Pith. sign in

REVIEW 3 major objections 5 minor 30 references

Automatic segmentation of kidney and liver tumors in CT images

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A standard LinkNet-34 trained carefully on LiTS matched or beat more complex networks for liver and kidney tumor segmentation in CT.

desk verdict A modest, useful baseline for liver and kidney tumor segmentation whose headline superiority claim is muddied by an unreported evaluation-mode choice. read the letter →

arxiv 1908.01279 v2 pith:45NJPF25 submitted 2019-08-04 eess.IV cs.CV

Dina B. Efremova , Dmitry A. Konovalov , Thanongchai Siriapisith , Worapan Kusakunniran , Peter Haddawy This is my paper · ORCID
classification eess.IVcs.CV
keywords automaticsegmentationlivertumorkidneyCTimagesconvolutionalneuralnetworksLinkNet-34transferlearningDICEscore
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

This paper argues that for automatic liver and kidney tumor segmentation in CT images, the training pipeline can matter as much as the CNN architecture. The authors took an off-the-shelf LinkNet-34 with a pretrained ResNet-34 encoder and added a simple set of learning techniques: a reduced learning rate for the encoder, a loss combining binary cross entropy with a negative log-DICE term, and data augmentation. DICE is a standard overlap measure between predicted and true masks, with 100% being perfect. Trained on the LiTS dataset and validated on 3DIRCADb, the method reached 96.2% liver DICE and 78.8% liver tumor DICE, which the authors read as outperforming at least one compound attention network, AHCNet. The same recipe applied to the KiTS-2019 challenge scored 96.38% kidney and 67.38% tumor DICE, placing 38th out of 106 submissions. If the result holds, well-resourced bespoke architectures are not required for competitive tumor segmentation; a disciplined training procedure on a simple network can get there at a fraction of the cost.

What carries the argument

The central object is LinkNet-34, a U-Net-style encoder-decoder segmentation network with ResNet-34 as its feature encoder, using the implementation of [26]. The method keeps the ImageNet-pretrained encoder weights unfrozen but lowers the encoder's learning rate by a factor of 10 relative to the randomly initialized decoder; the loss is binary cross entropy minus the logarithm of the DICE coefficient (Eq. 1). Training uses 58,638 image-mask pairs from the LiTS dataset (131 patients), and validation is on the 3DIRCADb dataset; a 2.5D variant that stacks adjacent slices was tried but did not improve DICE. This combination of a simple 2D CNN and a careful training recipe, rather than a bespoke deep architecture, is the mechanism the paper credits for its results.

What would settle it

Compute the paper's tumor DICE on 3DIRCADb under both stated evaluation modes, tumors larger than 100-pixel area and tumors of any size, and also compute AHCNet's DICE on the same set of test slices; if the 78.8% figure comes only from the larger-tumor subset while AHCNet's 73.4% was computed on any size, the claimed superiority would not hold.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that a customized training pipeline can make an off-the-shelf LinkNet-34 competitive with, and in one comparison better than, substantially more complex compound CNNs for liver and kidney tumor segmentation. On the 3DIRCADb validation set, the authors report liver DICE = 96.2% and tumor DICE = 78.8%, compared with AHCNet's 95.9% and 73.4% in their Table 1; both methods were trained on LiTS and evaluated on 3DIRCADb. When the same pipeline was applied to the KiTS-2019 challenge, it produced kidney DICE = 96.38%, tumor DICE = 67.38%, composite DICE = 81.88%, and 38th place among 106 submissions. The paper presents no new architecture; instead it attributes the gains to a combination of known techniques: transfer learning from a pretrained encoder, a 10x lower learning rate on that encoder, a BCE-minus-log-DICE loss, and data augmentation.

Load-bearing premise

The comparison assumes the DICE values in Table 1 were measured under compatible rules, but the paper never states whether its 78.8% tumor score counts all tumors or only those larger than 100 pixels.

