pith. sign in

arxiv: 2511.19367 · v2 · submitted 2025-11-24 · 💻 cs.CV · cs.AI

AnatomicalNets: A Multi-Structure Segmentation and Contour-Based Distance Estimation Pipeline for Clinically Grounded Lung Cancer T-Staging

Saniah Kayenat Chowdhury , Rusab Sarmun , Muhammad E. H. Chowdhury , Sohaib Bassam Zoghoul , Israa Al-Hashimi , Adam Mushtak , Amith Khandakar This is my paper

Pith reviewed 2026-05-17 05:39 UTC · model grok-4.3

classification 💻 cs.CV cs.AI
keywords lung cancerT-stagingimage segmentationmedical imagingIASLC guidelinescontour distancedeep learning pipelineCT analysis
0
0 comments X

The pith

AnatomicalNets segments lungs, tumors and mediastinum then measures distances to apply exact IASLC staging rules at 91.36 percent accuracy.

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

The paper establishes that lung cancer T-staging can be performed by first using separate networks to outline the lung tissue, tumor, and mediastinum, then calculating the tumor's largest dimension and its distances to those structures, and finally feeding the numbers into fixed clinical rules instead of training a classifier on images alone. This matters because staging decisions rest on precise quantitative thresholds that determine prognosis and treatment, and small measurement errors can shift a patient from one stage to another. By making the measurements explicit and the rules deterministic, the pipeline produces outputs that align with how guidelines are written and can be checked against the original images. Results on the Lung-PET-CT-Dx dataset show overall accuracy of 91.36 percent with per-stage F1 scores of 0.93 for T1, 0.89 for T2, 0.96 for T3, and 0.90 for T4.

Core claim

AnatomicalNets uses three dedicated encoder-decoder networks to segment the lung parenchyma, tumor, and mediastinum, estimates the diaphragm boundary from lung contours, computes tumor size and proximity to adjacent structures via contour-based distances, and passes these values through a deterministic decision module that follows the IASLC T-staging guidelines, reaching 91.36 percent overall accuracy on the Lung-PET-CT-Dx dataset.

What carries the argument

Three separate encoder-decoder networks for segmenting lung parenchyma, tumor, and mediastinum, followed by contour-based distance estimation and a deterministic rule module that applies IASLC quantitative criteria.

If this is right

  • Staging decisions trace directly to measurable anatomical features that can be inspected on the original images.
  • Per-stage F1 scores remain above 0.89 across T1 through T4, providing a more granular view than overall accuracy alone.
  • The performance gain comes from explicit feature design that encodes clinical criteria rather than from increased classifier capacity.
  • The pipeline can be audited stage by stage by verifying the intermediate segmentations and distance values.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same segmentation-plus-measurement structure could be adapted to other guideline-driven staging tasks where size and proximity thresholds matter.
  • Clinicians may prefer this approach in settings that require documented reasoning for each stage assignment.
  • Robustness testing on scans from different scanners or protocols would show how much the distance estimates depend on image quality.

Load-bearing premise

The segmentation boundaries produced by the networks are sufficiently accurate that the derived distances and sizes correctly reproduce the thresholds in the IASLC T-staging guidelines.

What would settle it

Direct comparison of the pipeline's computed tumor dimensions and distances against manual expert measurements on the same scans, checking whether stage assignments match in the large majority of cases.

Figures

Figures reproduced from arXiv: 2511.19367 by Adam Mushtak, Amith Khandakar, Israa Al-Hashimi, Muhammad E. H. Chowdhury, Rusab Sarmun, Saniah Kayenat Chowdhury, Sohaib Bassam Zoghoul.

Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗
read the original abstract

