Metastasis is facilitated by cancer-associated fibroblasts (CAF) in the tumor microenvironment through mechanisms yet to be elucidated. In this study, we used a size-based microfilter technology developed by our group to examine whether circulating CAF identified by FAP and a-SMA coexpression (cCAF) could be distinguished in the peripheral blood of patients with metastatic breast cancer. In a pilot study of patients with breast cancer, we detected the presence of cCAFs in 30/34 (88%) patients with metastatic disease (MET group) and in 3/13 (23%) patients with localized breast cancer (LOC group) with long-term disease-free survival. No cCAFs as defined were detected in healthy donors. Further, both cCAF and circulating tumor cells (CTC) were significantly greater in the MET group compared with the LOC group. Thus, the presence of cCAF was associated with clinical metastasis, suggesting that cCAF may complement CTC as a clinically relevant biomarker in metastatic breast cancer.

No abstract

Objectives: To apply deep learning algorithms using a conventional convolutional neural network (CNN) and a recurrent CNN to differentiate three breast cancer molecular subtypes on MRI.Methods: A total of 244 patients were analyzed, 99 in Training Dataset scanned at 1.5T, 83 in Testing-1 and 62 in Testing-2 scanned at 3T. Patients were classified into 3 subtypes based on hormonal receptor (HR) and HER2 receptor: (HR+/HER2−), HER2+ and triple negative (TN). Only images acquired in the DCE sequence were used in the analysis. The smallest bounding box covering tumor ROI was used as the input for deep learning to develop the model in the Training dataset, by using a conventional CNN and the convolutional long short term memory (CLSTM). Then, transfer learning was applied to re-tune the model using Testing-1(2) and evaluated in Testing-2(1). Results:In the Training dataset, the mean accuracy evaluated using 10-fold cross-validation was higher by using CLSTM (0.91) than CNN (0.79). When the developed model was applied to the independent Testing datasets, the accuracy was 0.4-0.5. With transfer learning by re-tuning parameters in Testing-1, the mean accuracy reached 0.91 by CNN and 0.83 by CLSTM, and improved accuracy in Testing-2 from 0.47 to 0.78 by CNN and from 0.39 to 0.74 by CLSTM. Overall, transfer learning could improve the classification accuracy by greater than 30%.

No abstract