Share：
Share this content in WeChat
X
[Chinese] [PDF] 71 1
• Review •
Research Progress of MRI Combined with Artificial Intelligence in Preoperative Prediction of Microscopic High-Risk Pathological Factors in Rectal Cancer
TANG Kunpeng  WANG Xiaoshan  ZHAO Jiayi  LIU Zhihui  PENG Zhongsong  LI Feng 

DOI:10.12015/issn.1674-8034.2026.05.032.


[Abstract] Rectal cancer is a common and highly prevalent malignant tumor of the digestive tract. Microscopic high-risk pathological features, including perineural invasion (PNI), lymphovascular invasion (LVI), and tumor budding (TB), are closely associated with tumor aggressiveness, patient survival prognosis, and individualized therapeutic decision-making. Conventional magnetic resonance imaging (MRI) relies on morphological assessment and has obvious limitations in identifying microscopic invasive behaviors at the submillimeter scale. Artificial intelligence approaches including radiomics, habitat imaging and deep learning have significantly improved the accuracy and stability of preoperative prediction of microscopic high-risk pathological features in rectal cancer by virtue of feature mining and modeling analysis based on MRI images. This review systematically summarizes the research progress of MRI combined with artificial intelligence in the preoperative prediction of PNI, LVI and TB in rectal cancer, sorts out the model construction, diagnostic efficiency and core bottlenecks of radiomics, habitat imaging and deep learning, analyzes the imaging-pathological correlation mechanism, clarifies the application value of new technologies such as habitat imaging and weakly-supervised learning, and prospects the future research directions such as model standardization, multicenter verification and clinical translation, aiming to provide a systematic reference for the precise imaging evaluation and individualized treatment of microscopic high-risk pathological factors in rectal cancer.
[Keywords] rectal cancer；perineural invasion；lymphovascular invasion；tumor budding；magnetic resonance imaging；artificial intelligence

TANG Kunpeng1   WANG Xiaoshan2   ZHAO Jiayi2   LIU Zhihui1   PENG Zhongsong2   LI Feng1*  

1 Department of Radiology and Imaging, Xiangyang Central Hospital Affiliated to Hubei University of Arts and Sciences, Xiangyang Central Hospital, Xiangyang 441021, China

2 School of Medicine, Wuhan University of Science and Technology, Wuhan 430081, China

Corresponding author: LI F, E-mail: xfkite@163.com

Conflicts of interest   None.

Received  2026-02-01
Accepted  2026-04-16
DOI: 10.12015/issn.1674-8034.2026.05.032
DOI:10.12015/issn.1674-8034.2026.05.032.

