Detecting microvascular invasion in HCC with contrast-enhanced MRI: Is it a good idea?

Deep learning-based accurate diagnosis and quantitative evaluation of microvascular invasion in hepatocellular carcinoma on whole-slide histopathology images

BACKGROUND

Microvascular invasion (MVI) is an independent prognostic factor that is associated with early recurrence and poor survival after resection of hepatocellular carcinoma (HCC). However, the traditional pathology approach is relatively subjective, time-consuming, and heterogeneous in the diagnosis of MVI. The aim of this study was to develop a deep-learning model that could significantly improve the efficiency and accuracy of MVI diagnosis.

MATERIALS AND METHODS

We collected H&E-stained slides from 753 patients with HCC at the First Affiliated Hospital of Zhejiang University. An external validation set with 358 patients was selected from The Cancer Genome Atlas database. The deep-learning model was trained by simulating the method used by pathologists to diagnose MVI. Model performance was evaluated with accuracy, precision, recall, F1 score, and the area under the receiver operating characteristic curve.

RESULTS

We successfully developed a MVI artificial intelligence diagnostic model (MVI-AIDM) which achieved an accuracy of 94.25% in the independent external validation set. The MVI positive detection rate of MVI-AIDM was significantly higher than the results of pathologists. Visualization results demonstrated the recognition of micro MVIs that were difficult to differentiate by the traditional pathology. Additionally, the model provided automatic quantification of the number of cancer cells and spatial information regarding MVI.

CONCLUSIONS

We developed a deep learning diagnostic model, which performed well and improved the efficiency and accuracy of MVI diagnosis. The model provided spatial information of MVI that was essential to accurately predict HCC recurrence after surgery.

A standardized pathological proposal for evaluating microvascular invasion of hepatocellular carcinoma: a multicenter study by LCPGC

BACKGROUND AND AIMS

Microvascular invasion (MVI) is a key pathological factor that severely affects the postoperative prognosis of patients with hepatocellular carcinoma (HCC). However, no MVI classification schemes based on standardized gross sampling protocols of HCC are available at present.

METHODS

119 HCC specimens were sampled at multiple sites (3-, 7-, and 13 points) for the optimum MVI detection rate. 16,144 resected HCCs were graded as M0, M1 or M2 by adopting three-tiered MVI grading (MVI-TTG) scheme based on the seven-point sampling protocol (SPSP). Survival analyses were performed on 2573 patients to explore the advantages of MVI-TTG.

RESULTS

The MVI detection rate determined by SPSP was significantly higher than that determined by the 3-point sampling method (34.5% vs. 47.1%, p = 0.048), but was similar to that determined by the 13-point sampling method (47.1% vs. 51.3%, p = 0.517). Among 16,144 resected HCCs, the proportions of M0, M1 and M2 specimens according to SPSP were 53.4%, 26.2% and 20.4%, respectively. Postoperative survival analysis in 2573 HCC patients showed that the 3-year recurrence rates in M0, M1 and M2 MVI groups were 62.5%, 71.6% and 86.1%, respectively (p < 0.001), and the corresponding 3-year overall survival (OS) rates were 94.1%, 87.5% and 67.0%, respectively (p < 0.001). M1 grade was associated with early recurrence, while M2 grade was associated with both early and late recurrence. MVI-TTG had a larger area under the curve and net benefit rate than the two-tiered MVI grading scheme for predicting time to recurrence and OS.

CONCLUSIONS

SPSP is a practical method to balance the efficacy of sampling numbers and MVI detection rates. MVI-TTG based on SPSP is a better prognostic predictor than the two-tiered MVI scheme. The combined use of SPSP and MVI-TTG is recommended for the routine pathological diagnosis of HCC.