Learning Segmentation from Radiology Reports

Alan L. Yuille; Andrea Cavalli; Jieneng Chen; Kang Wang; Pedro R. A. S. Bassi; Sergio Decherchi; Tianyu Lin; Wenxuan Li; Yang Yang; Zheren Zhu

REVIEW 3 major objections 4 minor 30 references

Learning Segmentation from Radiology Reports

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Radiology reports can directly supervise tumor segmentation, lifting F1 by up to 16 percentage points.

desk verdict Novel report-to-voxel supervision with a real external Dice gain, but the headline numbers are confounded by adding thousands of unlabeled CTs. read the letter →

arxiv 2507.05582 v1 pith:RVEEZGBP submitted 2025-07-08 eess.IV cs.CV

Pedro R. A. S. Bassi , Wenxuan Li , Jieneng Chen , Zheren Zhu , Tianyu Lin , Sergio Decherchi , Andrea Cavalli , Kang Wang

show 3 more authors

Yang Yang Alan L. Yuille Zongwei Zhou

This is my paper

classification eess.IVcs.CV

keywords tumorsegmentationradiologyreportsweaksupervisionlossfunctionCTimaginglargelanguagemodelpancreatickidney

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

Radiology reports, not just segmentation masks, can directly supervise tumor segmentation in CT scans. The paper proposes R-Super, a training paradigm that extracts tumor count, location, and diameters from free-text reports with a large language model and converts them into voxel-wise supervision through two new losses: Volume Loss and Ball Loss. Training with these report-derived losses alongside mask supervision raised F1 score by up to 16 percentage points over mask-only training, even when only 50 masks were available, and improved results on an external hospital not seen during training. The implication is that the huge archives of CT-report pairs in hospitals can be repurposed to scale segmentation AI.

What carries the argument

The central machinery is the pair of report-supervision losses. Volume Loss is applied as deep supervision: it sums predicted tumor probabilities inside pre-saved organ or sub-segment masks, multiplies by voxel volume to obtain a segmented volume, and penalizes the relative difference from the report-estimated volume with a 10% tolerance and a background cross-entropy term. Ball Loss is applied to the final layer: a fixed spherical kernel matching the reported tumor diameter is convolved over the output to find the highest-probability ball, the top-N voxels inside it (N derived from the report-estimated volume) are maximized, and all voxels not assigned to any reported tumor are minimized, repeating from largest to smallest tumor. Together these losses transform text-wise report information into voxel-wise labels.

What would settle it

Train R-Super on a dataset where ground-truth masks and paired reports are both available and compare report-estimated volumes against mask-derived volumes; if the losses improve segmentation only when the two agree, then on a cohort where reported diameters systematically underestimate true tumor volume (for example, single-slice measurements of elongated tumors), R-Super should underperform mask-only training. Alternatively, construct test reports that omit one of several visible tumors: if R-Super's false negatives match the omitted tumor, that confirms the losses enforce report completeness rather than true tumor burden.

Watch

Extended reading notes

Core claim

The central claim is that radiology reports contain enough quantitative detail—tumor count, organ or sub-segment location, and one-to-three diameters—to act as voxel-level training signal for tumor segmentation. The Volume Loss compares, per organ or sub-segment, the total segmented tumor volume against a volume estimated from report diameters using ball or ellipsoid formulas, with a 10% tolerance and a background penalty; the Ball Loss localizes each reported tumor with a spherical convolution whose kernel matches the reported diameter, then maximizes the top-N most probable voxels inside the highest-probability ball and minimizes all unassigned voxels, iterating from largest to smallest tumor. In internal and external validation, R-Super surpassed mask-only segmentation and four report-using baselines, with F1 gains up to 16 points with 344 masks, about 10 points with 50 masks, and 4.3 points with 1.7K masks, including gains for small (diameter at most 2 cm) and large tumors.

Load-bearing premise

The report-derived tumor attributes—count, location, and diameters converted to ellipsoid volumes—are accurate and complete enough to serve as voxel-level supervision, and the paper excludes the small fraction (under 11%) of reported tumors without size rather than modeling them.

