T-CACE: A Time-Conditioned Autoregressive Contrast Enhancement Multi-Task Framework for Contrast-Free Liver MRI Synthesis, Segmentation, and Diagnosis

Guanghui Wang; Jianfeng Zhao; Qinmin Vivian Hu; Xiaojiao Xiao

REVIEW 3 major objections 5 minor 40 references

T-CACE: A Time-Conditioned Autoregressive Contrast Enhancement Multi-Task Framework for Contrast-Free Liver MRI Synthesis, Segmentation, and Diagnosis

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read T-CACE claims that a time-conditioned autoregressive model can synthesize multi-phase contrast-enhanced liver MRI from non-contrast scans, and jointly segment and classify lesions, outperforming state-of-the-art baselines on two datasets.

desk verdict The mask leakage is real and it guts the central claim, despite a plausible architecture combination. read the letter →

arxiv 2508.09919 v1 pith:UGVQOXFO submitted 2025-08-13 eess.IV cs.AIcs.CV

Xiaojiao Xiao , Jianfeng Zhao , Qinmin Vivian Hu , Guanghui Wang This is my paper

classification eess.IVcs.AIcs.CV

keywords AutoregressivemodelMRIsynthesisLivertumorclassificationNon-contrastSegmentationMulti-phaseCEMRITemporalattentionMulti-tasklearning

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes T-CACE, a single time-conditioned autoregressive model that turns non-contrast liver MRI into the three contrast-enhanced phases used in diagnosis—arterial, portal venous, and delayed—and, from the same latent representation, produces a lesion segmentation mask and a lesion classification. The authors argue that encoding phase identity and elapsed time as tokens, and gating cross-phase attention with a Gaussian decay over temporal distance, keeps the synthesized enhancement sequence physiologically coherent, which in turn makes segmentation and classification more reliable. On two liver MRI datasets, they report that T-CACE outperforms seven synthesis baselines on fidelity metrics and also raises segmentation Dice and classification accuracy relative to dedicated models. If the results hold, routine non-contrast scans could carry diagnostic information that currently requires injecting gadolinium-based contrast agents, reducing cost and risk.

What carries the argument

The load-bearing mechanism is the time-conditioned autoregressive loop with the Dynamic Time-aware Attention Mask (DTAM). The non-contrast scan and tumor mask are encoded with a Swin UNETR backbone into conditional tokens, augmented by a learned phase token for arterial, portal venous, and delayed phases and by sinusoidal time embeddings. When generating a phase, the transformer attends to the conditional token and to image tokens from all earlier phases, with attention weights decayed by a Gaussian function of the time gap between phases. This enforces sequential, physiology-like enhancement while keeping every phase anchored to the same anatomy. A Temporal Consistency Constraint (TCC) then

What would settle it

Run the released model on non-contrast MRIs with the tumor mask withheld and let it use its own predicted segmentation as the conditional input; if Dice or classification accuracy collapses compared to the reported numbers, the contrast-free automated pipeline claim is not supported.

Watch

Extended reading notes

Core claim

T-CACE claims that multi-phase contrast enhancement can be modeled as a sequential, time-labelled autoregressive process rather than as independent image translations. At each step the model receives a conditional token built from the non-contrast scan, the lesion mask, a discrete phase token, and a continuous sinusoidal time encoding; it then synthesizes that phase's image while attending, through a Gaussian-decayed mask, to tokens of all previously generated phases. The same updated tokens feed segmentation and classification heads, and a temporal consistency constraint forces the image-based classifier to agree with a thresholded model of washout intensity derived from the synthesized pha

Load-bearing premise

The load-bearing premise is that the true tumor mask is available as an input at inference; remove it and the reported synthesis, segmentation, and classification numbers no longer describe a fully automated contrast-free pipeline.

Editorial extensions

If this is right

If the reported results transfer, a patient needing liver lesion assessment could skip contrast-agent injection and rely on synthetic multi-phase images, removing gadolinium-related risks.
Joint optimization means the synthesized phases are produced with segmentation and classification in mind, so image-quality metrics alone do not measure the framework's full value.
The autoregressive ordering ensures the delayed phase is informed by the arterial and portal venous phases, matching the washout pattern clinicians use to flag malignancy.
Phase-wise masks are fused by majority voting, so a segmentation error in one phase can be corrected by the other two.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The model's conditional token includes the true tumor mask; a test in which the model receives only its own predicted mask would show how much of the reported segmentation and classification gain depends on that oracle input.
The Gaussian decay rate and washout threshold are fixed constants; a learned or patient-specific schedule might matter under irregular acquisition intervals, which the authors list as untested.
A radiologist reader study comparing synthetic phases with true CEMRI for washout interpretation would test clinical equivalence more directly than pixel-level metrics like SSIM and FID.
The framework is modular, so adapting CTE and DTAM to kidney or pancreas timing is a natural next experiment, though enhancement kinetics differ from liver and would likely need organ-specific phase tokens.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

