Pith. sign in

REVIEW 2 major objections 4 minor 46 references

An open-source AI model segments whole-body FDG-PET/CT cancer lesions across multiple cancer types with fewer false positives than public benchmarks and approaches expert agreement.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review

2026-07-11 22:55 UTC pith:XLTCM72E

load-bearing objection Solid open multi-cancer FDG-PET/CT lesion segmenter with real external head-to-heads and fewer false positives; novelty is mostly data scale and release, not architecture. the 2 major comments →

arxiv 2607.03931 v1 pith:XLTCM72E submitted 2026-07-04 eess.IV cs.CVphysics.med-ph

GLOW-FDG: Generalized cancer LesiOn Whole-body segmentation model for ¹⁸F-FDG-PET/CT

Maksym Fritsak , Maximilian Rokuss , Hubert S. Gabry\'s , Yannick Kirchhoff , Benjamin Hamm , Sebastian M. Christ , Nicolas Martz , Isabelle Opitz
show 5 more authors
Rolf Stahel Martin H\"ullner Matthias Guckenberger Klaus Maier-Hein Stephanie Tanadini-Lang
This is my paper
classification eess.IV cs.CVphysics.med-ph
keywords FDG-PET/CTwhole-body lesion segmentationdeep learningtotal tumor burdentotal lesion glycolysisfalse-positive reductionmulti-cancer external validationopen-source model
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Manual outlining of cancer on whole-body FDG-PET/CT is slow and variable, so large-scale use of tumor burden biomarkers has stayed limited. This paper introduces GLOW-FDG, an open model trained on 1,563 multi-cancer scans and tested on 185 scans from other institutions covering breast, lung, head and neck, and melanoma. Across those cohorts it beats three public benchmark models on patient- and lesion-level detection, mainly by cutting false-positive lesions while keeping high recall. Total tumor burden and total lesion glycolysis stay close to human references with high agreement scores, and on a melanoma subset the model’s detection and overlap sit near the range of two radiation oncologists reading the same cases. The authors release weights and code so others can use automated pre-segmentation and biomarker extraction without fixed SUV thresholds.

Core claim

GLOW-FDG, trained on a diverse multi-cancer FDG-PET/CT corpus of 1,563 scans and evaluated on 185 external scans, consistently delivers the highest patient- and lesion-wise detection F1 among compared public models by raising precision (fewer false positives) while holding high recall, with robust total tumor burden and total lesion glycolysis quantification and performance that approaches inter-observer variability between expert readers on a metastatic melanoma subset.

What carries the argument

GLOW-FDG: a dual-headed residual U-Net (lesion head plus organ-supervision head) trained with PET–CT misalignment augmentation and multi-dataset pretraining, then fine-tuned so physiologic high-uptake organs are learned separately from pathology, reducing false positives without fixed SUV thresholds.

Load-bearing premise

That the five external cohorts—two used only for detection after partial re-check, one an institutional train/validation split, plus a ten-case dual-reader melanoma set—are enough to claim generalizable multi-cancer whole-body performance across scanners and disease types not fully represented in training.

What would settle it

Run the released model on a fully independent multi-center multi-cancer FDG-PET/CT set with complete expert lesion masks (including ultra-low-dose or scanners/protocols never seen in training) and check whether patient- and lesion-wise F1 and TTB/TLG agreement still beat the same public benchmarks and stay inside the dual-reader range.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. GLOW-FDG is an open-source dual-head ResEncL nnU-Net for whole-body FDG-PET/CT cancer lesion segmentation, pretrained MultiTalent-style then fine-tuned on 1,563 multi-cancer scans with organ supervision and PET–CT misalignment augmentation. On 185 external scans across breast, nonmetastatic and oligometastatic lung, head and neck, and metastatic melanoma, it reports the highest patient- and lesion-wise detection F1 among three public baselines (AutoPET DKFZ, AutoPET IKIM, onlyPET), mainly by higher precision (fewer FPs) at high recall, with strong TTB/TLG ICC and Dice, and performance near dual-reader variability on a 10-case melanoma subset.