Editorial extensions

If this is right

Hospitals can contribute existing CT-report archives to segmentation training without manual mask annotation, increasing data scale and diversity across centers, scanners, and contrast phases.
Tumor types with scarce public masks, including pancreatic and kidney tumors, can gain large F1 improvements from report supervision even when only 50 masks exist.
The method is architecture-agnostic and does not require perfect report-derived labels: the tolerance in Volume Loss and the dynamic mask refinement in Ball Loss accommodate imperfect diameter estimates.
Gains persist when mask counts are large (1.7K), so report supervision complements rather than replaces manual masks.
External validation on an unseen hospital shows the report-supervised model generalizes across institutions.

Reading between the lines

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

If report diameters are systematically recorded from a single axial slice, the ellipsoid volume estimates may be biased low, and the losses could steer the segmenter toward under-segmentation; a calibration study comparing report-derived volumes with mask-derived volumes on a paired dataset would test this.
The LLM extraction step is a potential bottleneck: prompt errors in diameter or location propagate directly into the losses, so end-to-end performance is upper-bounded by extraction accuracy; perturbing extracted attributes in a sensitivity analysis would quantify this.
The approach may transfer to other imaging modalities and lesion types where reports state dimensions and location, such as MRI or ultrasound, and to other anatomical regions beyond the abdomen.

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

Two things to know. The idea is genuinely new: report-extracted count, location, and diameters are turned into voxel-level losses, so the report supervises segmentation directly rather than via a side task. The external result is real: on JHH-Test, DSC for pancreatic tumors goes from 51 to 59 with 344 masks, and from 38 to 49 with 50 masks. That is a meaningful signal.

The soft spot is the attribution. R-Super trains on masks plus 2.2K/2.7K UCSF CT-report pairs, while the 'segmentation' baseline uses only masks. So the +16% F1 includes the effect of simply adding thousands of CTs, not just the report text. The paper does compare against Models Genesis, which also uses unlabeled CTs, and R-Super beats it, which suggests reports help beyond the extra images. But Models Genesis is a different pretraining scheme, not the same MedFormer fine-tuning recipe with the same CTs and no report losses. That leaves the central claim plausible but not fully isolated.

Other issues are secondary. UCSF-Test has no masks, so its F1 gains measure detection, not segmentation quality. The UCSF data is private, and the report-parsing details are described only briefly. The Volume and Ball losses assume diameters convert to ellipsoid volumes and that reported locations are accurate; the paper excludes <11% of reports without sizes rather than modeling them. No patient-level split is stated, so leakage within the private hospital data is a question the authors should answer.

Still, the paper is clearly written and the method is reproducible in principle: the losses are well-specified and the code is linked. The external Dice improvement is the strongest evidence and it is the right kind of evidence. I would send it to peer review, but the authors should add a control that trains on the UCSF CTs with the same architecture and masks but without report-derived losses, and report patient-level splits. If that control shows most of the gain remains, the claim is solid.

Referee Report

3 major / 4 minor

Summary. The paper introduces R-Super, a training loss that converts radiology reports into voxel-level supervision for tumor segmentation. An LLM extracts tumor count, organ/sub-segment location, and diameters from reports; the Volume Loss and Ball Loss enforce agreement between the predicted segmentation and tumor volumes, locations, and counts estimated from reports. Training combines public CT-mask pairs (AbdomenAtlas) with a large private UCSF CT-report dataset. The method is evaluated for pancreatic and kidney tumor segmentation, internally on UCSF-Test (no masks, so detection metrics) and externally on JHH-Test (with manual masks). R-Super reports F1 gains up to +16% internally and Dice gains of +8 to +11 points externally over mask-only training, and it outperforms several report-based and self-supervised baselines.

Significance. If the central claim holds, the contribution is significant: it offers a practical way to scale tumor segmentation by exploiting routinely available radiology reports, and it includes external validation with real masks. The paper provides ablations of the two proposed losses, comparisons with five state-of-the-art methods, and public code. The main caveats are that the few-mask result lacks a no-report control using the same extra CT volumes, and the internal headline metric measures tumor detection rather than voxel-level segmentation.

major comments (3)