You should know: the paper's central claim is dead on arrival because Algorithm 1 takes the ground-truth tumor mask as an input at inference, and the same mask is both the segmentation target and the conditioning for synthesis and classification. That is target leakage, and it directly contradicts the abstract and Fig. 1's promise of "fully automated, contrast-free... directly from non-contrast MRI."

What is actually new: the specific combination of a time-conditioned autoregressive generator, a Gaussian-decayed dynamic attention mask, and a temporal consistency constraint for classification is not in the cited literature. The idea of explicitly modeling inter-phase dynamics during synthesis is reasonable, and grounding it in continuous time tokens is a sensible move. Running two datasets and several baselines is more than many synthesis papers do, and the qualitative results look plausible at first glance.

Where it falls apart: the mask leakage is not a peripheral detail. It means the reported Dice, PSNR, and classification numbers are not evidence for the claimed capability. The model is effectively being told where the lesion is and then asked to segment and classify it. Second, the ablation table (Table III) has an internal inconsistency: "No DTAM" and "No T-Encoding" report higher PSNR than the full T-CACE while also reporting higher MSE. That is impossible under the standard PSNR formula, so the numbers are misreported or computed differently than stated. Third, the closest prior work, CAVM, is cited but never compared in any table. For a paper claiming SOTA, omitting the nearest autoregressive baseline is a big gap.

To be fair, the authors do acknowledge some limitations and promise code, but no commit or config details are given. The core problem is not a matter of adding a caveat; it requires redoing the inference so the mask is either predicted or not used at all.

Who this is for: a reader interested in contrast-free MRI synthesis who wants to see the latest architectural combinations. But as a published result, it does not hold. I would not cite it in its current form. I would bring it to a reading group to illustrate how easy it is to leak the target through the conditioning path.

Recommendation: a serious editor should send this to review, because the flaw deserves expert documentation and the authors deserve a chance to fix it. But the reviewer should be asked to check specifically for mask leakage and the ablation inconsistency. If the authors can rerun the experiments with a predicted mask or no mask, the underlying idea may be salvageable.

Referee Report

3 major / 5 minor

Summary. The paper proposes T-CACE, a multi-task framework that jointly synthesizes multi-phase contrast-enhanced liver MRI from non-contrast MRI, performs lesion segmentation, and outputs lesion classification. Methodologically, it combines a Conditional Token Encoding (CTE), a Dynamic Time-aware Attention Mask (DTAM), and a Temporal Classification Consistency (TCC) constraint. The authors report state-of-the-art synthesis, segmentation, and classification results on two datasets (MG-2021 and LLD-MMRI2023) and provide code. The central claim is that T-CACE enables fully automated, contrast-free liver-lesion diagnosis directly from non-contrast MRI. However, the paper's own algorithm requires the ground-truth tumor mask as an input, which is used to condition synthesis, segmentation, and classification, creating target leakage. In addition, the main ablation table contains mathematically inconsistent PSNR/MSE values. These issues undermine the validity of the reported results as evidence for the claimed fully automated diagnostic pipeline.

Significance. If the framework truly operated from non-contrast MRI alone and achieved the reported Dice, PSNR, and classification numbers, it would be a clinically valuable contribution: avoiding gadolinium-based contrast agents, jointly performing synthesis, segmentation, and classification, and releasing public code. The autoregressive phase-conditioned formulation is a reasonable and potentially productive design. However, the significance cannot be assessed from the present manuscript because the evaluation protocol leaks the ground-truth lesion mask into training and inference. The central claim of a fully automated contrast-free pipeline is therefore not supported by the presented experiments.

major comments (3)

