Share：
Share this content in WeChat
X
[Chinese] [PDF] 30 1
• Clinical Article •
Pre-treatment MRI radiomics and deep learning predict early progression of operable rectal cancer
QUAN Tao  LIU Yuanqing  WEI Mingxiang  CHEN Shuangqing 

Cite this article as QUAN T, LIU Y Q, WEI M X, et al. Pre-treatment MRI radiomics and deep learning predict early progression of operable rectal cancer[J]. Chin J Magn Reson Imaging, 2026, 17(6): 95-102. DOI:10.12015/issn.1674-8034.2026.06.012.


[Abstract] Objective The value of MRI-based radiomics and deep learning (DL) in predicting early progression (EP) after treatment for patients with operable rectal cancer: A pre-treatment observation.Materials and Methods Retrospectively included 255 rectal cancer patients from center 1 and 69 rectal cancer patients from center 2. Patients from center 1 were divided into a training set (n = 204) and an internal test set (n = 51) at a ratio of 8∶2. Patients from center 2 served as the external validation cohort (n = 69). Follow-up was conducted to record EP, defined as tumor recurrence or metastasis within two years post-treatment. Among all patients, 81 experienced EP, while 243 did not. From MRI images, features were extracted and selected to construct radiomics and DL-radiomics (DLR) models. The predictive performance of the models was evaluated and compared by plotting receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC). Additionally, decision curve analysis (DCA) was employed to assess the clinical utility and discriminative ability of the models.Results In this study, 6 radiomics features and 15 DLR features were selected. A k-nearest neighbor (KNN) classifier was used to construct the radiomics model, while a support vector machine (SVM) classifier was used to construct the DLR model. The DLR model (AUC = 0.866, 95% CI: 0.815 to 0.917) significantly outperformed the radiomics model (AUC = 0.724, 95% CI: 0.656 to 0.793). DCA showed that the DLR model provided higher clinical net benefit than the radiomics model. In the external validation set, the DeLong test revealed an extremely statistically significant difference in predictive performance between the two models (P < 0.001).Conclusions The DLR model based on pre-treatment MRI images shows better efficacy in predicting EP after treatment in patients with operable rectal cancer, and has good clinical application value, which can be used as an effective auxiliary tool for clinical prediction of patient prognosis and the formulation of individualized treatment strategies.
[Keywords] rectal cancer；early progression；magnetic resonance imaging；radiomics；deep learning；predictive model

QUAN Tao1, 2   LIU Yuanqing2   WEI Mingxiang1   CHEN Shuangqing1*  

1 Department of Radiology, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou 215001, China

2 Department of Radiology, The First Affiliated Hospital of Soochow University, Suzhou 215006, China

Corresponding author: CHEN S Q, E-mail: sznaonao@163.com

Conflicts of interest   None.

Received  2026-03-12
Accepted  2026-05-14
DOI: 10.12015/issn.1674-8034.2026.06.012
Cite this article as QUAN T, LIU Y Q, WEI M X, et al. Pre-treatment MRI radiomics and deep learning predict early progression of operable rectal cancer[J]. Chin J Magn Reson Imaging, 2026, 17(6): 95-102. DOI:10.12015/issn.1674-8034.2026.06.012.