Significance. If the external gains hold, this is a practically useful contribution: a released multi-cancer FDG-PET/CT lesion model with public weights/code, head-to-head comparison against recent AutoPET-class baselines, and clinically oriented metrics (lesion F1, TTB/TLG RPD/ARPD/ICC) rather than Dice alone. Strengths include open release, multi-institution external testing, bootstrap CIs, TP/FN/FP characterization, and an explicit inter-observer reference. The work addresses a real gap between challenge-trained models and broader multi-cancer whole-body use.

major comments (2)
  1. §4.3–4.4 and Results §2.1: Two of five external cohorts (QIN-Breast, ACRIN-NSCLC) support only detection after BAMF mask re-check, not full independent human re-segmentation; SINERGIA melanoma is a same-institution train/val split. The abstract’s “185 external scans from independent institutions” and “generalizable” framing should be tightened so claims rest primarily on fully human-segmented external cohorts (CHESS, HECKTOR-USZ, melanoma) and detection-only cohorts are labeled as such.
  2. §2.5 and §4.8: Inter-observer context is limited to 10 metastatic melanoma cases with deliberately different PET- vs CT-centric styles. The claim that performance “approached the variability observed between expert radiation oncologists” is directionally useful but over-extended for multi-cancer clinical-grade generalization; restrict or qualify this comparison and avoid treating it as a multi-reader multi-cancer standard.
minor comments (4)
  1. Figure 3 caption: clarify that breast and nonmetastatic lung Dice are omitted because only detection labels were used after BAMF re-check.
  2. Tables 7–9 vs 10–12: numbering and cross-references to TP/FN/FP distributions are slightly inconsistent in the text; align table numbers and in-text citations.
  3. Discussion: ultra-low-dose and non-FDG limitations are appropriately noted; a short explicit statement on scanner/protocol diversity in the external set would help readers bound expected domain shift.
  4. Minor typos and formatting: “publically avaliable,” “adress,” mixed SU V BW notation, and a few broken hyphenations (e.g., “H¨ ullner”) should be cleaned in production.

Circularity Check

0 steps flagged

No significant circularity: empirical supervised segmentation with standard held-out metrics; minor self-citations of design components are not load-bearing for the performance claims.

full rationale

GLOW-FDG is an empirical deep-learning segmentation paper. The central claims (higher patient-/lesion-wise F1 via better precision, robust TTB/TLG ICC, Dice near dual-reader variability) rest on direct evaluation of a trained nnU-Net ResEncL model against three public baselines on five external cohorts using standard detection (precision/recall/F1) and segmentation (Dice, RPD, ARPD, ICC) metrics. Training losses (Dice + CE on lesion and organ heads), MultiTalent-style pretraining, misalignment augmentation, and organ supervision are design choices that do not algebraically force the reported external F1 or ICC values. Benchmarks include the authors’ own AutoPET DKFZ model, but that is an independent public comparator whose weights are fixed; the head-to-head numbers are new measurements, not re-statements of prior self-results. No uniqueness theorem, fitted parameter renamed as prediction, or definitional identity appears. The only minor self-referential element is reuse of prior architectural motifs (MultiTalent, misalignment aug) from overlapping authors; these do not underwrite the external validation numbers. Score 1 reflects that ordinary self-citation of methods, not circular derivation of the claimed superiority.

Axiom & Free-Parameter Ledger

5 free parameters · 5 axioms · 1 invented entities

This is an empirical deep-learning methods paper. Load-bearing content is training data curation, architecture/training hyperparameters, and the assumption that FDG-visible lesions plus organ multitask labels yield generalizable segmentations. No new physical entities; free parameters are standard ML knobs and design choices rather than fitted physical constants.