Editorial extensions

If this is right

  • On the paper's own results, a single off-the-shelf LinkNet-34 trained on LiTS reaches 96.2% liver and 78.8% tumor DICE on 3DIRCADb, exceeding AHCNet's 95.9% and 73.4% in the same table.
  • The differential learning-rate scheme, with the encoder's learning rate reduced by a factor of 10, is reported to converge faster and to a better solution than uniform training, so transfer learning need not freeze the encoder.
  • Because data augmentation improved accuracy while 2.5D multi-slice input did not, slice-stacking is not necessary to achieve these numbers.
  • The same pipeline transfers to a different organ: on KiTS-2019 it achieved 96.38% kidney and 67.38% tumor DICE, finishing 38th of 106 submissions.

Reading between the lines

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

  • A fair reading of Table 1 should treat the 78.8% versus 73.4% comparison as conditional on the evaluation protocol: the paper does not state whether its tumor DICE used the 'any size' or 'larger than 100-pixel area' mode, and prior methods may have used different protocols; rescoring both sets of masks on the same subset would make the comparison decisive.
  • Because the same training recipe produced strong scores on two organs with different contrast characteristics, the recipe is a plausible candidate for other low-contrast lesion segmentation tasks; that is my extension, not a claim tested in the paper.
  • The negative result for 2.5D input suggests a testable hypothesis: in this regime, augmentation and loss shape matter more than inter-slice context; an ablation that drops the log-DICE term while keeping augmentation fixed would isolate that contribution.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript describes a segmentation system for liver, liver tumors, and kidney tumors from CT volumes. The method uses LinkNet-34 with an ImageNet-pretrained ResNet-34 encoder, a differential learning-rate schedule, a binary cross-entropy plus negative log-Dice loss, and data augmentation. Training is performed on the LiTS dataset and evaluation on 3DIRCADb, reporting a liver DICE of 96.2% and a tumor DICE of 78.8%; the same pipeline is applied to the KiTS-2019 challenge, reporting a kidney DICE of 96.38% and a tumor DICE of 67.38%. The paper's central claim is that a relatively simple CNN can match or beat more complex compound CNNs when attention is paid to the training pipeline rather than to the architecture.

Significance. If the results hold, the paper provides a useful data point: a straightforward 2D U-Net variant with transfer learning can reach competitive liver and liver-tumor segmentation on a public validation set and obtains a credible mid-ranking KiTS result. The evaluation design is a strength: training and validation are separated across public datasets, which avoids circularity, and the KiTS result was produced by the challenge's blinded test process. The contribution is, however, incremental. The reported advantage over AHCNet is not yet established at the level of statistical or protocol-wise rigor, and the 'fraction of time' claim lacks any supporting measurement.

major comments (3)
  1. [§2.2 and §3, Table 1] The paper defines two evaluation protocols for 3DIRCADb ('tumors larger than 100-pixel area' and 'tumors of any size') but reports only a single tumor DICE of 78.8%. Because the Conclusion's claim of outperforming AHCNet (73.4%) rests on this number, the missing protocol specification is load-bearing. Please report DICE for both evaluation modes and verify that the comparison methods' numbers were computed under the same mode; otherwise the superiority claim is unsupported.
  2. [§3, Table 1] No error bars, confidence intervals, or per-case statistics are reported for any method in Table 1. The liver DICE gap of 96.2 versus 95.9 is small and likely within patient-level noise, so the statement that the method 'outperformed at least one compound CNN [22] for liver and liver tumor segmentations' is not established by the data as presented. Add variance estimates such as per-volume DICE standard deviations, and ideally a paired or bootstrap test for the comparison.
  3. [§4, Conclusion] The claim that the pipeline achieves competitive results 'in fraction of time what would normally required for more complex CNNs' is not supported by any measured training or inference time. Please include hardware, epochs, and wall-clock measurements for the proposed method and, where possible, for the comparison methods, or remove the runtime claim from the conclusion.
minor comments (5)
  1. [§1.2] The word 'prepossessing' should read 'preprocessing'.
  2. [§2.1 and Table 1] The dataset name is written both as '3DIRCAD' and '3DIRCAB'; use '3DIRCADb' consistently throughout.
  3. [Table 1] The row for H-DenseUNet is ambiguous because '[18]' and 'H-DenseUNet' appear on adjacent lines with a single set of metrics; align the method names with their metric columns.
  4. [§2.2] The 2.5D input variant is mentioned but no details are given for the number of adjacent slices tested or the final configuration; a sentence with the chosen stack size would help reproducibility.
  5. [General] The paper uses 'DICE' and 'Dice' interchangeably; unify the notation.