Accurate tumor staging in lung cancer is crucial for prognosis and treatment planning and is governed by explicit anatomical criteria under fixed guidelines. However, most existing deep learning approaches treat this spatially structured clinical decision as an uninterpretable image classification problem. Tumor stage depends on predetermined quantitative criteria, including the tumor's dimensions and its proximity to adjacent anatomical structures, and small variations can alter the staging outcome. To address this gap, we propose AnatomicalNets, a medically grounded, multi-stage pipeline that reformulates tumor staging as a measurement and rule-based inference problem rather than a learned mapping. We employ three dedicated encoder-decoder networks to precisely segment the lung parenchyma, tumor, and mediastinum. The diaphragm boundary is estimated via a lung-contour heuristic, while the tumor's largest dimension and its proximity to adjacent structures are computed through a contour-based distance estimation method. These features are passed through a deterministic decision module following the international association for the study of lung cancer guidelines. Evaluated on the Lung-PET-CT-Dx dataset, AnatomicalNets achieves an overall classification accuracy of 91.36%. We report the per-stage F1-scores of 0.93 (T1), 0.89 (T2), 0.96 (T3), and 0.90 (T4), a critical evaluation aspect often omitted in prior literature. We highlight that the representational bottleneck in prior work lies in feature design rather than classifier capacity. This work establishes a transparent and reliable staging paradigm that bridges the gap between deep learning performance and clinical interpretability.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript presents AnatomicalNets, a multi-stage pipeline for lung cancer T-staging that uses three dedicated encoder-decoder networks to segment lung parenchyma, tumor, and mediastinum, estimates the diaphragm via a lung-contour heuristic, computes tumor size and proximity via contour-based distances, and applies a deterministic rule-based decision module following IASLC guidelines. On the Lung-PET-CT-Dx dataset the method reports 91.36% overall accuracy with per-stage F1 scores of 0.93 (T1), 0.89 (T2), 0.96 (T3), and 0.90 (T4). The central claim is that reformulating staging as explicit measurement plus rule inference yields clinically grounded, interpretable results superior to black-box classification.

Significance. If the intermediate segmentations and distance estimates prove faithful to IASLC quantitative thresholds, the work offers a transparent alternative to end-to-end classifiers and directly addresses the interpretability gap in automated staging. The explicit separation of feature extraction (segmentation + contour distances) from rule application is a methodological strength that could support clinical adoption and error analysis. However, the absence of any reported validation on the measurement stage leaves the clinical grounding of the 91.36% figure unconfirmed.

major comments (2)
  1. [Abstract and §3] Abstract and §3 (pipeline description): the reported staging accuracy and F1 scores rest on the assumption that the three segmentations plus contour-based distances faithfully reproduce IASLC cutoffs (3 cm, 5 cm, invasion distances). No Dice/Hausdorff scores for the lung/tumor/mediastinum segmentations, no MAE or Bland-Altman analysis of derived diameters/distances against radiologist ground truth, and no sensitivity analysis near decision thresholds are provided. Boundary errors of a few millimeters can flip T-stage labels, so end-to-end classification metrics alone do not establish measurement fidelity.
  2. [Evaluation] Evaluation section: the manuscript states results on the Lung-PET-CT-Dx dataset but supplies no information on train/validation/test split ratios, cross-validation procedure, or any baseline comparisons (e.g., against standard U-Net segmentation followed by rule-based staging or against published T-staging classifiers). Without these controls it is impossible to isolate whether the performance gain comes from the anatomical pipeline or from dataset-specific factors.
minor comments (2)
  1. [Abstract] The claim that 'the representational bottleneck in prior work lies in feature design rather than classifier capacity' is asserted without supporting ablation or citation of specific prior architectures; this should be either substantiated or removed.
  2. [§3] Notation for the contour-based distance estimation (e.g., how the largest tumor dimension is extracted from the segmented contour) is described at a high level; a short pseudocode or explicit formula would improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thorough review and constructive feedback. We address the major comments point by point below, and will revise the manuscript to incorporate additional validations and experimental details as suggested.

read point-by-point responses

  1. Referee: [Abstract and §3] Abstract and §3 (pipeline description): the reported staging accuracy and F1 scores rest on the assumption that the three segmentations plus contour-based distances faithfully reproduce IASLC cutoffs (3 cm, 5 cm, invasion distances). No Dice/Hausdorff scores for the lung/tumor/mediastinum segmentations, no MAE or Bland-Altman analysis of derived diameters/distances against radiologist ground truth, and no sensitivity analysis near decision thresholds are provided. Boundary errors of a few millimeters can flip T-stage labels, so end-to-end classification metrics alone do not establish measurement fidelity.

