Automated ITSS Extraction Combined with R2* Values Obtained from Enhanced T2*-Weighted Angiography in Magnetic Resonance Imaging: A Promising Approach for Differentiate Cervical Adenocarcinoma from Squamous Carcinoma

Anliang Chen^1,2Xing Meng³Changjun Ma¹Shifeng Tian¹Qingwei Song¹Ailian Liu^1,2*Mingrui Zhuang⁴Hongkai Wang^4,5

• ¹Department of Radiology, First Affiliated Hospital, Dalian Medical University, Dalian, Liaoning Province, China
• ²Dalian Medical Imaging artificial intelligence engineering technology research center, Dalian, Liaoning Province, China
• ³Dalian Women and Children’s Medical Center(Group), Dalian, Liaoning Province, China
• ⁴School of Biomedical Engineering, Faculty of Electronic Information and Electrical Engineering, Dalian University of Technology, Dalian, Liaoning Province, China
• ⁵Liaoning Key Laboratory of Integrated Circuit and Biomedical Electronic System, Dalian, China

【Abstract】 Word count: 277 Purpose This study aims to evaluate the efficacy of utilizing automated intertumoral susceptibility signal (ITSS) intensity extraction combined with R2* values derived from enhanced T2*-weighted angiography (ESWAN) in magnetic resonance imaging (MRI) to distinguish between cervical adenocarcinoma (CA) and cervical squamous carcinoma (CSC). Methods Seventy-eight patients who underwent ESWAN from 2014 to 2019 were stratified into two groups: CA (26 patients) and CSC (52 patients). R2* values of the lesions were measured, and ITSS ratios were automatically calculated using the Anatomy Sketch (AS) software. Independent samples t-tests or Mann-Whitney U-tests were utilized to evaluate disparities in the parameters. Binary logistic regression was conducted to identify independent predictors. The receiver operating characteristic curve was employed to assess diagnostic value, and the Delong test was applied to compare differences in the area under the curve (AUC). Results The CA group exhibited significantly higher values for the ITSSs, ITSSv and R2* value, lower alpha fetoprotein (AFP) and prognostic nutritional index (PNI) (ITSSs: 0.203±0.111; ITSSv:0.206±0.098; R2* value:20.340±5.572Hz; AFP: 1.73(1.33,2.99)ng/ml; PNI:49.150(45.825,51.775)) than that of the CSC group (ITSSs: 0.072±0.019; ITSSv: 0.076±0.030; R2* value: 13.233±4.083Hz; AFP: 2.99(1.88,2.99)ng/ml; PNI: 50.775(48.563,54.050)) (P<0.05). Among them, ITSSv and R2* value were independent risk predictors. The AUC values for ITSSv, R2* value and the combined model for differentiate between CA and CSC were 0.942, 0.851 and 0.950, respectively. The results of the Delong test indicated that the combined model exhibited superior diagnostic efficacy compared to R2* value (P<0.05), but no significant difference from ITSSv (P>0.05). Conclusion ITSSv and R2* values derived from ESWAN facilitate the quantitative differentiate between CA and CSC. The automated extraction of ITSSv is convenient and reliable, making it a promising candidate for clinical implementation.