[Section 3, Fig. 4, Table 2] The few-mask experiments (50 masks) compare R-Super trained with 2.2K/2.7K additional CT-report pairs against a segmentation baseline trained with only 50 masks. No baseline uses the same additional CT volumes without report supervision at the 50-mask setting; Models Genesis, the only method that learns from CTs without reports, is reported only at 344/1.7K masks. Consequently, the +9.7% F1 gain at 50 masks could be caused by the additional CT volumes rather than by the report-derived losses. Please add a no-report baseline at 50 masks using the same UCSF-Train CTs (e.g., Models Genesis pretraining or a pseudo-label semi-supervised method), or clearly restrict the few-mask claim to acknowledge this confound.
[Abstract, Section 3, Table 2] The headline '+16% F1' is measured on UCSF-Test, which has no segmentation masks, so F1 and AUC there evaluate tumor detection, not voxel-level segmentation. The abstract's phrase 'strongly improved tumor segmentation in internal and external validation' conflates detection and segmentation. Please rephrase the abstract to state that internal gains are in detection F1 while segmentation gains are demonstrated on JHH-Test via DSC/NSD, or obtain manual masks on a subset of UCSF-Test and report a segmentation metric there.
[Training R-Super, Table 1] The comparison to Models Genesis is intended to control for learning from CT volumes without reports, but the paper does not state explicitly whether the Models Genesis baseline was pre-trained on the same UCSF-Train CT volumes used by R-Super. Since the central claim is that report text, not merely extra CT data, drives the improvement, please specify the exact unlabeled CT data used for the Models Genesis baseline and, if it does not include UCSF-Train, add a control that does.

minor comments (4)

[Figure 1] The right panel of Figure 1 would be easier to read with labeled axes (for example, segmented volume V_s on the x-axis and loss value on the y-axis) and a clear indication of the tolerance region.
[Section 2.1 and Section 3] No sensitivity analysis is provided for the tolerance parameter τ, the stabilization constant E, or the relative loss weight of 0.1; a brief ablation or discussion of these choices would help readers understand the robustness of the method to hyperparameter settings.
[Table 2] The text reports that pancreas sub-segment masks improve R-Super and gives numerical results without them, but this comparison is not shown as a table row; consider adding a row or a supplementary table for this ablation.
[Section 2] The model name is written as 'LLama' in the text; it should be 'LLaMA' for consistency with the reference.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; external manual-mask validation independently supports the central claim.

full rationale

The R-Super derivation chain is not circular. Reports are parsed into tumor count, location, and diameters; Volume Loss (Eq. 2) and Ball Loss then penalize segmentations that disagree with these report-derived attributes, and the resulting model is evaluated on two test sets. The external JHH-Test uses manual segmentation masks and reports DSC/NSD/F1, so the claim that report supervision improves tumor segmentation is tested against independent anatomical ground truth rather than against the training objective. The internal UCSF-Test has no masks (as the table note states: 'DSC and NSD are not available in UCSF-Test, it does not have masks'), so its F1/AUC are report-based; this is a proxy limitation, not a circular identity, because the UCSF-Test reports are held out and the model must generalize to new patients. The cited RadGPT parser accuracy and organ nnU-Net are component tools from overlapping authors, but the external JHH-Test improvements empirically validate that the report-supervision signal is informative, so these self-citations are not load-bearing in a circular way. A separate concern, that the comparison lacks a no-report control on the added UCSF-Train CTs, is a confound about attributing gains to report text versus extra data; it is a correctness risk, not a by-construction equivalence.

Assumptions & free parameters 5 free parameters · 7 assumptions · 0 invented entities

The method introduces no new physical entities. Its load-bearing ingredients are hand-set hyperparameters, including tolerance, stabilization constant, loss weight, ball kernel width, and diameter completion rule, plus several domain assumptions about the reliability of report-derived tumor attributes and organ masks.

free parameters (5)

Volume loss tolerance tau = 10%
Set by hand in Eq. 3 to create a dead zone around report-derived volumes; not fitted to data.
Stabilization constant E = 500 mm^3
Added in Eq. 2 so the loss has a nonzero gradient when predicted volume is zero and the report says no tumor.
Weight of R-Super losses relative to segmentation loss = 0.1 vs 1.0
Chosen during fine-tuning to balance the influence of CT-mask and CT-report pairs.
Ball kernel Gaussian standard deviation = 0.75 times ball diameter
Specified in Section 2.2 to weight the center of the spherical kernel more heavily.
Third diameter estimate for two-diameter tumors = d3 = (d1 + d2)/2
Ad hoc way to complete ellipsoid volume when the report gives only two perpendicular diameters.

