Bit by bit, toward recovery. Step by step, toward lasting survival. Until the day we conquer kidney cancer.
涓滴之力，助患者康复；跬步之积，致长久生存；同心同行，必战胜肾癌。
A curated list of authoritative datasets, computational resources, and top-conference papers on kidney cancer (renal cell carcinoma, RCC) — from a computer science perspective. Our goal: accelerate computational research that helps defeat kidney cancer.
Kidney cancer accounts for ~2% of global cancer diagnoses; clear cell RCC (ccRCC) is the most common and lethal subtype. Computation — medical imaging, computational pathology, and multi-omics — is becoming a decisive weapon against it. This repo collects the resources you need to join that fight.
Contents
The KiTS challenge series is the standard benchmark for kidney and kidney tumor segmentation.
| Dataset | Scale | Modality / Annotations | Links |
|---|---|---|---|
| KiTS19 | 300 contrast-enhanced CTs (210 train / 90 test) | Late arterial phase CT; voxel-level kidney + tumor masks, clinical context, surgical outcomes | Challenge · Code · Paper |
| KiTS21 | 300 CTs | Corticomedullary-phase CT; adds renal cyst label; triple annotations per ROI; external-institution test set | Challenge · Code · Paper |
| KiTS23 | 599 CTs (489 train / 110 test) | Multi-phase CT; kidney / tumor / cyst masks with multi-reader majority vote | Challenge · Code |
| KNIGHT | CT + clinical (companion to KiTS21) | Preoperative risk-group classification of kidney tumor patients | Code |
| KIPA 2022 | 70+ contrast CTs | Kidney parsing: kidney, tumor, renal artery, renal vein | Challenge |
| AbdomenCT-1K | 1,112 CTs | Liver/kidney/spleen/pancreas organ masks + kidney tumor subtask (incorporates KiTS) | Code & Data |
| Resource | Contents | Links |
|---|---|---|
| TCGA-KIRC / KIRP / KICH | The three TCGA kidney projects (~530 clear cell, ~290 papillary, ~66 chromophobe tumors): RNA-seq, miRNA, WXS, methylation, CNV, clinical | GDC Portal · UCSC Xena mirror |
| CPTAC ccRCC | Proteogenomic cohort (110 tumor / 84 normal): ~10k proteins, phosphoproteome + genomics + transcriptome. Landmark paper: Clark et al., Cell 2019 | Proteomic Data Commons (PDC000127) · genomic side at GDC project CPTAC-3 |
Radiology collections matched to the genomic cohorts above, hosted on The Cancer Imaging Archive (also available via NCI Imaging Data Commons).
- TCGA-KIRC — 267 subjects, CT/MR/CR, linked to clinical & genomics. Collection · DOI
- TCGA-KIRP — papillary RCC imaging. Collection
- TCGA-KICH — chromophobe RCC imaging. Collection
- CPTAC-CCRCC — CT/MR + histopathology from the CPTAC-3 ccRCC cohort, linkable to PDC proteomics. Collection · IDC mirror
- MMIST-ccRCC — 618 ccRCC patients with CT + MRI, histopathology (WSI), genomics, and clinical data, with single-/multi-modal benchmarks. Dataset page · Paper (MICCAI 2024)
- CT KIDNEY DATASET: Normal-Cyst-Tumor-Stone — 12,446 CT slices (5,077 normal / 3,709 cyst / 2,283 tumor / 1,377 stone), axial + coronal, contrast and non-contrast. Open download on Kaggle.
- KAUH Kidney Tumor Dataset — 8,400 CT images from 120 patients (tumor vs normal, with benign/malignant labels). Paper: Alzu'bi et al., J. Healthcare Engineering 2022, DOI.
⚠️ No open download portal; available via the authors.