[1]
BRAY F, LAVERSANNE M, SUNG H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2024, 74(3): 229-263. DOI: 10.3322/caac.21834.
[2]
DE FELICE F, MICCINI M, BOTTICELLI A, et al. The multidisciplinary management of locally advanced rectal cancer[J]. Expert Rev Anticancer Ther, 2024, 24(7): 581-587. DOI: 10.1080/14737140.2024.2349137.
[3]
XIE P Y, ZENG Z M, LI Z H, et al. MRI-based radiomics for stratifying recurrence risk of early-onset rectal cancer: a multicenter study[J/OL]. ESMO Open, 2024, 9(10): 103735 [2026-04-06]. https://linkinghub.elsevier.com/retrieve/pii/S2059-7029(24)01505-9. DOI: 10.1016/j.esmoop.2024.103735.
[4]
SUN Z G, CHEN S X, SUN B L, et al. Important role of lymphovascular and perineural invasion in prognosis of colorectal cancer patients with N1c disease[J/OL]. World J Gastroenterol, 2025, 31(5): 102210 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC11718613/. DOI: 10.3748/wjg.v31.i5.102210.
[5]
RAFIEE A, NASRI P, MORADI A, et al. Tumor budding as an indicator of prognosis in locally advanced rectal cancer after neoadjuvant chemoradiotherapy: a systematic review and meta-analysis[J/OL]. Front Oncol, 2025, 15: 1429319 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC12014445/. DOI: 10.3389/fonc.2025.1429319.
[6]
VYAS A, KUMAR K, SHARMA A, et al. Advancing the frontier of artificial intelligence on emerging technologies to redefine cancer diagnosis and care[J/OL]. Comput Biol Med, 2025, 191: 110178 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/40228444/. DOI: 10.1016/j.compbiomed.2025.110178.
[7]
GONG T T, GAO Y, LI H, et al. Research progress in multimodal radiomics of rectal cancer tumors and peritumoral regions in MRI[J]. Abdom Radiol, 2025, 50(12): 5677-5689. DOI: 10.1007/s00261-025-04965-1.
[8]
WANG H, HUO R X, HE K X, et al. Perineural invasion in colorectal cancer: mechanisms of action and clinical relevance[J]. Cell Oncol (Dordr), 2024, 47(1): 1-17. DOI: 10.1007/s13402-023-00857-y.
[9]
ZHANG C Y, LU T, ZHANG H Y, et al. Relationship between apparent diffusion coefficient values and clinicopathologic features in rectal cancer: a cross-sectional study[J]. J Gastrointest Oncol, 2025, 16(2): 528-541. DOI: 10.21037/jgo-24-831.
[10]
ZHANG L J, DENG Y X, LIU S R, et al. Lymphovascular invasion represents a superior prognostic and predictive pathological factor of the duration of adjuvant chemotherapy for stage III colon cancer patients[J/OL]. BMC Cancer, 2023, 23(1): 3 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC9808960/. DOI: 10.1186/s12885-022-10416-7.
[11]
BAXTER N N, KENNEDY E B, BERGSLAND E, et al. Adjuvant therapy for stage II colon cancer: ASCO guideline update[J]. J Clin Oncol, 2022, 40(8): 892-910. DOI: 10.1200/JCO.21.02538.
[12]
LI H, CHEN G W, LIU Y S, et al. Assessment of histologic prognostic factors of resectable rectal cancer: comparison of diagnostic performance using various apparent diffusion coefficient parameters[J/OL]. Sci Rep, 2020, 10: 11554 [2026-04-06].https://pmc.ncbi.nlm.nih.gov/articles/PMC7360736/. DOI: 10.1038/s41598-020-68328-0.
[13]
Colorectal Cancer Working Group of the Chinese Society of Clinical Oncology. Guidelines of Chinese society of clinical oncology (CSCO) colorectal cancer[M]. Beijing: People's Medical Publishing House, 2025.
[14]
ZLOBEC I, BÄCHLI M, GALUPPINI F, et al. Refining the ITBCC tumor budding scoring system with a "zero-budding" category in colorectal cancer[J]. Virchows Arch, 2021, 479(6): 1085-1090. DOI: 10.1007/s00428-021-03090-w.
[15]
ZHU L P, ZHU Y M, DONG Y Y, et al. Application value of magnetic resonance IVIM in preoperative evaluation of tumor budding of rectal cancer[J]. Chin J CT MRI, 2025, 23(2): 175-177. DOI: 10.3969/j.issn.1672-5131.2025.02.054.
