KNIGHT Challenge
Kidney clinical Notes and Imaging to Guide and Help personalize Treatment and biomarkers discovery

Updates:

• The challenge submission deadline was extended to March 7
• Great news! We are happy to announce that ALL challenge participants are invited to submit their four-page short papers for a review to be featured in the ISBI conference! The review will focus primarily on the methods and clarity of presenting the approach, rather than strictly on scientific novelty.
• Submission is now open
• Test set sent to registered participants.



What is KNIGHT?

The goal of the KNIGHT challenge is to facilitate the development of Artificial Intelligence (AI) models for automatic preoperative prediction of risk class for patients with renal masses identified in clinical Computed Tomography (CT) imaging of the kidneys. The dataset, which we named the Kidney Classification (KiC) dataset, is based on the 2021 Kidney and Kidney Tumor Segmentation challenge (KiTS). This dataset was extended to include additional clinical information, as well as risk classification labels, deduced from postoperative pathology results. Some of the clinical information will also be available for inference. The patients are classified into five risk groups in accordance with American Urological Association (AUA) guidelines. These groups can be divided into two classes based on the follow-up treatment. The challenge consists of two quantitative tasks (1) binary patient classification as per the follow-up treatment, and (2) fine-grained classification into five risk groups, and a bonus task: (3) discovery of prognostic biomarkers.

Why the Challenge?

Renal cancer accounted for nearly 400,000 new cases and 180,000 deaths in 2020 [24]. Prognosis assessment and disease management are conducted by clinicians who classify the patients based on the tumor’s size, histopathology, and presence of local or distant extent, defining the Tumor, Node, and Metastases (TNM) stage and grade tools. According to the American Urological Association (AUA), the risk stratification of patients done by clinicians is based on the following risk groups: Benign (B); Low Risk (LR), which includes pT1 stage and grade 1/2; Intermediate Risk (IR), which includes pT1 stage and grade 3/4 or pT2 and any grade; High Risk (HR), which includes pT3 and any grade; and Very High Risk (VHR), which includes metastatic patients, pT4 or pN1, or sarcomatoid or rhabdoid features upon pathological examination, or macroscopic positive margin [1]. Diagnosis and management of localized sporadic renal masses typically require high quality, cross-sectional abdominal imaging such as CT to optimally characterize and clinically stage the renal mass [1].

The conspicuous appearance of renal tumors in CT imaging triggered the interest of radiologists and urologists to study the relationship between the tumor’s image characteristics and relevant patient outcomes such as prognosis, surgical approach, and need for subsequent treatment. Indeed, the tumor’s anatomical features inferred from CT or magnetic resonance imaging (MRI) were leveraged in the well-known standardized nephrometry Scoring Systems (SSs), such as RENAL [17], PADUA [7], or centrality-index (C-index) [23]. RENAL score was used in one of the first studies to objectify the relationship between tumor anatomy and pathology, where data of around 500 patients was used to quantify preoperative prediction of tumor histology and grade (Area Under the Curve (AUC) of 0.76 and 0.73, respectively) [16]. All three SSs were proven useful for risk stratification, and correlated well with patients’ outcomes [7,17,13,22].

Concerning the management of localized disease, Partial Nephrectomy (PN) is recommended over Radical Nephrectomy (RN) for surgical excision of the tumors, unless determined otherwise by the urologist. Other options of management include active surveillance and ablative techniques, addressed to patients with decreased life expectancy or comorbidities preventing them from undergoing invasive procedures [3]. Adjuvant therapy is considered in high-risk localized renal cancer. In recent years, clinical trials have shown that adjuvant administration of Sunitinib and Pembrolizumab led to an improvement in disease-free survival among patients presenting high-grade renal-cell carcinoma [20,6]. Hence, the five risk groups suggested by the AUA are naturally divided into two groups according to eligibility for adjuvant therapy (Patients with stage II (T2N0M0 Grade 4) cancer who are included in the IR group and are candidates for administration of Pembrolizumab are an exception that we do not consider for the sake of simplicity): No Adjuvant Therapy (NoAT) for masses classified as B, LR and IR; and Candidate for Adjuvant Therapy (CanAT) for masses classified as HR and VHR.