- Dartmouth Kidney Cancer Histology Dataset (DHMC) — 563 H&E-stained FFPE whole-slide images (484 resection + 79 biopsy slides, scanned at 20x, distributed at 5x), labeled by two pathologists' consensus for predominant RCC subtype: clear cell, papillary, chromophobe, and oncocytoma, with train/val/test splits. Companion paper: Zhu et al., Scientific Reports 2021. Dataset page (access via request form) · companion code: DeepSlide
- Advanced ccRCC single-cell atlas (SCP1288) — scRNA-seq of tumor + immune cells from advanced RCC patients on immunotherapy (Bi et al., Cancer Cell 2021). Broad Single Cell Portal
Challenge reports & datasets (KiTS series)
- The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging: Results of the KiTS19 Challenge — Heller, Isensee, Maier-Hein et al., MedIA 2021. Meta-analysis of 106 teams; the canonical KiTS19 analysis.
- The KiTS19 Challenge Data: 300 Kidney Tumor Cases with Clinical Context, CT Semantic Segmentations, and Surgical Outcomes — Heller et al., 2019. The KiTS19 dataset descriptor.
- The KiTS21 Challenge: Automatic segmentation of kidneys, renal tumors, and renal cysts in corticomedullary-phase CT — Heller et al., MedIA 2023. KiTS21 official report.
- KiTS21 Proceedings (LNCS, Springer 2022) — 21 participant papers. DOI
- KiTS23 Proceedings (LNCS 14540, Springer 2024) — KiTS23 participant papers. DOI
Segmentation methods
- An attempt at beating the 3D U-Net — Isensee & Maier-Hein, 2019. KiTS19 winner; residual-encoder 3D U-Net, precursor of nnU-Net.
- Kidney tumor segmentation using an ensembling multi-stage deep learning approach — Santini et al., 2019. Top KiTS19 submission; multi-stage 2.5D residual U-Net ensemble.
- A coarse-to-fine framework for the 2021 Kidney and Kidney Tumor Segmentation Challenge — Zhao et al. KiTS21 winner (avg. Dice 0.908). In KiTS21 Proceedings, pp. 53–58.
- Automated 3D Segmentation of Kidneys and Tumors in MICCAI KiTS 2023 Challenge — Myronenko et al. (NVIDIA). KiTS23 winner using MONAI Auto3DSeg.
- Exploring 3D U-Net Training Configurations and Post-Processing Strategies for MICCAI 2023 KiTS — Uhm et al. KiTS23 2nd place, nnU-Net-based.
Computational pathology
- Instance-Based Vision Transformer for Subtyping of Papillary Renal Cell Carcinoma in Histopathological Image — Gao et al., MICCAI 2021. i-ViT with nucleus-centered instance patches for type-1 vs type-2 pRCC subtyping on 171 WSIs. DOI
Foundation models and universal segmentation models where kidney tumor is an explicit task or training target.
- DoDNet: Learning to Segment Multi-Organ and Tumors from Multiple Partially Labeled Datasets — CVPR 2021. Dynamic on-demand network trained on 8 partially labeled datasets incl. KiTS; segments kidney + kidney tumor. arXiv
- CLIP-Driven Universal Model for Organ Segmentation and Tumor Detection — ICCV 2023. Text-prompted universal CT model: 25 organs + 6 tumor types including kidney tumor. Code
- SegVol: Universal and Interactive Volumetric Medical Image Segmentation — NeurIPS 2024 (spotlight). 3D foundation segmentation model; training data includes KiTS23. Proceedings
- ScribblePrompt: Fast and Flexible Interactive Segmentation for Any Biomedical Image — ECCV 2024. Interactive scribble/click/box segmentation trained on 65 datasets incl. KiTS. Proceedings
- VISTA3D: A Unified Segmentation Foundation Model for 3D Medical Imaging — CVPR 2025. NVIDIA/MONAI 3D CT foundation model (127 automatic classes + interactive + zero-shot), trained on 11,454 CT volumes incl. kidney/kidney-tumor collections. Proceedings
- AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation — NeurIPS 2022 (Datasets & Benchmarks). 500 CT + 100 MRI, 15 abdominal organs incl. kidneys. Data
IEEE ISBI
