Background: Evaluating the risk of metastasis and recurrence of a cervical cancer patient is critical for appropriate adjuvant therapy. However, current risk assessment models usually involve the testing of tens to thousands of genes from patients’ tissue samples, which is expensive and timeconsuming. Therefore, computer-aided diagnosis and prognosis prediction based on Hematoxylin and Eosin (H&E) pathological images have received much attention recently. Objective: The prognosis of whether patients will have metastasis and recurrence can support accurate treatment for patients in advance and help reduce patient loss. It is also important for guiding treatment after surgery to be able to quickly and accurately predict the risk of metastasis and recurrence of a cervical cancer patient. Method: To address this problem, we propose a hybrid method. Transfer learning is used to extract features, and it is combined with traditional machine learning in order to analyze and determine whether patients have the risks of metastasis and recurrence. First, the proposed model retrieved relevant patches using a color-based method from H&E pathological images, which were then subjected to image preprocessing steps such as image normalization and color homogenization. Based on the labeled patched images, the Xception model with good classification performance was selected, and deep features of patched pathological images were automatically extracted with transfer learning. After that, the extracted features were combined to train a random forest model to predict the label of a new patched image. Finally, a majority voting method was developed to predict the metastasis and recurrence risk of a patient based on the predictions of patched images from the whole-slide H&E image. Results: In our experiment, the proposed model yielded an area under the receiver operating characteristic curve of 0.82 for the whole-slide image. The experimental results showed that the high-level features extracted by the deep convolutional neural network from the whole-slide image can be used to predict the risk of recurrence and metastasis after surgical resection and help identify patients who might receive additional benefit from adjuvant therapy. Conclusion: This paper explored the feasibility of predicting the risk of metastasis and recurrence from cervical cancer whole slide H&E images through deep learning and random forest methods.

BackgroundLung cancer is a major health concern worldwide because of its increasing incidence and mortality. This study aimed to clarify the association between mesenchymal-epithelial transition (MET) genomic alterations and clinical characteristics of lung cancer.MethodWe collected data from 5,008 patients with lung cancer diagnosed and treated between January 2017 and July 2021 at the Affiliated Hospital of Qingdao University. Genomic alterations in the MET gene, including the exon 14 skipping mutation and amplification, were detected using amplification refractory mutation system-polymerase chain reaction (2,057 cases) and next-generation sequencing (2,951 cases). Clinical characteristics such as age, sex, tumor location, tumor stage, smoking, pleural invasion, and histology were statistically analyzed for MET exon 14 skipping mutation and amplification. The DNA splicing sites causing the MET exon 14 skipping mutation at the mRNA level were also investigated.ResultsThe incidence of the MET exon 14 skipping mutation was 0.90% (41/4,564) in adenocarcinoma, 1.02% (3/294) in squamous cell carcinoma, and 8.33% (1/12) in sarcomatoid carcinoma specimens. It was more frequently observed in patients over 60 years of age than the MET exon 14 skipping mutation wildtype. The MET exon 14 skipping mutation co-occurred with epidermal growth factor receptor (EGFR) L858R, EGFR 19-Del, and BRAF V600E mutations. At the DNA level, single nucleotide mutation and small fragment deletion (1–38 base pairs) upstream and downstream of MET exon 14 led to MET exon 14 skipping mutation at the mRNA level. MET amplification occurred in 0.78% (21/2,676) adenocarcinoma and 1.07% (2/187) squamous cell carcinoma specimens and was significantly associated with advanced tumor stages (III + IV) compared to the MET amplification wildtype. MET amplification primarily co-occurred with the EGFR mutation.ConclusionsOur study found that MET genomic alterations were statistically related to age and tumor stage and co-existed with mutations of other oncogenic driver genes, such as EGFR and BRAF. Moreover, various splicing site changes at the DNA level led to the exon 14 skipping mutation at the mRNA level. Further studies are required to clarify the association between MET genomic alterations and prognosis.

Human blood contains cell-free DNA (cfDNA), with circulating tumor-derived DNAs (ctDNAs) widely used in cancer diagnosis and treatment. However, it is still difficult to efficiently and accurately identify and distinguish specific ctDNAs from normal cfDNA in cancer patient blood samples. In this study, ctDNA fragment length distribution analysis showed that ctDNA fragments are frequently shorter than the normal cfDNAs, which is consistent with previous findings. Interestingly, the ctDNA fragment length was found to be partially associated with the mutant allele frequency, with a low mutant allele frequency (< ~0.6%) associated with a longer ctDNA fragment length when compared to normal cfDNAs. The findings of this study contribute to improving the detection of low-frequency tumor mutations.

Background: Microsatellite instability (MSI) is a prognostic biomarker used to guide medication selection in multiple cancers, such as colorectal cancer. Traditional PCR with capillary electrophoresis and next-generation sequencing using paired tumor tissue and leukocyte samples are the main approaches for MSI detection due to their high sensitivity and specificity. Currently, patient tissue samples are obtained through puncture or surgery, which causes injury and risk of concurrent disease, further illustrating the need for MSI detection by liquid biopsy. Methods: We propose an analytic method using paired plasma/leukocyte samples and MSI detection using next-generation sequencing technology. Based on the theoretical progress of oncogenesis, we hypothesized that the microsatellite site length in plasma equals the combination of the distribution of tumor tissue and leukocytes. Thus, we defined a window-judgement method to identify whether biomarkers were stable. Results: Compared to traditional PCR as the standard, we evaluated three methods in 20 samples (MSI-H:3/MSS:17): peak shifting method using tissue vs. leukocytes, peak shifting method using plasma vs. leukocytes, and our method using plasma vs. leukocytes. Compared to traditional PCR, we observed a sensitivity of 100%, 0%, and 100%, and a specificity of 100.00%, 94.12%, and 88.24%, respectively. Conclusion: Our method has the advantage of possibly detecting MSI in a liquid biopsy and provides a novel direction for future studies to increase the specificity of the method.