The use of deep learning methods for the automatic detection and quantification of bone metastases in bone scan images holds significant clinical value. A fast and accurate automated system for segmenting bone metastatic lesions can assist clinical physicians in diagnosis. In this study, a small internal dataset comprising 100 breast cancer patients (90 cases of bone metastasis and 10 cases of non-metastasis) and 100 prostate cancer patients (50 cases of bone metastasis and 50 cases of non-metastasis) was used for model training. Initially, all image labels were binary. We used the Otsu thresholding method or negative mining to generate a non-metastasis mask, thereby transforming the image labels into three classes. We adopted the Double U-Net as the baseline model and made modifications to its output activation function. We changed the activation function to SoftMax to accommodate multi-class segmentation. Several methods were used to enhance model performance, including background pre-processing to remove background information, adding negative samples to improve model precision, and using transfer learning to leverage shared features between two datasets, which enhances the model’s performance. The performance was investigated via 10-fold cross-validation and computed on a pixel-level scale. The best model we achieved had a precision of 69.96%, a sensitivity of 63.55%, and an F1-score of 66.60%. Compared to the baseline model, this represents an 8.40% improvement in precision, a 0.56% improvement in sensitivity, and a 4.33% improvement in the F1-score. The developed system has the potential to provide pre-diagnostic reports for physicians in final decisions and the calculation of the bone scan index (BSI) with the combination with bone skeleton segmentation.

Background: Lumbar degenerative disc disease (LDDD) is a leading cause of chronic lower back pain; however, a lack of clear diagnostic criteria and solid LDDD interventional therapies have made predicting the benefits of therapeutic strategies challenging. Our goal is to develop machine learning (ML)–based radiomic models based on pre-treatment imaging for predicting the outcomes of lumbar nucleoplasty (LNP), which is one of the interventional therapies for LDDD. Methods: The input data included general patient characteristics, perioperative medical and surgical details, and pre-operative magnetic resonance imaging (MRI) results from 181 LDDD patients receiving lumbar nucleoplasty. Post-treatment pain improvements were categorized as clinically significant (defined as a ≥80% decrease in the visual analog scale) or non-significant. To develop the ML models, T2-weighted MRI images were subjected to radiomic feature extraction, which was combined with physiological clinical parameters. After data processing, we developed five ML models: support vector machine, light gradient boosting machine, extreme gradient boosting, extreme gradient boosting random forest, and improved random forest. Model performance was measured by evaluating indicators, such as the confusion matrix, accuracy, sensitivity, specificity, F1 score, and area under the receiver operating characteristic curve (AUC), which were acquired using an 8:2 allocation of training to testing sequences. Results: Among the five ML models, the improved random forest algorithm had the best performance, with an accuracy of 0.76, a sensitivity of 0.69, a specificity of 0.83, an F1 score of 0.73, and an AUC of 0.77. The most influential clinical features included in the ML models were pre-operative VAS and age. In contrast, the most influential radiomic features had the correlation coefficient and gray-scale co-occurrence matrix. Conclusions: We developed an ML-based model for predicting pain improvement after LNP for patients with LDDD. We hope this tool will provide both doctors and patients with better information for therapeutic planning and decision-making.

Positron emission tomography and computed tomography with 18F-fluorodeoxyglucose (18F-FDG PET-CT) were used to predict outcomes after liver transplantation in patients with hepatocellular carcinoma (HCC). However, few approaches for prediction based on 18F-FDG PET-CT images that leverage automatic liver segmentation and deep learning were proposed. This study evaluated the performance of deep learning from 18F-FDG PET-CT images to predict overall survival in HCC patients before liver transplantation (LT). We retrospectively included 304 patients with HCC who underwent 18F-FDG PET/CT before LT between January 2010 and December 2016. The hepatic areas of 273 of the patients were segmented by software, while the other 31 were delineated manually. We analyzed the predictive value of the deep learning model from both FDG PET/CT images and CT images alone. The results of the developed prognostic model were obtained by combining FDG PET-CT images and combining FDG CT images (0.807 AUC vs. 0.743 AUC). The model based on FDG PET-CT images achieved somewhat better sensitivity than the model based on CT images alone (0.571 SEN vs. 0.432 SEN). Automatic liver segmentation from 18F-FDG PET-CT images is feasible and can be utilized to train deep-learning models. The proposed predictive tool can effectively determine prognosis (i.e., overall survival) and, thereby, select an optimal candidate of LT for patients with HCC.

Bone Scan Index (BSI) is an image biomarker for quantifying bone metastasis of cancers. To compute BSI, not only the hotspots (metastasis) but also the bones have to be segmented. Most related research focus on binary classification in bone scintigraphy: having metastasis or none. Rare studies focus on pixel-wise segmentation. This study compares three advanced convolutional neural network (CNN) based models to explore bone segmentation on a dataset in-house. The best model is Mask R-CNN, which reaches the precision, sensitivity, and F1-score: 0.93, 0.87, 0.90 for prostate cancer patients and 0.92, 0.86, and 0.88 for breast cancer patients, respectively. The results are the average of 10-fold cross-validation, which reveals the reliability of clinical use on bone segmentation.

Bone Scan Index (BSI) is an image biomarker for quantification on bone metastasis of cancers. To compute BSI, not only the hotspots (metastasis) but also the bones have to be segmented. Most related researches focus on binary classification on bone scintigraphy: having metastasis or none. Rare studies focus on pixelwise segmentation. In this study we compare three advanced convolutional neural network (CNN) based models to explore the bone segmentation on dataset in house. The best model is Mask R-CNN, which reaches the precision, sensitivity and F1-score: 0.93, 0.87, 0.90 for prostate cancer patients and 0.92, 0.86, 0.88 for breast cancer patients, respectively. The results are the average of 10-fold cross-validation, which reveals the reliability for clinical use on bone segmentation.

Bone metastasis detection and quantification on bone scintigraphy is challenging and clinical important for treatment and patient’s life quality. To develop a CNN based diagnostic system for automated segmentation on bone metastasis regions is non-trivial, especially for a small dataset. A dataset in house comprising 100 breast cancer patients and 100 prostate cancer patients is utilized for this research. The Double U-Net model is adapted through the integration of background removal, adding negative samples, and transfer learning methods for bone metastasis detection. The performance is investigated via 10-fold cross-validation and computed in pixel-wise scale. The best model we achieved has precision of 63.08%, sensitivity of 70.82%, and F1-score of 66.72%. The developed system has the potential to provide pre-diagnostic reports for physicians in final decisions and the calculation of the bone scan index (BSI) with the combination with bone skeleton segmentation.