- Towards an Interpretable Radiomics Model for Classifying Renal Cell Carcinomas Subtypes: A Radiogenomics Assessment — ISBI 2019. Interpretable radiomics differentiating ccRCC from non-ccRCC, with radiogenomics correlation. DOI
- A Triple-Stage Self-Guided Network for Kidney Tumor Segmentation — ISBI 2020. Coarse-to-fine CNN on KiTS19. DOI
- An Efficient Hybrid Model for Kidney Tumor Segmentation in CT Images — ISBI 2020. Memory-efficient hybrid 2D/3D model. DOI
- A New Computer-Aided Diagnostic (CAD) System for Precise Identification of Renal Tumors — ISBI 2021. DOI
- Automated Small Kidney Cancer Detection in Non-Contrast Computed Tomography — ISBI 2024. Detection pipeline for renal cancer in non-contrast CT. DOI
- Rethinking Intermediate Layers Design in Knowledge Distillation for Kidney and Liver Tumor Segmentation — ISBI 2024. DOI
- Leveraging Foundation Models for Clinically Instructed Tumor Image Synthesis in Renal Cell Carcinoma — ISBI 2025. Clinically-conditioned RCC image synthesis for augmentation. DOI
- Comparative Analysis of Unsupervised and Supervised Autoencoders for Nuclei Classification in Clear Cell Renal Cell Carcinoma Images — ISBI 2025. Automated nuclei classification for ISUP/WHO grading. DOI
AAAI
- Unpaired Multi-Domain Stain Transfer for Kidney Histopathological Images — AAAI 2022. Unpaired multi-stain virtual staining (H&E/MAS/PAS/PASM) for kidney pathology. Paper
Honest note: IJCAI and KDD (2019–2025) have no kidney-cancer-specific papers that we could verify — their "kidney" hits are kidney-exchange optimization. This is a gap, and arguably an opportunity.
The high-impact journal literature that computational kidney cancer research is built on — molecular atlases, single-cell maps, and clinical AI. (All entries verified against Crossref / PubMed metadata.)
- Comprehensive molecular characterization of clear cell renal cell carcinoma — TCGA Research Network, Nature 2013. The foundational multi-omics atlas of ccRCC; the reference dataset for nearly all downstream computational RCC work.
- Pan-Renal Cell Carcinoma classification and survival prediction from histopathology images using deep learning — Tabibu et al., Scientific Reports 2019. Early, widely cited CNN study: RCC subtyping + survival prediction directly from TCGA WSIs.
- nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation — Isensee et al., Nature Methods 2021. The de-facto standard segmentation framework; validated on (and winning) KiTS19.
- Data-efficient and weakly supervised computational pathology on whole-slide images — Lu et al., Nature Biomedical Engineering 2021. CLAM: attention-based MIL for gigapixel WSIs, with RCC subtyping (ccRCC/pRCC/chRCC) as a core benchmark.
- Development and evaluation of a deep neural network for histologic classification of renal cell carcinoma on biopsy and surgical resection slides — Zhu et al., Scientific Reports 2021. RCC subtype classification on biopsy + resection slides (the Dartmouth DHMC dataset).
- A single-cell map of dynamic chromatin landscapes of immune cells in renal cell carcinoma — Kourtis et al., Nature Cancer 2022. scATAC-seq atlas of the RCC immune microenvironment, linking chromatin dynamics to immunotherapy response.
- A multi-classifier system integrated by clinico-histology-genomic analysis for predicting recurrence of papillary renal cell carcinoma — Huang et al., Nature Communications 2024. Multimodal (WSI + lncRNA + clinicopathologic) classifier for pRCC recurrence; code publicly released.
- A multimodal AI model for precision prognosis in clear cell renal cell carcinoma: A multicenter study — Zang et al., npj Digital Medicine 2025. Fuses clinical data, contrast-enhanced CT, and WSI features; outperforms Leibovich/UISS and KEYNOTE-564 stratification.
