Radiomics and Its Feature Selection: A Review

Symmetry, Journal Year: 2023, Volume and Issue: 15(10), P. 1834 - 1834

Published: Sept. 27, 2023

Medical imaging plays an indispensable role in evaluating, predicting, and monitoring a range of medical conditions. Radiomics, specialized branch imaging, utilizes quantitative features extracted from images to describe underlying pathologies, genetic information, prognostic indicators. The integration radiomics with artificial intelligence presents innovative avenues for cancer diagnosis, prognosis evaluation, therapeutic choices. In the context oncology, offers significant potential. Feature selection emerges as pivotal step, enhancing clinical utility precision radiomics. It achieves this by purging superfluous unrelated features, thereby augmenting model performance generalizability. goal review is assess fundamental process progress feature methods, explore their applications challenges research, provide theoretical methodological support future investigations. Through extensive literature survey, articles pertinent were garnered, synthesized, appraised. paper provides detailed descriptions how applied challenged different types various stages. also comparative insights into strategies, including filtering, packing, embedding methodologies. Conclusively, broaches limitations prospective trajectories

Language: Английский

---

Radiomics and Deep Learning in Nasopharyngeal Carcinoma: A Review

IEEE Reviews in Biomedical Engineering, Journal Year: 2023, Volume and Issue: 17, P. 118 - 135

Published: April 25, 2023

Nasopharyngeal carcinoma is a common head and neck malignancy with distinct clinical management compared to other types of cancer. Precision risk stratification tailored therapeutic interventions are crucial improving the survival outcomes. Artificial intelligence, including radiomics deep learning, has exhibited considerable efficacy in various tasks for nasopharyngeal carcinoma. These techniques leverage medical images data optimize workflow ultimately benefit patients. In this review, we provide an overview technical aspects basic learning image analysis. We then conduct detailed review their applications seven typical diagnosis treatment carcinoma, covering synthesis, lesion segmentation, diagnosis, prognosis. The innovation application effects cutting-edge research summarized. Recognizing heterogeneity field existing gap between translation, potential avenues improvement discussed. propose that these issues can be gradually addressed by establishing standardized large datasets, exploring biological characteristics features, technological upgrades.

Language: Английский

---

Exploring non-invasive precision treatment in non-small cell lung cancer patients through deep learning radiomics across imaging features and molecular phenotypes

Biomarker Research, Journal Year: 2024, Volume and Issue: 12(1)

Published: Jan. 25, 2024

Abstract Background Accurate prediction of tumor molecular alterations is vital for optimizing cancer treatment. Traditional tissue-based approaches encounter limitations due to invasiveness, heterogeneity, and dynamic changes. We aim develop validate a deep learning radiomics framework obtain imaging features that reflect various changes, aiding first-line treatment decisions patients. Methods conducted retrospective study involving 508 NSCLC patients from three institutions, incorporating CT images clinicopathologic data. Two radiomic scores network feature were constructed on data sources in the 3D region. Using these features, we developed validated ‘Deep-RadScore,’ model predict prognostic factors, gene mutations, immune molecule expression levels. Findings The Deep-RadScore exhibits strong discrimination features. In independent test cohort, it achieved impressive AUCs: 0.889 lymphovascular invasion, 0.903 pleural 0.894 T staging; 0.884 EGFR ALK, 0.896 KRAS PIK3CA, TP53, 0.895 ROS1; 0.893 PD-1/PD-L1. Fusing yielded optimal predictive power, surpassing any single feature. Correlation interpretability analyses confirmed effectiveness customized capturing additional phenotypes beyond known Interpretation This proof-of-concept demonstrates new biomarkers across can be provided by fusing multiple sources. holds potential offer valuable insights radiological phenotyping characterizing diverse alterations, thereby advancing pursuit non-invasive personalized

Language: Английский

---

Radiomic signatures associated with tumor immune heterogeneity predict survival in locally recurrent nasopharyngeal carcinoma

JNCI Journal of the National Cancer Institute, Journal Year: 2024, Volume and Issue: 116(8), P. 1294 - 1302

Published: April 19, 2024

The prognostic value of traditional clinical indicators for locally recurrent nasopharyngeal carcinoma is limited because their inability to reflect intratumor heterogeneity. We aimed develop a radiomic signature reveal tumor immune heterogeneity and predict survival in carcinoma.

Language: Английский

---

Self-supervised tumor segmentation and prognosis prediction in osteosarcoma using multiparametric MRI and clinical characteristics

Computer Methods and Programs in Biomedicine, Journal Year: 2023, Volume and Issue: 244, P. 107974 - 107974

Published: Dec. 11, 2023

Language: Английский

---

Robust deep learning-based PET prognostic imaging biomarker for DLBCL patients: a multicenter study

European Journal of Nuclear Medicine and Molecular Imaging, Journal Year: 2023, Volume and Issue: 50(13), P. 3949 - 3960

Published: Aug. 22, 2023

Language: Английский

---

Application of Artificial Intelligence to the Diagnosis and Therapy of Nasopharyngeal Carcinoma

Journal of Clinical Medicine, Journal Year: 2023, Volume and Issue: 12(9), P. 3077 - 3077

Published: April 24, 2023

Artificial intelligence (AI) is an interdisciplinary field that encompasses a wide range of computer science disciplines, including image recognition, machine learning, human-computer interaction, robotics and so on. Recently, AI, especially deep learning algorithms, has shown excellent performance in the being able to automatically perform quantitative evaluation complex medical features improve diagnostic accuracy efficiency. AI wider deeper application diagnosis, treatment prognosis. Nasopharyngeal carcinoma (NPC) occurs frequently southern China Southeast Asian countries most common head neck cancer region. Detecting treating NPC early crucial for good This paper describes basic concepts traditional their clinical applications detecting assessing lesions, facilitating predicting The main limitations current technologies are briefly described, interpretability issues, privacy security need large amounts annotated data. Finally, we discuss remaining challenges promising future using diagnose treat NPC.