[Algorithm 1, §III-A, §III-B.1] Algorithm 1 explicitly lists the tumor mask y_mask as an input (line 1) and uses it to compute the conditional token (line 4: z_CT = proj(enc([x_T1, y_mask]))). The same y_mask is also the target of the segmentation head (§III-B.3) and conditions the synthesized phases used for classification (§III-B.4). This is target leakage: at inference the model is given the ground-truth lesion boundary, so the reported Dice scores, synthesis metrics, and classification accuracies are not independent measurements of a system that operates 'directly from non-contrast MRI' as claimed in the abstract and Fig. 1. No experiment evaluates the actual deployment setting without y_mask. This invalidates the central claim as stated.
[Table III and §III-C.3] The ablation table is internally inconsistent. For the full T-CACE model, MSE 0.307 corresponds to PSNR 25.19, while 'No DTAM' has MSE 0.328 but PSNR 25.43, and 'No T-Encoding' has MSE 0.319 but PSNR 25.72. Since PSNR = 10 log10(MAX^2/MSE), a lower MSE must give a higher PSNR for a fixed data range. These values violate that monotonic relation, suggesting the metrics were not computed consistently. This casts doubt on the reliability of all quantitative comparisons in the paper, not only the ablation.
[§III-B.4, Eq. (6)] The Temporal Consistency Constraint derives binary diagnostic labels by thresholding a learned scalar signal intensity with a fixed threshold tau=0.5, described as 'clinically relevant'. No evidence is provided that this threshold corresponds to actual clinical washout or malignancy patterns; it appears to be an arbitrarily chosen mid-range value. Moreover, the image-based classifier is trained to match this signal-derived label, and the signal is itself predicted from the non-contrast MRI plus a time embedding. This creates an additional circular dependency: classification performance is partially evaluated against a label generated from the model's own auxiliary prediction rather than an independent clinical ground truth. The authors should validate the threshold against radiologist annotations or report classification without the TCC alignment.

minor comments (5)

[Fig. 1 vs. Fig. 2] Fig. 1 promises 'fully automated, contrast-free synthesis, segmentation, and classification directly from non-contrast MRI', while Fig. 2 and Algorithm 1 show tumor mask as an input. This discrepancy should be resolved by either removing the mask from the pipeline or clearly labeling the method as mask-guided.
[Tables I–II] The baseline name 'AVG-GAN' appears with an extra space in the table rows; also the caption unit for MSE is inconsistent with the text. Minor typographical issues, but they should be corrected.
[§III-C.1] MG-2021 is described as an 'in-house dataset' but later in §III-E.2 as a 'public liver MRI dataset'. This is contradictory and should be clarified.
[§III-E.2] Paired t-tests are mentioned and annotated, but exact p-values, test statistics, or confidence intervals are not reported. Given the small number of cross-validation folds, more detail is needed to support the significance claims.
[Fig. 4 and §III-D.1] The text refers to 'PV' phases, but the figure caption and some labels use 'PC'. Please make the phase nomenclature consistent.

Circularity Check

1 steps flagged · score 8.0 of 10

Ground-truth tumor mask y_mask is a required input at inference, so the claimed fully automated NCMRI-only segmentation and classification results are not independent predictions.

self definitional [Algorithm 1, lines 1–4; Section III-B.1; Fig. 3 caption]
"1: Input: Non-contrast MRI x_T1, tumor mask y_mask ... 4: z_proj = proj(enc(x_T1, y_mask)) // Organ-level token. Section III-B.1: "The original token z_0 is computed from the inputs x_T1 and y_mask using a hierarchical feature encoder, z_proj = proj(enc(x_T1, y_mask))." Fig. 3 caption: "the left pink-shaded region denotes fixed conditional tokens (e.g., non-contrast images and tumor masks), which remain constant throughout the process.""

The model's segmentation output is decoded from latent tokens that are computed from the ground-truth tumor mask y_mask at inference, while y_mask is also the target of the segmentation head and the claimed fully automated pipeline is supposed to produce segmentation 'directly from non-contrast MRI.' Thus the reported Dice/HD95 scores do not measure prediction of the mask from NCMRI alone; they measure the model's ability to reproduce a mask it was given as input. The same mask conditions the synthesized CEMRI phases that feed the classifier, so the classification results also leak the ground-truth lesion annotation. This is target leakage by construction, not an independent multi-task prediction.

full rationale

The central claim in the abstract and Fig. 1 is that T-CACE enables 'fully automated, contrast-free synthesis, segmentation, and classification directly from non-contrast MRI.' However, the paper's own Algorithm 1 lists 'tumor mask y_mask' as an input, and Section III-B.1 defines the conditional token as a function of both x_T1 and y_mask. The same mask is the segmentation target and is used to condition the autoregressive synthesis that is then fed to the classifier. Consequently, the reported segmentation, synthesis, and classification numbers are not evidence for the stated NCMRI-only diagnostic pipeline. This is a severe internal inconsistency that makes the central evaluation circular: the model is given the ground-truth lesion boundary at inference. I did not find other circularity patterns: the TCC loss is an internal consistency constraint between two learned heads, and the authors' self-citations are not load-bearing for the main derivation. The score is 8 because the central quantitative claims reduce, to a substantial degree, to the model being conditioned on the label it is supposed to predict.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