Circularity Check

0 steps flagged · score 0.0 of 10

This is an empirical benchmark report with no fitted-input-as-prediction or self-citation-driven derivation; the central comparison is against held-out public CT datasets.

full rationale

The paper reports application of an off-the-shelf LinkNet-34 with transfer learning to two public segmentation benchmarks: training on LiTS and evaluating on the held-out 3DIRCADb dataset, and applying the same pipeline to the KiTS-2019 challenge test set. There is no derivation chain that reduces a predicted quantity to a fitted parameter. The reported DICE values are produced by running a fixed CNN on external validation data, not by optimizing a parameter against those same labels. The authors do cite their own prior semi-automatic segmentation work (references [3] and [28]), but those citations appear only in the related-work review and do not carry any load-bearing argument; the central result does not depend on them. The only substantive weakness is that the tumor DICE of 78.8% is reported without specifying whether it uses the 'tumors larger than 100-pixel area' or 'any size' evaluation mode described in Section 2.2, which makes the comparison against AHCNet's 73.4% ambiguous. That is an evaluation-reporting and comparability issue, not a circularity issue: no equation or construction makes the output equal to its input, and no self-citation supplies the conclusion. The paper is self-contained against external benchmarks, so the circularity score is 0.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

This is an empirical machine learning paper with no mathematical derivation. The free parameters are hand-chosen training choices that directly affect the reported performance. The axioms are standard domain assumptions for transfer learning and medical image segmentation.

free parameters (4)
  • Encoder learning rate reduction factor = 0.1 (stated as factor of 10)
    Chosen by hand to balance pretrained encoder fine-tuning with decoder training; no experiments shown for other values (Section 2.2).
  • Loss weighting (BCE and negative log Dice) = unweighted sum
    The loss is defined as bc - log(dice); the relative weighting is implicit and not tuned or justified (Eq. 1).
  • Tumor size threshold for evaluation = 100-pixel area (one of two modes)
    The paper tests on tumors larger than 100 pixels and on any size, but does not report which threshold corresponds to the headline 78.8% DICE (Section 2.2/3).
  • Data augmentation set = elastic transformation, shifting, scaling, rotating
    Augmentation types are listed but not quantitatively compared; the paper only notes data augmentation had a positive contribution (Section 2.1).
assumptions (3)
  • domain assumption The LiTS training annotations are correct and representative enough for transfer to 3DIRCADb and KiTS-19.
    The paper trains on LiTS and evaluates on 3DIRCADb and KiTS without domain adaptation, implicitly assuming anatomical and imaging consistency (Section 2.1).
  • domain assumption ImageNet-pretrained ResNet-34 features transfer to grayscale CT images.
    The encoder is initialized with ImageNet weights and not frozen; the paper assumes these features are useful despite the domain gap between natural images and CT (Section 2.2).
  • standard math DICE and VOE metrics are computed as defined in Eqs. 2-3.
    Standard definitions from the literature are used; no issue but listed for completeness.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Automatic segmentation of kidney and liver tumors in CT images." pith.science (2026). https://pith.science/paper/45NJPF25

@misc{pith2026190801279,
  author       = {Pith},
  title        = {Pith review of: Automatic segmentation of kidney and liver tumors in CT images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/45NJPF25}},
  note         = {Machine review of arXiv:1908.01279}
}
read the original abstract

Automatic segmentation of hepatic lesions in computed tomography (CT) images is a challenging task to perform due to heterogeneous, diffusive shape of tumors and complex background. To address the problem more and more researchers rely on assistance of deep convolutional neural networks (CNN) with 2D or 3D type architecture that have proven to be effective in a wide range of computer vision tasks, including medical image processing. In this technical report, we carry out research focused on more careful approach to the process of learning rather than on complex architecture of the CNN. We have chosen MICCAI 2017 LiTS dataset for training process and the public 3DIRCADb dataset for validation of our method. The proposed algorithm reached DICE score 78.8% on the 3DIRCADb dataset. The described method was then applied to the 2019 Kidney Tumor Segmentation (KiTS-2019) challenge, where our single submission achieved 96.38% for kidney and 67.38% for tumor Dice scores.

