Abstract No. 294 Multimodal Vision Transformer Modeling of Survival and Transplant Eligibility Following Radioembolization for Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) with portal vein tumor thrombus is considered an advanced stage disease. Non-surgical local and systemic therapies are the only treatment options available. To analyze the survival and toxicity outcomes of systemic treatment concurrent with yttrium-90 transarterial radioembolization in HCC with liver-limited disease and portal vein involvement with Child–Pugh B liver reserve. The medical records of 22 patients who underwent yttrium-90 transarterial radioembolization concomitant with capecitabine chemotherapy as first-line treatment between 2014 and 2019 were retrospectively reviewed. Twenty-two patients were included in the study. Grade 3 to 4 side effects were evaluated, and hepatic encephalopathy developed in 1 patient after yttrium-90 transarterial radioembolization. In the fourth month of radiological evaluation, 11 patients had a partial response (50%), 5 patients had stable disease (22.7%), and 6 patients (27.3%) developed progressive disease. The median survival time was 21 months. Combined treatment with yttrium-90 transarterial radioembolization and capecitabine may be an effective and safe treatment option. Treatment was associated with a median overall survival of 21 months and a disease control rate of 72.7% at 4 months in patients with inoperable HCC.

No. 168 - Radioembolization for Hepatocellular Carcinoma in Patients with Hyperbilirubinemia

No. 188: Impact of C-Arm CT in Patients with Hepatocellular Carcinoma Undergoing Transhepatic Arterial Chemoembolization

Radioembolization for hepatocellular carcinoma: The time has come.

The referring video object segmentation task (RVOS) involves segmentation of a text-referred object instance in the frames of a given video. Due to the complex nature of this multimodal task, which combines text reasoning, video understanding, instance segmentation and tracking, existing approaches typically rely on sophisticated pipelines in order to tackle it. In this paper, we propose a simple Transformer-based approach to RVOS. Our framework, termed Multimodal Tracking Transformer (MTTR), models the RVOS task as a sequence prediction problem. Following recent advancements in computer vision and natural language processing, MTTR is based on the realization that video and text can be processed together effectively and elegantly by a single multimodal Transformer model. MTTR is end-to-end trainable, free of text-related inductive bias components and requires no additional mask-refinement post-processing steps. As such, it simplifies the RVOS pipeline considerably compared to existing methods. Evaluation on standard benchmarks reveals that MTTR significantly outperforms previous art across multiple metrics. In particular, MTTR shows impressive +5.7 and +5.0 mAP gains on the A2D-Sentences and JHMDB-Sentences datasets respectively, while processing 76 frames per second. In addition, we report strong results on the public validation set of Refer-YouTube-VOS, a more challenging RVOS dataset that has yet to receive the attention of researchers. The code to reproduce our experiments is avail-able at https://github.com/mttr2021/MTTR.

Accelerating climate change and the intensifying global food security crisis have increased the importance of reliable crop classification across diverse environmental conditions. Existing crop classification models have primarily focused on improving accuracy by learning spectral and temporal patterns from satellite imagery; however, their black-box nature makes it difficult to understand the rationale behind each prediction, limiting their applicability in real-world agricultural decision-making. To address this issue, this study introduces a multimodal Transformer model that incorporates a BERTbased bidirectional attention mechanism, aiming to retain classification performance while enhancing interpretability. The proposed BERT Hybrid model employs a PVT backbone to extract spatial features from Sentinel-2 satellite imagery and integrates them with meteorological time-series embeddings; bidirectional self-attention is then used to jointly model cross-temporal and cross-modal interactions. We further conduct comparative experiments under the same conditions as the MMST-ViT(Multi-Modal Spatial-Temporal Vision Transformer) baseline, evaluating not only overall accuracy but also temporal attention patterns across crop growth stages and the relative importance of different weather variables. Experimental results show that bidirectional attention alleviates excessive focus on specific timestamps or single variables, producing more consistent and interpretable attention distributions. This study highlights the performance– interpretability trade-off in multimodal agricultural AI models and provides a foundation for building trustworthy deeplearning systems for crop monitoring. In addition, because the proposed approach relies solely on globally accessible Sentinel-2 satellite imagery and publicly available meteorological data, it demonstrates the potential for constructing large-scale crop monitoring systems at low cost, aligning with the principles of appropriate technology.