No new physical entities or forces are introduced. The latent variables, attention masks, and learned intensity signals are internal mechanisms of the network, not independently falsifiable entities. The most load-bearing assumption is the availability of the true tumor mask as model input, which creates a target-leakage circularity.

free parameters (4)

Gaussian decay rate sigma = 0.7
Eq. (3); empirically selected via sensitivity analysis for optimal cross-phase interaction.
TCC washout threshold tau = 0.5
Section III-B.4; empirically set to the mid-range normalized signal intensity used to differentiate washout patterns.
Loss weights lambda_dice, lambda_ce, lambda_cls, lambda_TCC = 1.0 each
Section III-B.5; empirically set to 1.0 because this configuration yielded stable optimization.
Continuous-time frequency modulation parameter f = unspecified
Section III-B.1; the sinusoidal time encoding requires a hand-selected frequency, but no value or sensitivity analysis is reported.

assumptions (5)

ad hoc to paper The ground-truth tumor mask y_mask is available as an input to the conditional token encoder at inference.
Algorithm 1 and Section III-B.1; no test-time mask-free inference is described, yet the claimed clinical pipeline is fully automated.
domain assumption Paired NCMRI and CEMRI volumes are spatially registered and phase labels define a valid temporal order.
MG-2021 used manual registration by radiologists; LLD-MMRI provides pre-cropped lesion patches. The autoregressive sequence depends on this alignment.
domain assumption Gaussian decay with sigma=0.7 approximates the physiological contrast enhancement dynamics between phases.
Eq. (3) and Eq. (2); this is a modeling choice, not derived from physiology or validated against measured enhancement curves.
ad hoc to paper A thresholded learned scalar intensity p_t is a valid proxy for clinical washout and malignancy.
Section III-B.4; the threshold tau=0.5 is empirically chosen and F_signal is trained jointly within the same framework.
standard math Standard deep learning components and loss functions perform as expected.
Swin UNETR, Transformer attention, Dice and cross-entropy losses, and Adam optimization are treated as reliable background tools.

how reviews work

0 comments

Cite this review

Pith. "Pith review of T-CACE: A Time-Conditioned Autoregressive Contrast Enhancement Multi-Task Framework for Contrast-Free Liver MRI Synthesis, Segmentation, and Diagnosis." pith.science (2026). https://pith.science/paper/UGVQOXFO

@misc{pith2026250809919,
  author       = {Pith},
  title        = {Pith review of: T-CACE: A Time-Conditioned Autoregressive Contrast Enhancement Multi-Task Framework for Contrast-Free Liver MRI Synthesis, Segmentation, and Diagnosis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UGVQOXFO}},
  note         = {Machine review of arXiv:2508.09919}
}

read the original abstract

Magnetic resonance imaging (MRI) is a leading modality for the diagnosis of liver cancer, significantly improving the classification of the lesion and patient outcomes. However, traditional MRI faces challenges including risks from contrast agent (CA) administration, time-consuming manual assessment, and limited annotated datasets. To address these limitations, we propose a Time-Conditioned Autoregressive Contrast Enhancement (T-CACE) framework for synthesizing multi-phase contrast-enhanced MRI (CEMRI) directly from non-contrast MRI (NCMRI). T-CACE introduces three core innovations: a conditional token encoding (CTE) mechanism that unifies anatomical priors and temporal phase information into latent representations; and a dynamic time-aware attention mask (DTAM) that adaptively modulates inter-phase information flow using a Gaussian-decayed attention mechanism, ensuring smooth and physiologically plausible transitions across phases. Furthermore, a constraint for temporal classification consistency (TCC) aligns the lesion classification output with the evolution of the physiological signal, further enhancing diagnostic reliability. Extensive experiments on two independent liver MRI datasets demonstrate that T-CACE outperforms state-of-the-art methods in image synthesis, segmentation, and lesion classification. This framework offers a clinically relevant and efficient alternative to traditional contrast-enhanced imaging, improving safety, diagnostic efficiency, and reliability for the assessment of liver lesion. The implementation of T-CACE is publicly available at: https://github.com/xiaojiao929/T-CACE.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

