Cancer Imaging最新文献

Background: The Breast Imaging Reporting and Data System (BI-RADS) is a widely accepted standardized framework for breast imaging interpretation including ultrasound, mammogram and magnetic resonance. Intermediate BI-RADS categories nodules currently require further biopsy or surgical resection to obtain pathological information. Notably, many such nodules are ultimately diagnosed as benign, prompting us to question whether intermediate BI-RADS categories nodules truly need invasive procedures. Additionally, malignancy rates of intermediate BI-RADS nodules vary across age groups and are influenced by clinical/biochemical factors. Therefore, a pressing challenge is to leverage current diagnostic tools for more precise identification of nodules that truly require biopsy, thereby reducing unnecessary invasive interventions. This study aims to address these challenges by integrating radiomics features with clinical and biochemical data to improve diagnostic accuracy.

Methods: This retrospective study enrolled 384 breast nodule patients from two medical centers with preoperative dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and blood biochemical tests, allocated into training and external test sets. A total of 3,948 radiomic features were extracted from DCE-MRI images and integrated with clinical characteristics. After 5-fold cross-validation for high-frequency feature selection, a malignancy-predicting nomogram was developed. Diagnostic performance was evaluated via area under receiver operator characteristic curve (AUC) with DeLong test against BI-RADS. Under the sensitivity threshold of > 95%, McNemar's test compared the specificity between the nomogram and BI-RADS to evaluate their biopsy reduction capabilities.

Results: The nomogram yielded an AUC of 0.89 [95% confidence interval (CI), 0.85-0.92] in the training cohort and an AUC of 0.89 (95% CI, 0.81-0.96) in the test cohort. When applying the cut-off value with ≥ 95% sensitivity, the nomogram can reduce unnecessary biopsies by 12.8% (16/125) in the training cohort and 25% (9/36) in the test cohort when compared with BI-RADS (p = 0.068 in training cohort and p = 0.078 in test cohort).

Conclusions: We have established a nomogram based on DCE-MRI radiomics and clinical risk factors to distinguish malignant from benign breast lesions, and demonstrated potential to reduce unnecessary biopsies, serving as a supplementary tool for BI-RADS-based clinical decision-making.

Primary lung cancer (LC) and breast cancer (BC) are among the most common malignancies and are highly prone to brain metastasis (BM). This study aimed to identify risk factors for brain metastasis-free survival in patients with primary LC or BC and to construct clinically simple nomograms. Our study analyzed the independent factors for the occurrence of BM by univariate and multivariate Cox regression based on the training set and then developed nomograms. The performance of the nomogram was determined by the C-index and calibration curve. The results were verified with a validation set. A total of 1739 patients with primary LC and 1150 with primary BC were included in our retrospective study. In primary LC, pathological staging, N stage, targeted therapy, and chemotherapy treatment were significantly associated with BM. In primary BC, the factors significantly associated with BM were TNBC, Ki-67 index, targeted therapy, radiotherapy, and surgery. These two nomograms had discriminatory ability, with C-indices of 0.786 and 0.783 in the training set and 0.809 and 0.843 in the validation set, respectively. We constructed and validated predictive nomograms for the development of BM in patients with primary LC or BC. The proposed nomograms certainly have good performance.

Objectives: The study aims to develop a deep learning (DL) model based on multiparametric magnetic resonance imaging (MRI) for distinguishing between benign and malignant breast lesions.

Methods: A total of 556 lesions (307 malignant, 249 benign) in 509 patients were pooled in the training/validation datasets between November 2018 and October 2019 in this retrospective study. A combined DL model based on the dynamic contrast enhanced-MRI (DCE-MRI) and apparent diffusion coefficient (ADC) map was developed to characterize breast lesions. Model performance was evaluated by using areas under the receiver operating characteristic curve (AUC) in the validation dataset and an independent testing dataset consisting of 243 lesions in 225 patients, and compared with other combined and single-parametric DL models. The predictive performance for malignancy was also compared between the DCE-ADC combined DL model and human readers.

Results: The DCE-ADC combined DL model achieved the highest diagnostic efficiency with the AUC, accuracy, sensitivity, and specificity of 0.889, 82.5%, 80.7%, and 84.1% for predicting malignant breast lesions, surpassing other combined and single-parametric DL models. The DCE-ADC combined DL model achieved good performance (accuracy:82%) and outperformed both the junior radiologists (82% vs. 70%, p = 0.073; 82% vs. 72%, p = 0.142). The diagnostic performance of two junior radiologists was improved after artificial intelligence assistance with AUCs increased to 0.798 and 0.772 from 0.689 to 0.708, respectively.

Conclusion: The DCE-ADC combined DL model shows promising diagnostic performance and has good potential to assist junior radiologists in improving diagnostic efficacy, which can facilitate clinical decision-making. Further studies will validate these findings in prospective, larger cohorts, multicenter, multiscanner and multinational studies.

Background: Primary central nervous system lymphoma (PCNSL) and glioblastoma (GBM) are two distinct types of malignant brain tumors, each requiring specific therapeutic approaches. Accurate differentiation between these tumors is crucial for selecting appropriate treatments.

Methods: We developed and validated a 3D DenseNet264 convolutional neural network (CNN) to automatically differentiate PCNSL and GBM. A total of 141 patients initially admitted to Tiantan Hospital underwent preoperative T1Gd-MRI and were confirmed by histopathology. These patients were randomly divided into training and validation groups at a 7:3 ratio. Subsequently, the DenseNet264 was trained and validated using these datasets. External validation was performed using additional datasets from the Radiological Society of North America (RSNA) and patients previously admitted to Tiantan Hospital. Standardized image preprocessing was conducted following the RSNA-ASNR-MICCAI BraTS 2021 guidelines.

Result: A total of 623 patients (Tiantan Hospital: 535, RSNA: 88) were initially enrolled, of whom 316 patients (Tiantan Hospital: 228 [141 patients enrolled between December 2015 and December 2021, and 87 patients enrolled before November 2015], RSNA: 88; GBM: 159, PCNSL: 157) met the inclusion criteria. The DenseNet264 achieved optimal classification performance in the training set (AUC: 0.98) and validation set (AUC: 0.90). In held-out data from RSNA and patients enrolled earlier at Tiantan Hospital, the model showed similarly consistent performance (C-statistic: 0.77).

Conclusions: We successfully developed and validated a robust deep-learning model capable of accurately differentiating PCNSL from GBM. This model provides a reliable, efficient, and cost-effective clinical decision-support tool for differential diagnosis.