KiPA22 (Regular Challenge) Banner
Join

KiPA22 Regular Challenge! 

Kidney PArsing Challenge 2022:

Multi-Structure Segmentation for Renal Cancer Treatment


[NEWS!]2022/10/01 - The regular challenge of KiPA22 starts now! Training and open testing will be provided continuously! Let's continue to challenge the multi-structure segmentation for renal cancer treatment!

2022/09/18 - The KiPA session has achieved complete success! Thanks to every participant and contributor for their support, KiPA22 Challenge is over, and we hope to see you again in the future!
2022/09/02 - The KiPA session will be held on September 18th (SGT)! The challenge is now closed, it will be re-open after KiPA Session. Please refer to session plan for details.
2022/08/10 - The open and closed testing phase has now ended. The final results will come out in a few days. Thanks for participating in KiPA22 Challenge!
2022/07/16 - The closed test phase is now open for submission! Please go to the guidelines page for details. 
2022/07/08 - A new submission mechanism for the open test phase is now applied. The open test leaderborad will be refreshed at 9:00 and 21:00 UTC+8 every day. Please go to the evaluation page for details.
2022/07/01 - The open test dataset is now available! And the open test submission phase is now open!
2022/04/22 - The evaluation code and baseline models are available!
2022/04/10 - The training dataset is now available!
2022/03/26 - KiPA challenge 2022  is now open for registration!

 Clinical Significance

Three-dimensional (3D) kidney parsing on computed tomography angiography (CTA) images is one of the most important tasks for surgery-based renal cancer treatment (e.g., laparoscopic partial nephrectomy [1]).

It targets segmenting 3D kidneys, renal tumors, arteries, and veins. Once successful, clinicians will benefit from the 3D visual model of renal structures for accurate preoperative planning [2]. Preoperatively, the renal arteries will help estimate the renal perfusion model [3], so that the clinicians will select the tumor-feeding arterial branches and locate the arterial clamping position easily [4]. The tumor and kidney models will visually show the lesion regions, thus helping the pre-plan of the tumor resection surface. Intraoperatively, the preoperative plan will be displayed on the screen together with laparoscopic videos to guide the surgery [5]. Renal vessels (veins, arteries) outside the hilum will show a clear arterial clamping region visually, thus the clinicians will select arterial clamping branches quickly. The 3D visual model will also guide the clinicians in making appropriate decisions. Therefore, the costs of treatment will be reduced, the quality of surgery will be improved, and the pain of patients will be relieved.

[1] Shao, P., Qin, C., Yin, C., Meng, X., Ju, X., Li, J., Lv, Q., Zhang, W., Xu,Z., 2011. Laparoscopic partial nephrectomy with segmental renal artery clamping: technique and clinical outcomes. European urology 59, 849–855.
[2] Porpiglia, F., Fiori, C., Checcucci, E., Amparore, D., Bertolo, R., 2018. Hyperaccuracy three-dimensional reconstruction is able to maximize the efficacy of selective clamping during robot-assisted partial nephrectomy for complex renal masses. European urology 74, 651–660.
[3] Zhang, S., Yang, G., Tang, L., Lv, Q., Li, J., Xu, Y., Zhu, X., Li, P., Shao, P., Wang, Z., 2019. Application of a functional3-dimensional perfusion model in laparoscopic partial nephrectomy with precise segmental renal artery clamping. Urology 125, 98–103.
[4] Shao, P., Tang, L., Li, P., Xu, Y., Qin, C., Cao, Q., Ju, X., Meng, X., Lv, Q., Li, J., et al., 2012. Precise segmental renal artery clamping under the guidance of dual-source computed tomography angiography during laparoscopic partial nephrectomy. European urology 62, 1001–1008
[5] Nicolau, S., Soler, L., Mutter, D., Marescaux, J., 2011. Augmented reality in laparoscopic surgical oncology. Surgical oncology 20, 189–201

Task

Based on the great clinical significance, we organize a new challenge named Kidney Parsing (KiPA) for Renal Cancer Treatment 2022 Challenge.

The KiPA challenge is an important step in the development of reliable, valid, and reproducible methods that segment four kidney-related structures on CTA images to promote surgery-based renal cancer treatment.

