A Novel Approach for Predicting the Survival of Colorectal Cancer Patients Using Machine Learning Techniques and Advanced Parameter Optimization Methods

Cancers, Journal Year: 2024, Volume and Issue: 16(18), P. 3205 - 3205

Published: Sept. 20, 2024

Background: Colorectal cancer is one of the most prevalent forms and associated with a high mortality rate. Additionally, an increasing number adults under 50 are being diagnosed disease. This underscores importance leveraging modern technologies, such as artificial intelligence, for early diagnosis treatment support. Methods: Eight classifiers were utilized in this research: Random Forest, XGBoost, CatBoost, LightGBM, Gradient Boosting, Extra Trees, k-nearest neighbor algorithm (KNN), decision trees. These algorithms optimized using frameworks Optuna, RayTune, HyperOpt. study was conducted on public dataset from Brazil, containing information tens thousands patients. Results: The models developed demonstrated classification accuracy predicting one-, three-, five-year survival, well overall cancer-specific mortality. Forest delivered best performance, achieving approximately 80% across all evaluated tasks. Conclusions: research enabled development effective that can be applied clinical practice.

Language: Английский

---

Lung Cancer Detection and Severity Analysis with a 3D Deep Learning CNN Model Using CT-DICOM Clinical Dataset

Indian Journal of Science and Technology, Journal Year: 2024, Volume and Issue: 17(10), P. 899 - 910

Published: March 1, 2024

Objectives: To propose a new AI based CAD model for early detection and severity analysis of pulmonary (lung) cancer disease. A deep learning artificial intelligence-based approach is employed to maximize the discrimination power in CT images minimize dimensionality order boost accuracy. Methods: The AI-based 3D Convolutional Neural Network (3D-DLCNN) method learn complex patterns features robust way efficient classification. nodules are identified by Mask-R-CNN at initial level, classification done 3D-DLCNN. Kernel Density Estimation (KDE) used discover error data points extracted removal before candidate screening. study uses CT-DICOM dataset, which includes 355 instances 251135 with target attributes cancer, healthy, condition (if positive). Statistical outlier utilized measure z-score each feature reduce point deviation. intensity pixel masking CT-DOCIM measured using ER-NCN identify performance 3D-DLCNN MATLAB R2020a tool comparative prevailing approaches such as GA-PSO, SVM, KNN, BPNN. Findings: suggested outperforms models promising results 93% accuracy rate, 92.7% sensitivity, 93.4% specificity, 0.8 AUC-ROC, 6.6% FPR, 0.87 C-Index, helps pulmonologists detect PC diagnosis. Novelty: novel hybrid has ability disease analyze score patient an stage during screening process candidates. It overcomes limitations machine models, Keywords: Artificial Intelligence, Disease Prediction, Lung Cancer, Deep Learning, Cancer Detection, Computational Model,

Language: Английский

---

Biomarker-driven drug repurposing for NAFLD-associated hepatocellular carcinoma using machine learning integrated ensemble feature selection

Frontiers in Bioinformatics, Journal Year: 2025, Volume and Issue: 5

Published: April 17, 2025

The incidence of non-alcoholic fatty liver disease (NAFLD), encompassing the more severe steatohepatitis (NASH), is rising alongside surges in diabetes and obesity. Increasing evidence indicates that NASH responsible for a significant share idiopathic hepatocellular carcinoma (HCC) cases, fatal cancer with 5-year survival rate below 22%. Biomarkers can facilitate early screening monitoring at-risk NAFLD/NASH patients assist identifying potential drug candidates treatment. This study utilized an ensemble feature selection framework to analyze transcriptomic data, biomarker genes associated stage-wise progression NAFLD-related HCC. Seven machine learning algorithms were assessed stage classification. Twelve methods including correlation-based techniques, mutual information-based methods, embedded techniques rank top as features, through this approach, multiple combined yield robust features important progression. Cox regression-based analysis was carried out evaluate potentiality these genes. Furthermore, multiphase repurposing strategy molecular docking employed identify against biomarkers. Among seven models initially evaluated, DISCR resulted most accurate classifier. Ensemble identified ten genes, among which eight recognized biomarkers based on analysis. These include ABAT, ABCB11, MBTPS1, ZFP1 mostly involved alanine glutamate metabolism, butanoate ER protein processing. Through repurposing, 81 candidate drugs found be effective markers Diosmin, Esculin, Lapatinib, Phenelzine best screened MMGBSA. consensus derived from enhances accuracy relevant NAFLD-associated use highlights therapeutic options intervention, essential stop improve outcomes.

