This study aimed to investigate the predictive efficacy of positron emission tomography/computed tomography (PET/CT) and magnetic resonance imaging (MRI) for the pathological response of advanced breast cancer to neoadjuvant chemotherapy (NAC). The breast PET/MRI image deep learning model was introduced and compared with the conventional methods. PET/CT and MRI parameters were evaluated before and after the first NAC cycle in patients with advanced breast cancer [n = 56; all women; median age, 49 (range 26–66) years]. The maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) were obtained with the corresponding baseline values (SUV0, MTV0, and TLG0, respectively) and interim PET images (SUV1, MTV1, and TLG1, respectively). Mean apparent diffusion coefficients were obtained from baseline and interim diffusion MR images (ADC0 and ADC1, respectively). The differences between the baseline and interim parameters were measured (ΔSUV, ΔMTV, ΔTLG, and ΔADC). Subgroup analysis was performed for the HER2-negative and triple-negative groups. Datasets for convolutional neural network (CNN), assigned as training (80%) and test datasets (20%), were cropped from the baseline (PET0, MRI0) and interim (PET1, MRI1) images. Histopathologic responses were assessed using the Miller and Payne system, after three cycles of chemotherapy. Receiver operating characteristic curve analysis was used to assess the performance of the differentiating responders and non-responders. There were six responders (11%) and 50 non-responders (89%). The area under the curve (AUC) was the highest for ΔSUV at 0.805 (95% CI 0.677–0.899). The AUC was the highest for ΔSUV at 0.879 (95% CI 0.722–0.965) for the HER2-negative subtype. AUC improved following CNN application (SUV0:PET0 = 0.652:0.886, SUV1:PET1 = 0.687:0.980, and ADC1:MRI1 = 0.537:0.701), except for ADC0 (ADC0:MRI0 = 0.703:0.602). PET/MRI image deep learning model can predict pathological responses to NAC in patients with advanced breast cancer.

Osteosarcoma (OS) is the most common primary bone tumor affecting humans and it has extreme heterogeneity. Despite modern therapy, it recurs in approximately 30–40% of patients initially diagnosed with no metastatic disease, with the long-term survival rates of patients with recurrent OS being generally 20%. Thus, early prediction of metastases in OS management plans is crucial for better-adapted treatments and survival rates. In this study, a radiomics model for metastasis risk prediction in OS was developed and evaluated using metabolic imaging phenotypes. The subjects were eighty-three patients with OS, and all were treated with surgery and chemotherapy for local control. All patients underwent a pretreatment ¹⁸F-FDG-PET scan. Forty-five features were extracted from the tumor region. The incorporation of features into multivariable models was performed using logistic regression. The multivariable modeling strategy involved cross validation in the following four steps leading to final prediction model construction: (1) feature set reduction and selection; (2) model coefficients computation through train and validation processing; and (3) prediction performance estimation. The multivariable logistic regression model was developed using two radiomics features, SUVmax and GLZLM-SZLGE. The trained and validated multivariable logistic model based on probability of endpoint (P) = 1/ (1+exp (-Z)) was Z = -1.23 + 1.53*SUVmax + 1.68*GLZLM-SZLGE with significant p-values (SUVmax: 0.0462 and GLZLM_SZLGE: 0.0154). The final multivariable logistic model achieved an area under the curve (AUC) receiver operating characteristics (ROC) curve of 0.80, a sensitivity of 0.66, and a specificity of 0.88 in cross validation. The SUVmax and GLZLM-SZLGE from metabolic imaging phenotypes are independent predictors of metastasis risk assessment. They show the association between ¹⁸F-FDG-PET and metastatic colonization knowledge. The multivariable model developed using them could improve patient outcomes by allowing aggressive treatment in patients with high metastasis risk.

Purpose. Patients with high-grade osteosarcoma undergo several chemotherapy cycles before surgical intervention. Response to chemotherapy, however, is affected by intratumor heterogeneity. In this study, we assessed the ability of a machine learning approach using baseline ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emitted tomography (PET) textural features to predict response to chemotherapy in osteosarcoma patients. Materials and Methods. This study included 70 osteosarcoma patients who received neoadjuvant chemotherapy. Quantitative characteristics of the tumors were evaluated by standard uptake value (SUV), total lesion glycolysis (TLG), and metabolic tumor volume (MTV). Tumor heterogeneity was evaluated using textural analysis of ¹⁸F-FDG PET scan images. Assessments were performed at baseline and after chemotherapy using ¹⁸F-FDG PET; ¹⁸F-FDG textural features were evaluated using the Chang-Gung Image Texture Analysis toolbox. To predict the chemotherapy response, several features were chosen using the principal component analysis (PCA) feature selection method. Machine learning was performed using linear support vector machine (SVM), random forest, and gradient boost methods. The ability to predict chemotherapy response was evaluated using the area under the receiver operating characteristic curve (AUC). Results. AUCs of the baseline ¹⁸F-FDG features SUVmax, TLG, MTV, 1st entropy, and gray level co-occurrence matrix entropy were 0.553, 0538, 0.536, 0.538, and 0.543, respectively. However, AUCs of the machine learning features linear SVM, random forest, and gradient boost were 0.72, 0.78, and 0.82, respectively. Conclusion. We found that a machine learning approach based on ¹⁸F-FDG textural features could predict the chemotherapy response using baseline PET images. This early prediction of the chemotherapy response may aid in determining treatment plans for osteosarcoma patients.