Figures

Figures reproduced from arXiv: 1908.01279 by the authors.

Figure 1
Figure 1. An example of contrast-enhanced CT scan showing the difficult case of lesion segmentation when edges of the tumor barely distinguishable in the original image (on the left). The red region denotes the ground truth mask of liver lesion (on the right). 2 Methods 2.1 Datasets Liver For this research two datasets have been used: MICCAI 2017 Liver Tumor Segmentation (LiTS) Challenge [20] and 3DIRCADb [15] (3D Image Recon… view at source ↗
Figure 2
Figure 2. An example of the method segmentation: liver (green) and liver tumor (yellow) segmentation. For the purpose of this paper we have selected a variation of U-Net [8], LinkNet-34 [24], where ResNet-34 [25] has been used as the feature encoder and PyTorch implementation was from [26]. LinkNet-34 has a reasonable number of parameters and a good balance between running time and accuracy. One of the problems of deep learni… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 26 canonical work pages

  1. [22]

    AHCNet: An Applica- tion of Attention Mechanism and Hybrid Connection for Liver Tumor Segmentation in CT Volumes

    Jiang, Huiyan, Tianyu Shi, Zhiqi Bai and Liangliang Huan g. AHCNet: An Applica- tion of Attention Mechanism and Hybrid Connection for Liver Tumor Segmentation in CT Volumes. IEEE Access 7 (2019): 24898-24909

    work page 2019

  2. [1]

    Fast approxim ate energy minimiza- tion via graph cuts

    Boykov, Yuri, Olga Veksler and Ramin Zabih. Fast approxim ate energy minimiza- tion via graph cuts. Proceedings of the Seventh IEEE Interna tional Conference on Computer Vision 1 (1999): 377-384 vol.1

    work page 1999

  3. [2]

    Schwartz

    Liu, Fan Shuo, Binsheng Zhao, Peter Klaus Kijewski, Liang Wang and Lawrence H. Schwartz. Liver segmentation for CT images using GVF snake. Medical physics 32 12 (2005): 3699-706 . Automatic segmentation of kidney and liver tumors in CT imag es 9

    work page 2005

  4. [3]

    A General Approach to Segmentation in CT Grayscale Images using Varia ble Neighborhood Search

    Siriapisith, Thanongchai, Worapan Kusakunniran and Pet er Haddawy. A General Approach to Segmentation in CT Grayscale Images using Varia ble Neighborhood Search. 2018 Digital Image Computing: Techniques and Appli cations (DICTA) (2018): 1-7

    work page 2018

  5. [4]

    Xu, Yingying, Lanfen Lin, Hongjie Hu, Dan Wang, Wenchao Zh u, J. J. Wang, Xian- hua Han and Yen-Wei Chen. Texture-specific bag of visual word s model and spatial cone matching-based method for the retrieval of focal liver lesions using multiphase contrast-enhanced CT images. International Journal of Com puter Assisted Radiol- ogy and Surgery 13 (2017): 151-164

    work page 2017

  6. [5]

    A Unified Level Set Framework Combining Hybrid Algorith ms for Liver and Liver Tumor Segmentation in CT Images

    Zheng, Zhou, Xuechang Zhang, Huafei Xu, Wang Liang, Simin g Zheng and Yueding Shi. A Unified Level Set Framework Combining Hybrid Algorith ms for Liver and Liver Tumor Segmentation in CT Images. BioMed research inte rnational (2018)

    work page 2018

  7. [6]