Language: Английский

---

An integration of meta-heuristic approach utilizing kernel principal component analysis for multimodal medical image registration

Cluster Computing, Journal Year: 2024, Volume and Issue: 27(5), P. 6223 - 6246

Published: Feb. 26, 2024

Language: Английский

---

Lung Nodule Segmentation Using Machine Learning and Deep Learning Techniques

Studies in big data, Journal Year: 2024, Volume and Issue: unknown, P. 289 - 316

Published: Jan. 1, 2024

Language: Английский

---

Enhancing Lung Cancer Detection and Localization with a Hybrid VGG-19 and Adaptive Neuro-Fuzzy Inference System (ANFIS) Approach on Imaging Data

Published: May 3, 2024

Timely and precise identification categorization of lung cancer are crucial for improving patient survival rates. Although diagnostic technologies have made progress, the complex characteristics malignant tumors provide considerable difficulties in analyzing images. In this study, a groundbreaking approach is introduced, merging Adaptive Neuro-Fuzzy Inference System (ANFIS) with VGG-19 deep learning framework to effectively address these complexities. The model, trained on varied imaging data from IQ-OTH/NCCD dataset, has exceptional performance precisely recognizing, predicting, pinpointing symptoms. model utilized because its extensive capacity extract intricate information pictures, accurately identifying possible cancers. After extracting data, ANFIS utilizes fuzzy logic analyze features, enabling detailed patterns that enhances both accuracy interpretability. This technique combines advanced capabilities neural networks valuable insights logic, establishing new standard medical diagnoses. results our research show improvements important measures, such as classifying, reliability precision locating. These advancements represent major progress compared current methods, underscore transformative potential integrating AI traditional techniques.

Language: Английский

---

AI-Powered Lung Cancer Detection From CT Imaging

VFAST Transactions on Software Engineering, Journal Year: 2024, Volume and Issue: 12(2), P. 241 - 249

Published: June 30, 2024

Lung cancer is one of the deadliest forms cancer, witnessing thousands new diagnoses annually. Early detection remains paramount; without it, survival rates plummet drastically. This underscores critical role employing artificial intelligence (AI) for early diagnosis, a pivotal step in combating this devastating illness. study introduces sophisticated computer-aided system, aiming to revolutionize lung through state-of-the-art convolutional neural network (CNN) technology. By harnessing capabilities AI and CNN's, enabling precise categorization patients into those exhibiting normal tissue, benign nodules, or malignant cancer.The primary objective streamline diagnosis efforts, thereby facilitating prompt intervention treatment initiation enhance patient outcomes bolster rates. Leveraging cutting-edge technology, innovative approach aims transform landscape offering hope more effective strategies deadly disease. Furthermore, by CNN bridge existing gaps insights opportunities advancements medical research clinical practice. Ultimately, successful implementation has potential significantly impact field treatment, improved increased Through continued development, further AI-based diagnostic tools can be achieved, paving way brighter future fight against cancer.

Language: Английский

---

A Hybrid RNN-based Deep Learning Model for Lung Cancer and COPD Detection

Engineering Technology & Applied Science Research, Journal Year: 2024, Volume and Issue: 14(5), P. 16847 - 16853

Published: Oct. 9, 2024

In the last ten years, lung cancer and chronic pulmonary diseases have become prominent respiratory that require significant attention. This increase in prominence underscores their widespread impact on public health urgent need for better understanding, detection, management strategies. Accurate identification of Chronic Obstructive Pulmonary Disease (COPD) is crucial preserving human life. differentiation between two disorders administration necessary treatment are very important. study focuses effectively discriminating deadliest chest using X-ray images. Recurrent neural networks help to classify accurately by improving feature extraction from radiographs. The proposed algorithm performs more when analyzing image datasets showing alterations a patient's chest, including development tiny lobes or thicker capillaries system among other details, compared standard imaging.

Language: Английский

---

A Novel Hybrid Dehazing and Illumination based Approach for Preprocessing, Enhancement and Segmentation of Lung Images using Deep Learning

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 29, 2024

Abstract Medical images are affected by various complications such as noise and deficient contrast. To increase the quality of an image, it is highly important to contrast eliminate noise. In field image processing, enhancement one essential methods for recovering visual aspects image. However segmentation medical brain MRI lungs CT scans properly difficult. this article, a novel hybrid method proposed lung images. The suggested article includes two steps. 1st step, were enhanced. During enhancement, gone through many steps de-hazing, complementing, channel stretching, course illumination, fusion principal component analysis (PCA). second modified U-Net model was applied segment We evaluated entropy input output images, mean square error (MSE), peak signal-to-noise ratio (PSNR), gradient magnitude similarity deviation (GMSD), multi-scale (MCSD) after process. we used both original enhanced calculated accuracy. found that Dice-coefficient 0.9695 0.9797

Language: Английский

---

A Review on DeepLungNet: CNN-Based Lung Cancer Detection Techniques Using CT Images

Published: April 18, 2024

The disease known as lung cancer, which is common and frequently deadly, starts in the cells of lungs causes symptoms like exhaustion, chest pain, persistent coughing. Since small cell cancer (SCLC) less frequent but more dangerous than non-small (NSCLC), knowledge its causes, including smoking, pollution, genetic factors, essential. Survival rates are greatly increased by early identification about 20%. This study comprehensively analyzes deep learning machine models-based prediction methods from 2016 to 2023. review emphasizes how well these models work achieve greater accuracy. Expanding on this, a new method suggested that combines data augmentation denoising preprocessing with CNN ResNet architecture for classification. goal this hybrid model maintain improve accuracy levels benign malignant pictures. methodology offers promising promises accurate reliable utilizing advances both classification techniques.

Language: Английский

---