Hepatocellular carcinoma (HCC) is the third most common cause of cancer-related mortality. Radioembolization is a form of selective internal radiation therapy that is increasingly used to treat patients with HCC, particularly those with more advanced disease. This review will try to answer some of the most frequently asked questions regarding the use of radioembolization to treat HCC patients and provide supporting evidence. Rather than a new form of transarterial chemoembolization (TACE), radioembolization is a form of brachytherapy that has a highly localized effect on liver tumors. The two devices that are available (glass and resin microspheres) are similar in size (25 to 35 microns), but differ in the amount of isotope loaded onto each microsphere and the number of spheres injected in a single treatment. Despite this, the evidence seems to indicate that the antitumor effect and safety profiles of these two devices in HCC are similar. Liver cirrhosis frequently underlies HCC. Despite the higher chance for relevant liver toxicity, there is now good evidence from large studies to show that radioembolization can be safely and effectively performed in cirrhotic patients with HCC. With no randomized controlled trials published so far, there is recent scientific evidence that allows comparison between radioembolization and other treatment options including TACE and the systemic, agent sorafenib. Radioembolization appears to have similar efficacy to TACE in patients that are ideal candidates for locoregional therapy and has shown encouraging results in patients that have failed TACE or who are poor candidates for this therapy. Survival in comparable sorafenib- and radioembolizationtreated HCC patients is quite similar. The indication for radioembolization has to be balanced against the risk of liver decompensation and the natural history of the disease, based on tumor burden and liver function. Patients with inadequate liver functional reserve and diffuse tumors affecting either lobes, or portal vein thrombosis that reaches the main trunk should probably not be treated with this procedure.

Radioembolization for hepatocellular carcinoma with portal vein thrombosis: Impact of liver function on systemic treatment options at disease progression

366 Background: The Childs-Pugh class (C-P) is a commonly used scoring system to measure liver function in patients with hepatocellular carcinoma (HCC). While originally developed for patients undergoing esophageal resection for varices in the setting of cirrhosis and portal hypertension, it is now a widely accepted measure of liver function. The Albumin-Bilirubin (ALBI) grading system is a model developed in 2015 to assess liver function that has eliminated the subjective measures present in C-P. We correlate the ALBI grading system and C-P classes to overall survival in our HCC patients receiving radioembolization. Methods: With IRB approval, we retrospectively evaluated patients who received radioembolization for HCC between 2009-2014. We evaluated the albumin and bilirubin levels in our patients prior to receiving their first radioembolization (n = 121). The ALBI grades were calculated from these data with the formula (log10 bilirubin x 0.66) + (albumin x -0.085) and separated into grades based on the thresholds set in the original manuscript. These grades were then correlated to outcomes using Mantel-Cox Log analysis. The statistical comparisons were duplicated with C-P classes. Results: Median survival for C-P class A and B were 13.1 and 8.4 months (p < 0.0005), respectively while median survival for ALBI grades 1 and 2 were 20.9 and 11.0 months (p < 0.0005), respectively. Our C-P class A patient group was re-stratified using the ALBI formula and found to have two cohorts of patients with statistically significant differences in survival. The median survival between ALBI grades 1 and 2 within this patient group was 20.9 and 11.0 months (p = 0.002), respectively. Too few C-P class C or ALBI grade 3 patients were within our dataset for statistically significant results. Conclusions: We demonstrated that the ALBI grading system is a more sensitive marker of liver function than the C-P classes in the setting of mild to moderate dysfunction. Using the ALBI grade, we identified a subset of patients that have significantly better outcomes from radioembolization for HCC within C-P class A.

