Feasibility and effectiveness of automatic deep learning network and radiomics models for differentiating tumor stroma ratio in pancreatic ductal adenocarcinoma

被引:2
|
作者
Liao, Hongfan [1 ,2 ]
Yuan, Jiang [3 ]
Liu, Chunhua [4 ]
Zhang, Jiao [5 ]
Yang, Yaying [6 ]
Liang, Hongwei [2 ]
Jiang, Song [7 ]
Chen, Shanxiong [3 ]
Li, Yongmei [2 ]
Liu, Yanbing [1 ]
机构
[1] Chongqing Med Univ, Coll Med Informat, Chongqing 400016, Peoples R China
[2] Chongqing Med Univ, Dept Orthoped, Affiliated Hosp 1, Chongqing 400016, Peoples R China
[3] Southwest Univ, Coll Comp & Informat Sci, Chongqing 400715, Peoples R China
[4] Army Med Univ, Daping Hosp, Dept Radiol, Chongqing, Peoples R China
[5] Chongqing Med Univ, Dept Orthoped, Affiliated Hosp 3, Chongqing, Peoples R China
[6] Chongqing Med Univ, Mol Med & Canc Res Ctr, Dept Pathol, Chongqing 400016, Peoples R China
[7] Chongqing Ping Med Imaging Diag Ctr, Dept Radiol, Chongqing, Peoples R China
关键词
Machine learning; Radiomics; Deep learning; Tumor stroma ratio; Pancreatic ductal adenocarcinoma; CANCER; IMAGES;
D O I
10.1186/s13244-023-01553-z
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
ObjectiveThis study aims to compare the feasibility and effectiveness of automatic deep learning network and radiomics models in differentiating low tumor stroma ratio (TSR) from high TSR in pancreatic ductal adenocarcinoma (PDAC).MethodsA retrospective analysis was conducted on a total of 207 PDAC patients from three centers (training cohort: n = 160; test cohort: n = 47). TSR was assessed on hematoxylin and eosin-stained specimens by experienced pathologists and divided as low TSR and high TSR. Deep learning and radiomics models were developed including ShuffulNetV2, Xception, MobileNetV3, ResNet18, support vector machine (SVM), k-nearest neighbor (KNN), random forest (RF), and logistic regression (LR). Additionally, the clinical models were constructed through univariate and multivariate logistic regression. Kaplan-Meier survival analysis and log-rank tests were conducted to compare the overall survival time between different TSR groups.ResultsTo differentiate low TSR from high TSR, the deep learning models based on ShuffulNetV2, Xception, MobileNetV3, and ResNet18 achieved AUCs of 0.846, 0.924, 0.930, and 0.941, respectively, outperforming the radiomics models based on SVM, KNN, RF, and LR with AUCs of 0.739, 0.717, 0.763, and 0.756, respectively. Resnet 18 achieved the best predictive performance. The clinical model based on T stage alone performed worse than deep learning models and radiomics models. The survival analysis based on 142 of the 207 patients demonstrated that patients with low TSR had longer overall survival.ConclusionsDeep learning models demonstrate feasibility and superiority over radiomics in differentiating TSR in PDAC. The tumor stroma ratio in the PDAC microenvironment plays a significant role in determining prognosis.Critical relevance statementThe objective was to compare the feasibility and effectiveness of automatic deep learning networks and radiomics models in identifying the tumor-stroma ratio in pancreatic ductal adenocarcinoma. Our findings demonstrate deep learning models exhibited superior performance compared to traditional radiomics models.Key points center dot Deep learning demonstrates better performance than radiomics in differentiating tumor-stroma ratio in pancreatic ductal adenocarcinoma.center dot The tumor-stroma ratio in the pancreatic ductal adenocarcinoma microenvironment plays a protective role in prognosis.center dot Preoperative prediction of tumor-stroma ratio contributes to clinical decision-making and guiding precise medicine.Key points center dot Deep learning demonstrates better performance than radiomics in differentiating tumor-stroma ratio in pancreatic ductal adenocarcinoma.center dot The tumor-stroma ratio in the pancreatic ductal adenocarcinoma microenvironment plays a protective role in prognosis.center dot Preoperative prediction of tumor-stroma ratio contributes to clinical decision-making and guiding precise medicine.Key points center dot Deep learning demonstrates better performance than radiomics in differentiating tumor-stroma ratio in pancreatic ductal adenocarcinoma.center dot The tumor-stroma ratio in the pancreatic ductal adenocarcinoma microenvironment plays a protective role in prognosis.center dot Preoperative prediction of tumor-stroma ratio contributes to clinical decision-making and guiding precise medicine.
