Overall survival in patients with FIGO stage IVA cervical cancer

Schiff, Joshua; Mintz, Rachel; Cohen, Alexander C.; Huang, Yi; Thaker, Premal H.; Massad, L. Stewart; Powell, Matthew A.; Mutch, David G.; Schwarz, Julie K.; Markovina, Stephanie; Grigsby, Perry W.

doi:10.1016/j.ygyno.2022.05.022

Cited by 5 publications

(1 citation statement)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, the staging methods of the International Federation of Gynecology and Obstetrics (FIGO) are the most commonly used for cervical cancer ( 23 ). Imaging or pathological examination are the major components of the FIGO staging system.…”

Section: Discussionmentioning

confidence: 99%

A novel cervix carcinoma biomarker: Pathological-epigenomics, integrated analysis of MethylMix algorithm and pathology for predicting response to cancer immunotherapy

Shi²,

Gu³

et al. 2022

Front. Oncol.

View full text Add to dashboard Cite

Herein, A non-invasive pathomics approach was developed to reveal the methylation status in patients with cervical squamous cell carcinoma and predict clinical outcomes and treatment response. Using the MethylMix algorithm, 14 methylation-driven genes were selected for further analysis. We confirmed that methylation-driven genes were differentially expressed in immune, stromal, and tumor cells. In addition, we constructed a methylation-driven model and explored the alterations in immunocyte infiltration between the different models. The methylation-driven subtypes identified in our investigation could effectively predict the clinical outcomes of cervical cancer. To further evaluate the level of methylation-driven patterns, we constructed a risk model with four genes. Significant correlations were observed between the score and immune response markers, including PD1 and CTLA4. Multiple immune infiltration algorithms evaluated the level of immunocyte infiltration between the high- and low-risk groups, while the components of anti-tumor immunocytes in the low-risk group were significantly increased. Subsequently, a total of 205 acquired whole-slide imaging (WSI) images were processed to capture image signatures, and the pathological algorithm was employed to construct an image prediction model based on the risk score classification. The model achieved an area under the curve (AUC) of 0.737 and 0.582 for the training and test datasets, respectively. Moreover, we conducted vitro assays for validation of hub risk gene. The proposed prediction model is a non-invasive method that combines pathomics features and genomic profiles and shows satisfactory performance in predicting patient survival and treatment response. More interdisciplinary fields combining medicine and electronics should be explored in the future.

show abstract

Section: Discussionmentioning

confidence: 99%