Search

BackgroundBreast cancer is the most common cancer type that affects women. In hormone receptor–positive (HR+), human epidermal growth factor receptor 2−negative (HER2–) advanced breast cancer (ABC), phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) is the most frequently mutated gene associated with poor prognosis. This study evaluated the frequency of PIK3CA mutations in the Taiwanese breast cancer population.MethodologyThis is a retrospective study; patient data were collected for 2 years from a next-generation sequencing database linked to electronic health records (EHRs). The primary endpoint was the regional prevalence of PIK3CA mutation. The secondary endpoints were to decipher the mutation types across breast cancer subtype, menopausal status, and time to treatment failure after everolimus (an mTOR inhibitor) or cyclin-dependent kinase 4/6 (CDK4/6) inhibitor treatment.ResultsPIK3CA mutations were identified in 278 of 728 patients (38%). PIK3CA mutations were reported in 43% of patients with HR−/HER2+ subtype and 42% of patients with HR+/HER2– postmenopausal status. A lower prevalence of PIK3CA mutations was observed in triple-negative (27%) and HR+/HER2– premenopausal patients (29%). The most common mutation was at exon 20 (H1047R mutation, 41.6%), followed by exon 9 (E545K mutation, 18.9% and E542K mutation, 10.3%). Among patients treated with CDK4/6 inhibitors, the median time to treatment failure was 12 months (95% CI: 7-21 months) in the PIK3CA mutation cohort and 16 months (95% CI: 11-23 months) in the PIK3CA wild-type cohort, whereas patients receiving an mTOR inhibitor reported a median time to treatment failure of 20.5 months (95% CI: 8-33 months) in the PIK3CA mutation cohort and 6 months (95% CI: 2-9 months) in the PIK3CA wild-type cohort.ConclusionA high frequency of PIK3CA mutations was detected in Taiwanese patients with breast cancer, which was consistent with previous studies. Early detection of PIK3CA mutations might influence therapeutic decisions, leading to better treatment outcomes.

Taxanes-containing chemotherapy constitutes an essential backbone for both early and metastatic breast cancer (mBC). However, the two major taxane drugs—paclitaxel and docetaxel—have distinct safety profiles. In this review, we summarize the safety outcome and management following treatment with both taxanes from selected clinical trials. We utilized PubMed to perform literature search before April 2021. Five phase III randomized controlled trials with reports of individual taxane adverse events (AEs) were included in this review. Grade 3/4 AEs were summarized and discussed extensively. The rates of grade 3/4 neutropenia were higher with docetaxel than with paclitaxel. For non-hematologic grade 3/4 AEs, peripheral neuropathy was more frequent with paclitaxel while fluid retention was more frequent with docetaxel. Compared to paclitaxel, docetaxel had a higher rate of grade 3/4 gastrointestinal AEs. Grade 3/4 myalgia were generally comparable between the two taxanes. Except for neutropenia, the incidence rate of grade 3/4 AEs of taxanes was generally manageable. Peripheral neuropathy was more common with paclitaxel while grade 3/4 neutropenia was more common with docetaxel.

We proposed a highly versatile two-step transfer learning pipeline for predicting the gene signature defining the intrinsic breast cancer subtypes using unannotated pathological images. Deciphering breast cancer molecular subtypes by deep learning approaches could provide a convenient and efficient method for the diagnosis of breast cancer patients. It could reduce costs associated with transcriptional profiling and subtyping discrepancy between IHC assays and mRNA expression. Four pretrained models such as VGG16, ResNet50, ResNet101, and Xception were trained with our in-house pathological images from breast cancer patient with recurrent status in the first transfer learning step and TCGA-BRCA dataset for the second transfer learning step. Furthermore, we also trained ResNet101 model with weight from ImageNet for comparison to the aforementioned models. The two-step deep learning models showed promising classification results of the four breast cancer intrinsic subtypes with accuracy ranging from 0.68 (ResNet50) to 0.78 (ResNet101) in both validation and testing sets. Additionally, the overall accuracy of slide-wise prediction showed even higher average accuracy of 0.913 with ResNet101 model. The micro- and macro-average area under the curve (AUC) for these models ranged from 0.88 (ResNet50) to 0.94 (ResNet101), whereas ResNet101_imgnet weighted with ImageNet archived an AUC of 0.92. We also show the deep learning model prediction performance is significantly improved relatively to the common Genefu tool for breast cancer classification. Our study demonstrated the capability of deep learning models to classify breast cancer intrinsic subtypes without the region of interest annotation, which will facilitate the clinical applicability of the proposed models.

PurposeThe present study aimed to assign a risk score for breast cancer recurrence based on pathological whole slide images (WSIs) using a deep learning model.MethodsA total of 233 WSIs from 138 breast cancer patients were assigned either a low-risk or a high-risk score based on a 70-gene signature. These images were processed into patches of 512x512 pixels by the PyHIST tool and underwent color normalization using the Macenko method. Afterward, out of focus and pixelated patches were removed using the Laplacian algorithm. Finally, the remaining patches (n=294,562) were split into 3 parts for model training (50%), validation (7%) and testing (43%). We used 6 pretrained models for transfer learning and evaluated their performance using accuracy, precision, recall, F1 score, confusion matrix, and AUC. Additionally, to demonstrate the robustness of the final model and its generalization capacity, the testing set was used for model evaluation. Finally, the GRAD-CAM algorithm was used for model visualization.ResultsSix models, namely VGG16, ResNet50, ResNet101, Inception_ResNet, EfficientB5, and Xception, achieved high performance in the validation set with an overall accuracy of 0.84, 0.85, 0.83, 0.84, 0.87, and 0.91, respectively. We selected Xception for assessment of the testing set, and this model achieved an overall accuracy of 0.87 with a patch-wise approach and 0.90 and 1.00 with a patient-wise approach for high-risk and low-risk groups, respectively.ConclusionsOur study demonstrated the feasibility and high performance of artificial intelligence models trained without region-of-interest labeling for predicting cancer recurrence based on a 70-gene signature risk score.