    Authors: We agree that the fidelity of the segmentation and distance measurements is critical to substantiate the clinical grounding of the 91.36% accuracy. The manuscript emphasizes the rule-based inference but does not include quantitative validation of the intermediate steps. In the revised manuscript, we will add Dice and Hausdorff scores for all three segmentations (lung parenchyma, tumor, mediastinum). We will also report MAE and include Bland-Altman analysis for the tumor size and proximity distances, using available annotations in the Lung-PET-CT-Dx dataset. Furthermore, we will perform and report a sensitivity analysis around the IASLC thresholds to evaluate the impact of small boundary errors on staging decisions. These results will be added to the Evaluation section. revision: yes

  2. Referee: [Evaluation] Evaluation section: the manuscript states results on the Lung-PET-CT-Dx dataset but supplies no information on train/validation/test split ratios, cross-validation procedure, or any baseline comparisons (e.g., against standard U-Net segmentation followed by rule-based staging or against published T-staging classifiers). Without these controls it is impossible to isolate whether the performance gain comes from the anatomical pipeline or from dataset-specific factors.

    Authors: We acknowledge the importance of detailing the experimental protocol for reproducibility and fair comparison. The current version reports aggregate results without specifying the data splits or including baselines. In the revision, we will provide the exact train/validation/test split ratios used for the Lung-PET-CT-Dx dataset and clarify the evaluation procedure, including whether cross-validation was applied. We will also introduce baseline experiments: (1) a standard U-Net for multi-structure segmentation followed by the same contour-based distance estimation and rule-based staging, and (2) direct comparisons to published end-to-end T-staging methods. This will allow readers to assess the contribution of the AnatomicalNets pipeline. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical pipeline with external rules

full rationale

The paper's chain is segmentation (three encoder-decoders) → contour-based measurements (heuristic diaphragm + distance estimation) → deterministic rule application (IASLC guidelines). Reported accuracy and F1 scores are end-to-end empirical results on the Lung-PET-CT-Dx dataset, not quantities forced by construction from fitted parameters or self-referential definitions. No self-citations, uniqueness theorems, or ansatzes appear in the load-bearing steps. The approach is self-contained against external clinical criteria and does not reduce any claimed result to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The pipeline rests on the domain assumption that IASLC size-and-invasion criteria can be faithfully recovered from automated segmentations and contour distances; no free parameters, invented entities, or additional axioms are stated.

axioms (1)
  • domain assumption IASLC guidelines supply the definitive quantitative criteria for T-staging based on tumor dimensions and proximity to adjacent structures.
    The deterministic decision module is described as following these guidelines directly.

pith-pipeline@v0.9.0 · 5634 in / 1182 out tokens · 42184 ms · 2026-05-17T05:39:53.919254+00:00 · methodology

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

48 extracted references · 48 canonical work pages · 1 internal anchor

  1. [1]

    Confronting the Boundaries of Human Longevity: Many people now live beyond their natural lifespans through the intervention of medical technology and improved lifestyles—a form of

    S. J. Olshansky, B. A. Carnes, and D. Grahn, "Confronting the Boundaries of Human Longevity: Many people now live beyond their natural lifespans through the intervention of medical technology and improved lifestyles—a form of" manufactured time"," American Scientist, vol. 86, no. 1, pp. 52-61, 1998

    work page 1998

  2. [2]

    The global burden of lung cancer: current status and future trends,

    A. Leiter, R. R. Veluswamy, and J. P . Wisnivesky, "The global burden of lung cancer: current status and future trends," Nature reviews Clinical oncology, vol. 20, no. 9, pp. 624-639, 2023

    work page 2023

  3. [3]

    Computer aided lung cancer diagnosis with deep learning algorithms,

    W. Sun, B. Zheng, and W. Qian, "Computer aided lung cancer diagnosis with deep learning algorithms," in Medical imaging 2016: computer-aided diagnosis, 2016, vol. 9785: SPIE, pp. 241-248

    work page 2016

  4. [4]

    Optimal deep learning model for classification of lung cancer on CT images,

    S. Lakshmanaprabu, S. N. Mohanty, K. Shankar, N. Arunkumar, and G. Ramirez, "Optimal deep learning model for classification of lung cancer on CT images," Future Generation Computer Systems, vol. 92, pp. 374-382, 2019

    work page 2019

  5. [5]

    Deep learning techniques to diagnose lung cancer,

    L. Wang, "Deep learning techniques to diagnose lung cancer," Cancers, vol. 14, no. 22, p. 5569, 2022

    work page 2022

  6. [6]