[1]
SUNG H, FERLAY J, SIEGEL R L, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2021, 71(3): 209-249. DOI: 10.3322/caac.21660.
[2]
XIA C F, DONG X S, LI H, et al. Cancer statistics in China and United States, 2022: profiles, trends, and determinants[J]. Chin Med J, 2022, 135(5): 584-590. DOI: 10.1097/CM9.0000000000002108.
[3]
ZHANG Y C, YANG Z H, FENG Y Z, et al. Development and validation of nomogram models incorporating the inflammatory nutritional index CALLY for predicting survival in locally advanced rectal cancer after neoadjuvant chemoradiotherapy[J]. Cancer Manag Res, 2025, 17: 2961-2975. DOI: 10.2147/CMAR.S555346.
[4]
JIN T, ZHU Y S, LIU C C, et al. Epidemiological characteristics of early-onset colorectal cancer: a prospective cohort study from a single center[J]. Chin J Gastrointest Surg, 2024, 27(5): 457-463. DOI: 10.3760/cma.j.cn441530-20240222-00069.
[5]
LI M, ZHU Y Z, ZHANG Y C, et al. Radiomics of rectal cancer for predicting distant metastasis and overall survival[J]. World J Gastroenterol, 2020, 26(33): 5008-5021. DOI: 10.3748/wjg.v26.i33.5008.
[6]
YOSHINO T, CERVANTES A, BANDO H, et al. Pan-Asian adapted ESMO Clinical Practice Guidelines for the diagnosis, treatment and follow-up of patients with metastatic colorectal cancer[J/OL]. ESMO Open, 2023, 8(3): 101558 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/37236086/. DOI: 10.1016/j.esmoop.2023.101558.
[7]
RYU H S, LEE J L, KIM C W, et al. Correlative significance of tumor regression grade and ypT category in patients undergoing preoperative chemoradiotherapy for locally advanced rectal cancer[J]. Clin Colorectal Cancer, 2022, 21(3): 212-219. DOI: 10.1016/j.clcc.2022.02.001.
[8]
LIU Z Y, LIANG C H. A critical analysis of the development, challenges, and future prospects of artificial intelligence in medical imaging[J]. Chin J Radiol, 2024, 58(11): 1365-1370. DOI: 10.3760/cma.j.cn112149-20240726-00440.
[9]
SHARMA P, HASSAN C. Artificial intelligence and deep learning for upper gastrointestinal neoplasia[J]. Gastroenterology, 2022, 162(4): 1056-1066. DOI: 10.1053/j.gastro.2021.11.040.
[10]
FAN L F, WU H Z, WU Y M, et al. Preoperative prediction of rectal Cancer staging combining MRI deep transfer learning, radiomics features, and clinical factors: accurate differentiation from stage T2 to T3[J/OL]. BMC Gastroenterol, 2024, 24(1): 247 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/39103772/. DOI: 10.1186/s12876-024-03316-6.
[11]
WU S J, YU Y M, FAN L F, et al. Radiomics based on deep learning to predict T2 and T3 staging of rectal cancer[J]. Chin J Magn Reson Imaging, 2023, 14(11): 84-89, 102. DOI: 10.12015/issn.1674-8034.2023.11.014.
[12]
SHANG X F, MENG L L, REN S H, et al. Predictive value of the radiomics model based on T2WI combined with ADC for tumour deposition in colorectal cancer[J]. Radiol Pract, 2025, 40(9): 1139-1146. DOI: 10.13609/j.cnki.1000-0313.2025.09.011.
[13]
WU B J, WANG L W, WANG Y, et al. Utilizing baseline multiregional MRI radiomics for prediction of tumor deposition and prognosis following neoadjuvant therapy in resectable rectal cancer[J]. Eur J Radiol, 2026, 196: 112676. DOI: 10.1016/j.ejrad.2026.112676.
[14]
QIN S Y, LIU K, CHEN Y Y, et al. Prediction of pathological response and lymph node metastasis after neoadjuvant therapy in rectal cancer through tumor and mesorectal MRI radiomic features[J/OL]. Sci Rep, 2024, 14(1): 21927 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/39304726/. DOI: 10.1038/s41598-024-72916-9.
[15]