- Artificial Intelligence Links CT Images to Pathologic Features and Survival Outcomes of Renal Masses — Xiong et al., Nature Communications 2025. Multicenter deep-learning system predicting malignancy, WHO/ISUP grade, and survival from preoperative CT.
- A disease-centric vision-language foundation model for precision oncology in kidney cancer — Tao et al., Nature Communications 2026. Kidney-cancer-specific vision-language foundation model for radiologic diagnosis and prognostication.
- Genomic correlates of response to immune checkpoint therapies in clear cell renal cell carcinoma — Miao et al., Science 2018. Landmark WES study linking PBRM1 loss to checkpoint-blockade response; one of the most reused immunogenomics datasets in the field.
- Single-cell transcriptomes from human kidneys reveal the cellular identity of renal tumors — Young et al., Science 2018. First large scRNA-seq atlas of fetal/adult kidney and renal tumors; established the proximal-tubule cell-of-origin of ccRCC.
- VHL substrate transcription factor ZHX2 as an oncogenic driver in clear cell renal cell carcinoma — Zhang et al., Science 2018. Integrative genomic screening identifying ZHX2 as a VHL-regulated driver.
- Aberrant activation of m6A demethylase FTO renders HIF2α-low/− clear cell renal cell carcinoma sensitive to BRD9 inhibitors — Zhang et al., Science Translational Medicine 2021. Epigenomic/m6A profiling revealing a synthetic-lethal therapeutic axis.
- Accurate detection of benign and malignant renal tumor subtypes with MethylBoostER: An epigenetic marker-driven learning framework — Rossi et al., Science Advances 2022. ML classifier on DNA-methylation signatures for renal tumor subtyping, with cell-free DNA early-detection application.
- Intratumoral mycobiome heterogeneity influences the tumor microenvironment and immunotherapy outcomes in renal cell carcinoma — Mou et al., Science Advances 2025. Multi-cohort computational analysis of fungal communities in RCC and immunotherapy response.
- Single-cell epigenetic profiling reveals a tumor-intrinsic interferon response program in ccRCC tied to poor prognosis and BAP1 loss — Camp et al., Science Advances 2026. Single-cell multi-omics linking BAP1 loss to an interferon program in ccRCC.
- An Immune Atlas of Clear Cell Renal Cell Carcinoma — Chevrier et al., Cell 2017. Landmark high-dimensional CyTOF immune atlas of ccRCC; foundational TME reference.
- Multilevel Genomics-Based Taxonomy of Renal Cell Carcinoma — Chen et al., Cell Reports 2016. TCGA integrative multi-platform clustering defining nine RCC subtypes.
- Timing the Landmark Events in the Evolution of Clear Cell Renal Cell Cancer: TRACERx Renal — Mitchell et al., Cell 2018. Multi-region phylogenetic reconstruction timing ccRCC driver events.
- Deterministic Evolutionary Trajectories Influence Primary Tumor Growth: TRACERx Renal — Turajlic et al., Cell 2018. Multi-region genomics defining evolutionary subtypes that predict clinical course.
- The Cancer Genome Atlas Comprehensive Molecular Characterization of Renal Cell Carcinoma — Ricketts et al., Cell Reports 2018. Pan-RCC TCGA analysis across ccRCC, papillary, and chromophobe.
- Integrated Proteogenomic Characterization of Clear Cell Renal Cell Carcinoma — Clark et al., Cell 2019. The CPTAC proteogenomic landmark (103 treatment-naive ccRCC); data via PDC/cBioPortal/LinkedOmics.
- Tumor and Immune Reprogramming during Immunotherapy in Advanced Renal Cell Carcinoma — Bi et al., Cancer Cell 2021. scRNA-seq of metastatic RCC pre/post immune checkpoint blockade (dataset: SCP1288).
