Radiomics Models for Predicting Microvascular Invasion in Hepatocellular Carcinoma: A Systematic Review and Radiomics Quality Score Assessment

Abstract

Preoperative prediction of microvascular invasion (MVI) is of importance in hepatocellular carcinoma (HCC) patient treatment management. Plenty of radiomics models for MVI prediction have been proposed. This study aimed to elucidate the role of radiomics models in the prediction of MVI and to evaluate their methodological quality. The methodological quality was assessed by the Radiomics Quality Score (RQS), and the risk of bias was evaluated by the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). Twenty-two studies using CT, MRI, or PET/CT for MVI prediction were included. All were retrospective studies, and only two had an external validation cohort. The AUC values of the prediction models ranged from 0.69 to 0.94 in the test cohort. Substantial methodological heterogeneity existed, and the methodological quality was low, with an average RQS score of 10 (28% of the total). Most studies demonstrated a low or unclear risk of bias in the domains of QUADAS-2. In conclusion, a radiomics model could be an accurate and effective tool for MVI prediction in HCC patients, although the methodological quality has so far been insufficient. Future prospective studies with an external validation cohort in accordance with a standardized radiomics workflow are expected to supply a reliable model that translates into clinical utilization.

A Multi-Parametric Radiomics Nomogram for Preoperative Prediction of Microvascular Invasion Status in Intrahepatic Cholangiocarcinoma

“…Radiomics is a powerful and sophisticated image mining tool, and it can improve diagnostic accuracy and predict prognosis by high-throughput selecting imaging features from medical images ( 15 ). Also, several studies have constructed radiomics-based nomogram in distinguishing different pathological types of primary liver cancer ( 16 ) and predicting MVI of hepatocellular carcinoma preoperatively ( 17 , 18 ). Recently, radiomics nomograms have been established for the prediction of lymph node metastasis ( 19 ), early recurrence ( 20 ), and prognosis after hepatectomy ( 21 ) in ICC patients.…”

Reply

“…This value is significantly higher than the RQS for the majority of retrospective studies dealing with traditional volumetric imaging (i.e. computed tomography and magnetic resonance imaging), these being reported as, on average, 11 (lungs), 7 (heart) and 10 (liver) [3][4][5] .…”

“…This value is significantly higher than the RQS for the majority of retrospective studies dealing with traditional volumetric imaging (i.e. computed tomography and magnetic resonance imaging), these being reported as, on average, 11 (lungs), 7 (heart) and 10 (liver) [3][4][5] .We are fully aware that image preprocessing is a crucial step prior to feature extraction in the framework of traditional radiomics, and our group was among the first to describe the variability between different ultrasound equipment used for radiomics purposes in gynecology 6 . Nevertheless, in this first large, multicenter study, we decided to focus on unfiltered real-world data images, excluding only very low-quality images that did not permit clear and reliable identification of the endometrial lesion, as we chose to rely on examiner expertise for…”

A Radiomics Model Based on Gd-EOB-DTPA-Enhanced MRI for the Prediction of Microvascular Invasion in Solitary Hepatocellular Carcinoma ≤ 5 cm

“…This makes it important to validate the model on different cohorts. Fourth, the optimal dilation of the tumor needs to be evaluated as we just dilated the tumor VOIs to 10 mm of the margin as most previously published studies did ( 20 , 21 ). Future research can be designed to compare different dilations of the tumor diameter when predicting MVI.…”

“…Tumor delineation was performed manually using the open-source software ITK-SNAP (version 3.8.0, ). The delineated tumor was further expanded at a radius of 10 mm ( 20 , 21 ) using a topologic algorithm in Python (version 3.8), and the expansion would cease automatically if it reached the liver edge for the marginal liver tumors. The expanded volume of interest (VOI) was then used for radiomics feature extraction ( Figure 3 ).…”