A Transformer-Based microvascular invasion classifier enhances prognostic stratification in HCC following radiofrequency ablation

MRI-based prediction of microvascular invasion/high tumor grade and adjuvant therapy benefit for solitary HCC ≤ 5 cm: a multicenter cohort study

OBJECTIVES

To develop and externally validate an MRI-based diagnostic model for microvascular invasion (MVI) or Edmondson-Steiner G3/4 (i.e., high-risk histopathology) in solitary BCLC 0/A hepatocellular carcinoma (HCC) ≤ 5 cm and to assess its performance in predicting adjuvant therapy benefits.

MATERIALS AND METHODS

This multicenter retrospective cohort study included 577 consecutive adult patients who underwent contrast-enhanced MRI and subsequent curative resection or ablation for solitary BCLC 0/A HCC ≤ 5 cm (December 2011 to January 2024) from four hospitals. For resection-treated patients, a diagnostic model integrating clinical and 50 semantic MRI features was developed against pathology with logistic regression analyses on the training set (center 1) and externally validated on the testing dataset (centers 2-4), with its utilities in predicting posttreatment recurrence-free survival (RFS) and adjuvant therapy benefit evaluated by Cox regression analyses.

RESULTS

Serum α-fetoprotein > 100 ng/mL (odds ratio (OR), 1.94; p = 0.006), non-simple nodular growth subtype (OR, 1.69; p = 0.03), and the VICT2 trait (OR, 4.49; p < 0.001) were included in the MVI or high-grade (MHG) trait, with testing set AUC, sensitivity, and specificity of 0.832, 74.0%, and 82.5%, respectively. In the multivariable Cox analysis, the MHG-positive status was associated with worse RFS (resection testing set HR, 3.55, p = 0.02; ablation HR, 3.45, p < 0.001), and adjuvant therapy was associated with improved RFS only for the MHG-positive patients (resection HR, 0.39, p < 0.001; ablation HR, 0.30, p = 0.005).

CONCLUSION

The MHG trait effectively predicted high-risk histopathology, RFS and adjuvant therapy benefit among patients receiving curative resection or ablation for solitary BCLC 0/A HCC ≤ 5 cm.

KEY POINTS

Question Despite being associated with increased recurrence and potential benefit from adjuvancy in HCC, microvascular invasion or Edmondson-Steiner grade 3/4 are hardly assessable noninvasively. Findings We developed and externally validated an MRI-based model for predicting high-risk histopathology, post-resection/ablation recurrence-free survival, and adjuvant therapy benefit in solitary HCC ≤ 5 cm. Clinical relevance Among patients receiving curative-intent resection or ablation for solitary HCC ≤ 5 cm, noninvasive identification of high-risk histopathology (MVI or high-grade) using our proposed MRI model may help improve individualized prognostication and patient selection for adjuvant therapies.

Development and validation of a user-friendly prediction tool for preoperative T-Staging in gallbladder Cancer: A multicenter study using contrast-enhanced CT-Based fusion models

INTRODUCTION

Accurate T-staging of gallbladder cancer (GBC) is critical for surgical planning; however, existing imaging techniques have limited identification accuracy. This study aimed to develop a robust model and prediction tool to address these limitations.

MATERIALS AND METHODS

A retrospective cohort of 189 GBC patients from two institutions between January 2014 and December 2023 were analyzed. Patients were randomly assigned to internal training (ITC, n = 111), internal validation (IVC, n = 48), and temporal validation (TVC, n = 30) cohorts. Radiomics (Rad) and deep learning (DL) features were extracted from arterial and portal venous sequences, alongside clinical data, were used to construct pre- and post-fusion models. A weighted GBC T-staging (wGBCT) model was developed by combining probabilities from four modalities in the TVC: Clinical, Clinical + Rad(AP + PVP), Clinical + DL(AP + PVP), and Clinical + Rad + DL(CRDL), using a weighted averaging method. This model was validated and implemented as a user-friendly prediction tool.

RESULTS

The CRDL model achieved AUCs of 1.0 in the ITC and 0.913 in the IVC. In the TVC, the prediction tool attained an accuracy of 0.867, while the wGBCT model outperformed the CRDL model with an AUC of 0.910 (95 % CI: 0.792-1.000) compared to 0.869 (95 % CI:0.729-1.000). The wGBCT model also demonstrated superior sensitivity (1.0) and F1-score (0.867). Calibration curve analysis confirmed strong alignment, and decision curve analysis indicated the highest clinical net benefit at risk thresholds below 0.6.

CONCLUSIONS

The wGBCT model, integrating multimodal features and a user-friendly prediction tool, demonstrated high predictive accuracy and stability for preoperative T-staging of GBC, providing a valuable reference for individualized surgical planning.

Preoperative assessment in lymph node metastasis of pancreatic ductal adenocarcinoma: a transformer model based on dual-energy CT

BACKGROUND

Deep learning(DL) models can improve significantly discrimination of lymph node metastasis(LNM) of pancreatic ductal adenocarcinoma(PDAC), but have not been systematically assessed.