free parameters (5)
  • Pretraining epochs / patch / batch (4000 epochs, 192^3, batch 24)
    Hand-chosen MultiTalent-style pretraining schedule that shapes the foundation features before FDG fine-tuning.
  • Fine-tuning schedule (1500 epochs, SGD lr 1e-2, Nesterov 0.99, batch 3)
    Optimization hyperparameters that control final lesion/organ performance; not derived from theory.
  • Equal Dice+CE weights on lesion and organ heads
    Ad hoc multitask loss balance that trades lesion accuracy against false-positive control.
  • Misalignment augmentation bounds (±5° rotation, ≤2 voxel xy translation)
    Chosen rigid perturbation range claimed to improve small-lesion robustness; not learned from first principles.
  • Organ set for auxiliary supervision
    Hand-selected high-uptake organs (spleen, kidneys, liver, bladder, lung, brain, heart, stomach, prostate, salivary glands) define the multitask prior.
axioms (5)
  • domain assumption Clinically relevant lesions for this task are those visually detectable on FDG-PET; PET-invisible disease is out of scope.
    Stated in Discussion and enforced by manual review of training masks; defines what success means.
  • ad hoc to paper Auxiliary organ segmentation of physiologic FDG uptake reduces false-positive lesion calls and improves generalization.
    Core design hypothesis of the dual-head model; supported by lower FP rates vs baselines but not independently proven as the sole cause.
  • domain assumption nnU-Net ResEncL with residual blocks is an adequate capacity/architecture for whole-body PET/CT lesion segmentation.
    Inherited from AutoPET-era practice; paper does not re-derive architecture optimality.
  • domain assumption External cohort labels (after manual review / re-evaluation) are sufficiently accurate ground truth for detection and segmentation metrics.
    All F1/Dice/ICC claims rest on these annotations despite known inter-observer variability.
  • domain assumption SUV_BW conversion and intensity checks make multi-vendor PET intensities comparable enough for joint training.
    Preprocessing section; vendor-specific conversion via Z-Rad when DICOM available.
invented entities (1)
  • GLOW-FDG model (trained dual-head ResEncL weights) independent evidence
    purpose: Provide open whole-body multi-cancer FDG-PET/CT lesion segmentation and biomarker extraction.
    The named artifact is the paper’s product; independent evidence is the released weights and external metrics, not a new physical object.

reviewed 2026-07-11 · how reviews work

0 comments
Cite this review

Pith. "Pith review of GLOW-FDG: Generalized cancer LesiOn Whole-body segmentation model for $^{18}$F-FDG-PET/CT." pith.science (2026). https://pith.science/paper/XLTCM72E

@misc{pith2026260703931,
  author       = {Pith},
  title        = {Pith review of: GLOW-FDG: Generalized cancer LesiOn Whole-body segmentation model for $^18$F-FDG-PET/CT},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XLTCM72E}},
  note         = {Machine review of arXiv:2607.03931}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Whole-body fluorodeoxyglucose positron emission tomography combined with computed tomography is widely used in cancer care, but manual lesion delineation is slow, subjective, and difficult to scale. We present GLOW-FDG, an open-source artificial intelligence model for whole-body cancer lesion segmentation in fluorodeoxyglucose positron emission tomography and computed tomography. The model was trained on 1,563 scans spanning multiple cancer types and evaluated on 185 external scans from independent institutions. Across breast cancer, nonmetastatic and oligometastatic lung cancer, head and neck cancer, and metastatic melanoma, GLOW-FDG consistently outperformed publicly available benchmark models in lesion detection, while reducing false positives and maintaining strong segmentation accuracy. Quantification of total tumor burden and total lesion glycolysis was robust across cohorts, and performance approached the variability observed between expert radiation oncologists. These results support GLOW-FDG as a generalizable tool for automated cancer segmentation and quantitative imaging biomarker extraction in whole-body imaging.

Figures

Figures reproduced from arXiv: 2607.03931 by Benjamin Hamm, Hubert S. Gabry\'s, Isabelle Opitz, Klaus Maier-Hein, Maksym Fritsak, Martin H\"ullner, Matthias Guckenberger, Maximilian Rokuss, Nicolas Martz, Rolf Stahel, Sebastian M. Christ, Stephanie Tanadini-Lang, Yannick Kirchhoff.

Figure 1
Figure 1. Figure 1: Overview of the GLOW-FDG training pipeline. The model first undergoes large-scale pretraining using a [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of validation, benchmarking, and analysis workflow. Five independent FDG-PET/CT valida [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: A) Patient-wise F1 score and B) Dice similarity score. Nonmetastatic lung cancer and breast cancer [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: GLOW-FDG: lesion-wise Dice. Only human-segmented datasets were used for Dice score calculation. [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Whole-body oligometastatic lung cancer: AI–GT comparison of lesion-wise volume and glycolysis. [PITH_FULL_IMAGE:figures/full_fig_p018_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Head and neck cancer: AI–GT comparison of lesion-wise volume and glycolysis. [PITH_FULL_IMAGE:figures/full_fig_p019_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Whole-body metastatic melanoma: AI–GT comparison of lesion-wise volume and glycolysis. [PITH_FULL_IMAGE:figures/full_fig_p021_7.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