The management of kidney cancer disease must be personalized and guided by the patient’s comorbidities as well as the tumor’s behavior, size, and histopathology, defining its prognosis. Few studies performed risk stratification analysis to identify prognostic biomarkers of localized renal cancer (e.g., [18,10,9,5]). Thus, directing the management of renal cancer may benefit from additional prognostic factors automatically detected through AI models by using clinical CT images, as well as complementary information such as gender, age, and smoking history. An open research subject would be the accuracy of pre-surgical risk group stratification and the identification of the contributing factors to this assessment.

What's in it for you?

• You get to play a part in advancing scientific efforts to help cancer treatment
• You get to use a collection of incredible data for your algorithms
• ALL challenge participants are invited to submit their 4-page short papers for review to be featured in the ISBI conference! The review will focus primarily on the performance, methods, and clarity of presenting the approach, rather than strictly on scientific novelty.

References

1. Guidelines of the American Urological Association. https://www.auanet.org/guidelines/guidelines/renal- mass-and-localized-renal-cancer-evaluation-management-and-follow-up (2021)
2. KiTS21 challenge website.(2021)
3. National Comprehensive Cancer Network. Kidney Cancer, (version 2.022), (2021)
4. Maier-Hein, Lena, e.t.al.: Bias: Transparent reporting of biomedical image analysis challenges (2020)
5. Chang, T.W., Cheng, W.M., Fan, Y.H., Lin, C.C., Lin, T.P., Huang, E.Y.H., Chung, H.J., Huang, W.J., Weng, S.H.: Predictive factors for disease recurrence in patients with locally advanced renal cell carcinoma treated with curative surgery. Journal of the Chinese Medical Association 84(4), 405–409 (2021)
6. Choueiri, T.K., Tomczak, P., Park, S.H., Venugopal, B., Ferguson, T., Chang, Y.H., Hajek, J., Symeonides, S.N., Lee, J.L., Sarwar, N., et al.: Adjuvant pembrolizumab after nephrectomy in renal-cell carcinoma. New England Journal of Medicine 385(8), 683–694 (2021)
7. Ficarra, V., Novara, G., Secco, S., Macchi, V., Porzionato, A., De Caro, R., Artibani, W.: Preoperative aspects and dimensions used for an anatomical (padua) classification of renal tumors in patients who are candidates for nephron-sparing surgery. European Urology 56(5), 786–793 (2009)
8. Golts, A., Khapun, D., Shats, D., Shoshan, Y., Gilboa-Solomon, F.: An ensemble of 3d u-net based models for segmentation of kidney and masses in CT scans (2021)
9. Gu, L., Wang, Z., Chen, L., Ma, X., Li, H., Nie, W., Peng, C., Li, X., Gao, Y., Zhang, X.: A proposal of post-operative nomogram for overall survival in patients with renal cell carcinoma and venous tumor thrombus. Journal of Surgical Oncology 115(7), 905–912 (2017)
10. Haddad, A.Q., Leibovich, B.C., Abel, E.J., Luo, J.H., Krabbe, L.M., Thompson, R.H., Heckman, J.E., Merrill, M.M., Gayed, B.A., Sagalowsky, A.I., et al.: Preoperative multivariable prognostic models for prediction of survival and major complications following surgical resection of renal cell carcinoma with suprahepatic caval tumor thrombus. In: Urologic Oncology: Seminars and Original Investigations. vol. 33, pp. 388–e1. Elsevier (2015)
11. Heller, N., Isensee, F., Maier-Hein, K.H., Hou, X., Xie, C., Li, F., Nan, Y., Mu, G., Lin, Z., Han, M., et al.: The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging: Results of the KiTS19 challenge. Medical Image Analysis p. 101821 (2020)