    Deep learning classification of lung cancer histology using CT images,

    T. L. Chaunzwa et al., "Deep learning classification of lung cancer histology using CT images," Scientific reports, vol. 11, no. 1, pp. 1-12, 2021

    work page 2021

  7. [7]

    Global epidemiology of lung cancer,

    J. A. Barta, C. A. Powell, and J. P . Wisnivesky, "Global epidemiology of lung cancer," Annals of global health, vol. 85, no. 1, p. 8, 2019

    work page 2019

  8. [8]

    Global burden of lung cancer in 2022 and projections to 2050: Incidence and mortality estimates from GLOBOCAN,

    J. Zhou, Y . Xu, J. Liu, L. Feng, J. Yu, and D. Chen, "Global burden of lung cancer in 2022 and projections to 2050: Incidence and mortality estimates from GLOBOCAN," Cancer Epidemiology, vol. 93, p. 102693, 2024

    work page 2022

  9. [9]

    PET/CT imaging in different types of lung cancer: an overview,

    V. Ambrosini et al., "PET/CT imaging in different types of lung cancer: an overview," European journal of radiology, vol. 81, no. 5, pp. 988-1001, 2012

    work page 2012

  10. [10]

    Emerging strategies for the treatment of small cell lung cancer: a review,

    W. J. Petty and L. Paz-Ares, "Emerging strategies for the treatment of small cell lung cancer: a review," JAMA oncology, vol. 9, no. 3, pp. 419-429, 2023

    work page 2023

  11. [11]

    R. U. Osarogiagbon et al., "The International Association for the Study of Lung Cancer Lung Cancer Staging Project: overview of challenges and opportunities in revising the nodal classification of lung cancer," Journal of Thoracic Oncology, vol. 18, no. 4, pp. 410- 418, 2023

    work page 2023

  12. [12]

    Lung cancer (staging - IASLC 8th edition)

    K. H. Weerakkody Y , Worsley C. "Lung cancer (staging - IASLC 8th edition)." https://radiopaedia.org/articles/lung-cancer-staging-iaslc-8th-edition (accessed

    work page

  13. [13]

    Deep learning for prediction of N2 metastasis and survival for clinical stage I non–small cell lung cancer,

    Y . Zhong et al., "Deep learning for prediction of N2 metastasis and survival for clinical stage I non–small cell lung cancer," Radiology, vol. 302, no. 1, pp. 200-211, 2022

    work page 2022

  14. [14]

    Survival analysis and factors affecting survival in patients who presented to the medicaloncology unit with non-small cell lung cancer,

    Ö. Önal, M. Kocer, H. N. Eroğlu, S. D. Yilmaz, I. Eroğlu, and D. Karadoğan, "Survival analysis and factors affecting survival in patients who presented to the medicaloncology unit with non-small cell lung cancer," Turkish Journal of Medical Sciences, vol. 50, no. 8, pp. 1838-1850, 2020

    work page 2020

  15. [15]

    Factors influencing CNN performance,

    I. M. Israel, S. A. Israel, and J. M. Irvine, "Factors influencing CNN performance," in 2021 IEEE Applied Imagery Pattern Recognition Workshop (AIPR), 2021: IEEE, pp. 1-4

    work page 2021

  16. [16]

    CNN Distance Estimation Based on Received Signal Strength Indicator,

    H.-C. Chu, H.-C. Lai, M.-Y . Wang, and S.-J. Wu, "CNN Distance Estimation Based on Received Signal Strength Indicator," in 2024 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2024: IEEE, pp. 2103-2107

    work page 2024

  17. [17]

    Lung cancer prediction by Deep Learning to identify benign lung nodules,

    M. A. Heuvelmans et al., "Lung cancer prediction by Deep Learning to identify benign lung nodules," Lung cancer, vol. 154, pp. 1-4, 2021

    work page 2021

  18. [18]

    Deep learning for the detection of benign and malignant pulmonary nodules in non-screening chest CT scans,

    W. Hendrix et al., "Deep learning for the detection of benign and malignant pulmonary nodules in non-screening chest CT scans," Communications medicine, vol. 3, no. 1, p. 156, 2023

    work page 2023

  19. [19]

    Automated lung tumor detection and diagnosis in CT Scans using texture feature analysis and SVM,