Neuroinflammation involves activation of glial cells in the brain, and activated microglia play a particularly important role in neurodegenerative diseases such as Alzheimer’s disease (AD). In this study, we developed 5-cyano-N-(4-(4-(2-[¹⁸F]fluoroethyl)piperazin-1-yl)-2-(piperidin-1-yl)phenyl)furan-2-carboxamide ([¹⁸F]1) for PET imaging of colony-stimulating factor 1 receptor (CSF1R), an emerging target for neuroinflammation imaging. Non-radioactive ligand 1 exhibited binding affinity comparable to that of a known CSF1R inhibitor, 5-cyano-N-(4-(4-methylpiperazin-1-yl)-2-(piperidin-1-yl)phenyl)furan-2-carboxamide (CPPC). Therefore, we synthesized radioligand [¹⁸F]1 by radiofluorination of chlorine-substituted precursor 7 in 13–15% decay-corrected radiochemical yield. Dynamic PET/CT images showed higher uptake in the lipopolysaccharide (LPS)-treated mouse brain than in control mouse brain. Ex vivo biodistribution study conducted at 45 min after radioligand injection showed that the brain uptake in LPS mice increased by 78% compared to that of control mice and was inhibited by 22% in LPS mice pretreated with CPPC, indicating specificity of [¹⁸F]1 for CSF1R. A metabolism study demonstrated that the radioligand underwent little metabolism in the mouse brain. Taken together, these results suggest that [¹⁸F]1 may hold promise as a radioligand for CSF1R imaging.

The image texture features obtained from ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) images of non-small cell lung cancer (NSCLC) have revealed tumor heterogeneity. A combination of genomic data and radiomics may improve the prediction of tumor prognosis. This study aimed to predict NSCLC metastasis using a graph neural network (GNN) obtained by combining a protein–protein interaction (PPI) network based on gene expression data and image texture features. ¹⁸F-FDG PET/CT images and RNA sequencing data of 93 patients with NSCLC were acquired from The Cancer Imaging Archive. Image texture features were extracted from ¹⁸F-FDG PET/CT images and area under the curve receiver operating characteristic curve (AUC) of each image feature was calculated. Weighted gene co-expression network analysis (WGCNA) was used to construct gene modules, followed by functional enrichment analysis and identification of differentially expressed genes. The PPI of each gene module and genes belonging to metastasis-related processes were converted via a graph attention network. Images and genomic features were concatenated. The GNN model using PPI modules from WGCNA and metastasis-related functions combined with image texture features was evaluated quantitatively. Fifty-five image texture features were extracted from ¹⁸F-FDG PET/CT, and radiomic features were selected based on AUC (n = 10). Eighty-six gene modules were clustered by WGCNA. Genes (n = 19) enriched in the metastasis-related pathways were filtered using DEG analysis. The accuracy of the PPI network, derived from WGCNA modules and metastasis-related genes, improved from 0.4795 to 0.5830 (p < 2.75 × 10⁻¹²). Integrating PPI of four metastasis-related genes with ¹⁸F-FDG PET/CT image features in a GNN model elevated its accuracy over a without image feature model to 0.8545 (95% CI = 0.8401–0.8689, p-value < 0.02). This model demonstrated significant enhancement compared to the model using PPI and ¹⁸F-FDG PET/CT derived from WGCNA (p-value < 0.02), underscoring the critical role of metastasis-related genes in prediction model. The enhanced predictive capability of the lymph node metastasis prediction GNN model for NSCLC, achieved through the integration of comprehensive image features with genomic data, demonstrates promise for clinical implementation.

We investigated predictions from ¹⁸F-FDG PET/CT using machine learning (ML) to assess the neoadjuvant CCRT response of patients with stage III non-small cell lung cancer (NSCLC) and compared them with predictions from conventional PET parameters and from physicians. A retrospective study was conducted of 430 patients. They underwent ¹⁸F-FDG PET/CT before initial treatment and after neoadjuvant CCRT followed by curative surgery. We analyzed texture features from segmented tumors and reviewed the pathologic response. The ML model employed a random forest and was used to classify the binary outcome of the pathological complete response (pCR). The predictive accuracy of the ML model for the pCR was 93.4%. The accuracy of predicting pCR using the conventional PET parameters was up to 70.9%, and the accuracy of the physicians’ assessment was 80.5%. The accuracy of the prediction from the ML model was significantly higher than those derived from conventional PET parameters and provided by physicians (p < 0.05). The ML model is useful for predicting pCR after neoadjuvant CCRT, which showed a higher predictive accuracy than those achieved from conventional PET parameters and from physicians.