Evaluation of the performance of both machine learning models using PET and CT radiomics for predicting recurrence following lung stereotactic body radiation therapy: A single‐institutional study

Journal of Applied Clinical Medical Physics, Journal Year: 2024, Volume and Issue: 25(7)

Published: March 4, 2024

Predicting recurrence following stereotactic body radiotherapy (SBRT) for non-small cell lung cancer provides important information the feasibility of individualized and allows to select appropriate treatment strategy based on risk recurrence. In this study, we evaluated performance both machine learning models using positron emission tomography (PET) computed (CT) radiomic features predicting after SBRT.

Language: Английский

---

Prediction of the Ki-67 expression level in head and neck squamous cell carcinoma with machine learning-based multiparametric MRI radiomics: a multicenter study

BMC Cancer, Journal Year: 2024, Volume and Issue: 24(1)

Published: April 5, 2024

Abstract Background This study aimed to develop and validate a machine learning (ML)-based fusion model preoperatively predict Ki-67 expression levels in patients with head neck squamous cell carcinoma (HNSCC) using multiparametric magnetic resonance imaging (MRI). Methods A total of 351 pathologically proven HNSCC from two medical centers were retrospectively enrolled the divided into training ( n = 196), internal validation 84), external 71) cohorts. Radiomics features extracted T2-weighted images contrast-enhanced T1-weighted screened. Seven ML classifiers, including k-nearest neighbors (KNN), support vector (SVM), logistic regression (LR), random forest (RF), linear discriminant analysis (LDA), naive Bayes (NB), eXtreme Gradient Boosting (XGBoost) trained. The best classifier was used calculate radiomics (Rad)-scores combine clinical factors construct model. Performance evaluated based on calibration, discrimination, reclassification, utility. Results Thirteen combining MRI finally selected. SVM showed performance, highest average area under curve (AUC) 0.851 incorporating SVM-based Rad-scores T stage MR-reported lymph node status achieved encouraging predictive performance (AUC 0.916), 0.903), 0.885) Furthermore, better benefit higher classification accuracy than Conclusions ML-based exhibited promise for predicting patients, which might be helpful prognosis evaluation decision-making.

Language: Английский

---

Ct-based intratumoral and peritumoral radiomics for predicting prognosis in osteosarcoma: A multicenter study

European Journal of Radiology, Journal Year: 2024, Volume and Issue: 172, P. 111350 - 111350

Published: Feb. 2, 2024

Language: Английский

---

Development and Validation of a Radiomics Model Based on 18F-FDG PET of Primary Gastric Cancer for Predicting Peritoneal Metastasis

Frontiers in Oncology, Journal Year: 2021, Volume and Issue: 11

Published: Oct. 26, 2021

The aim of this study was to develop a preoperative positron emission tomography (PET)-based radiomics model for predicting peritoneal metastasis (PM) gastric cancer (GC).In study, total 355 patients (109PM+, 246PM-) who underwent fluorine-18-fludeoxyglucose (18F-FDG) PET images were retrospectively analyzed. According 7:3 ratio, randomly divided into training set and validation set. Radiomics features metabolic parameters data extracted from images. selected by logistic regression after using maximum relevance minimum redundancy (mRMR) the least shrinkage selection operator (LASSO) method. models based on rest these features. performance determined their discrimination, calibration, clinical usefulness in sets.After dimensionality reduction, 12 feature obtained construct signatures. results multivariate analysis, only carbohydrate antigen 125 (CA125), standardized uptake value (SUVmax), signature showed statistically significant differences between (P<0.05). A developed analyses with an AUC 0.86 cohort 0.87 cohort. prediction CA125 SUVmax 0.76 0.69 comprehensive model, which contained rad-score factor (CA125) as well parameter promising 0.90 0.88 cohort, respectively. calibration curve actual rate nomogram-predicted probability metastasis. Decision analysis (DCA) also demonstrated good utility nomogram.The other factors (SUVmax, CA125) can provide novel tool preoperatively.

Language: Английский

---

Contrast-Enhanced CT-Based Radiomics Analysis in Predicting Lymphovascular Invasion in Esophageal Squamous Cell Carcinoma

Frontiers in Oncology, Journal Year: 2021, Volume and Issue: 11

Published: May 14, 2021

To develop a radiomics model based on contrast-enhanced CT (CECT) to predict the lymphovascular invasion (LVI) in esophageal squamous cell carcinoma (ESCC) and provide decision-making support for clinicians.This retrospective study enrolled 334 patients with surgically resected pathologically confirmed ESCC, including 96 LVI 238 without LVI. All were randomly divided into training cohort testing at ratio of 7:3, containing 234 (68 166 LVI) 100 (28 72 LVI). underwent preoperative CECT scans within 2 weeks before operation. Quantitative features extracted from images, least absolute shrinkage selection operator (LASSO) method was applied select features. Logistic regression (Logistic), vector machine (SVM), decision tree (Tree) methods separately used establish models status best selected calculate Radscore, which combined two clinical predictors build model. The also developed by using logistic regression. receiver characteristic curve (ROC) (DCA) analysis evaluate performance predicting ESCC.In model, Sphericity gray-level non-uniformity (GLNU) most significant In maximum tumor thickness (cThick) significantly greater than that (P<0.001). Patients had higher N stage (cN stage) ROC showed both (AUC values 0.847 0.826 cohort, respectively) (0.876 0.867, performed better (0.775 0.798, respectively), exhibiting performance.The incorporating may potentially ESCC treatment decisions.