40 extracted references · 31 canonical work pages

[1]

Estimates of worldwide burden of cancer in 2008: Globocan 2008,

J. Ferlay, H.-R. Shin, F. Bray, D. Forman, C. Mathers, and D. M. Parkin, “Estimates of worldwide burden of cancer in 2008: Globocan 2008,” International journal of cancer, vol. 127, no. 12, pp. 2893–2917, 2010

work page 2008
[2]

Focal liver lesions: comparison of dual-phase ct and multisequence multiplanar mr imaging including dynamic gadolinium enhancement,

R. C. Semelka, D. R. Martin, C. Balci, and T. Lance, “Focal liver lesions: comparison of dual-phase ct and multisequence multiplanar mr imaging including dynamic gadolinium enhancement,”Journal of Magnetic Resonance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine, vol. 13, no. 3, pp. 397– 401, 2001

work page 2001
[3]

Nephrogenic systemic fibrosis: suspected causative role of gadodiamide used for contrast-enhanced magnetic resonance imaging,

P. Marckmann, L. Skov, K. Rossen, A. Dupont, M. B. Damholt, J. G. Heaf, and H. S. Thomsen, “Nephrogenic systemic fibrosis: suspected causative role of gadodiamide used for contrast-enhanced magnetic resonance imaging,”Journal of the American Society of Nephrology, vol. 17, no. 9, pp. 2359–2362, 2006

work page 2006
[4]

Radiomics-based classification of hepatocellular carcinoma and hepatic haemangioma on precontrast magnetic resonance images,

J. Wu, A. Liu, J. Cui, A. Chen, Q. Song, and L. Xie, “Radiomics-based classification of hepatocellular carcinoma and hepatic haemangioma on precontrast magnetic resonance images,”BMC medical imaging, vol. 19, no. 1, p. 23, 2019

work page 2019
[5]

Synthesis of gadolinium-enhanced liver tumors on nonenhanced liver mr images using pixel-level graph reinforcement learning,

C. Xu, D. Zhang, J. Chong, B. Chen, and S. Li, “Synthesis of gadolinium-enhanced liver tumors on nonenhanced liver mr images using pixel-level graph reinforcement learning,”Medical image analysis, vol. 69, p. 101976, 2021

work page 2021
[6]

Tripartite-gan: Synthesizing liver contrast-enhanced mri to improve tumor detection,

J. Zhao, D. Li, Z. Kassam, J. Howey, J. Chong, B. Chen, and S. Li, “Tripartite-gan: Synthesizing liver contrast-enhanced mri to improve tumor detection,”Medical image analysis, vol. 63, p. 101667, 2020

work page 2020
[7]

Hepatobiliary versus extracellular mri contrast agents in hepatocellular carcinoma detection: hepatobiliary phase features in relation to disease-free survival,

D. K. Kim, C. An, Y . E. Chung, J.-Y . Choi, J. S. Lim, M.-S. Park, and M.-J. Kim, “Hepatobiliary versus extracellular mri contrast agents in hepatocellular carcinoma detection: hepatobiliary phase features in relation to disease-free survival,”Radiology, vol. 293, no. 3, pp. 594– 604, 2019

work page 2019
[8]

Quantitatively defining washout in hepatocellular carcinoma,

Y . I. Liu, L. K. Shin, R. B. Jeffrey, and A. Kamaya, “Quantitatively defining washout in hepatocellular carcinoma,”American Journal of Roentgenology, vol. 200, no. 1, pp. 84–89, 2013

work page 2013

Show all 40 references

[9]

Fgc2f-udiff: Frequency-guided and coarse-to-fine unified diffusion model for multi-modality missing mri synthesis,

X. Xiao, Q. V . Hu, and G. Wang, “Fgc2f-udiff: Frequency-guided and coarse-to-fine unified diffusion model for multi-modality missing mri synthesis,”IEEE Transactions on Computational Imaging, 2024

2024
[10]

Contrast-enhanced liver magnetic resonance image synthesis using gradient regularized multi-modal multi-discrimination sparse attention fusion gan,

C. Jiao, D. Ling, S. Bian, A. Vassantachart, K. Cheng, S. Mehta, D. Lock, Z. Zhu, M. Feng, H. Thomaset al., “Contrast-enhanced liver magnetic resonance image synthesis using gradient regularized multi-modal multi-discrimination sparse attention fusion gan,”Cancers, vol. 15, no...

