No abstract

Background: The prevalence and types of fibroblast growth factor receptor (FGFR) mutations vary significantly among different ethnic groups. The optimal application of FGFR inhibitors depends on these variations being comprehensively understood. However, such an analysis has yet to be conducted in Chinese patients. Methods:We retrospectively screened the genomic profiling results of 10,582 Chinese cancer patients across 16 cancer types to investigate the frequency and distribution of FGFR aberrations.Results: FGFR aberrations were identified in 745 patients, equating to an overall prevalence of 7.0%. A majority of the aberrations occurred on FGFR1 (56.8%), which was followed by FGFR3 (17.7%), FGFR2 (14.4%), and FGFR4 (2.8%). Further, 8.5% of patients had aberrations of more than 1 FGFR gene. The most common types of aberrations were amplification (53.7%), other mutations (38.8%), and fusions (5.6%). FGFR fusion and amplification occurred concurrently in 1.9% of the patients. FGFR aberrations were detected in 12 of the 16 cancers, with the highest prevalence belonging to colorectal cancer (CRC) (31%). Other FGFR-aberrant cancer types included stomach (16.8%), breast (14.3%), and esophageal (12.7%) cancer. Breast tumors were also more likely than other cancer types to have concurrent FGFR rearrangements and amplifications (P<0.001). In comparison with the public dataset, our cohort had a significantly higher number of FGFR aberrations in colorectal (P<0.001) and breast cancer (P=0.05).Conclusions: Among the Chinese cancer patients in our study, the overall prevalence of FGFR aberrations was 7.0%. FGFR1 amplification was the most common genetic alteration in CRC, breast cancer, and lung cancer; while FGFR2 amplification was more commonly observed in gastric cancer than in other cancers in our cohort. Our study advances the understanding of the distribution of FGFR aberrations in various cancer types in the Chinese population, which will facilitate the further development of FGFR inhibitors.

No abstract

Imbalanced classes and dimensional disasters are critical challenges in medical image classification. As a classical machine learning model, the n -gram model has shown excellent performance in addressing this issue in text classification. In this study, we proposed an algorithm to classify medical images by extracting their n -gram semantic features. This algorithm first converts an image classification problem to a text classification problem by building an n -gram corpus for an image. After that, the algorithm was based on the n -gram model to classify images. The algorithm was evaluated by two independent public datasets. The first experiment is to diagnose benign and malignant thyroid nodules. The best area under the curve (AUC) is 0.989. The second experiment is to diagnose the type of fundus lesion. The best result is that it correctly identified 86.667% of patients with dry age-related macular degeneration (AMD), 93.333% of patients with diabetic macular edema (DME), and 93.333% of normal individuals.