate and 100 % accuracy in predicting CT doses. Conclusions: The LR was shown to be the best approach out of those used in the ML estimation of CT-induced doses.

This study evaluates the dosimetric accuracy of PLA and ABS 3D-printed phantoms compared to real tissues using Monte Carlo simulations in radionuclide therapy. Materials and methods: A phantom representing average liver and lung volumes, with a 10 mm tumor mimic in the liver, was simulated for radioembolization using 1 mCi Tc-99 m and 1 mCi Y-90. The dose distribution (DD) was compared across PLA, ABS, and real organ densities. Results: For Tc-99 m, PLA showed a + 5.6% DD difference in the liver, and ABS showed − 35.3% and − 40.9% differences in the lungs. For Y-90, PLA had a + 1.7% DD difference in the liver, while ABS showed − 34.2% and − 34.9% differences in the lungs. Conclusion: In MC simulation, PLA is suitable for representing high-density tissues, while ABS is appropriate for simulating moderately low-density tissues.

Maritime activities pose significant safety risks, particularly with the growing presence of nuclear-powered vessels (NPVs) alongside traditional fossil-powered vessels (FPVs). This study employs a probabilistic risk assessment (PRA) approach to evaluate and compare accident hazards involving NPVs and FPVs. By analyzing historical data from 1960 to 2024, this study identifies risk patterns, accident frequency (probability), and severity levels. The methodology focuses on incidents such as marine incidents, marine casualties, and very serious cases with sub-causes. Key findings reveal that Russia exhibits the highest risk for very serious incidents involving both NPVs and FPVs, with a significant 100% risk for NPVs. China has the highest FPV risk, while France and the USA show above-average risks, particularly for marine casualties and very serious incidents. Moreover, collision is the most significant global risk, with a 26% risk for NPVs and 34% for FPVs, followed by fire hazards, which also pose a major concern, with a 17% risk for NPVs and 16% for FPVs, highlighting the need for enhanced safety and fire-prevention measures. In conclusion, comparative analysis highlights the need for enhanced stability improvements, fire prevention, and maintenance practices, particularly in the UK, France, Russia, and China. This study underscores the importance of targeted safety measures to mitigate risks, improve ship design, and promote safer maritime operations for both nuclear- and fossil-fueled vessels.

There are various treatment modalities for prostate cancer, which has a high incidence. In this study, it is aimed to make predictions with machine learning in order to determine the optimal treatment option for prostate cancer patients. The study included 88 male patients diagnosed with prostate cancer. Independent variables were determined as Gleason scores, biopsy, PSA, SUV_max, and age. Prostate cancer treatments, which are dependent variables, were determined as hormone therapy(n = 30), radiotherapy(n = 28) and radiotherapy + hormone therapy(n = 30). Machine learning was carried out in the Python with SVM, RF, DT, ETC and XGBoost. Metrics such as accuracy, ROC curve, and AUC were used to evaluate the performance of multi-class predictions. The model with the highest number of successful predictions was the XGBoost. False negative rates for hormone therapy, radiotherapy, and radiotherapy + hormone therapy treatments were, respectively, 12.5, 33.3, and 0%. The accuracy values were computed as 0.61, 0.83, 0.83, 0.72 and 0.89 for SVM, RF, DT, ETC and XGBoost, respectively. The three features that had the greatest influence on the treatment model prediction for prostate cancer with XGBoost were biopsy, Gleason score (3 + 3), and PSA level, respectively. According to the AUC, ROC and accuracy, it was determined that the XGBoost was the model that made the best estimation of prostate cancer treatment. Among the variables biopsy, Gleason score, and PSA level are identified as key variables in prediction of treatment.