Language: Английский

---

Deep learning model based on primary tumor to predict lymph node status in clinical stage IA lung adenocarcinoma: a multicenter study

Journal of the National Cancer Center, Journal Year: 2024, Volume and Issue: 4(3), P. 233 - 240

Published: Feb. 2, 2024

To develop a deep learning model to predict lymph node (LN) status in clinical stage IA lung adenocarcinoma patients. This diagnostic study included 1,009 patients with pathologically confirmed T1N0M0 from two independent datasets (699 Cancer Hospital of Chinese Academy Medical Sciences and 310 PLA General Hospital) between January 2005 December 2019. The dataset was randomly split into training cohort (559 patients) validation (140 train tune based on residual network (ResNet). used as testing evaluate the generalization ability model. Thoracic radiologists manually segmented tumors interpreted high-resolution computed tomography (HRCT) features for predictive performance assessed by area under curves (AUCs), accuracy, precision, recall, F1 score. Subgroup analysis performed potential bias population. A total were this study; 409 (40.5%) male 600 (59.5%) female. median age 57.0 years (inter-quartile range, IQR: 50.0–64.0). achieved AUCs 0.906 (95% CI: 0.873–0.938) 0.893 0.857–0.930) predicting pN0 disease non-pure ground glass nodule (non-pGGN) cohort, respectively. No significant difference detected non-pGGN (P = 0.622). precisions 0.979 0.963–0.995) 0.983 0.967–0.998) 0.848 0.798–0.898) 0.831 0.776–0.887) pN2 0.657). recalls 0.903 0.870–0.936) 0.931 0.901–0.961) superior will help target extension dissection reduce ineffective early-stage

Language: Английский

---

Habitat radiomics analysis for progression free survival and immune-related adverse reaction prediction in non-small cell lung cancer treated by immunotherapy

Journal of Translational Medicine, Journal Year: 2025, Volume and Issue: 23(1)

Published: April 3, 2025

Non-small cell lung cancer (NSCLC) is highly heterogeneous, leading to varied treatment responses and immune-related adverse reactions (irAEs) among patients. Habitat radiomics allows non-invasive quantitative assessment of intratumor heterogeneity (ITH). Therefore, our objective employ habitat techniques develop a robust approach for predicting the efficacy Immune checkpoint inhibitors (ICIs) likelihood irAEs in advanced NSCLC In this retrospective two center study, independent cohorts patients with were used (n = 248) validate signatures 95). After applying four kinds machine learning algorithms select key preoperative CT radiomic features, we clinical, features clinical signature, signature ICIs prognostics prediction. By combining corresponding clinicopathologic information, nomogram was constructed training cohort. Next, internal validation cohort 75) patients, external 20) treated included evaluate predictive value signatures, their performance assessed by area under operating characteristic curve (AUC). Our study introduces model that integrates identify who may benefit from or experience irAEs. The Radiomics Nomogram exhibited superior training, validation, sets, AUCs 0.923, 0.817, 0.899, respectively. This outperformed both Whole-tumor Signature (AUCs 0.870, 0.736, 0.626) 0.900, 0.804, 0.808). focusing on tumor sub-regional showed better than derived entire tumor. Decision Curve Analysis (DCA) calibration curves confirmed nomogram's effectiveness. leveraging predict outcomes ICIs, can move closer achieving tailored cancer. advancement will assist physicians selecting managing subsequent strategies, thereby facilitating decision-making.

Language: Английский

---

Neural Network-based Automated Classification of 18F-FDG PET/CT Lesions and Prognosis Prediction in Nasopharyngeal Carcinoma Without Distant Metastasis

Clinical Nuclear Medicine, Journal Year: 2025, Volume and Issue: unknown

Published: May 9, 2025

Purpose: To evaluate the diagnostic performance of PET Assisted Reporting System (PARS) in nasopharyngeal carcinoma (NPC) patients without distant metastasis, and to investigate prognostic significance metabolic parameters. Patients Methods: Eighty-three NPC who underwent pretreatment 18 F-FDG PET/CT were retrospectively collected. First, sensitivity, specificity, accuracy PARS for diagnosing malignant lesions calculated, using histopathology as gold standard. Next, parameters primary tumor derived both manual segmentation. The differences consistency between 2 methods analyzed. Finally, value was evaluated. Prognostic analysis progression-free survival (PFS) overall (OS) conducted. Results: demonstrated high patient-based (97.2%), sensitivity (88.9%), specificity (97.4%), 96.7%, 84.0%, 96.9% based on lesions. Manual segmentation yielded higher volume (MTV) total lesion glycolysis (TLG) than PARS. Metabolic from highly correlated consistent. ROC showed exhibited prediction, but generally performed well predicting 3-year PFS OS overall. MTV age independent factors; Cox proportional-hazards models incorporating them significant predictive improvements when combined. Kaplan-Meier confirmed better prognosis low-risk group combined indicators (χ² = 42.25, P < 0.001; χ² 20.44, 0.001). Conclusions: Preliminary validation metastasis shows identification classification, correlate with manual. reflects prognosis, its combination enhances prediction risk stratification.

Language: Английский

---