A systematic review of prognosis predictive role of radiomics in pancreatic cancer: heterogeneity markers or statistical tricks?

Gao, Yuhan; Cheng, Sihang; Zhu, Liang; Wang, Qin; Deng, Wenyi; Sun, Hansong; Wang, Shitian; Xue, Huadan

doi:10.1007/s00330-022-08922-0

Cited by 21 publications

(14 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared with the OPC and H&N 1 datasets, we think the difference may be due to the differences in the difficulty of predicting outcomes in different cancer types. 75,76 Compared with LUNG 4, we believe there are two main factors that play a role: the fact that LUNG 4 is more homogeneous (it only includes patients with advanced cancer) and has a smaller sample size. 74 Results of the survival prediction experiments seem to indicate that radiomic features outperformed deep features.…”

Section: Discussionmentioning

confidence: 99%

Using 3D deep features from CT scans for cancer prognosis based on a video classification model: A multi‐dataset feasibility study

et al. 2023

View full text Add to dashboard Cite

Background Cancer prognosis before and after treatment is key for patient management and decision making. Handcrafted imaging biomarkers—radiomics—have shown potential in predicting prognosis. Purpose However, given the recent progress in deep learning, it is timely and relevant to pose the question: could deep learning based 3D imaging features be used as imaging biomarkers and outperform radiomics? Methods Effectiveness, reproducibility in test/retest, across modalities, and correlation of deep features with clinical features such as tumor volume and TNM staging were tested in this study. Radiomics was introduced as the reference image biomarker. For deep feature extraction, we transformed the CT scans into videos, and we adopted the pre‐trained Inflated 3D ConvNet (I3D) video classification network as the architecture. We used four datasets—LUNG 1 (n = 422), LUNG 4 (n = 106), OPC (n = 605), and H&N 1 (n = 89)—with 1270 samples from different centers and cancer types—lung and head and neck cancer—to test deep features’ predictiveness and two additional datasets to assess the reproducibility of deep features. Results Support Vector Machine–Recursive Feature Elimination (SVM–RFE) selected top 100 deep features achieved a concordance index (CI) of 0.67 in survival prediction in LUNG 1, 0.87 in LUNG 4, 0.76 in OPC, and 0.87 in H&N 1, while SVM‐RFE selected top 100 radiomics achieved CIs of 0.64, 0.77, 0.73, and 0.74, respectively, all statistically significant differences (p < 0.01, Wilcoxon's test). Most selected deep features are not correlated with tumor volume and TNM staging. However, full radiomics features show higher reproducibility than full deep features in a test/retest setting (0.89 vs. 0.62, concordance correlation coefficient). Conclusion The results show that deep features can outperform radiomics while providing different views for tumor prognosis compared to tumor volume and TNM staging. However, deep features suffer from lower reproducibility than radiomic features and lack the interpretability of the latter.

show abstract

Section: Discussionmentioning

confidence: 99%

Using 3D deep features from CT scans for cancer prognosis based on a video classification model: A multi‐dataset feasibility study

et al. 2023

View full text Add to dashboard Cite

show abstract

“…51 In general, the majority of radiomics studies that used CT data found that first-order entropy is significantly associated with OS. 52…”

Section: Survival Prediction After Curative Surgerymentioning

confidence: 99%

Imaging of Pancreatic Ductal Adenocarcinoma: An Update on Recent Advances

et al. 2022

View full text Add to dashboard Cite

Pancreatic ductal carcinoma (PDAC) is one of the leading causes of cancer-related death worldwide. Computed tomography (CT) remains the primary imaging modality for diagnosis of PDAC. However, CT has limitations for early pancreatic tumor detection and tumor characterization so that it is currently challenged by magnetic resonance imaging. More recently, a particular attention has been given to radiomics for the characterization of pancreatic lesions using extraction and analysis of quantitative imaging features. In addition, radiomics has currently many applications that are developed in conjunction with artificial intelligence (AI) with the aim of better characterizing pancreatic lesions and providing a more precise assessment of tumor burden. This review article sums up recent advances in imaging of PDAC in the field of image/data acquisition, tumor detection, tumor characterization, treatment response evaluation, and preoperative planning. In addition, current applications of radiomics and AI in the field of PDAC are discussed.

show abstract

“…The authors identified several features in the cancerous patch with lower values reflecting image intensity (first order: median, mean and 90th percentile) and higher values for features reflecting heterogeneity (NGTDM busyness, GLDM gray-level nonuniformity and GLDM dependence nonuniformity), consistent with the notion that PDACs typically manifest with heterogenous hypoenhancement on CT. Multiple other studies investigated survival prediction in PDAC using radiomics [ 49 , 50 , 51 , 52 ]. For example, Khalvati et al [ 51 ] conducted a multicenter study, including 98 patients, aiming to assess the reproducibility and prognostic value of CT-derived radiomics features for resectable PDAC.…”

Section: Discussionmentioning

confidence: 99%

CT Radiomics and Whole Genome Sequencing in Patients with Pancreatic Ductal Adenocarcinoma: Predictive Radiogenomics Modeling

Hinzpeter

Kulanthaivelu

Kohan

et al. 2022

Cancers

View full text Add to dashboard Cite

We investigate whether computed tomography (CT) derived radiomics may correlate with driver gene mutations in patients with pancreatic ductal adenocarcinoma (PDAC). In this retrospective study, 47 patients (mean age 64 ± 11 years; range: 42–86 years) with PDAC, who were treated surgically and who underwent preoperative CT imaging at our institution were included in the study. Image segmentation and feature extraction was performed semi-automatically with a commonly used open-source software platform. Genomic data from whole genome sequencing (WGS) were collected from our institution’s web-based resource. Two statistical models were then built, in order to evaluate the predictive ability of CT-derived radiomics feature for driver gene mutations in PDAC. 30/47 of all tumor samples harbored 2 or more gene mutations. Overall, 81% of tumor samples demonstrated mutations in KRAS, 68% of samples had alterations in TP53, 26% in SMAD4 and 19% in CDKN2A. Extended statistical analysis revealed acceptable predictive ability for KRAS and TP53 (Youden Index 0.56 and 0.67, respectively) and mild to acceptable predictive signal for SMAD4 and CDKN2A (Youden Index 0.5, respectively). Our study establishes acceptable correlation of radiomics features and driver gene mutations in PDAC, indicating an acceptable prognostication of genomic profiles using CT-derived radiomics. A larger and more homogenous cohort may further enhance the predictive ability.

show abstract

A systematic review of prognosis predictive role of radiomics in pancreatic cancer: heterogeneity markers or statistical tricks?

Cited by 21 publications

References 52 publications

Using 3D deep features from CT scans for cancer prognosis based on a video classification model: A multi‐dataset feasibility study

Using 3D deep features from CT scans for cancer prognosis based on a video classification model: A multi‐dataset feasibility study

Imaging of Pancreatic Ductal Adenocarcinoma: An Update on Recent Advances

CT Radiomics and Whole Genome Sequencing in Patients with Pancreatic Ductal Adenocarcinoma: Predictive Radiogenomics Modeling

Contact Info

Product

Resources

About