- Progressive Immune Dysfunction with Advancing Disease Stage in Renal Cell Carcinoma — Braun et al., Cancer Cell 2021. Large scRNA-seq + TCR atlas of immune dysfunction across RCC stages.
- Single-Cell Sequencing Links Multiregional Immune Landscapes and Tissue-Resident T Cells in ccRCC to Tumor Topology and Therapy Efficacy — Krishna et al., Cancer Cell 2021. Multiregional scRNA + TCR sequencing (167k cells) linking immune topology to ICB response.
- Single-Cell Protein Activity Analysis Identifies Recurrence-Associated Renal Tumor Macrophages — Obradovic et al., Cell 2021. VIPER-based protein activity inference on a ccRCC TME atlas.
- Mapping Single-Cell Transcriptomes in the Intra-Tumoral and Associated Territories of Kidney Cancer — Li et al., Cancer Cell 2022. Multiregional single-cell atlas across tumor core, interface, normal tissue, and blood.
- Histopathologic and Proteogenomic Heterogeneity Reveals Features of Clear Cell Renal Cell Carcinoma Aggressiveness — Li et al., Cancer Cell 2023. CPTAC follow-up: 305 tumor segments integrating histopathology, proteogenomics, and metabolomics.
- Clear Cell Renal Cell Carcinoma: Discrimination from Other Renal Cell Carcinoma Subtypes and Oncocytoma at Multiphasic Multidetector CT — Young et al., Radiology 2013. The foundational quantitative-imaging paper for non-invasive renal mass characterization.
- Application of Machine Learning Models to Predict Recurrence After Surgical Resection of Nonmetastatic Renal Cell Carcinoma — Khene et al., European Urology Oncology 2023. Multicenter ML models outperforming standard clinicopathologic scores for post-surgery recurrence.
- Multimodal Recurrence Scoring System for Prediction of Clear Cell Renal Cell Carcinoma Outcome: A Discovery and Validation Study — Gui et al., The Lancet Digital Health 2023. Multimodal (clinical + histopathology) deep-learning recurrence score with external validation.
- Deep Learning Assessment of Small Renal Masses at Contrast-enhanced Multiphase CT — Dai et al., Radiology 2024. Deep-learning identification of benign small renal masses, externally tested at five independent hospitals.
Honest note: we could not verify any RCC-central computational paper in JAMA Oncology or Lancet Oncology (2019–2025); Cell Genomics, Patterns, and Med likewise have none we could confirm. Another set of gaps — and opportunities.
Segmentation frameworks
- nnU-Net — Self-configuring segmentation framework (Nature Methods 2021). An nnU-Net ensemble won KiTS19; it dominates the KiTS lineage.
- MONAI — PyTorch-based medical imaging AI framework (Core / Label / Deploy); its Auto3DSeg won KiTS23.
Computational pathology (WSI)
- CLAM — Weakly-supervised attention-based MIL for whole-slide classification (Nature BME 2021); includes RCC subtyping (ccRCC/pRCC/chRCC) experiments on TCGA.
- dlrcc — Deep learning for RCC: subtype classification + ccRCC survival prediction from 6,340 TCGA WSIs, with demo notebooks.
Genomics
- TCGAbiolinks — The standard R/Bioconductor package to query, download, and analyze GDC/TCGA data — the usual entry point for TCGA-KIRC multi-omics.
Diagnosis & prognosis
- Survival_CTplusClinical — Multimodal deep learning for personalized RCC prognosis, integrating CT with clinical data (Cancers 2023).
- deep-kidney-cancer — Deep learning for end-to-end kidney cancer diagnosis on multi-phase abdominal CT.
Contributions welcome — datasets, papers, tools, or corrections. Please open an issue or PR. Guidelines:
- Verify before you add: every entry must have a working link to an official source (arXiv, DOI, conference proceedings, or official repo).
- Kidney cancer must be a real component of the work (task, dataset, or evaluation) — not an incidental mention.
- Follow the existing format: title — authors, venue + year, one-line description, links.