PURPOSE

To develop and test a transformer model utilizing dual-energy computed tomography (DECT) for predicting LNM in patients with PDAC.

MATERIALS AND METHODS

This retrospective study examined patients who had undergone surgical resection and had pathologically confirmed PDAC, with DECT performed between August 2016 and October 2022. Six predictive models were constructed: a DECT report model, a clinical model, 100 keV DL model, 150 keV DL model, a combined 100 + 150 keV DL model, and a model that integrated clinical information with DL-derived signatures. Multivariable logistic regression analysis was employed to develop the integrated model. The efficacy of these models was assessed by comparing their areas under the receiver operating characteristic curve (AUC) using the Delong test. Survival analysis was conducted using Kaplan-Meier curves.

RESULTS

In brief, 223 patients (mean age, 57 years ± 11 standard deviation; 93 men) were evaluated. All patients were divided into training (n = 160) and test (n = 63) sets. Patients with LNM accounted for 96 of the 223 patients (43%). In the test set, the integrated model, which integrated DECT parameters such as IC and Z, CA- 199 levels, DECT reports, and DL signatures, demonstrated the highest performance in predicting LNM, with an AUC of 0.93. In contrast, the radiologists'assessment and the clinical model yielded AUCs of 0.60 and 0.62, respectively. The integrated model-predicted positive LNM was associated with worse overall survival (hazard ratio, 1.75; 95% confidence interval: 1.22 - 2.83; P =.023).

CONCLUSION

A transformer-based model outperformed radiologists and clinical model for prediction of LNM at DECT in patients with PDAC.

Development and validation of MRI-derived deep learning score for non-invasive prediction of PD-L1 expression and prognostic stratification in head and neck squamous cell carcinoma

BACKGROUND

Immunotherapy has revolutionized the treatment landscape for head and neck squamous cell carcinoma (HNSCC) and PD-L1 combined positivity score (CPS) scoring is recommended as a biomarker for immunotherapy. Therefore, this study aimed to develop an MRI-based deep learning score (DLS) to non-invasively assess PD-L1 expression status in HNSCC patients and evaluate its potential effeciency in predicting prognostic stratification following treatment with immune checkpoint inhibitors (ICI).

METHODS

In this study, we collected data from four patient cohorts comprising a total of 610 HNSCC patients from two separate institutions. We developed deep learning models based on the ResNet-101 convolutional neural network to analyze three MRI sequences (T1WI, T2WI, and contrast-enhanced T1WI). Tumor regions were manually segmented, and features extracted from different MRI sequences were fused using a transformer-based model incorporating attention mechanisms. The model's performance in predicting PD-L1 expression was evaluated using the area under the curve (AUC), sensitivity, specificity, and calibration metrics. Survival analyses were conducted using Kaplan-Meier survival curves and log-rank tests to evaluate the prognostic significance of the DLS.

RESULTS

The DLS demonstrated high predictive accuracy for PD-L1 expression, achieving an AUC of 0.981, 0.860 and 0.803 in the training, internal and external validation cohort. Patients with higher DLS scores demonstrated significantly improved progression-free survival (PFS) in both the internal validation cohort (hazard ratio: 0.491; 95% CI, 0.270-0.892; P = 0.005) and the external validation cohort (hazard ratio: 0.617; 95% CI, 0.391-0.973; P = 0.040). In the ICI-treated cohort, the DLS achieved an AUC of 0.739 for predicting durable clinical benefit (DCB).

CONCLUSIONS

The proposed DLS offered a non-invasive and accurate approach for assessing PD-L1 expression in patients with HNSCC and effectively stratified HNSCC patients to benefit from immunotherapy based on PFS.

Artificial intelligence applied to 'omics data in liver disease: towards a personalised approach for diagnosis, prognosis and treatment

Advancements in omics technologies and artificial intelligence (AI) methodologies are fuelling our progress towards personalised diagnosis, prognosis and treatment strategies in hepatology. This review provides a comprehensive overview of the current landscape of AI methods used for analysis of omics data in liver diseases. We present an overview of the prevalence of different omics levels across various liver diseases, as well as categorise the AI methodology used across the studies. Specifically, we highlight the predominance of transcriptomic and genomic profiling and the relatively sparse exploration of other levels such as the proteome and methylome, which represent untapped potential for novel insights. Publicly available database initiatives such as The Cancer Genome Atlas and The International Cancer Genome Consortium have paved the way for advancements in the diagnosis and treatment of hepatocellular carcinoma. However, the same availability of large omics datasets remains limited for other liver diseases. Furthermore, the application of sophisticated AI methods to handle the complexities of multiomics datasets requires substantial data to train and validate the models and faces challenges in achieving bias-free results with clinical utility. Strategies to address the paucity of data and capitalise on opportunities are discussed. Given the substantial global burden of chronic liver diseases, it is imperative that multicentre collaborations be established to generate large-scale omics data for early disease recognition and intervention. Exploring advanced AI methods is also necessary to maximise the potential of these datasets and improve early detection and personalised treatment strategies.