Multiparametric MRI-based radiomics signature for preoperative estimation of tumor-stroma ratio in rectal cancer
- PDF / 3,268,665 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 102 Downloads / 177 Views
GASTROINTESTINAL
Multiparametric MRI-based radiomics signature for preoperative estimation of tumor-stroma ratio in rectal cancer Chongpeng Cai 1,2 & Tingdan Hu 1,2 & Jing Gong 1,2 & Dan Huang 3 & Fangqi Liu 4 & Caixia Fu 5 & Tong Tong 1,2 Received: 21 May 2020 / Revised: 26 September 2020 / Accepted: 9 October 2020 # European Society of Radiology 2020
Abstract Objective To determine whether a radiomics signature (rad-score) outperforms ADC in TSR estimation by developing a radiomics biomarker for preoperative TSR diagnosis in rectal cancer. Methods This study included 149 patients (119 and 30 in the training and validation cohorts, respectively). All patients underwent T2-weighted, diffusion-weighted, and contrast-enhanced T1-weighted imaging. A rad-score was generated using the least absolute shrinkage and selection operator (LASSO) algorithm and stepwise multivariate logistic regression. Meanwhile, the mean ADCs were calculated from ADC maps. For both the mean ADC and rad-score, binary logistic regression and Spearman correlation coefficients were used to determine associations with the TSR, and the area under the receiver operating characteristic (ROC) curve was used to assess the diagnostic performance. The reliability of the rad-score was quantified by comparing the imaging-estimated TSR with the actual TSR of each patient. Results Both the mean ADC and rad-score were positively correlated with the TSR in the training cohort (mean ADC: p < 0.001, r = 0.566; rad-score: p < 0.001, r = 0.559) and validation cohort (mean ADC: p < 0.001, r = 0.671; rad-score: p = 0.002, r = 0.536). The rad-score, with AUCs of 0.917 (95% CI 0.869–0.965) and 0.787 (95% CI 0.602–0.972) in the training and validation cohorts, respectively, outperformed the mean ADC (training cohort: AUC = 0.776, 95% CI 0.693–0.859; validation cohort: AUC = 0.764, 95% CI 0.592–0.936) in TSR estimation. Conclusion The ADC possesses potential diagnostic value for TSR estimation in rectal cancer, and the rad-score shows increased diagnostic value over the ADC and may be a promising supplemental tool for patient stratification and informing decisionmaking. Key Points • Tumor-stroma ratio has been verified as an independent prognostic factor for various solid tumors including rectal cancer. • The ADC and multiparametric MRI-based radiomics features were significantly and positively correlated with the tumorstroma ratio in rectal cancer. • The radiomics signature outperformed the ADC in discriminating TSR in rectal cancer. Keywords Rectal neoplasms . Magnetic resonance imaging . Tumor microenvironment . Apparent diffusion coefficient
Chongpeng Cai and Tingdan Hu contributed equally to this work. Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00330-020-07403-6) contains supplementary material, which is available to authorized users. * Tong Tong [email protected] 1
Department of Radiology, Fudan University Shanghai Cancer Center, Shanghai, China
2
Department of Oncology, Shanghai Medical College, Fudan Univer
Data Loading...