46 extracted references · 1 linked inside Pith

  1. [1]

    Role of fdg-pet scans in staging, response assessment, and follow-up care for non-small cell lung cancer.Frontiers in Oncology, 2:208, 2012

    John Cuaron, Mark Dunphy, and Andreas Rimner. Role of fdg-pet scans in staging, response assessment, and follow-up care for non-small cell lung cancer.Frontiers in Oncology, 2:208, 2012

    2012

  2. [2]

    Fletcher, Benjamin Djulbegovic, Heloisa P

    James W. Fletcher, Benjamin Djulbegovic, Heloisa P. Soares, Barry A. Siegel, Val J. Lowe, Gary H. Lyman, R. Edward Coleman, Richard Wahl, John Christopher Paschold, Norbert Avril, Lawrence H. Einhorn, W. War- ren Suh, David Samson, Dominique Delbeke, Mark Gorman, and Anthony F. Shields. Recommendations on the use of 18f-fdg pet in oncology.Journal of Nucl...

    2008

  3. [3]

    Wahl, Heather Jacene, Yvette Kasamon, and Martin A

    Richard L. Wahl, Heather Jacene, Yvette Kasamon, and Martin A. Lodge. From recist to percist: Evolving considerations for pet response criteria in solid tumors.Journal of Nuclear Medicine, 50(Suppl 1):122S–150S, May 2009

    2009

  4. [4]

    A role of fdg-pet/ct for response evaluation in metastatic breast cancer?Seminars in Nuclear Medicine, 52(5):520–530, Sep 2022

    Malene Grubbe Hildebrandt, Mohammad Naghavi-Behzad, and Marianne Vogsen. A role of fdg-pet/ct for response evaluation in metastatic breast cancer?Seminars in Nuclear Medicine, 52(5):520–530, Sep 2022

    2022

  5. [5]

    Edmund Kim, and Dong Soo Lee

    Kyoungjune Pak, Gi Jeong Cheon, Hyun-Yeol Nam, Seong-Jang Kim, Keon Wook Kang, June-Key Chung, E. Edmund Kim, and Dong Soo Lee. Prognostic value of metabolic tumor volume and total lesion glycolysis in head and neck cancer: a systematic review and meta-analysis.Journal of Nuclear Medicine, 55(6):884–890, Jun 2014

    2014

  6. [6]

    (18)f-fdg pet dissemination features in diffuse large b-cell lymphoma are predictive of outcome.Journal of Nuclear Medicine, 61(1):40–45, Jan 2020

    Anne-S´ egol` ene Cottereau, Christophe Nioche, Anne-Sophie Dirand, J´ erˆ ome Clerc, Franck Morschhauser, Olivier Casasnovas, Michel Meignan, and Ir` ene Buvat. (18)f-fdg pet dissemination features in diffuse large b-cell lymphoma are predictive of outcome.Journal of Nuclear Medicine, 61(1):40–45, Jan 2020

    2020

  7. [7]

    Quantitative fdg pet assessment for oncology therapy.Cancers, 13(4):869, Feb 2021

    Kenji Hirata and Nagara Tamaki. Quantitative fdg pet assessment for oncology therapy.Cancers, 13(4):869, Feb 2021

    2021

  8. [8]

    Use of fdg pet for staging and re-staging of head and neck squamous cell carcinoma.Cancers, 17(19):3140, Sep 2025

    Charles Marcus. Use of fdg pet for staging and re-staging of head and neck squamous cell carcinoma.Cancers, 17(19):3140, Sep 2025

    2025

  9. [9]

    The first miccai challenge on pet tumor segmentation.Medical Image Analysis, 44:177–195, 2018

    Mathieu Hatt, Baptiste Laurent, Anouar Ouahabi, Hadi Fayad, Shan Tan, Laquan Li, Wei Lu, Vincent Jaouen, Clovis Tauber, Jakub Czakon, Filip Drapejkowski, Witold Dyrka, Sorina Camarasu-Pop, Fr´ ed´ eric Cervenansky, Pascal Girard, Tristan Glatard, Michael Kain, Yao Yao, Christian Barillot, Assen Kirov, and Dimitris Visvikis. The first miccai challenge on p...