    T. Adams, J. Dörpinghaus, M. Jacobs, and V. Steinhage, "Automated lung tumor detection and diagnosis in CT Scans using texture feature analysis and SVM," in FedCSIS (Communication Papers), 2018, pp. 13-20

    work page 2018

  20. [20]

    VGG19 network assisted joint segmentation and classification of lung nodules in CT images,

    M. A. Khan et al., "VGG19 network assisted joint segmentation and classification of lung nodules in CT images," Diagnostics, vol. 11, no. 12, p. 2208, 2021

    work page 2021

  21. [21]

    A transformer- based deep neural network for detection and classification of lung cancer via PET/CT images,

    K. Barbouchi, D. El Hamdi, I. Elouedi, T. B. Aïcha, A. K. Echi, and I. Slim, "A transformer- based deep neural network for detection and classification of lung cancer via PET/CT images," International Journal of Imaging Systems and Technology, vol. 33, no. 4, pp. 1383-1395, 2023

    work page 2023

  22. [22]

    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

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  23. [23]

    Li, Wang, S., Li, T., Lu, J., HuangFu, Y ., & Wang, D

    P . Li, Wang, S., Li, T., Lu, J., HuangFu, Y ., & Wang, D. . A Large-Scale CT and PET/CT Dataset for Lung Cancer Diagnosis (Lung-PET-CT-Dx) [Data set], doi: https://www.cancerimagingarchive.net/collection/lung-pet-ct-dx/#citations

    work page

  24. [24]

    S. G. Armato III, McLennan, G., Bidaut et al. LIDC-IDRI | Data from The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): A completed reference database of lung nodules on CT scans, doi: https://www.cancerimagingarchive.net/collection/lidc-idri/#citations

    work page

  25. [25]

    Enhanced Lung Cancer Detection and TNM Staging Using YOLOv8 and TNMClassifier: An Integrated Deep Learning Approach for CT Imaging,

    A. Wehbe, S. Dellepiane, and I. Minetti, "Enhanced Lung Cancer Detection and TNM Staging Using YOLOv8 and TNMClassifier: An Integrated Deep Learning Approach for CT Imaging," IEEE Access, 2024

    work page 2024

  26. [26]

    & Daoudi, A.Real-Time Flying Ob- ject Detection with YOLOv8 2023

    D. Reis, J. Kupec, J. Hong, and A. Daoudi, "Real-time flying object detection with YOLOv8," arXiv preprint arXiv:2305.09972, 2023

    work page arXiv 2023

  27. [27]

    Dynamic detr: End-to-end object detection with dynamic attention,

    X. Dai, Y . Chen, J. Yang, P . Zhang, L. Yuan, and L. Zhang, "Dynamic detr: End-to-end object detection with dynamic attention," in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 2988-2997

    work page 2021

  28. [28]

    Automated lung cancer T-Stage detection and classification using improved U-Net model,

    B. K. Sathiyamurthy and V. K. Madhaiyan, "Automated lung cancer T-Stage detection and classification using improved U-Net model," International Journal of Electrical & Computer Engineering (2088-8708), vol. 14, no. 6, 2024

    work page 2088

  29. [29]

    T-stage diagnosis of lung cancer based on deep learning in CT images,

    R. Fan et al., "T-stage diagnosis of lung cancer based on deep learning in CT images," Digital Medicine, vol. 10, no. 4, p. e00017, 2024

    work page 2024

  30. [30]

    A Bayesian neural network model based on CT images for staging non-small cell lung cancer,

    J. Zhang and H. Zhang, "A Bayesian neural network model based on CT images for staging non-small cell lung cancer," in 2023 8th International Conference on Computer and Communication Systems (ICCCS), 2023: IEEE, pp. 884-893

    work page 2023

  31. [31]

    Medical Segmentation Decathlon (MSD), doi: https://doi.org/10.1109/ISBI.2019.00030

    work page doi:10.1109/isbi.2019.00030 2019

  32. [32]

    [Online]

    COVID-19 CT segmentation dataset. [Online]. Available: https://medicalsegmentation.com/covid19/

    work page

  33. [33]

    COVID-19 CT lung and infection segmentation dataset,

    M. Jun et al., "COVID-19 CT lung and infection segmentation dataset," (No Title), 2020

    work page 2020

  34. [34]