    Fully convolutional networks for semantic segmentation

    Long, Jonathan, Evan Shelhamer and Trevor Darrell. Fully convolutional networks for semantic segmentation. CVPR (2015)

    work page 2015

  8. [7]

    Fully Convolutional Networks for Semantic Segmentation

    Shelhamer, Evan, Jonathan Long and Trevor Darrell. Fully Convolutional Networks for Semantic Segmentation. 2015 IEEE Conference on Compute r Vision and Pattern Recognition (CVPR) (2014): 3431-3440

    work page 2014

Show all 30 references
  1. [8]

    U-Ne t: Convolutional Net- works for Biomedical Image Segmentation

    Ronneberger, Olaf, Philipp Fischer and Thomas Brox. U-Ne t: Convolutional Net- works for Biomedical Image Segmentation. ArXiv abs/1505.0 4597 (2015): n. pag

    2015

  2. [9]

    Automatic Segment ation of Liver Tu- mor in CT Images with Deep Convolutional Neural Networks

    Li, Wen Jung, Fucang Jia and Qingmao Hu. Automatic Segment ation of Liver Tu- mor in CT Images with Deep Convolutional Neural Networks. (2 015)

  3. [10]

    Measures of the Amount of Ecologic Ass ociation Between Species

    Dice, Lee Raymond. Measures of the Amount of Ecologic Ass ociation Between Species. (1945)

    1945

  4. [11]

    Very Deep Convolu tional Networks for Large-Scale Image Recognition

    Simonyan, Karen and Andrew Zisserman. Very Deep Convolu tional Networks for Large-Scale Image Recognition. CoRR abs/1409.1556 (2014) : n. pag

    2014 arXiv

  5. [12]

    3D Deeply Supervised Network for Automatic Liver Segmentatio n from CT Volumes

    Dou, Qi, Hao Chen, Yueming Jin, Lequan Yu, Jing Qin and Phe ng Ann Heng. 3D Deeply Supervised Network for Automatic Liver Segmentatio n from CT Volumes. MICCAI (2016)

    2016

  6. [13]

    Aut omatic 3D liver location and segmentation via convolutional neural networ k and graph cut

    Lu, Fang, Fa Wu, Peijun Hu, Zhiyi Peng and Dexing Kong. Aut omatic 3D liver location and segmentation via convolutional neural networ k and graph cut. Inter- national Journal of Computer Assisted Radiology and Surger y 12 (2016): 171-182

    2016

  7. [14]

    Christ, Patrick Ferdinand, Florian Ettlinger, Felix Gr n, Mohamed Ezzeldin A. Elshaer, Jana Lipkov, Sebastian Schlecht, Freba Ahmaddy, S unil Tatavarty, Marc Bickel, Patrick Bilic, Markus Rempfler, Felix Hofmann, Melv in D’Anastasi, Seyed- Ahmad Ahmadi, Georgios A Kaissis, Juli...

    2017 arXiv

  8. [15]

    and others

    Soler, L. and others. 3D Image reconstruction for compar ison of algo- rithm database: A patient specific anatomical and medical im age database, http://www.ircad.fr/softwares/3Dircadb/3Dircadb.php?lng=en

  9. [16]

    Automat ic segmentation of liver tumors from multiphase contrast-enhanced CT images b ased on FCNs

    Sun, Changjian, Shuxu Guo, Huimao Zhang, Jing Li, Meimei Chen, Shuzhi Ma, Lanyi Jin, Xiaoming Liu, Xueyan Li and Xiaohua Qian. Automat ic segmentation of liver tumors from multiphase contrast-enhanced CT images b ased on FCNs. Artifi- cial intelligence in medicine 83 (2017): 58-66

    2017

  10. [17]

    Krizhevsky, Alex, Ilya Sutskever and Geoffrey E. Hinton. ImageNet Classification with Deep Convolutional Neural Networks. Commun. ACM 60 (20 12): 84-90

  11. [18]

    H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tu mor Segmentation From CT Volumes

    Li, Xiaomeng, Hao Chen, Xiaojuan Qi, Qi Dou, Chi-Wing Fu a nd Pheng Ann Heng. H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tu mor Segmentation From CT Volumes. IEEE Transactions on Medical Imaging 37 (20 17): 2663-2674. 10 D. Efremova et al