LI X M, ZHOU Y F, WANG A Y, et al. Prediction of treatment response to neoadjuvant chemoradiotherapy in patients with locally advanced rectal cancer by interpretable model based on multiparametric MRI[J]. Chin J Magn Reson Imaging, 2026, 17(1): 59-68, 122. DOI: 10.12015/issn.1674-8034.2026.01.009.
[16]
SUN Y, LI F X, CHEN X M, et al. Interpretable machine learning model based on DCE-MRI habitat imaging radiomics for predicting lymph node metastasis in rectal cancer[J]. Chin J Magn Reson Imaging, 2026, 17(1): 69-78. DOI: 10.12015/issn.1674-8034.2026.01.010.
[17]
ZHUANG Z X, ZHANG Y, YANG X Y, et al. T2WI-based texture analysis predicts preoperative lymph node metastasis of rectal cancer[J]. Abdom Radiol (NY), 2024, 49(6): 2008-2016. DOI: 10.1007/s00261-024-04209-8.
[18]
MAO J W, YE W L, MA W L, et al. Prediction by a multiparametric magnetic resonance imaging-based radiomics signature model of disease-free survival in patients with rectal cancer treated by surgery[J/OL]. Front Oncol, 2024, 14: 1255438 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/38454930/. DOI: 10.3389/fonc.2024.1255438.
[19]
ZHU Y, WEI Y R, CHEN Z W, et al. Different radiomics annotation methods comparison in rectal cancer characterisation and prognosis prediction: a two-centre study[J/OL]. Insights Imaging, 2024, 15(1): 211 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/39186173/. DOI: 10.1186/s13244-024-01795-5.
[20]
LIU Z L, MENG R Q, MA Q, et al. Predicting recurrence in locally advanced rectal cancer using multitask deep learning and multimodal MRI[J/OL]. Radiol Imaging Cancer, 2025, 7(3): e240359 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/39186173/. DOI: 10.1148/rycan.240359.
[21]
Medical Imaging Big Data and Artificial Intelligence Working Committee of Chinese Society of Radiology, Medical Imaging and Artificial Intelligence Committee of Chinese Research Hospital Society, Clinical Precision Medicine Committee of Chinese Medical Doctor Association, et al. Expert consensus on methodology of radiomics multimodal MRI quantitative analysis for efficacy and prognosis evaluation after neoadjuvant therapy in rectal cancer[J]. Chin J Radiol, 2022, 56 (6):608-615. DOI: 10.3760/cma.j.cn112149-20211220-01122.
[22]
LIU H H, ZHANG C Y, WANG L J, et al. MRI radiomics analysis for predicting preoperative synchronous distant metastasis in patients with rectal cancer[J]. Eur Radiol, 2019, 29(8): 4418-4426. DOI: 10.1007/s00330-018-5802-7.
[23]
LI H, CHEN X L, LIU H, et al. MRI-based multiregional radiomics for predicting lymph nodes status and prognosis in patients with resectable rectal cancer[J/OL]. Front Oncol, 2022, 12: 1087882 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/36686763/. DOI: 10.3389/fonc.2022.1087882.
[24]
HU S X, YANG K, WANG X R, et al. Application of MRI-based radiomics models in the assessment of hepatic metastasis of rectal cancer[J]. J Sichuan Univ Med Sci, 2021, 52(2): 311-318. DOI: 10.12182/20210360202.
[25]
LI Z H, QIN Y Y, LIAO X Q, et al. Comparison of clinical, radiomics, deep learning, and fusion models for predicting early recurrence in locally advanced rectal cancer based on multiparametric MRI: a multicenter study[J/OL]. Eur J Radiol, 2025, 189: 112173 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/40403678/. DOI: 10.1016/j.ejrad.2025.112173.
[26]
AO W Q, WANG N, CHEN X, et al. Multiparametric MRI-based deep learning models for preoperative prediction of tumor deposits in rectal cancer and prognostic outcome[J]. Acad Radiol, 2025, 32(3): 1451-1464. DOI: 10.1016/j.acra.2024.10.004.
[27]