[16]
CHEN A L, XIE S L, WANG Y, et al. The value of quantitative parameters of diffusion kurtosis imaging in preoperative prediction of tumor budding grade of rectal cancer[J]. Chin J Magn Reson Imaging, 2025, 16(2): 59-64, 99. DOI: 10.12015/issn.1674-8034.2025.02.009.
[17]
ZHANG X Y, LIU N J, ZHANG Y M, et al. Research progress on multimodal MRI and radiomics for assessing tumor budding in rectal cancer[J]. Chin J Magn Reson Imaging, 2025, 16(8): 221-227. DOI: 10.12015/issn.1674-8034.2025.08.033.
[18]
WANG Y F, XUE W C. Correct understanding of pathological evaluation and prognostic value of the invasion depth of early colorectal cancer (pT1)[J]. Chin J Pathol, 2025, 54(11): 1117-1123. DOI: 10.3760/cma.j.cn112151-20250723-00500.
[19]
WENG J Y, YE Z L, ZHANG R X, et al. Exploring the guiding role of the number of adverse pathological features in risk stratification for recurrence of stage Ⅰ-Ⅲ colorectal cancer: a retrospective cohort study of 9875 cases[J]. China Oncol, 2024, 34(6): 527-536. DOI: 10.19401/j.cnki.1007-3639.2024.06.001.
[20]
SCOTT A J, KENNEDY E B, BERLIN J, et al. Management of locally advanced rectal cancer: ASCO guideline[J]. J Clin Oncol, 2024, 42(28): 3355-3375. DOI: 10.1200/jco.24.01160.
[21]
TOMASZEWSKI M R, GILLIES R J. The biological meaning of radiomic features[J/OL]. Radiology, 2021, 299(2): E256 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC7924519/. DOI: 10.1148/radiol.2021219005.
[22]
YANG Y S, QIU Y J, ZHENG G H, et al. High resolution MRI-based radiomic nomogram in predicting perineural invasion in rectal cancer[J/OL]. Cancer Imaging, 2021, 21(1): 40 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC8157664/. DOI: 10.1186/s40644-021-00408-4.
[23]
LIU Q X, WU S J, YUAN Q, et al. Explainable machine learning model based on DKI, IVIM, and clinical features for preoperative prediction of lymphovascular invasion in rectal cancer[J]. Chin J Magn Reson Imaging, 2025, 16(11): 129-134, 141. DOI: 10.12015/issn.1674-8034.2025.11.019.
[24]
ZHANG Y, PENG J X, LIU J, et al. Preoperative prediction of perineural invasion status of rectal cancer based on radiomics nomogram of multiparametric magnetic resonance imaging[J/OL]. Front Oncol, 2022, 12: 828904 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC9036372/. DOI: 10.3389/fonc.2022.828904.
[25]
BOEHM K M, KHOSRAVI P, VANGURI R, et al. Harnessing multimodal data integration to advance precision oncology[J]. Nat Rev Cancer, 2022, 22(2): 114-126. DOI: 10.1038/s41568-021-00408-3.
[26]
WU J J, XIA Y W, WANG X C, et al. Radiomics++: review of habitat imaging analysis for decoding tumor heterogeneity[J/OL]. Annu Rev Biomed Eng, 2026 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/41591802/. DOI: 10.1146/annurev-bioeng-031825-040442.
[27]
YANG X L, NIU W J, WU K, et al. Diffusion kurtosis imaging-based habitat analysis identifies high-risk molecular subtypes and heterogeneity matching in diffuse gliomas[J]. Ann Clin Transl Neurol, 2024, 11(8): 2073-2087. DOI: 10.1002/acn3.52128.
[28]
WU M S, QUE Z L, LAI S J, et al. Predicting the early therapeutic response to hepatic artery infusion chemotherapy in patients with unresectable HCC using a contrast-enhanced computed tomography-based habitat radiomics model: a multi-center retrospective study[J]. Cell Oncol (Dordr), 2025, 48(3): 709-723. DOI: 10.1007/s13402-025-01041-0.
[29]