    2018

  10. [10]

    Totalsegmentator: robust segmentation of 104 anatomic structures in CT images.Radiology: Artificial Intelligence, 5(5), 2023

    Jakob Wasserthal, Hanns-Christian Breit, Manfred T Meyer, Maurice Pradella, Daniel Hinck, Alexander W Sauter, Tobias Heye, Daniel T Boll, Joshy Cyriac, Shan Yang, et al. Totalsegmentator: robust segmentation of 104 anatomic structures in CT images.Radiology: Artificial Intelligence, 5(5), 2023

    2023

  11. [11]

    Christ, Ezequiel de la Rosa, Julian Deseoe, Robert Graf, Hendrik M¨ oller, Anjany Sekuboyina, Jan C

    Murong Xu, Tamaz Amiranashvili, Fernando Navarro, Maksym Fritsak, Ibrahim Ethem Hamamci, Suprosanna Shit, Bastian Wittmann, Sezgin Er, Sebastian M. Christ, Ezequiel de la Rosa, Julian Deseoe, Robert Graf, Hendrik M¨ oller, Anjany Sekuboyina, Jan C. Peeken, Sven Becker, Giulia Baldini, Johannes Haubold, Felix Nensa, Ren´ e Hosch, Nikhil Mirajkar, Saad Khal...

    2025

  12. [12]

    John Lucido, Todd A

    J. John Lucido, Todd A. DeWees, Todd R. Leavitt, Aman Anand, Chris J. Beltran, Mark D. Brooke, Justine R. Buroker, Robert L. Foote, Olivia R. Foss, Angela M. Gleason, Teresa L. Hodge, C´ ıan O. Hughes, Ashley E. Hunzeker, Nadia N. Laack, Tamra K. Lenz, Michelle Livne, Megumi Morigami, Douglas J. Moseley, Lisa M. Undahl, Yojan Patel, Erik J. Tryggestad, Me...

    2023

  13. [13]

    Gabry´ s, Simon Burgermeister, Daniel Abler, Maksym Fritsak, Maiwand Ahmad- sei, Ciro Franzese, Adrien Depeursinge, Michel A

    Lorenzo Lo Faro, Hubert S. Gabry´ s, Simon Burgermeister, Daniel Abler, Maksym Fritsak, Maiwand Ahmad- sei, Ciro Franzese, Adrien Depeursinge, Michel A. Cuendet, Stephanie Tanadini-Lang, Panagiotis Balermpas, Marta Scorsetti, Matthias Guckenberger, and Sebastian M. Christ. Deep learning [(18)f]-fdg-pet/ct-based al- gorithm for tumor burden estimation in m...

    2026

  14. [14]

    F. Yang, G. Simpson, L. Young, J. Ford, N. Dogan, and L. Wang. Impact of contouring variability on oncological pet radiomics features in the lung.Scientific Reports, 10(1):369, Jan 2020

    2020

  15. [15]

    A whole-body fdg-pet/ct dataset with manually annotated tumor lesions (fdg-pet-ct-lesions), 2022

    Sergios Gatidis and Thomas Kuestner. A whole-body fdg-pet/ct dataset with manually annotated tumor lesions (fdg-pet-ct-lesions), 2022

    2022

  16. [16]

    Naser, Clifton D

    Valentin Oreiller, Vincent Andrearczyk, Mario Jreige, Sarah Boughdad, Hesham Elhalawani, Joel Castelli, Martin Valli` eres, Simeng Zhu, Juanying Xie, Ying Peng, Andrei Iantsen, Mathieu Hatt, Yading Yuan, Jun Ma, Xiaoping Yang, Chinmay Rao, Suraj Pai, Kanchan Ghimire, Xue Feng, Mohamed A. Naser, Clifton D. Fuller, Fereshteh Yousefirizi, Arman Rahmim, Huai ...