12. Heller, N., Sathianathen, N., Kalapara, A., Walczak, E., Moore, K., Kaluzniak, H., Rosenberg, J., Blake, P., Rengel, Z., Oestreich, M., Dean, J., Tradewell, M., Shah, A., Tejpaul, R., Edgerton, Z., Peterson, M., Raza, S., Regmi, S., Papanikolopoulos, N., Weight, C.: The KiTS19 challenge data: 300 kidney tumor cases with clinical context, CT semantic segmentations, and surgical outcomes (2020)
13. Hsieh, P.F., Wang, Y.D., Huang, C.P., Wu, H.C., Yang, C.R., Chen, G.H., Chang, C.H.: A mathematical method to calculate tumor contact surface area: An effective parameter to predict renal function after partial nephrectomy. The Journal of Urology 196(1), 33–40 (2016)
14. Huang, J., Ling, C.X.: Using AUC and accuracy in evaluating learning algorithms. IEEE Transactions on Knowledge and Data Engineering 17(3), 299–310 (2005)
15. IBM Research, Haifa: FuseMedML. https://github.com/IBM/fuse-med-ml (2021). https://doi.org/10.5281/ZENODO.5146491, https://zenodo.org/record/5146491
16. Kutikov, A., Smaldone, M.C., Egleston, B.L., Manley, B.J., Canter, D.J., Simhan, J., Boorjian, S.A., Viterbo, R., Chen, D.Y., Greenberg, R.E., et al.: Anatomic features of enhancing renal masses predict ma- lignant and high-grade pathology: a preoperative nomogram using the renal nephrometry score. European Urology 60(2), 241–248 (2011)
17. Kutikov, A., Uzzo, R.G.: The renal nephrometry score: A comprehensive standardized system for quanti- tating renal tumor size, location and depth. The Journal of Urology 182(3), 844–853 (2009)
18. Nakayama, T., Saito, K., Fujii, Y., Abe-Suzuki, S., Nakanishi, Y., Kijima, T., Yoshida, S., Ishioka, J., Matsuoka, Y., Numao, N., et al.: Pre-operative risk stratification for cancer-specific survival in patients with renal cell carcinoma with venous involvement who underwent nephrectomy. Japanese Journal of Clinical Oncology 44(8), 756–761 (2014)
19. Park, J., Shoshan, Y., Martí, R., Gómez del Campo, P., Ratner, V., Khapun, D., Zlotnick, A., Barkan, E., Gilboa-Solomon, F., Chłędowski, J., Rosen-Zvi, M., Geras, K., et al.: Lessons from the first DBTex challenge. Nature Machine Intelligence 3(8), 735–736 (2021)
20. Ravaud, A., Motzer, R.J., Pandha, H.S., George, D.J., Pantuck, A.J., Patel, A., Chang, Y.H., Escudier, B., Donskov, F., Magheli, A., et al.: Adjuvant sunitinib in high-risk renal-cell carcinoma after nephrectomy. New England Journal of Medicine 375(23), 2246–2254 (2016)
21. Schaffter, T., Buist, D.S.M., Lee, C.I., Nikulin, Y., Ribli, D., Guan, Y., Lotter, W., Jie, Z., Du, H., Wang, S., Feng, J., Feng, M., Kim, H.E., Albiol, F., Albiol, A., Morrell, S., Wojna, Z., Ah- sen, M.E., Asif, U., Jimeno Yepes, A., Yohanandan, S., Rabinovici-Cohen, S., Yi, D., Hoff, B., Yu, T., Chaibub Neto, E., Rubin, D.L., Lindholm, P., Margolies, L.R., McBride, R.B., Rothstein, J.H., Sieh, W., Ben-Ari, R., Harrer, S., Trister, A., Friend, S., Norman, T., Sahiner, B., Strand, F., Guin- ney, J., Stolovitzky, G., and the DM DREAM Consortium: Evaluation of combined artificial intelli- gence and radiologist assessment to interpret screening mammograms. JAMA Network Open 3(3), e200265–e200265 (03 2020). https://doi.org/10.1001/jamanetworkopen.2020.0265, https://doi.org/10.1001/jamanetworkopen.2020.0265
22. Simhan, J., Smaldone, M.C., Tsai, K.J., Canter, D.J., Li, T., Kutikov, A., Viterbo, R., Chen, D.Y., Green- berg, R.E., Uzzo, R.G.: Objective measures of renal mass anatomic complexity predict rates of major complications following partial nephrectomy. European Urology 60(4), 724–730 (2011)
23. Simmons, M.N., Ching, C.B., Samplaski, M.K., Park, C.H., Gill, I.S.: Kidney tumor location measurement using the C index method. The Journal of Urology 183(5), 1708–1713 (2010)
24. Sung, H., Ferlay, J., Siegel, R.L., Laversanne, M., Soerjomataram, I., Jemal, A., Bray, F.: Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer Journal for Clinicians 71(3), 209–249 (2021)