ZHU Z C, LI Y J, WANG P, et al. Machine learning-based quantitative prediction of spread through air spaces in primary lung adenocarcinoma using intratumoural heterogeneity scores[J/OL]. Int J Surg, 2025 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/41376342/. DOI: 10.1097/JS9.0000000000004447.
[30]
FU Y, MA C Y, ZHANG L, et al. Research progress of habitat analysis in radiomics of malignant tumors[J]. J Int Oncol, 2024, 51(5): 292-297. DOI: 10.3760/cma.j.cn371439-20240108-00049.
[31]
ZHONG J Y, HUANG T, JIANG R J, et al. MRI-based habitat, intra-, and peritumoral machine learning model for perineural invasion prediction in rectal cancer[J]. Abdom Radiol, 2026, 51(2): 545-557. DOI: 10.1007/s00261-025-05095-4.
[32]
PENG L P, LI F X, ZHANG F, et al. An interpretable machine learning model combining MRI-DKI habitat radiomic features and clinical biomarkers for noninvasive prediction of lymphatic metastasis in rectal cancer: a prospective study[J/OL]. Insights Imaging, 2026, 17(1): 79 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC13018487/. DOI: 10.1186/s13244-026-02243-2.
[33]
SU Y X, ZHAO H Y, LYU Z H, et al. Quantification of intratumoral heterogeneity based on habitat analysis for preoperative assessment of lymphovascular invasion in colorectal cancer[J]. Acad Radiol, 2025, 32(8): 4532-4543. DOI: 10.1016/j.acra.2025.03.014.
[34]
CHEN Q L, ZHANG Q W, LI Z H, et al. MRI-based habitat analysis for pathologic response prediction after neoadjuvant chemoradiotherapy in rectal cancer: a multicenter study[J]. Eur Radiol, 2026, 36(3): 1671-1685. DOI: 10.1007/s00330-025-11997-0.
[35]
YANG Z T, WU H, GAO H Y, et al. Progress in the application of habitat imaging in multi-system tumors[J]. Chin J Magn Reson Imaging, 2025, 16(3): 222-227. DOI: 10.12015/issn.1674-8034.2025.03.038.
[36]
AVANZO M, WEI L S, STANCANELLO J, et al. Machine and deep learning methods for radiomics[J/OL]. Med Phys, 2020, 47(5) [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC8965689/. DOI: 10.1002/mp.13678.
[37]
DEMIRCIOĞLU A. Are deep models in radiomics performing better than generic models A systematic review[J/OL]. Eur Radiol Exp, 2023, 7(1): 11 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC8965689/. DOI: 10.1186/s41747-023-00325-0.
[38]
SHI S M, XIAO L Q, MA J Q, et al. Progress in deep learning based on magnetic resonance imaging for rectal cancer research[J]. Chin J Magn Reson Imaging, 2024, 15(3): 218-222. DOI: 10.12015/issn.1674-8034.2024.03.036.
[39]
WAN L J, HU J S, CHEN S, et al. Prediction of lymph node metastasis in stage T1-2 rectal cancers with MRI-based deep learning[J]. Eur Radiol, 2023, 33(5): 3638-3646. DOI: 10.1007/s00330-023-09450-1.
[40]
WANG T J, CHEN C Y, LIU C, et al. A 3D deep learning model based on MRI for predicting lymphovascular invasion in rectal cancer[J/OL]. Med Phys, 2025, 52(7): e17882 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/40391614/. DOI: 10.1002/mp.17882.
[41]
PENG L, WANG D Q, ZHUANG Z J, et al. Research progress of radiomics and deep learning in rectal cancer[J]. Chin J CT MRI, 2024, 22(5): 177-180. DOI: 10.3969/j.issn.1672-5131.2024.05.056.
[42]
ZHANG H T, YANG X T, LI D A, et al. Dual parallel net: a novel deep learning model for rectal tumor segmentation via CNN and transformer with Gaussian Mixture prior[J]. J Biomed Inform, 2023, 139: 104304 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/36736447/. DOI: 10.1016/j.jbi.2023.104304.
[43]
LIU Z H, YANG H, NIE L, et al. Prediction of tumor budding grading in rectal cancer using a multiparametric MRI radiomics combined with a 3D vision transformer deep learning approach[J]. Acad Radiol, 2025, 32(8): 4512-4523. DOI: 10.1016/j.acra.2025.03.046.
[44]