JIANG X F, ZHAO H Y, SALDANHA O L, et al. An MRI deep learning model predicts outcome in rectal cancer[J/OL]. Radiology, 2023, 307(5): e222223 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/37278629/. DOI: 10.1148/radiol.222223.
[28]
YANG Y J, HAN K T, XU Z Y, et al. Development and validation of multiparametric MRI-based interpretable deep learning radiomics fusion model for predicting lymph node metastasis and prognosis in rectal cancer: a two-center study[J]. Acad Radiol, 2025, 32(5): 2642-2654. DOI: 10.1016/j.acra.2024.11.045.
[29]
LI X, ZHU Y, WEI Y R, et al. Development of anMRI-based comprehensive model fusing clinical, habitat radiomics, and deep learning models for preoperative identification of tumor deposits in rectal cancer[J]. Magnetic Resonance Imaging, 2025, 62(6): 1812-1823. DOI: 10.1002/jmri.70075.
[30]
LI M, JIN Y M, ZHANG Y C, et al. Radiomics for predicting perineural invasion status in rectal cancer[J]. World J Gastroenterol, 2021, 27(33): 5610-5621. DOI: 10.3748/wjg.v27.i33.5610.
[31]
LI J, ZHOU Y, WANG X X, et al. An MRI-based multi-objective radiomics model predicts lymph node status in patients with rectal cancer[J]. Abdom Radiol (NY), 2021, 46(5): 1816-1824. DOI: 10.1007/s00261-020-02863-2.
[32]
DOU Y, LIU Y Q, LI Y J. Prediction of lower-grade glioma IDH-1 mutation status using a combined model of radiomics and transformer deep learning features based on multi-parametric MRI of intratumoral and peritumoral edema[J]. Chin J Magn Reson Imaging, 2025, 16(9): 46-52, 59. DOI: 10.12015/issn.1674-8034.2025.09.008.
[33]
ZHANG Y C, LIAO Q C, DING L, et al. Bridging 2D and 3D segmentation networks for computation-efficient volumetric medical image segmentation: an empirical study of 2.5D solutions[J/OL]. Comput Med Imaging Graph, 2022, 99: 102088 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/35780703/. DOI: 10.1016/j.compmedimag.2022.102088.
[34]
ZHU J H, GAO Y, WU Y B, et al. Application of deep learning on MRI for prognostic prediction in rectal cancer[J/OL]. Eur Radiol, 2026 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/41537781/. DOI: 10.1007/s00330-025-12246-0.
[35]
PU H J, XIE P Y, CHEN Y X, et al. Relationship between preoperative and postoperative serum carcinoembryonic antigen and prognosis of patients with stage I-III rectal cancer: a retrospective study of a multicentre cohort of 1022 rectal cancer patients[J]. Cancer Manag Res, 2021, 13: 2643-2651. DOI: 10.2147/CMAR.S290416.
[36]
GELARDI F, CAVINATO L, DE SANCTIS R, et al. The predictive role of radiomics in breast cancer patients imaged by [18F] FDG PET: preliminary results from a prospective cohort[J/OL]. Diagnostics (Basel), 2024, 14(20): 2312 [2026-03-11]. https://pubmed.ncbi.nlm.nih.gov/39451637/. DOI: 10.3390/diagnostics14202312.
[37]
CAO X Y, LI R, WANG W Q, et al. Predictive value of multimodal MRI histology for mediastinal lymph node metastasis in non-small cell lung cancer[J]. Chin J Magn Reson Imaging, 2024, 15(4): 72-77. DOI: 10.12015/issn.1674-8034.2024.04.012.
[38]
LI H, LUO Y, WANG T F, et al. Predictive factors of pathological complete response after neoadjuvant therapy for locally advanced rectal cancer[J]. J Surg Concepts Pract, 2025, 30(1): 47-53. DOI: 10.16139/j.1007-9610.2025.01.09.

PREV Nomogram model based on MRI and clinical features to early predict the surgical timing of laparoscopic cholecystectomy in non-severe acute biliary pancreatitis
NEXT Value of intratumoral and peritumoral DCE-MRI combined with mDixon-Quant parameters in predicting lymphovascular invasion of rectal cancer
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn