RESUMO
Objectives: The diagnosis and treatment of brain tumors have greatly benefited from extensive research in traditional radiomics, leading to improved efficiency for clinicians. With the rapid development of cutting-edge technologies, especially deep learning, further improvements in accuracy and automation are expected. In this study, we explored a hybrid deep learning scheme that integrates several advanced techniques to achieve reliable diagnosis of primary brain tumors with enhanced classification performance and interpretability. Methods: This study retrospectively included 230 patients with primary brain tumors, including 97 meningiomas, 66 gliomas and 67 pituitary tumors, from the First Affiliated Hospital of Yangtze University. The effectiveness of the proposed scheme was validated by the included data and a commonly used data. Based on super-resolution reconstruction and dynamic learning rate annealing strategies, we compared the classification results of several deep learning models. The multi-classification performance was further improved by combining feature transfer and machine learning. Classification performance metrics included accuracy (ACC), area under the curve (AUC), sensitivity (SEN), and specificity (SPE). Results: In the deep learning tests conducted on two datasets, the DenseNet121 model achieved the highest classification performance, with five-test accuracies of 0.989 ± 0.006 and 0.967 ± 0.013, and AUCs of 0.999 ± 0.001 and 0.994 ± 0.005, respectively. In the hybrid deep learning tests, LightGBM, a promising classifier, achieved accuracies of 0.989 and 0.984, which were improved from the original deep learning scheme of 0.987 and 0.965. Sensitivities for both datasets were 0.985, specificities were 0.988 and 0.984, respectively, and relatively desirable receiver operating characteristic (ROC) curves were obtained. In addition, model visualization studies further verified the reliability and interpretability of the results. Conclusions: These results illustrated that deep learning models combining several advanced technologies can reliably improve the performance, automation, and interpretability of primary brain tumor diagnosis, which is crucial for further brain tumor diagnostic research and individualized treatment.
RESUMO
Cetuximab therapy, which heavily relies on the activation of Ras pathway, has been used in KRAS, NRAS, BRAF, and PIK3CA wild-type colorectal cancer (CRC) (Ras-normal). However, the response rate only reached 60%, due to false-negative mutation detection and mutation-like transcriptome features in wild-type patients. Herein, by integrating RNA-seq, microarray, and mutation data, we developed a Ras pathway signature by characterizing KRAS/NRAS/BRAF/PIK3CA mutations to identify the hidden nonresponders from the Ras-normal patients by mutation detection. Using public and in-house data of CRC patients treated with cetuximab, discovery of the signature could identify cetuximab-resistant samples from the Ras-normal samples. Cetuximab resistance-related genes, such as PTEN, were significantly and frequently mutated in the identified Ras-activated samples, whereas two cetuximab sensitivity-related genes, APC and TP53, showed comutation and significantly higher mutation frequencies in the remaining Ras-normal samples. Furthermore, all the NF1- and BCL2L1-mutated samples were identified as Ras-activated from the Ras-normal samples by the Ras pathway signature with significantly under-regulated expression. Genes co-expressed with the two genes were both involved in Ras signaling pathway, the out-of-control of which could be attributed by the genes' loss-of-function mutations. To improve the treatment of cetuximab in CRC, NF1 and BCL2L1 could be used as complementary detection technique to those applied in clinical. In conclusion, the proposed Ras pathway signature could identify the hidden CRC patients resistant to cetuximab therapy and help to reveal resistance mechanisms.