2023
[11]

Contrast-enhanced mri synthesis from non-contrast mri using attention cyclegan,

T. Wang, Y . Lei, W. J. Curran, T. Liu, and X. Yang, “Contrast-enhanced mri synthesis from non-contrast mri using attention cyclegan,” inMed- ical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging, vol. 11600. SPIE, 2021, pp. 388–393

2021
[12]

Non-contrast ct liver segmentation using cyclegan data augmentation from contrast enhanced ct,

C. Song, B. He, H. Chen, S. Jia, X. Chen, and F. Jia, “Non-contrast ct liver segmentation using cyclegan data augmentation from contrast enhanced ct,” inInterpretable and Annotation-Efficient Learning for Medical Image Computing: Third International Workshop, iMIMIC 2020, Seco...

2020
[13]

United multi-task learning for abdominal contrast-enhanced ct synthesis through joint deformable registration,

L. Zhong, P. Huang, H. Shu, Y . Li, Y . Zhang, Q. Feng, Y . Wu, and W. Yang, “United multi-task learning for abdominal contrast-enhanced ct synthesis through joint deformable registration,”Computer Methods and Programs in Biomedicine, vol. 231, p. 107391, 2023

2023
[14]

Accurate segmentation of liver tumor from multi-modality non-contrast images using a dual-stream multi-level fusion framework,

C. Xu, X. Wu, B. Wang, J. Chen, Z. Gao, X. Liu, and H. Zhang, “Accurate segmentation of liver tumor from multi-modality non-contrast images using a dual-stream multi-level fusion framework,”Computer- ized Medical Imaging and Graphics, vol. 116, p. 102414, 2024

2024
[15]

Gan-based synthetic medical image augmentation for increased cnn performance in liver lesion classification,

M. Frid-Adar, I. Diamant, E. Klang, M. Amitai, J. Goldberger, and H. Greenspan, “Gan-based synthetic medical image augmentation for increased cnn performance in liver lesion classification,”Neurocomput- ing, vol. 321, pp. 321–331, 2018

2018
[16]

Evoprompting: Language models for code-level neural architecture search,

A. Chen, D. Dohan, and D. So, “Evoprompting: Language models for code-level neural architecture search,”Advances in neural information processing systems, vol. 36, pp. 7787–7817, 2023

2023
[17]

Au- toregressive sequence modeling for 3d medical image representation,

S. Wang, C. Wang, F. Gao, L. Su, F. Zhang, Y . Wang, and Y . Yu, “Au- toregressive sequence modeling for 3d medical image representation,” arXiv preprint arXiv:2409.08691, 2024

2024 arXiv
[18]

Morphology-preserving autoregressive 3d generative modelling of the brain,

P.-D. Tudosiu, W. H. L. Pinaya, M. S. Graham, P. Borges, V . Fer- nandez, D. Yang, J. Appleyard, G. Novati, D. Mehra, M. Vellaet al., “Morphology-preserving autoregressive 3d generative modelling of the brain,” inInternational Workshop on Simulation and Synthesis in Med- ical ...

2022
[19]

Cavm: Conditional autoregressive vision model for contrast-enhanced brain tumor mri synthesis,

L. Gui, C. Ye, and T. Yan, “Cavm: Conditional autoregressive vision model for contrast-enhanced brain tumor mri synthesis,” inInternational Conference on Medical Image Computing and Computer-Assisted Inter- vention. Springer, 2024, pp. 161–170

2024
[20]

Causal recurrent variational autoencoder for medical time series generation,

H. Li, S. Yu, and J. Principe, “Causal recurrent variational autoencoder for medical time series generation,” inProceedings of the AAAI confer- ence on artificial intelligence, vol. 37, no. 7, 2023, pp. 8562–8570

2023
[21]

Physics-driven autoregressive state space models for medical image reconstruction,

B. Kabas, F. Arslan, V . A. Nezhad, S. Ozturk, E. U. Saritas, and T. C ¸ ukur, “Physics-driven autoregressive state space models for medical image reconstruction,”arXiv preprint arXiv:2412.09331, 2024

2024 arXiv
[22]

Autoregressive medical image segmentation via next-scale mask prediction,

T. Chen, C. Wang, Z. Chen, and H. Shan, “Autoregressive medical image segmentation via next-scale mask prediction,”arXiv preprint arXiv:2502.20784, 2025

2025 arXiv
[23]

Deep learning with con- volutional neural network for differentiation of liver masses at dynamic contrast-enhanced ct: a preliminary study,

K. Yasaka, H. Akai, O. Abe, and S. Kiryu, “Deep learning with con- volutional neural network for differentiation of liver masses at dynamic contrast-enhanced ct: a preliminary study,”Radiology, vol. 286, no. 3, pp. 887–896, 2018

2018
[24]

Extending 2-d convolutional neural networks to 3-d for advancing deep learning cancer classification with application to mri liver tumor differentiation,

E. Trivizakis, G. C. Manikis, K. Nikiforaki, K. Drevelegas, M. Con- stantinides, A. Drevelegas, and K. Marias, “Extending 2-d convolutional neural networks to 3-d for advancing deep learning cancer classification with application to mri liver tumor differentiation,”IEEE journa...

