2026-10-06T19:50:48Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34048952 2026-10-01T17:40:33Z category_7 category_12 category_21 category_734 category_811 category_64 portal_5 item_type_1 month_year_10_2026

<p>Heatmap of inter-reader agreement between the AI-mrEMVI model and radiologists in mrEMVI assessment.</p> Haitao Huang (182705) Lili Feng (154280) Min-Er Zhong (6758177) Huifen Ye (11271543) Zhenhui Li (282794) Su Yao (3627038) Yunrui Ye (13909343) Yulin Liu (457083) Minning Zhao (13909337) Weixiong Xu (25157608) Lifen Yan (1404280) ChuanMiao Xie (25157611) Changhong Liang (215520) Zaiyi Liu (611024) Tong Tong (532210) Yanfen Cui (2146873) Xin-Juan Fan (329948) Ke Zhao (248186) Medicine Cell Biology Biotechnology Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Cancer sup >+</ sup observer variability inherent multivariable cox regression magnetic resonance imaging evaluated using kaplan dice similarity coefficient 9 %, hr evaluate segmentation performance dilated vessel segmentation observer agreement analysis meier survival analysis important prognostic biomarker achieved substantial agreement significantly lower 3 75 – 3 reflecting tumor invasiveness intravascular tumor signal year overall survival div >< p 7 %– 88 7 %– 85 mremvi assessment based manual mremvi assessment based segmentation model 1 %, hr mremvi status identified comprehensive prognostic validation 713 – 0 95 – 3 level segmentation reader agreement prognostic value free survival year disease work establishes vs .</ treatment individualization training cohort study trained senior radiologists risk stratification rectal cancer patients identified p </ objective framework metastatic potential level localization internal five fold cross <p>The color intensity in the heatmap represents the pairwise Cohen’s kappa values between each pair of readers. R1 and R2 denote junior radiologists (H.H. and Y.Y.); R3 and R4, mid-level radiologists (H.Y. and Y.L.); and R5 and R6, senior radiologists (L.Y. and W.X.). R, reader. AI, artificial intelligence; mr, magnetic resonance; EMVI, extramural vascular invasion.</p> 2026-10-01T17:40:25Z Image Figure 10.1371/journal.pdig.0001763.g004 https://figshare.com/articles/figure/_p_Heatmap_of_inter-reader_agreement_between_the_AI-mrEMVI_model_and_radiologists_in_mrEMVI_assessment_p_/34048952 CC BY 4.0