Deep learning radiomics model based on PET/CT predicts PD-L1 expression in non-small cell lung cancer

European Journal of Radiology Open, Journal Year: 2024, Volume and Issue: 12, P. 100549 - 100549

Published: Jan. 21, 2024

Abstract

Purpose

Programmed cell death protein-1 ligand (PD-L1) is an important prognostic predictor for immunotherapy of non-small lung cancer (NSCLC). This study aimed to develop a non-invasive deep learning and radiomics model based on positron emission tomography computed (PET/CT) predict PD-L1 expression in NSCLC.

Methods

A total 136 patients with NSCLC between January 2021 September 2022 were enrolled this study. The randomly divided into the training dataset validation ratio 7:3. Radiomics feature extracted from their PET/CT images. Mann-whitney U-test, Least Absolute Shrinkage Selection Operator algorithm Spearman correlation analysis used select top significant features. Then we developed model, fusion selected performance three models compared by area under curve (AUC), sensitivity, specificity, accuracy, positive predictive value, negative value.

Results

Of patients, 42 94 positive. 2446 features 4096 per patient. In dataset, achieved highest AUC (0.954, 95% confident internal [CI]: 0.890–0.986) (0.829, 95%CI: 0.738–0.898) (0.935, 0.865–0.975). (0.910, CI: 0.779–0.977) was also higher than that (0.785, 0.628–0.897) (0.867, 0.724–0.952).

Conclusion

PET/CT-based can accurately provides tool clinicians patients.

Language: Английский

---

Nanoparticle-mediated blockade of CXCL12/CXCR4 signaling enhances glioblastoma immunotherapy: Monitoring early responses with MRI radiomics

Acta Biomaterialia, Journal Year: 2024, Volume and Issue: 177, P. 414 - 430

Published: Feb. 14, 2024

Language: Английский

---

Explainable PET-Based Habitat and Peritumoral Machine Learning Model for Predicting Progression-free Survival in Clinical Stage IA Pure-Solid Non-small Cell Lung Cancer: A Two-center Study

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

Published: Jan. 1, 2025

Language: Английский

---

Deep learning-based CT radiomics predicts prognosis of unresectable hepatocellular carcinoma treated with TACE-HAIC combined with PD-1 inhibitors and tyrosine kinase inhibitors

BMC Gastroenterology, Journal Year: 2025, Volume and Issue: 25(1)

Published: Jan. 21, 2025

To develop and validate a computed tomography (CT)-based deep learning radiomics model to predict treatment response progression-free survival (PFS) in patients with unresectable hepatocellular carcinoma (uHCC) treated transarterial chemoembolization (TACE)-hepatic arterial infusion chemotherapy (HAIC) combined PD-1 inhibitors tyrosine kinase (TKIs). This retrospective study included 172 uHCC who underwent combination therapy of TACE-HAIC TKIs inhibitors. Among them, 122 were from the Interventional Department Harbin Medical University Cancer Hospital, 92 randomly assigned training cohort 30 cases testing cohort. The remaining 50 Affiliated Fourth Hospital used for external validation. All liver enhanced CT examination before treatment. Residual convolutional neural network (ResNet) technology was extract image features. A predictive PFS established based on features clinical Model effectiveness evaluated using metrics such as area under receiver operating characteristic (ROC) curve (AUC), concordance index (C-index), accuracy, precision, F1-score. had median follow-up 25.2 months (95% CI 24.4–26.0), 14.0 8.5–19.4) overall (OS) 26.2 15.9–36.4) achieved. Objective rate (ORR) disease control (DCR) 41.0% 55.7%, respectively. In prediction model, AUC reached 0.96, an accuracy 89.5%, precision 85.6%, F1-score 0.896; 0.87, 80.4%, 74.5%, 0.802. validation 0.85, 79.1%, 73.6%, f1-score 0.784. predicted 12 months, 18 24 months-PFS 0.874, 0.809, 0.801, 0.762, 0.804, 0.792. 0.764, 0.796, 0.773. C-index 0.75, 0.591, 0.655, calibration demonstrated that significantly superior both models. provides CT-based can TKIs.

Language: Английский

---

Variações na Captação de FDG Miocárdica e Uso de Metformina: Implicações para a Sobrevida Durante a Imunoterapia

ARQUIVOS BRASILEIROS DE CARDIOLOGIA - IMAGEM CARDIOVASCULAR, Journal Year: 2025, Volume and Issue: 38(1)

Published: Jan. 30, 2025

Introdução: O aumento do uso de inibidores checkpoint imunológicos (ICIs) melhorou significativamente os resultados no câncer pulmão; entanto, ainda há falta protocolos para prever a resposta ao tratamento. Além disso, estudos pré-clínicos indicaram uma associação promissora entre metformina, β-bloqueadores (BBs) e melhores em pacientes com câncer. Objetivos: objetivo principal deste estudo foi investigar o impacto da metformina nos desfechos sobrevida. Os objetivos secundários incluíram avaliação variação na captação FDG miocárdio (alteração valor padronizado [ΔSUV]) durante tratamento ICIs dos efeitos tabagismo, diabetes, hipertensão BBs Métodos: Este coorte retrospectivo unicêntrico braço único avaliou pulmão que começaram usar julho 2016 dezembro 2021. critérios inclusão foram: idade superior 18 anos, tratado (inibidores CTLA-4, PD-1 PD-L1) realização pelo menos dois exames tomografia por emissão pósitrons combinada à computadorizada (PET-CT). Resultados: Cinquenta oito preencheram todos inclusão. usuários apresentaram um 759 dias sobrevida global (SG) (p = 0,015). Uma tendência 161 livre progressão (SLP) observada ΔSUV miocárdica positiva comparação grupo negativa 0,066), juntamente 285 favor (p=0,886). Conclusão: A significativa SG sugere é adjuvante promissor terapia ICI. pode sugerir papel potencial PET-CT previsão resposta, porém, maiores são necessários solidificar essa hipótese.

---