JIA J Y, KANG Y, WANG J H, et al. Attention mechanism-based multi-parametric MRI ensemble model for predicting tumor budding grade in rectal cancer patients[J]. Abdom Radiol, 2025, 50(10): 4483-4494. DOI: 10.1007/s00261-025-04886-z.
[45]
LIU Y W, TANG L P, LIAO C, et al. Optimized dropkey-based grad-CAM: toward accurate image feature localization[J/OL]. Sensors, 2023, 23(20): 8351 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC10611172/. DOI: 10.3390/s23208351.
[46]
TEOH J R, DONG J, ZUO X W, et al. Advancing healthcare through multimodal data fusion: a comprehensive review of techniques and applications[J/OL]. PeerJ Comput Sci, 2024, 10: e2298 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC11623190/. DOI: 10.7717/peerj-cs.2298.
[47]
LIU Z Y, JIA J Y, BAI F, et al. Predicting rectal cancer tumor budding grading based on MRI and CT with multimodal deep transfer learning: a dual-center study[J/OL]. Heliyon, 2024, 10(7): e28769 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC11000007/. DOI: 10.1016/j.heliyon.2024.e28769.
[48]
WANG H R, GUO X Y, SONG K W, et al. GO-MAE: Self-supervised pre-training via masked autoencoder for OCT image classification of gynecology[J/OL]. Neural Netw, 2025, 181: 106817 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/39500244/. DOI: 10.1016/j.neunet.2024.106817.
[49]
GOCERI E. Medical image data augmentation: techniques, comparisons and interpretations[J]. Artif Intell Rev, 2023: 1-45. DOI: 10.1007/s10462-023-10453-z.
[50]
YEUNG M, SALA E, SCHÖNLIEB C B, et al. Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation[J]. Comput Med Imaging Graph, 2022, 95: 102026 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC8785124/. DOI: 10.1016/j.compmedimag.2021.102026.
[51]
OWUSU-ADJEI M, HAYFRON-ACQUAH J BEN, FRIMPONG T, et al. Imbalanced class distribution and performance evaluation metrics: a systematic review of prediction accuracy for determining model performance in healthcare systems[J]. PLoS Digit Health, 2023, 2(11): 1-19. DOI: 10.1371/journal.pdig.0000290.
[52]
BRADSHAW T J, HUEMANN Z, HU J J, et al. A guide to cross-validation for artificial intelligence in medical imaging[J/OL]. Radiol Artif Intell, 2023, 5(4): e220232 [2026-04-06]. https://pmc.ncbi.nlm.nih.gov/articles/PMC10388213/. DOI: 10.1148/ryai.220232.
[53]
HOU M, ZHOU L, SUN J H. Deep-learning-based 3D super-resolution MRI radiomics model: superior predictive performance in preoperative T-staging of rectal cancer[J]. Eur Radiol, 2023, 33(1): 1-10. DOI: 10.1007/s00330-022-08952-8.
[54]
LI H, CHEN X L, LIU H, et al. MRI-based multiregional radiomics for preoperative prediction of tumor deposit and prognosis in resectable rectal cancer: a bicenter study[J]. Eur Radiol, 2023, 33(11): 7561-7572. DOI: 10.1007/s00330-023-09723-9.
[55]
MA Q X, KALADJI A, SHU H Z, et al. Beyond strong labels: Weakly-supervised learning based on Gaussian pseudo labels for the segmentation of ellipse-like vascular structures in non-contrast CTs[J/OL]. Med Image Anal, 2025, 99: 103378 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/39500029/. DOI: 10.1016/j.media.2024.103378.
[56]
GHASSEMI M, OAKDEN-RAYNER L, BEAM A L. The false hope of current approaches to explainable artificial intelligence in health care[J/OL]. Lancet Digit Health, 2021, 3(11): e745-e750 [2026-04-06]. https://pubmed.ncbi.nlm.nih.gov/34711379/. DOI: 10.1016/S2589-7500(21)00208-9.

PREV Research advances in habitat analysis for the diagnosis and treatment of hepatocellular carcinoma
NEXT Research progress in multimodal MRI for infrapatellar fat pad injuries in knee osteoarthritis
  


LINK


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