Mining organization-related topics is helpful to analyze and monitor the information dissemination situation. Existing methods based on heterogeneous graph neural networks mainly consider the association between words and documents, they ignore the semantic interactions between documents, and do not consider the heterogeneity of edges, which are difficult to solve the challenge of blurred topic boundaries in real scenarios, resulting in performance loss. This paper proposes a BERT-based Heterogeneous Graph Convolution Network (BERT-HGCN) approach for semi-supervised topic mining that comprehensively considers multi-semantic relations between words and documents. It deeply combines the advantages of transductive learning with labeled examples and pre-training models. We model documents as graph-structured data and capture the multiple semantic dependencies among word-word, word-doc, and doc-doc via neighborhood propagation. During the model learning process, a two-channel encoding mechanism is used to learn the structure and semantic representations, which fuses a hierarchical graph convolution network (HGCN) and a BERT-based DNN autoencoder. It simultaneously considers edges heterogeneity and semantics of original documents. Finally, a dual-supervised loss function is used to train a classifier based on graph nodes and semantic representations for topic mining. We empirically evaluate the performance of the proposed model on a real-world organization-related dataset, the experimental results demonstrate the efficacy of the model.