Hepatocellular carcinoma (HCC) is the most common type of liver cancer and is the second cause of death due to malignancy in the world. The treatment of HCC is complex and includes potentially curative and palliative approaches. However, both curative and palliative treatments for HCC are often associated with a not-completely favorable safety/efficacy ratio. Therefore, other treatment options appear necessary in clinical practice. Transarterial radioembolization has shown a promising efficacy in terms of disease control and is associated with a good safety profile. This review discusses the use of transarterial radioembolization in HCC, with a focus on the clinical aspects of this therapeutic strategy.

To compare changes in RECIST, anatomical volume, mRECIST, and volumetric diffusion-weighted Imaging parameters (3D apparent diffusion coefficient (ADC) measurements), with pathological analysis of hepatocellular carcinoma (HCC) treated by (90)Yttrium radioembolization (Y90). 21 patients were treated by Y90 as a sole treatment modality for solitary, >2 cm HCC that underwent liver transplantation. MRI at baseline, 1 and 3 months post-Y90, and tumor pathological findings on explants were reviewed in all patients. Compared to baseline (RECIST/volume: 3.6 cm/17.7 cm(3)), RECIST and volume were not modified after Y90 (1 month, p = 0.28/0.09 RECIST/tumor volume; 3 months, p = 0.28/0.54). In contrast, mRECIST (3.3-1.4 cm, p < 0.001), mean ADC (0.185-1.093 mm(2)/s × 10(-3), p = 0.04), and ADC standard deviation (STD) (0.041-0.201 mm(2)/s × 10(-3), p = 0.0496) changed as earlier as 1 month post-Y90. ADC STD % change was higher in ADC responding lesions than non-responding lesions at 1 month (p = 0.002) and 3 months (p = 0.008). All lesions exhibited necrosis on pathological analysis (11 partially viable, 10 complete pathological necrosis (CPN)) but no imaging criterion was able to predict CPN. mRECIST (±ADC) at 1 (κ ± ADC = 0.08/0.06) or 3 months (κ = -0.06/-0.06) were poor predictors of pathological response. As soon as 1 month post-treatment, mRECIST and volumetric ADC performed better than traditional size RECIST or volumetric parameters in detecting imaging response to Y90; however, CPN cannot be predicted by any criteria. Improvements in methodologies to assess response and identification of better surrogates are awaited.

Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer deaths worldwide. Despite the high incidence of HCC, mortality remains high, with an estimated 5-year survival rate of less than 20%. Surgical resection represents a potential curative treatment for HCC; however, less than 20% of patients with HCC are candidates for surgical resection. In patients with unresectable HCC, Yttrium-90 (Y90) transarterial radioembolization (TARE) has emerged as an innovative treatment option. This locoregional therapy delivers high doses of radiation directly to liver tumors via intra-arterial injection, allowing for the targeted destruction of malignant cells while sparing surrounding healthy tissue. In this review, we will explore the latest advances in Y90 TARE for the treatment of HCC, focusing on key developments such as the following: (1) improvements in radiation lobectomy and segmentectomy techniques, (2) the introduction of personalized dosimetry, (3) the integration of combination therapies, (4) the use of imageable microspheres, (5) pressure-enabled Y90 delivery systems, and (6) the application of Y90 surrogates.