The target structures are:

  • Kidney-Abnormal organ: Kidney structures are damaged by the tumors, losing their normal shape and making a large challenge in shape variations. 
  • Tumor-Multi-subtype lesion: Five sub-tumor types with different appearances are in the dataset, leading to a large challenge in distribution variation. 
  • Renal Artery-Very-thin structure: Renal arteries are thin structures that only account for 0.27% of the whole image, leading to a large challenge in class imbalance
  • Renal Vein-Low-significant region: Only a small amount of contrast agent has entered the veins making low-significant vein regions, leading to a large challenge to distinguish them from the background.


We have collected 130 images, 70 for the training dataset, 30 for the closed testing dataset, and 30 for the open testing dataset. In the regular challenge, only the training and open testing datasets are available! Dice, HD, and AVD are adopted as evaluation metrics. This challenge will promote renal cancer treatment, interactions between researchers, and interdisciplinary communication.

 Dataset (regular challenge)

  • Training cases: 70 (No specific validation cases are provided, any manner for validation is allowed.)
  • Opening testing cases: 30. (The results of opened test cases can be evaluated by submitting to the opening testing leaderboard.)
  • Closed testing cases: 30. (The results of closed test cases can be obtained after the submission deadline.)
  • The images are saved as .nii.gz files.

For more details,  please refer to Dataset.

 Rules (regular challenge)

More details are in the participant Rules.

Registration

All participants should click the Join button to take part in the KiPA22 Regular Challenge, and fill in the information in the application form (Dataset page) to apply the challenge datasets.  All teams have 5 opportunities per week to submit the evaluation of the open test dataset.

One team only needs to register once. All members of the team need to jointly sign one form.

Leaderboard

There are two leaderboards, the opening testing leaderboard, and the challenge leaderboard. But only the opening testing leaderboard is available during the KiPA22 Regular Challenge.

  • For the open testing leaderboard, the prediction results of the opened testing dataset should be submitted. Each team has 5 opportunities per week. 

Submission limitation

  • At most 5 submissions per week are allowed for the open testing leaderboard.  Wrong submissions also will be counted due to the automated process. After the submission, the scores and leaderboard will be automatically updated after a while. 
  • All results should be from an automatic framework without any manual part.
  • [Note!] Once any violation of the rules is found, we will cancel the score, and pull the institution of the violated team into the KiPA blacklist. In the future, any KiPA-related challenges and activities will refer to this blacklist.

Evaluation metrics

  • DSC: Dice Similarity Coefficient
  • HD:  Hausdorff Distance
  • AVD:  Average Hausdorff Distance

    For more details, please refer to Evaluation.

     Citation

    If using our dataset, you must cite the following papers:


    [1] He, Y., Yang, G., Yang, J., Ge, R., Kong, Y., Zhu, X., Shao,P., Shu H., Dillenseger, J.L., Li, S., 2021. Meta grayscale adaptive network for 3D integrated renal structures segmentation. Medical image analysis 71, 102055. [paper]
    [2] He, Y., Yang, G., Yang, J., Chen, Y., Kong, Y., Wu, J., Tang, L., Zhu, X., Dillenseger, J.L., Shao, P., Zhang, S., Shu, H., Coatrieux, J.L., Li, S., 2020. Dense biased networks with deep priori anatomy and hard region adaptation: Semisupervised learning for fine renal artery segmentation. Medical Image Analysis 63, 101722. [paper]
    [3] Shao, P., Qin, C., Yin, C., Meng, X., Ju, X., Li, J., Lv, Q., Zhang, W., Xu,Z., 2011. Laparoscopic partial nephrectomy with segmental renal artery clamping: technique and clinical outcomes. European urology 59, 849–855. [paper]
    [4] Shao, P., Tang, L., Li, P., Xu, Y., Qin, C., Cao, Q., Ju, X., Meng, X., Lv, Q., Li, J., et al., 2012. Precise segmental renal artery clamping under the guidance of dual-source computed tomography angiography during laparoscopic partial nephrectomy. European urology 62, 1001–1008. [paper]


    Any questions, please email us.