    2022

  17. [17]

    Is automatic tumor segmentation on whole-body (18)f-fdg pet images a clinical reality?Journal of Nuclear Medicine, 65(7):995–997, 2024

    Lalith Kumar Shiyam Sundar and Thomas Beyer. Is automatic tumor segmentation on whole-body (18)f-fdg pet images a clinical reality?Journal of Nuclear Medicine, 65(7):995–997, 2024

    2024

  18. [18]

    Maier-Hein, and Fabian Isensee

    Maximilian Rokuss, Balint Kovacs, Yannick Kirchhoff, Shuhan Xiao, Constantin Ulrich, Klaus H. Maier-Hein, and Fabian Isensee. From fdg to psma: A hitchhiker’s guide to multitracer, multicenter lesion segmentation in pet/ct imaging, 2024

    2024

  19. [19]

    Autopet iii challenge: Incorporating anatomical knowledge into nnunet for lesion segmentation in pet/ct, 2024

    Hamza Kalisch, Fabian H¨ orst, Ken Herrmann, Jens Kleesiek, and Constantin Seibold. Autopet iii challenge: Incorporating anatomical knowledge into nnunet for lesion segmentation in pet/ct, 2024

    2024

  20. [20]

    Rotkopf, Heinz-Peter Schlemmer, and Klaus Maier-Hein

    Maximilian Rokuss, Yannick Kirchhoff, Seval Akbal, Balint Kovacs, Saikat Roy, Constantin Ulrich, Tassilo Wald, Lukas T. Rotkopf, Heinz-Peter Schlemmer, and Klaus Maier-Hein. Lesionlocator: Zero-shot universal tumor segmentation and tracking in 3d whole-body imaging. InProceedings of the Computer Vision and Pattern Recognition Conference (CVPR), pages 3087...

    2025

  21. [21]

    nninteractive: Redefining 3d promptable segmentation, 2025

    Fabian Isensee, Maximilian Rokuss, Lars Kr¨ amer, Stefan Dinkelacker, Ashis Ravindran, Florian Stritzke, Benjamin Hamm, Tassilo Wald, Moritz Langenberg, Constantin Ulrich, Jonathan Deissler, Ralf Floca, and Klaus Maier-Hein. nninteractive: Redefining 3d promptable segmentation, 2025

    2025

  22. [22]

    Voxtell: Free-text promptable universal 3d medical image segmentation, 2025

    Maximilian Rokuss, Moritz Langenberg, Yannick Kirchhoff, Fabian Isensee, Benjamin Hamm, Constantin Ulrich, Sebastian Regnery, Lukas Bauer, Efthimios Katsigiannopulos, Tobias Norajitra, and Klaus Maier-Hein. Voxtell: Free-text promptable universal 3d medical image segmentation, 2025

    2025

  23. [23]

    Glance and focus reinforcement for pan-cancer screening

    Linshan Wu, Jiaxin Zhuang, and Hao Chen. Glance and focus reinforcement for pan-cancer screening. In ICLR, 2026

    2026

  24. [24]

    A synthetic data-driven radiology foundation model for pan-tumor clinical diagnosis, 2026

    Wenhui Lei, Hanyu Chen, Zitian Zhang, Luyang Luo, Qiong Xiao, Yannian Gu, Peng Gao, Yankai Jiang, Ci Wang, Guangtao Wu, Tongjia Xu, Yingjie Zhang, Pranav Rajpurkar, Xiaofan Zhang, Shaoting Zhang, and Zhenning Wang. A synthetic data-driven radiology foundation model for pan-tumor clinical diagnosis, 2026

    2026

  25. [25]

    Stahel, Martin W

    Maksym Fritsak, Hubert Gabrys, Maximilian Rokuss, Sebastian Christ, Nicolas Martz, Alina Paunoiu, Isabelle Opitz, Rolf A. Stahel, Martin W. Huellner, Matthias Guckenberger, and Stephanie Tanadini-Lang. Generalizing beyond training data: Performance assessment of the best autopet iii model on diverse cancer types, 2026. Available at SSRN

    2026

  26. [26]

    Lesion-segmentation, n.d

    Yazdan Salimi. Lesion-segmentation, n.d

  27. [27]

    Fully automated whole-body tumor segmentation on pet/ct using deep transfer learning.Journal of Nuclear Medicine, 65(supplement 2):241979–241979, 2024