assumptions (7)

domain assumption Radiology report descriptions of tumor count, location, and size are accurate and complete enough to serve as voxel-level training targets.
The Volume and Ball losses in Sections 2.1 and 2.2 treat report-extracted attributes as ground truth for segmentation supervision.
domain assumption Tumor volume can be estimated from reported diameters using ball and ellipsoid formulas.
Section 2.1 converts one, two, or three diameters into volumes with d1^3*pi/6 or d1*d2*d3*pi/6; only a 10% tolerance absorbs the estimation error.
domain assumption Summing predicted tumor probabilities over organ voxels gives a usable estimate of segmented tumor volume.
Eq. 1 defines Vs,o as the probability-weighted voxel sum, which ignores spatial correlation and thresholding effects.
domain assumption The highest response of the spherical convolution identifies the true tumor center.
Ball Loss selects the highest-probability ball to place each reported tumor; if the model's initial probability mass is misplaced, the loss can reinforce the wrong location.
domain assumption Organ and organ sub-segment masks from a separately trained nnU-Net are accurate enough for tumor localization.
Eqs. 1 and 5 and the Ball Loss multiply predictions by organ masks, so organ segmentation errors propagate into the tumor losses.
domain assumption The LLM extraction of tumor attributes has 96% accuracy for presence and location.
Stated in Section 2 and cited from the authors' prior RadGPT work; no independent verification is provided in this paper.
ad hoc to paper Tumors mentioned in reports without size information can be ignored during training.
Section 2.2 excludes such organs (<11% of cases), which could remove difficult or incomplete supervision signals.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Learning Segmentation from Radiology Reports." pith.science (2026). https://pith.science/paper/RVEEZGBP

@misc{pith2026250705582,
  author       = {Pith},
  title        = {Pith review of: Learning Segmentation from Radiology Reports},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RVEEZGBP}},
  note         = {Machine review of arXiv:2507.05582}
}

read the original abstract

Tumor segmentation in CT scans is key for diagnosis, surgery, and prognosis, yet segmentation masks are scarce because their creation requires time and expertise. Public abdominal CT datasets have from dozens to a couple thousand tumor masks, but hospitals have hundreds of thousands of tumor CTs with radiology reports. Thus, leveraging reports to improve segmentation is key for scaling. In this paper, we propose a report-supervision loss (R-Super) that converts radiology reports into voxel-wise supervision for tumor segmentation AI. We created a dataset with 6,718 CT-Report pairs (from the UCSF Hospital), and merged it with public CT-Mask datasets (from AbdomenAtlas 2.0). We used our R-Super to train with these masks and reports, and strongly improved tumor segmentation in internal and external validation--F1 Score increased by up to 16% with respect to training with masks only. By leveraging readily available radiology reports to supplement scarce segmentation masks, R-Super strongly improves AI performance both when very few training masks are available (e.g., 50), and when many masks were available (e.g., 1.7K). Project: https://github.com/MrGiovanni/R-Super

Figures

Figures reproduced from arXiv: 2507.05582 by the authors.

**Figure 2.** The Ball Loss enforces segmented tumors to match detailed tumor [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** UCSF-Train has more pancreatic and kidney tumor CTs than any [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: By learning from reports, R-Super strongly surpassed 6 state-of-theart methods in internal (UCSF-Test) and external validation (JHH-Test). It surpassed standard segmentation (no report) by up to 16% in tumor detection F1- Score (green arrows). Each plot corresponds to…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

30 extracted references · 16 canonical work pages

[1]

arXiv preprint arXiv:2106.05735 (2021)

Antonelli, M., Reinke, A., Bakas, S., Farahani, K., Landman, B.A., Litjens, G., Menze, B., Ronneberger, O., Summers, R.M., van Ginneken, B., et al.: The medical segmentation decathlon. arXiv preprint arXiv:2106.05735 (2021)

arXiv 2021
[2]

Bassi, P.R., Li, W., Tang, Y., Isensee, F., Wang, Z., Chen, J., Chou, Y.C., Kirch- hoff, Y., Rokuss, M., Huang, Z., Ye, J., He, J., Wald, T., Ulrich, C., Baumgartner, M., Roy, S., Maier-Hein, K.H., Jaeger, P., Ye, Y., Xie, Y., Zhang, J., Chen, Z., Xia, Y., Xing, Z., Zhu, L., Sadegheih, Y., Bozorgpour, A., Kumari, P., Azad, R., Mer- hof, D., Shi, P., Ma, T...