2018
[25]

Synthetic data augmentation using gan for improved liver lesion clas- sification,

M. Frid-Adar, E. Klang, M. Amitai, J. Goldberger, and H. Greenspan, “Synthetic data augmentation using gan for improved liver lesion clas- sification,” in2018 IEEE 15th international symposium on biomedical imaging (ISBI 2018). IEEE, 2018, pp. 289–293. AUTHORet al.: TITLE 13

2018
[26]

Mix-domain contrastive learning for unpaired h&e-to-ihc stain translation,

S. Wang, Z. Zhang, H. Yan, M. Xu, and G. Wang, “Mix-domain contrastive learning for unpaired h&e-to-ihc stain translation,” in2024 IEEE International Conference on Image Processing (ICIP). IEEE, 2024, pp. 2982–2988

2024
[27]

Swin-unet: Unet-like pure transformer for medical image segmenta- tion,

H. Cao, Y . Wang, J. Chen, D. Jiang, X. Zhang, Q. Tian, and M. Wang, “Swin-unet: Unet-like pure transformer for medical image segmenta- tion,” inEuropean conference on computer vision. Springer, 2022, pp. 205–218

2022
[28]

Pixel recurrent neural networks,

A. Van Den Oord, N. Kalchbrenner, and K. Kavukcuoglu, “Pixel recurrent neural networks,” inInternational conference on machine learning. PMLR, 2016, pp. 1747–1756

2016
[29]

Diagnosis of hepatic tumors with texture analysis in nonenhanced computed tomography images,

Y .-L. Huang, J.-H. Chen, and W.-C. Shen, “Diagnosis of hepatic tumors with texture analysis in nonenhanced computed tomography images,” Academic radiology, vol. 13, no. 6, pp. 713–720, 2006

2006
[30]

Common-unique decomposition driven diffusion model for contrast- enhanced liver mr images multi-phase interconversion,

C. Xu, S. Tian, B. Wang, J. Zhang, K. Polat, A. Alhudhaif, and S. Li, “Common-unique decomposition driven diffusion model for contrast- enhanced liver mr images multi-phase interconversion,”IEEE Journal of Biomedical and Health Informatics, 2024

2024
[31]

Medical vision generalist: Unifying medical imaging tasks in context,

S. Ren, X. Huang, X. Li, J. Xiao, J. Mei, Z. Wang, A. Yuille, and Y . Zhou, “Medical vision generalist: Unifying medical imaging tasks in context,”arXiv preprint arXiv:2406.05565, 2024

2024 arXiv
[32]

U-net: Convolutional networks for biomedical image segmentation,

O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” inMedical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III

2015
[33]

Springer, 2015, pp. 234–241

2015
[34]

Transunet: Transformers make strong encoders for medical image segmentation,

J. Chen, Y . Lu, Q. Yu, X. Luo, E. Adeli, Y . Wang, L. Lu, A. L. Yuille, and Y . Zhou, “Transunet: Transformers make strong encoders for medical image segmentation,”arXiv preprint arXiv:2102.04306, 2021

2021 arXiv
[35]

Edge-aware multi-task network for integrating quantification segmentation and uncertainty prediction of liver tumor on multi-modality non-contrast mri,

X. Xiao, Q. V . Hu, and G. Wang, “Edge-aware multi-task network for integrating quantification segmentation and uncertainty prediction of liver tumor on multi-modality non-contrast mri,” inInternational con- ference on medical image computing and computer-assisted intervention...