    Kevin Leung, Steven Rowe, Moe Sadaghiani, Jeffrey Leal, Esther Mena, Peter Choyke, Yong Du, and Martin Pomper. Fully automated whole-body tumor segmentation on pet/ct using deep transfer learning.Journal of Nuclear Medicine, 65(supplement 2):241979–241979, 2024

    2024

  28. [28]

    Whole-body tumor segmentation from pet/ct images using a two-stage cascaded neural network with camouflaged object detection mechanisms.Medical Physics, 50(10):6151–6162, 2023

    Jiangping He, Yangjie Zhang, Maggie Chung, Michael Wang, Kun Wang, Yan Ma, Xiaoyang Ding, Qiang Li, and Yonglin Pu. Whole-body tumor segmentation from pet/ct images using a two-stage cascaded neural network with camouflaged object detection mechanisms.Medical Physics, 50(10):6151–6162, 2023

    2023

  29. [29]

    Xuyao Yu, Lian He, Yuwen Wang, Yang Dong, Yongchun Song, Zhiyong Yuan, Ziye Yan, and Wei Wang. A deep learning approach for automatic tumor delineation in stereotactic radiotherapy for non-small cell lung cancer using diagnostic pet-ct and planning ct.Frontiers in Oncology, Volume 13 - 2023, 2023. 24

    2023

  30. [30]

    nnU-Net: a self- configuring method for deep learning-based biomedical image segmentation.Nature methods, 18(2):203–211, 2021

    Fabian Isensee, Paul F Jaeger, Simon AA Kohl, Jens Petersen, and Klaus H Maier-Hein. nnU-Net: a self- configuring method for deep learning-based biomedical image segmentation.Nature methods, 18(2):203–211, 2021

    2021

  31. [31]

    nnU-Net revisited: A call for rigorous validation in 3d medical image segmentation.arXiv preprint arXiv:2404.09556, 2024

    Fabian Isensee, Tassilo Wald, Constantin Ulrich, Michael Baumgartner, Saikat Roy, Klaus Maier-Hein, and Paul F Jaeger. nnU-Net revisited: A call for rigorous validation in 3d medical image segmentation.arXiv preprint arXiv:2404.09556, 2024

    Pith/arXiv arXiv 2024

  32. [32]

    https://deep-psma.grand-challenge.org/, 2025

    Deep-psma challenge: Deep-learning evaluation for enhanced prognostics – prostate-specific membrane antigen. https://deep-psma.grand-challenge.org/, 2025. Grand Challenge platform, accessed 2026-03-17

    2025

  33. [33]

    Data from the acrin 6685 trial hnscc-fdg-pet/ct, 2019

    Paul Kinahan, Marc Muzi, Bartosz Bialecki, and Lisa Coombs. Data from the acrin 6685 trial hnscc-fdg-pet/ct, 2019

    2019

  34. [34]

    Data from head-neck-pet-ct, 2017

    Martin Valli` eres, Emily Kay-Rivest, L´ eo Jean Perrin, Xavier Liem, Christophe Furstoss, Nader Khaouam, Phuc F´ elix Nguyen-Tan, Chang-Shu Wang, and Khalil Sultanem. Data from head-neck-pet-ct, 2017

    2017

  35. [35]

    Kadoch, Jeff Shrager, Andrew Quon, Daniel Rubin, Sylvia Plevritis, and Sandy Napel

    Shaimaa Bakr, Olivier Gevaert, Sebastian Echegaray, Kyle Ayers, Mu Zhou, Muhammad Shafiq, Hong Zheng, Weili Zhang, Albert Leung, M. Kadoch, Jeff Shrager, Andrew Quon, Daniel Rubin, Sylvia Plevritis, and Sandy Napel. Data for nsclc radiogenomics (version 4), 2017

    2017

  36. [36]

    Induction chemoradiation in stage iiia/n2 non-small-cell lung cancer: a phase 3 randomised trial.The Lancet, 386(9998):1049–1056, 2015

    Miklos Pless, Roger Stupp, Hans-Beat Ris, Rolf A Stahel, Walter Weder, Sandra Thierstein, Marie-Aline Ger- ard, Alexandros Xyrafas, Martin Fr¨ uh, Richard Cathomas, Alfred Zippelius, Arnaud Roth, Milorad Bijelovic, Adrian Ochsenbein, Urs R Meier, Christoph Mamot, Daniel Rauch, Oliver Gautschi, Daniel C Betticher, Ren´ e- Olivier Mirimanoff, and Solange Pe...