work page 2024
[3]

arXiv preprint arXiv:2501.04678 (2025), https://github.com/ MrGiovanni/RadGPT

Bassi, P.R., Yavuz, M.C., Wang, K., Chen, X., Li, W., Decherchi, S., Cav- alli, A., Yang, Y., Yuille, A., Zhou, Z.: Radgpt: Constructing 3d image-text tumor datasets. arXiv preprint arXiv:2501.04678 (2025), https://github.com/ MrGiovanni/RadGPT

arXiv 2025
[4]

arXiv preprint arXiv:1901.04056 (2019)

Bilic, P., Christ, P.F., Vorontsov, E., Chlebus, G., Chen, H., Dou, Q., Fu, C.W., Han, X., Heng, P.A., Hesser, J., et al.: The liver tumor segmentation benchmark (lits). arXiv preprint arXiv:1901.04056 (2019)

arXiv 2019
[5]

Research Square pp

Blankemeier, L., Cohen, J.P., Kumar, A., Van Veen, D., Gardezi, S.J.S., Paschali, M., Chen, Z., Delbrouck, J.B., Reis, E., Truyts, C., et al.: Merlin: A vision language foundation model for 3d computed tomography. Research Square pp. rs–3 (2024)

2024
[6]

Radiology: Artificial Intelligence 5(5), e230031 (2023)

Bosma, J.S., Saha, A., Hosseinzadeh, M., Slootweg, I., de Rooij, M., Huisman, H.: Semisupervised learning with report-guided pseudo labels for deep learning–based prostate cancer detection using biparametric mri. Radiology: Artificial Intelligence 5(5), e230031 (2023)

work page 2023
[7]

In: 2019 IEEE 16th In- ternational Symposium on Biomedical Imaging (ISBI 2019)

Chen, E.Z., Dong, X., Li, X., Jiang, H., Rong, R., Wu, J.: Lesion attributes seg- mentation for melanoma detection with multi-task u-net. In: 2019 IEEE 16th In- ternational Symposium on Biomedical Imaging (ISBI 2019). pp. 485–488. IEEE (2019)

work page 2019
[8]

Machine Intelligence Research pp

Chou, Y.C., Li, B., Fan, D.P., Yuille, A., Zhou, Z.: Acquiring weak annotations for tumor localization in temporal and volumetric data. Machine Intelligence Research pp. 1–13 (2024),https://github.com/johnson111788/Drag-Drop

work page 2024

Show all 30 references

[9]

Chou,Y.C.,Zhou,Z.,Yuille,A.:Embracingmassivemedicaldata.In:International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 24–35. Springer (2024),https://github.com/MrGiovanni/OnlineLearning

2024
[10]

arXiv preprint arXiv:2407.21783 (2024)

Dubey, A., Jauhri, A., Pandey, A., Kadian, A., et al.: The llama 3 herd of models. arXiv preprint arXiv:2407.21783 (2024)

2024 arXiv
[11]

arXiv preprint arXiv:2203.00131 (2022)

Gao, Y., Zhou, M., Liu, D., Yan, Z., Zhang, S., Metaxas, D.N.: A data-scalable transformer for medical image segmentation: architecture, model efficiency, and benchmark. arXiv preprint arXiv:2203.00131 (2022)

2022 arXiv
[12]

Hamamci, I.E., Er, S., Menze, B.: Ct2rep: Automated radiology report generation for 3d medical imaging (2024),https://arxiv.org/abs/2403.06801

2024 arXiv
[13]

arXiv preprint arXiv:1904.00445 (2019) 12 P

Heller, N., Sathianathen, N., Kalapara, A., Walczak, E., Moore, K., Kaluzniak, H., Rosenberg, J., Blake, P., Rengel, Z., Oestreich, M., et al.: The kits19 challenge data: 300 kidney tumor cases with clinical context, ct semantic segmentations, and surgical outcomes. arXiv prep...

2019 arXiv
[14]

Nature Methods18(2), 203–211 (2021)

Isensee, F., Jaeger, P.F., Kohl, S.A., Petersen, J., Maier-Hein, K.H.: nnu-net: a self-configuring method for deep learning-based biomedical image segmentation. Nature Methods18(2), 203–211 (2021)

2021
[15]

arXiv preprint arXiv:2404.09556 (2024)

Isensee, F., Wald, T., Ulrich, C., Baumgartner, M., Roy, S., Maier-Hein, K., Jaeger, P.F.: nnu-net revisited: A call for rigorous validation in 3d medical image segmen- tation. arXiv preprint arXiv:2404.09556 (2024)

2024 arXiv
[16]

arXiv preprint arXiv:2307.13375 (2023)

