Preoperative multiparameter MRI‐based prediction of Ki‐67 expression in primary central nervous system lymphoma

Precision Radiation Oncology, Journal Year: 2025, Volume and Issue: unknown

Published: March 18, 2025

Abstract Background Ki‐67 is a key marker of tumor proliferation. This study aimed to develop machine learning models using single‐ and multi‐parameter MRI radiomic features for the preoperative prediction expression in primary central nervous system lymphoma (PCNSL), aiding prognosis individualized treatment planning. Methods A retrospective analysis 74 patients was conducted scans, including T1, contrast‐enhanced T2, T2‐FLAIR, DWI, ADC sequences. Patients were categorized into high‐expression (Ki‐67 > 70%) low‐expression ≤ groups. Tumor volumes interest (VOIs) manually delineated by radiologists, 851 extracted 3DSlicer. After preprocessing, bias field correction normalization, feature selection performed SelectKBest ANOVA. Eight classifiers, Logistic Regression, Random Forest, SVM, applied datasets. Results Multiparameter models, particularly Naive Bayes demonstrated superior predictive performance (AUC: 0.78, 0.73; AP: 0.90, 0.83) compared single‐parameter models. Decision curve highlighted that Regression provides highest net benefit, followed Bayes. Conclusion are more accurate stable predicting PCNSL, supporting clinical decision‐making.

Language: Английский

---

Machine learning models for prediction of NPVR ≥80% with HIFU ablation for uterine fibroids

International Journal of Hyperthermia, Journal Year: 2025, Volume and Issue: 42(1)

Published: March 23, 2025

Background Currently high-intensity focused ultrasound (HIFU) is widely used to treat uterine fibroids (UFs). The aim of this study develop a machine learning model that can accurately predict the efficacy HIFU ablation for UFs, assisting preoperative selection suitable patients with UFs.

Language: Английский

---

Noninvasive prediction of Ki-67 expression level in IDH-wildtype glioblastoma using MRI histogram analysis: comparison and combination of MRI morphological features

Frontiers in Oncology, Journal Year: 2025, Volume and Issue: 15

Published: May 16, 2025

Purpose To assess and compare the effectiveness of magnetic resonance imaging (MRI) morphological features MRI histogram analysis in noninvasively predicting Ki-67 expression levels patients with IDH-wildtype glioblastoma. Methods Forty-six cases glioblastoma measured from January 2022 to July 2024 were retrospectively collected. They divided into low-level group (Ki-67<20%, n=20) high-level (Ki-67≥20%, n=26) according level. assessed recorded. performed on contrast-enhanced T1-weighted images. Differences between these parameters compared two groups. The diagnostic performance was by area under receiver operating characteristic curve (AUC). Spearman correlation used evaluate relationship Results Hemorrhage more prone occur ( P =0.017). min, P01, P50, P75 higher than those <0.00357). There a significant positive min r =0.774), P01 =0.729), P50 =0.625), =0.591), level <0.05). optimal obtained combining parameters, an AUC 0.867. Conclusion Both could predict glioblastoma, combined model integrating can be excellent biomarker for

Language: Английский

---

Deep Learning and Habitat Radiomics for the Prediction of Glioma Pathology Using Multiparametric MRI: A Multicenter Study

Academic Radiology, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 1, 2024

Language: Английский

---

Clinicopathological and radiological characteristics of false-positive and false-negative results in T2-FLAIR mismatch sign of IDH-mutated gliomas

Clinical Neurology and Neurosurgery, Journal Year: 2024, Volume and Issue: 246, P. 108579 - 108579

Published: Oct. 2, 2024

Language: Английский

---

A systematic review of radiological prediction of ki 67 proliferation index of meningioma

Neurosurgical Review, Journal Year: 2024, Volume and Issue: 47(1)

Published: Dec. 2, 2024

Language: Английский

---