    2015

  37. [37]

    Freeman, Sonia R

    Martin Valli` eres, Carolyn R. Freeman, Sonia R. Skamene, and Issam El Naqa. A radiomics model from joint fdg-pet and mri texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities (soft-tissue-sarcoma), 2015

    2015

  38. [38]

    Gabry´ s, Sabrina A

    Lucas Basler, Hubert S. Gabry´ s, Sabrina A. Hogan, Matea Pavic, Marta Bogowicz, Diem Vuong, Stephanie Tanadini-Lang, Robert F¨ orster, Ken Kudura, Martin W. Huellner, Reinhard Dummer, Matthias Guckenberger, and Mitchell P. Levesque. Radiomics, tumor volume, and blood biomarkers for early prediction of pseudopro- gression in patients with metastatic melan...

    2020

  39. [39]

    Xiaoxiao Li, R. G. Abramson, L. R. Arlinghaus, A. B. Chakravarthy, V. G. Abramson, M. Sanders, and T. E. Yankeelov. Data from qin-breast (version 2), 2016

    2016

  40. [40]

    Data from the acrin 6685 trial hnscc-fdg- pet/ct, 2019

    Paul Kinahan, Megan Muzi, Barbara Bialecki, and Lisa Coombs. Data from the acrin 6685 trial hnscc-fdg- pet/ct, 2019

    2019

  41. [41]

    Matthias Guckenberger, Isabelle Opitz, Tereza Dellaporta, Alessandra Curioni-Fontecedro, Thomas Frauen- felder, Karin Ribi, Ferdinando Cerciello, Ivana Sullivan, Lizza Hendriks, Miriam Dorta, Ana Callejo, Joachim Aerts, Alfredo Addeo, Anne-Marie C. Dingemans, Giulia Pasello, Mariano Provencio, Filippo de Marinis, Nuria Mederos-Alfonso, Heidi Roschitzki-Vo...

    2025

  42. [42]

    Van Oss, G

    J. Van Oss, G. K. Murugesan, D. McCrumb, and R. Soni. Image segmentations produced by bamf under the aimi annotations initiative (v2.0.2), 2024

    2024

  43. [43]

    Gabry´ s, Preethi Mohan, Matthias Guckenberger, and Stephanie Tanadini-Lang

    Maksym Fritsak, Hubert S. Gabry´ s, Preethi Mohan, Matthias Guckenberger, and Stephanie Tanadini-Lang. Technical note: Vendor-specific approach for standardized uptake value calculation, 2025

    2025

  44. [44]

    Gabry´ s

    Maksym Fritsak, Marta Bogowicz, Diem Vuong, Florian Dietsche, Anja Joye, Matthias Guckenberger, Stephanie Tanadini-Lang, and Hubert S. Gabry´ s. 264 z-rad: the swiss pocket knife for radiomics.Radio- therapy and Oncology, 206:S3712–S3714, 2025. ESTRO 2025, 2-6 May 2025, Vienna, Austria

    2025

  45. [45]

    Multitalent: A multi-dataset approach to medical image segmentation

    Constantin Ulrich, Fabian Isensee, Tassilo Wald, Maximilian Zenk, Michael Baumgartner, and Klaus H Maier- Hein. Multitalent: A multi-dataset approach to medical image segmentation. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention, pages 648–658. Springer, 2023. 25

    2023

  46. [46]

    Addressing image misalignments in multi-parametric prostate MRI for enhanced computer-aided diagnosis of prostate cancer.Scientific Reports, 13(1):19805, 2023

    Balint Kovacs, Nils Netzer, Michael Baumgartner, Adrian Schrader, Fabian Isensee, Cedric Weißer, Ivo Wolf, Magdalena G¨ ortz, Paul F Jaeger, Victoria Sch¨ utz, et al. Addressing image misalignments in multi-parametric prostate MRI for enhanced computer-aided diagnosis of prostate cancer.Scientific Reports, 13(1):19805, 2023. 26

    2023

This paper was first reviewed by grok-4.5 on July 11, 2026.