2023
[36]

Uncertainty-guided and cross-modality attention network for liver tumor segmentation and quantification via integrating dynamic mri,

J. Zhao and S. Li, “Uncertainty-guided and cross-modality attention network for liver tumor segmentation and quantification via integrating dynamic mri,”Knowledge-Based Systems, vol. 310, p. 113021, 2025

2025
[37]

Very deep convolutional networks for large-scale image recognition,

K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,”arXiv preprint arXiv:1409.1556, 2014

2014 arXiv
[38]

Efficientnet: Rethinking model scaling for con- volutional neural networks,

M. Tan and Q. Le, “Efficientnet: Rethinking model scaling for con- volutional neural networks,” inInternational conference on machine learning. PMLR, 2019, pp. 6105–6114

2019
[39]

Medvit: a robust vision transformer for generalized medical image classification,

O. N. Manzari, H. Ahmadabadi, H. Kashiani, S. B. Shokouhi, and A. Ayatollahi, “Medvit: a robust vision transformer for generalized medical image classification,”Computers in biology and medicine, vol. 157, p. 106791, 2023

2023
[40]

Transmed: Transformers advance multi- modal medical image classification,

Y . Dai, Y . Gao, and F. Liu, “Transmed: Transformers advance multi- modal medical image classification,”Diagnostics, vol. 11, no. 8, p. 1384, 2021

2021

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Estimates of worldwide burden of cancer in 2008: Globocan 2008,

J. Ferlay, H.-R. Shin, F. Bray, D. Forman, C. Mathers, and D. M. Parkin, “Estimates of worldwide burden of cancer in 2008: Globocan 2008,” International journal of cancer, vol. 127, no. 12, pp. 2893–2917, 2010

work page 2008

[2] [2]

Focal liver lesions: comparison of dual-phase ct and multisequence multiplanar mr imaging including dynamic gadolinium enhancement,

R. C. Semelka, D. R. Martin, C. Balci, and T. Lance, “Focal liver lesions: comparison of dual-phase ct and multisequence multiplanar mr imaging including dynamic gadolinium enhancement,”Journal of Magnetic Resonance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine, vol. 13, no. 3, pp. 397– 401, 2001

work page 2001

[3] [3]

Nephrogenic systemic fibrosis: suspected causative role of gadodiamide used for contrast-enhanced magnetic resonance imaging,

P. Marckmann, L. Skov, K. Rossen, A. Dupont, M. B. Damholt, J. G. Heaf, and H. S. Thomsen, “Nephrogenic systemic fibrosis: suspected causative role of gadodiamide used for contrast-enhanced magnetic resonance imaging,”Journal of the American Society of Nephrology, vol. 17, no. 9, pp. 2359–2362, 2006

work page 2006

[4] [4]

Radiomics-based classification of hepatocellular carcinoma and hepatic haemangioma on precontrast magnetic resonance images,

J. Wu, A. Liu, J. Cui, A. Chen, Q. Song, and L. Xie, “Radiomics-based classification of hepatocellular carcinoma and hepatic haemangioma on precontrast magnetic resonance images,”BMC medical imaging, vol. 19, no. 1, p. 23, 2019

work page 2019

[5] [5]

Synthesis of gadolinium-enhanced liver tumors on nonenhanced liver mr images using pixel-level graph reinforcement learning,

C. Xu, D. Zhang, J. Chong, B. Chen, and S. Li, “Synthesis of gadolinium-enhanced liver tumors on nonenhanced liver mr images using pixel-level graph reinforcement learning,”Medical image analysis, vol. 69, p. 101976, 2021

work page 2021

[6] [6]

Tripartite-gan: Synthesizing liver contrast-enhanced mri to improve tumor detection,

J. Zhao, D. Li, Z. Kassam, J. Howey, J. Chong, B. Chen, and S. Li, “Tripartite-gan: Synthesizing liver contrast-enhanced mri to improve tumor detection,”Medical image analysis, vol. 63, p. 101667, 2020

work page 2020

[7] [7]

Hepatobiliary versus extracellular mri contrast agents in hepatocellular carcinoma detection: hepatobiliary phase features in relation to disease-free survival,

D. K. Kim, C. An, Y . E. Chung, J.-Y . Choi, J. S. Lim, M.-S. Park, and M.-J. Kim, “Hepatobiliary versus extracellular mri contrast agents in hepatocellular carcinoma detection: hepatobiliary phase features in relation to disease-free survival,”Radiology, vol. 293, no. 3, pp. 594– 604, 2019

work page 2019

[8] [8]

Quantitatively defining washout in hepatocellular carcinoma,

Y . I. Liu, L. K. Shin, R. B. Jeffrey, and A. Kamaya, “Quantitatively defining washout in hepatocellular carcinoma,”American Journal of Roentgenology, vol. 200, no. 1, pp. 84–89, 2013

work page 2013