  12. [19]

    Li, Bing Nan, Chee Kong Chui, Stephen K. Y. Chang and Sim He ng Ong. A new unified level set method for semi-automatic liver tumor segm entation on contrast- enhanced CT images. Expert Syst. Appl. 39 (2012): 9661-9668

    2012

  13. [20]

    LiTS—Liver Tumor Segmentation Challenge (2017), https://competitions.codalab.org/competitions/17094

    2017

  14. [21]

    RA- UNet: A hybrid deep attention-aware network to extract live r and tumor in CT scans

    Jin, Qiangguo, Zhao-Peng Meng, Changming Sun, Leyi Wei a nd Ran Su. RA- UNet: A hybrid deep attention-aware network to extract live r and tumor in CT scans. ArXiv abs/1811.01328 (2018): n. pag

    2018 arXiv

  15. [23]

    The KiTS19 Challenge Data: 300 Kidn ey Tumor Cases with Clinical Context, CT Semantic Segmentations, and Surg ical Outcomes ArXiv abs/1904.00445 (2019): n

    Nicholas Heller and Niranjan Sathianathen and Arveen Ka lapara and Edward Walczak and Keenan Moore and Heather Kaluzniak and Joel Rose nberg and Paul Blake and Zachary Rengel and Makinna Oestreich and Joshua De an and Michael Tradewell and Aneri Shah and Resha Tejpaul and Zach...

    2019 arXiv

  16. [24]

    LinkNet: Exploiting encoder repre- sentations for efficient semantic segmentation

    Chaurasia, Abhishek and Eugenio Culurciello. LinkNet: Exploiting encoder repre- sentations for efficient semantic segmentation. 2017 IEEE Vi sual Communications and Image Processing (VCIP) (2017): 1-4

    2017

  17. [25]

    De ep Residual Learning for Image Recognition

    He, Kaiming, Xiangyu Zhang, Shaoqing Ren and Jian Sun. De ep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vis ion and Pattern Recognition (CVPR) (2015): 770-778

    2015

  18. [26]

    Kalin in and Vladimir I

    Shvets, Alexey A., Alexander Rakhlin, Alexandr A. Kalin in and Vladimir I. Iglovikov. Automatic Instrument Segmentation in Robot-As sisted Surgery using Deep Learning. 2018 17th IEEE International Conference on M achine Learning and Applications (ICMLA) (2018): 624-628

    2018

  19. [27]

    Automatic Liver Lesion Segmentation Usin g A Deep Convolutional Neural Network Method

    Han, Xiguang. Automatic Liver Lesion Segmentation Usin g A Deep Convolutional Neural Network Method. ArXiv abs/1704.07239 (2017): n. pag

    2017 arXiv

  20. [28]

    Outer Wall Segmentation of Abdominal Aortic Aneurysm by Variable Neig hborhood Search Through Intensity and Gradient Spaces

    Siriapisith, Thanongchai, Worapan Kusakunniran and Pe ter Haddawy. Outer Wall Segmentation of Abdominal Aortic Aneurysm by Variable Neig hborhood Search Through Intensity and Gradient Spaces. Journal of Digital I maging 31 (2018): 490- 504

    2018

  21. [29]

    Dynamic Regu lation of Level Set Parameters Using 3D Convolutional Neural Network for Liver Tumor Segmentation

    Deng, Zhuofu, Qingzhe Guo and Zhiliang Zhu. Dynamic Regu lation of Level Set Parameters Using 3D Convolutional Neural Network for Liver Tumor Segmentation. Journal of healthcare engineering (2019)

    2019

  22. [30]

    fig1.png

    Huang, Qing, Hui Ding, Xiaodong Wang and Guangzhi Wang. R obust extraction for low-contrast liver tumors using modified adaptive likel ihood estimation. Interna- tional Journal of Computer Assisted Radiology and Surgery 1 3 (2018): 1565-1578. This figure "fig1.png" is available...

    2018 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.