Jaus, A., Seibold, C., Hermann, K., Walter, A., Giske, K., Haubold, J., Kleesiek, J., Stiefelhagen, R.: Towards unifying anatomy segmentation: Automated generation of a full-body ct dataset via knowledge aggregation and anatomical guidelines. arXiv preprint arXiv:2307.13375 (2023)

2023 arXiv
[17]

arXiv preprint arXiv:2206.08023 (2022)

Ji, Y., Bai, H., Yang, J., Ge, C., Zhu, Y., Zhang, R., Li, Z., Zhang, L., Ma, W., Wan, X., et al.: Amos: A large-scale abdominal multi-organ benchmark for versatile medical image segmentation. arXiv preprint arXiv:2206.08023 (2022)

2022 arXiv
[18]

arXiv preprint arXiv:2501.03410 (2025),https://github

Li, W., Bassi, P.R., Lin, T., Chou, Y.C., Zhou, X., Tang, Y., Isensee, F., Wang, K., Chen, Q., Xu, X., et al.: Scalemai: Accelerating the development of trusted datasets and ai models. arXiv preprint arXiv:2501.03410 (2025),https://github. com/MrGiovanni/ScaleMAI

2025
[19]

Medical Image Analysis p

Li, W., Qu, C., Chen, X., Bassi, P.R., Shi, Y., Lai, Y., Yu, Q., Xue, H., Chen, Y., Lin, X., et al.: Abdomenatlas: A large-scale, detailed-annotated, & multi- center dataset for efficient transfer learning and open algorithmic benchmark- ing. Medical Image Analysis p. 103285 (...

2024
[20]

Li, W., Yuille, A., Zhou, Z.: How well do supervised models transfer to 3d image segmentation? In: International Conference on Learning Representations (2024), https://github.com/MrGiovanni/SuPreM

2024
[21]

arXiv preprint arXiv:2507.01291 (2025),https://github.com/MrGiovanni/PanTS

Li, W., Zhou, X., Chen, Q., Lin, T., Bassi, P.R., Plotka, S., Cwikla, J.B., Chen, X., Ye, C., Zhu, Z., et al.: Pants: The pancreatic tumor segmentation dataset. arXiv preprint arXiv:2507.01291 (2025),https://github.com/MrGiovanni/PanTS

2025 arXiv
[22]

Ma, J., Zhang, Y., Gu, S., Ge, C., Wang, E., Zhou, Q., Huang, Z., Lyu, P., He, J., Wang, B.: Automatic organ and pan-cancer segmentation in abdomen ct: the flare 2023 challenge (2024),https://arxiv.org/abs/2408.12534

2024 arXiv
[23]

Cancer47(1), 207–214 (1981)

Miller, A., Hoogstraten, B., Staquet, M., Winkler, A.: Reporting results of cancer treatment. Cancer47(1), 207–214 (1981)

1981
[24]

Diagnostic and interventional imaging101(1), 35–44 (2020)

Park, S., Chu, L., Fishman, E., Yuille, A., Vogelstein, B., Kinzler, K., Horton, K., Hruban, R., Zinreich, E., Fouladi, D.F., et al.: Annotated normal ct data of the ab- domen for deep learning: Challenges and strategies for implementation. Diagnostic and interventional imagin...

2020
[25]

In: Conference on Neural Information Processing Systems

Qu, C., Zhang, T., Qiao, H., Liu, J., Tang, Y., Yuille, A., Zhou, Z.: Abdomenatlas- 8k: Annotating 8,000 abdominal ct volumes for multi-organ segmentation in three weeks. In: Conference on Neural Information Processing Systems. vol. 21 (2023), https://github.com/MrGiovanni/Abd...

2023
[26]

Radiology: Artificial Intelligence 5(5) (2023)

Wasserthal, J., Breit, H.C., Meyer, M.T., Pradella, M., Hinck, D., Sauter, A.W., Heye, T., Boll, D.T., Cyriac, J., Yang, S., et al.: Totalsegmentator: robust segmen- tation of 104 anatomic structures in ct images. Radiology: Artificial Intelligence 5(5) (2023)

2023
[27]

medRxiv (2022)

Xia, Y., Yu, Q., Chu, L., Kawamoto, S., Park, S., Liu, F., Chen, J., Zhu, Z., Li, B., Zhou, Z., et al.: The felix project: Deep networks to detect pancreatic neoplasms. medRxiv (2022)

2022
[28]

IEEE transactions on medical imaging40(6), 1618–1631 (2021)

Zhang, Y., Li, H., Du, J., Qin, J., Wang, T., Chen, Y., Liu, B., Gao, W., Ma, G., Lei, B.: 3d multi-attention guided multi-task learning network for automatic Learning Segmentation from Radiology Reports 13 gastric tumor segmentation and lymph node classification. IEEE transac...