引用
收藏
页数:16
相关论文
共 29 条
  • [1] Feasibility and effectiveness of automatic deep learning network and radiomics models for differentiating tumor stroma ratio in pancreatic ductal adenocarcinoma
    Hongfan Liao
    Jiang Yuan
    Chunhua Liu
    Jiao Zhang
    Yaying Yang
    Hongwei Liang
    Song Jiang
    Shanxiong Chen
    Yongmei Li
    Yanbing Liu
    Insights into Imaging, 14
  • [2] CT Radiomics and Machine-Learning Models for Predicting Tumor-Stroma Ratio in Patients With Pancreatic Ductal Adenocarcinoma
    Meng, Yinghao
    Zhang, Hao
    Li, Qi
    Liu, Fang
    Fang, Xu
    Li, Jing
    Yu, Jieyu
    Feng, Xiaochen
    Zhu, Mengmeng
    Li, Na
    Jing, Guodong
    Wang, Li
    Ma, Chao
    Lu, Jianping
    Bian, Yun
    Shao, Chengwei
    FRONTIERS IN ONCOLOGY, 2021, 11
  • [3] Magnetic Resonance Radiomics and Machine-learning Models: An Approach for Evaluating Tumor-stroma Ratio in Patients with Pancreatic Ductal Adenocarcinoma
    Meng, Yinghao
    Zhang, Hao
    Li, Qi
    Liu, Fang
    Fang, Xu
    Li, Jing
    Yu, Jieyu
    Feng, Xiaochen
    Lu, Jianping
    Bian, Yun
    Shao, Chengwei
    ACADEMIC RADIOLOGY, 2022, 29 (04) : 523 - 535
  • [4] Tenascin C, Fibronectin, and Tumor-Stroma Ratio in Pancreatic Ductal Adenocarcinoma
    Leppaenen, Joni
    Lindholm, Ville
    Isohookana, Joel
    Haapasaari, Kirsi-Maria
    Karihtala, Peeter
    Lehenkari, Petri P.
    Saarnio, Juha
    Kauppila, Joonas H.
    Karttunen, Tuomo J.
    Helminen, Olli
    Huhta, Heikki
    PANCREAS, 2019, 48 (01) : 43 - 48
  • [5] The tumor–stroma ratio and the immune microenvironment improve the prognostic prediction of pancreatic ductal adenocarcinoma
    Mei Lu
    Yi Zou
    Peiling Fu
    Yuyang Li
    Pengcheng Wang
    Guoping Li
    Sheng Luo
    Yupeng Chen
    Guoping Guan
    Sheng Zhang
    Linying Chen
    Discover Oncology, 14
  • [6] The tumor-stroma ratio and the immune microenvironment improve the prognostic prediction of pancreatic ductal adenocarcinoma
    Lu, Mei
    Zou, Yi
    Fu, Peiling
    Li, Yuyang
    Wang, Pengcheng
    Li, Guoping
    Luo, Sheng
    Chen, Yupeng
    Guan, Guoping
    Zhang, Sheng
    Chen, Linying
    DISCOVER ONCOLOGY, 2023, 14 (01)
  • [7] One novel transfer learning-based CLIP model combined with self-attention mechanism for differentiating the tumor-stroma ratio in pancreatic ductal adenocarcinoma
    Liao, Hongfan
    Yuan, Jiang
    Liu, Chunhua
    Zhang, Jiao
    Yang, Yaying
    Liang, Hongwei
    Liu, Haotian
    Chen, Shanxiong
    Li, Yongmei
    RADIOLOGIA MEDICA, 2024, 129 (11): : 1559 - 1574
  • [8] Radiomics Combined with Multiple Machine Learning Algorithms in Differentiating Pancreatic Ductal Adenocarcinoma from Pancreatic Neuroendocrine Tumor: More Hands Produce a Stronger Flame
    Zhang, Tao
    Xiang, Yu
    Wang, Hang
    Yun, Hong
    Liu, Yichun
    Wang, Xing
    Zhang, Hao
    JOURNAL OF CLINICAL MEDICINE, 2022, 11 (22)
  • [9] Deep learning for analysis of tumor-lymphocyte interactions in pancreatic ductal adenocarcinoma
    Kobayashi, Soma
    Le, Han
    Chrastecka, Lucie
    Gupta, Rajarsi
    Hou, Le
    Abousamra, Shahira
    Fassler, Danielle
    Shroyer, Kenneth R.
    Samaras, Dimitris
    Kurc, Tahsin
    Moffitt, Richard A.
    Saltz, Joel H.
    CANCER RESEARCH, 2019, 79 (24)
  • [10] Fully Automatic Deep Learning Framework for Pancreatic Ductal Adenocarcinoma Detection on Computed Tomography
    Alves, Natalia
    Schuurmans, Megan
    Litjens, Geke
    Bosma, Joeran S.
    Hermans, John
    Huisman, Henkjan
    CANCERS, 2022, 14 (02)