    SAROS: A dataset for whole-body region and organ segmentation in CT imaging,

    S. Koitka et al., "SAROS: A dataset for whole-body region and organ segmentation in CT imaging," Scientific Data, vol. 11, no. 1, p. 483, 2024

    work page 2024

  35. [35]

    K. S. Mader. Finding and Measuring Lungs in CT Data [Online]. Available: https://www.kaggle.com/datasets/kmader/finding-lungs-in-ct-data

    work page

  36. [36]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P . Fischer, and T. Brox, "U-net: Convolutional networks for biomedical image segmentation," in Medical image computing and computer-assisted intervention– MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18, 2015: Springer, pp. 234-241

    work page 2015

  37. [37]

    Enhancing intima-media complex segmentation with a multi-stage feature fusion-based novel deep learning framework,

    R. Sarmun et al., "Enhancing intima-media complex segmentation with a multi-stage feature fusion-based novel deep learning framework," Engineering Applications of Artificial Intelligence, vol. 133, p. 108050, 2024

    work page 2024

  38. [38]

    Convolutional neural network based encoder-decoder architectures for semantic segmentation of plants,

    S. Kolhar and J. Jagtap, "Convolutional neural network based encoder-decoder architectures for semantic segmentation of plants," Ecological Informatics, vol. 64, p. 101373, 2021

    work page 2021

  39. [39]

    DefED-Net: Deformable encoder-decoder network for liver and liver tumor segmentation,

    T. Lei, R. Wang, Y . Zhang, Y . Wan, C. Liu, and A. K. Nandi, "DefED-Net: Deformable encoder-decoder network for liver and liver tumor segmentation," IEEE Transactions on Radiation and Plasma Medical Sciences, vol. 6, no. 1, pp. 68-78, 2021

    work page 2021

  40. [40]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770-778

    work page 2016

  41. [41]

    Densely connected convolutional networks,

    G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, "Densely connected convolutional networks," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 4700-4708

    work page 2017

  42. [42]

    Feature pyramid networks for object detection,

    T.-Y . Lin, P . Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, "Feature pyramid networks for object detection," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117-2125

    work page 2017

  43. [43]

    Unet++: A nested u-net architecture for medical image segmentation,

    Z. Zhou, M. M. Rahman Siddiquee, N. Tajbakhsh, and J. Liang, "Unet++: A nested u-net architecture for medical image segmentation," in Deep learning in medical image analysis and multimodal learning for clinical decision support: 4th international workshop, DLMIA 2018, and 8th international workshop, ML-CDS 2018, held in conjunction with MICCAI 2018, Grana...

    work page 2018

  44. [44]

    Variation in diaphragm position and shape in adults with normal pulmonary function,

    T. Suwatanapongched, D. S. Gierada, R. M. Slone, T. K. Pilgram, and P . G. Tuteur, "Variation in diaphragm position and shape in adults with normal pulmonary function," Chest, vol. 123, no. 6, pp. 2019-2027, 2003

    work page 2019

  45. [45]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, "Imagenet: A large-scale hierarchical image database," in 2009 IEEE conference on computer vision and pattern recognition, 2009: Ieee, pp. 248-255

    work page 2009

  46. [46]

    The IASLC lung cancer staging project: proposals for revision of the TNM stage groupings in the forthcoming (eighth) edition of the TNM classification for lung cancer,

    P . Goldstraw et al., "The IASLC lung cancer staging project: proposals for revision of the TNM stage groupings in the forthcoming (eighth) edition of the TNM classification for lung cancer," Journal of Thoracic Oncology, vol. 11, no. 1, pp. 39-51, 2016

    work page 2016

  47. [47]

    The eighth edition TNM stage classification for lung cancer: What does it mean on main street?,

    F. C. Detterbeck, "The eighth edition TNM stage classification for lung cancer: What does it mean on main street?," The Journal of thoracic and cardiovascular surgery, vol. 155, no. 1, pp. 356-359, 2018

    work page 2018

  48. [48]

    The eighth edition AJCC cancer staging manual: continuing to build a bridge from a population-based to a more “personalized

    M. B. Amin et al., "The eighth edition AJCC cancer staging manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging," CA: a cancer journal for clinicians, vol. 67, no. 2, pp. 93-99, 2017

    work page 2017