2021
[29]

Zhou,Z.,Sodha,V.,Pang,J.,Gotway,M.B.,Liang,J.:Modelsgenesis.MedicalIm- ageAnalysis 67,101840(2021), https://github.com/MrGiovanni/ModelsGenesis

2021
[30]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention

Zhou, Z., Sodha, V., Siddiquee, M.M.R., Feng, R., Tajbakhsh, N., Gotway, M.B., Liang, J.: Models genesis: Generic autodidactic models for 3d medical image analysis. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 384–393. Springe...

2019

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

arXiv preprint arXiv:2106.05735 (2021)

Antonelli, M., Reinke, A., Bakas, S., Farahani, K., Landman, B.A., Litjens, G., Menze, B., Ronneberger, O., Summers, R.M., van Ginneken, B., et al.: The medical segmentation decathlon. arXiv preprint arXiv:2106.05735 (2021)

arXiv 2021

[2] [2]

Bassi, P.R., Li, W., Tang, Y., Isensee, F., Wang, Z., Chen, J., Chou, Y.C., Kirch- hoff, Y., Rokuss, M., Huang, Z., Ye, J., He, J., Wald, T., Ulrich, C., Baumgartner, M., Roy, S., Maier-Hein, K.H., Jaeger, P., Ye, Y., Xie, Y., Zhang, J., Chen, Z., Xia, Y., Xing, Z., Zhu, L., Sadegheih, Y., Bozorgpour, A., Kumari, P., Azad, R., Mer- hof, D., Shi, P., Ma, T...

work page 2024

[3] [3]

arXiv preprint arXiv:2501.04678 (2025), https://github.com/ MrGiovanni/RadGPT

Bassi, P.R., Yavuz, M.C., Wang, K., Chen, X., Li, W., Decherchi, S., Cav- alli, A., Yang, Y., Yuille, A., Zhou, Z.: Radgpt: Constructing 3d image-text tumor datasets. arXiv preprint arXiv:2501.04678 (2025), https://github.com/ MrGiovanni/RadGPT

arXiv 2025

[4] [4]

arXiv preprint arXiv:1901.04056 (2019)

Bilic, P., Christ, P.F., Vorontsov, E., Chlebus, G., Chen, H., Dou, Q., Fu, C.W., Han, X., Heng, P.A., Hesser, J., et al.: The liver tumor segmentation benchmark (lits). arXiv preprint arXiv:1901.04056 (2019)

arXiv 2019

[5] [5]

Research Square pp

Blankemeier, L., Cohen, J.P., Kumar, A., Van Veen, D., Gardezi, S.J.S., Paschali, M., Chen, Z., Delbrouck, J.B., Reis, E., Truyts, C., et al.: Merlin: A vision language foundation model for 3d computed tomography. Research Square pp. rs–3 (2024)

2024

[6] [6]

Radiology: Artificial Intelligence 5(5), e230031 (2023)

Bosma, J.S., Saha, A., Hosseinzadeh, M., Slootweg, I., de Rooij, M., Huisman, H.: Semisupervised learning with report-guided pseudo labels for deep learning–based prostate cancer detection using biparametric mri. Radiology: Artificial Intelligence 5(5), e230031 (2023)

work page 2023

[7] [7]

In: 2019 IEEE 16th In- ternational Symposium on Biomedical Imaging (ISBI 2019)

Chen, E.Z., Dong, X., Li, X., Jiang, H., Rong, R., Wu, J.: Lesion attributes seg- mentation for melanoma detection with multi-task u-net. In: 2019 IEEE 16th In- ternational Symposium on Biomedical Imaging (ISBI 2019). pp. 485–488. IEEE (2019)

work page 2019

[8] [8]

Machine Intelligence Research pp

Chou, Y.C., Li, B., Fan, D.P., Yuille, A., Zhou, Z.: Acquiring weak annotations for tumor localization in temporal and volumetric data. Machine Intelligence Research pp. 1–13 (2024),https://github.com/johnson111788/Drag-Drop

work page 2024