Language: Английский

---

The value of radiomics and deep learning based on PET/CT in predicting perineural nerve invasion in rectal cancer

Abdominal Radiology, Journal Year: 2025, Volume and Issue: unknown

Published: March 7, 2025

Language: Английский

---

Quantification of Intratumoral Heterogeneity Based on Habitat Analysis for Preoperative Assessment of Lymphovascular Invasion in Colorectal Cancer

Academic Radiology, Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Preoperative knowledge of the status lymphovascular invasion (LVI) in colorectal cancer (CRC) patients can provide valuable information for choosing appropriate treatment strategies. This study aimed to explore value heterogeneity features derived from habitat analysis 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) images predicting LVI. Pretreatment 18F-FDG PET/computed (CT) 177 diagnosed with CRC were retrospectively obtained (training cohort, n=106; validation n=71). Conventional radiomics and habitat-derived tumor extracted PET scans. The output probabilities imaging-based random forest model used generate a score (Radscore) intratumoral (ITHscore). Multivariate logistic regression was determine independent risk factors On this basis, four LVI classification models developed using (a) clinical variables (Clinical model), (b) (ITHscore (c) (Radscore (d) variables, features, (Combined model). area under curve (AUC) decision evaluate performance. Among all PET/CT-reported lymph node status, ITHscore, Radscore retained as predictors related (P<0.05). predictive effect ITHscore (AUC: 0.712) better than that 0.650) Clinical 0.652) cohort. Combined achieved effects usefulness, AUCs training cohorts 0.857 0.798, respectively. A nomogram established, calibration plot well fitted (P>0.05). In addition, results Spearman's rank correlation tests showed there no significant between (R=0.044, P=0.655 cohort; R=0.067, P=0.580 cohort). Our is novel stable quantitative indicator helpful effectively facilitating stratification after integrating features.

Language: Английский

---

A novel nomogram based on PET/CT to predict lymph nodal metastasis for lung adenocarcinoma with normal size lymph node

Journal of Cardiothoracic Surgery, Journal Year: 2025, Volume and Issue: 20(1)

Published: April 16, 2025

Language: Английский

---

The Role of 18F-FDG PET/CT in Guiding Precision Medicine for Invasive Bladder Carcinoma

Frontiers in Oncology, Journal Year: 2020, Volume and Issue: 10

Published: Oct. 6, 2020

Bladder cancer (BC) is the tenth most common worldwide. Approximately one quarter of patients with BC have muscle-invasive disease (MIBC). MIBC carries a poor prognosis and choosing optimal treatment option critical to improve patients' outcomes. Ongoing research supports role 2-deoxy-2-(18F)fluoro-D-glucose positron emission tomography (18F-FDG PET) in guiding patient-specific management decisions throughout course MIBC. As an imaging modality, 18F-FDG PET acquired simultaneously either CT or MRI offer hybrid approach combining anatomical metabolic information that complement each other. At initial staging, PET/CT enhances detection extravesical disease, particularly classified as oligometastatic by conventional imaging. has value monitoring response chemotherapy, well localizing relapse after treatment. In new era immunotherapy, may also be useful monitor efficacy detect immune related adverse events. With advent artificial intelligence techniques such radiomics deep-learning, these medical images can mined for quantitative data, providing incremental over current standard-of-care clinical biological data. This potential produce major paradigm shift towards data-driven precision medicine ultimate goal personalized medicine. this review we highlight literature reporting supporting Specific topics reviewed include prognostication, pre-operative planning, assessment, prediction recurrence, diagnosing drug toxicity.

Language: Английский

---

A clinical‐radiomics model incorporating T2‐weighted and diffusion‐weighted magnetic resonance images predicts the existence of lymphovascular invasion / perineural invasion in patients with colorectal cancer

Medical Physics, Journal Year: 2021, Volume and Issue: 48(9), P. 4872 - 4882

Published: May 27, 2021

Lymphovascular invasion (LVI) and perineural (PNI) are independent prognostic factors in patients with colorectal cancer (CRC). In this study, we aimed to develop validate a preoperative predictive model based on high-throughput radiomic features clinical for accurate prediction of LVI/PNI these patients.Two hundred sixty-three who underwent resection histologically confirmed CRC between 1 February 2011 30 June 2020 were retrospectively enrolled. Between September 2018, 213 randomly divided into training cohort (n = 149) validation 64) by ratio 7:3. We used 10000-iteration bootstrap analysis estimate the error confidence interval two cohorts. The test consisted 50 October 2018 2020. Regions interest (ROIs) manually delineated high-resolution T2-weighted diffusion-weighted images using ITK-SNAP software each tumor slice. total, 3356 extracted from ROI. Next, maximum relevance minimum redundancy least absolute shrinkage selection operator algorithms select strongest establish clinical-radiomics predicting LVI/PNI. Receiver-operating characteristic calibration curves then plotted evaluate performance training, validation, cohorts.A multiparametric combining MRI-reported extramural vascular (EMVI) status Radiomics score estimation was established. This had significant power (area under curve [AUC] 0.91; 95% [CI]: 0.85-0.97), (AUC: 0.88; CI: 0.79-89), cohorts (AUC 0.83, CI 0.72-0.95). performed well sensitivity 0.818, specificity 0.714, accuracy 0.760. Calibration decision demonstrated benefits.Multiparametric models can accurately predict CRC. Our has ability that should improve diagnostic allow more individualized treatment decisions.

Language: Английский

---