Transarterial radioembolization (TARE) with β-emitting radionuclides is widely used for hepatocellular carcinoma (HCC), but its clinical efficacy remains to be further improved. α-particle-emitting radionuclides possess high linear energy transfer (LET) and unique advantages in cancer therapy, motivating α-particle based composite platform. Accordingly, we engineer the first clinically mimetic α-TARE microsphere by in-situ ²²³Ra-doped calcium-alginate composite microsphere (²²³Ra/Ca-ALG MS) using a hydrogel matrix, in which alginate "egg-box" coordination captures Ra²⁺ to provide stable radiolabeling, delivered via selective hepatic arterial injection to HCC. The microspheres exhibited excellent radiolabeling stability (88% retention after 384h) and potent, dose-dependent cytotoxicity against HCC cells under hypoxia. In an orthotopic rat HCC model, 223Ra/Ca-ALG MS-based TARE achieves precise intratumoral localization and sustained retention on SPECT/CT; ¹⁸F-FDG PET/CT and histopathology indicate a robust antitumor response, while serum biochemistry and histology support a favorable safety profile. Moreover, ²²³Ra/Ca-ALG MS provide powerful immune-activating capacity. Transcriptomics reveals activation of DNA-damage response, immunogenic cell death, and antigen-presentation pathways, flow cytometry and immunohistochemistry show increased dendritic-cell maturation and CD8⁺ T-cell infiltration. Collectively, 223Ra/Ca-ALG MS demonstrates hypoxia-tolerant cytotoxicity, immune-activating potential, offering new insights for the development of immune-based TARE strategies in HCC and showing promising prospects for clinical translation.

e16633 Background: Yttrium-90 (Y90) radioembolization is a catheter-based therapy utilized in the treatment of hepatic tumors such as hepatocellular carcinoma (HCC). Multiple trials have evaluated the efficacy of Y90 in HCC; few have assessed long term hepatic function. These studies reported PFS ranging from 7.9 to 18 months (m), and median OS of 17.2 to 18 m for Child-Pugh (CP) Class A and 6.0 to 7.7 m for CP Class B. This study aimed to evaluate a clinical real-world experience of Y90 therapeutic effectiveness and long-term impact on hepatic function. Methods: A single-center retrospective chart review was performed for patients with a CP score of A or B who received Y90 for primary HCC between 2008 and 2016. Model for end stage liver disease (MELD) and CP scores were calculated for the day of treatment and at 1, 3, 6, 12, and 24 m post procedure. OS and PFS were characterized using a Kaplan-Meier survival analysis and a multivariate model. Results: 134 patients were included. Mean age was 60 years old, ranging from 44-90. Underlying liver disease included HCV – 66%, ETOH cirrhosis – 23%, NAFLD – 19%, HBV – 5%. CP Class A patients (79%) had a median OS of 16.59 m (95% CI: 8.90 to 23.10m) from date of Y90 treatment compared to a median of 7.90 m (95% CI: 4.60 to 15.64 m) for CP Class B. CP Class A patients had a median PFS of 3.45 m (95% CI: 2.99 to 5.55 m) from date of Y90 treatment compared to a median of 3.71 m (95% CI: 2.07 to 8.28 m) for CP Class B patients. These differences were not statistically significant. Median OS from date of diagnosis for the entire study population was 29.37 m (95% CI: 23.00 to 38.77 m). MELD scores were statistically significantly higher at 1, 3, 6 and 12 m post-treatment than on Day 0, with significant recovery at 24 vs 3 m. Older age, higher T stage, higher MELD and CP scores, and patients with a history of portal vein thrombosis (PVT) had significantly lower OS. PFS was significantly shorter in those with advanced stage at diagnosis and higher CP score. Conclusions: While our study supports the literature for OS in Y90 patients, we found a shorter PFS in this population. This may reflect the utilization of RECIST in clinical trials vs clinical radiology practice in determining progression. Over time, the most significant factors associated with OS were age, MELD and CP scores and PVT. For PFS, CP score and stage at diagnosis were significant. The increasing MELD and CP scores over time reflect a possible mix of radio embolization‐induced liver disease and progression of HCC. The downtrend at 24 m is likely due to durable survivors with significant benefit from therapy and no long-term complications from Y90.

BackgroundDiagnosing mediastinal tumours, including incidental lesions, using low-dose CT (LDCT) performed for lung cancer screening, is challenging. It often requires additional invasive and costly tests for proper characterisation and surgical...