Task Objective

This task extends Task 2 by integrating histopathology and transcriptomics to predict recurrence in HR-NMIBC patients. The aim is to model patient-level time-to-recurrence using both morphological and molecular data. Schematic overview of the multimodal prediction pipeline. Histopathology, RNA-seq and clinical data are encoded using pretrained networks and combined for prediction. Please note that RNA-seq data is derived from a selected tumor region within the histopathology slide.

Data Details - Training Data

• 🧠 Histopathology: A single H&E-stained whole slide image (WSI) per patient, with 0.25 µm/pixel resolution at its highest resolution. Note that this WSI is either of an adjacent section of the H&E slide used for bulk RNA-seq ("adjacent"), the same H&E slide with a punched cavity on the tissue section ("used in RNA-seq"), or an H&E slide of another tumor of the same patient ("another tumor").

• 🧠 Histopathology: Binary tissue mask outlining the tissue section

• (Upcoming) A list detailing the type of tissue present in the H&E WSI (i.e. adjacent section, used in RNA-seq, another tumor)

• (Upcoming) Tumor-specific tissue mask corresponding to the RNA-seq region. This can be used for explainability, as pathologists annotate the most aggressive region.

• 🧬 Transcriptomics: Bulk RNA-seq data extracted from selected tumor regions, normalized using DESeq2

• 📋 Clinical Data: Same variables as Task 2.

Feature Type / Values Description
age Integer (years) Age of the patient in years
sex m = male, f = female Biological sex of the patient
smoking Yes / No / Missing Smoking history
tumor Primary / Recurrence Indicates whether the tumor is primary or recurrent
stage TaHG / T1HG / T2HG Tumor stage: Ta (inner lining), T1 (connective tissue), T2 (muscle invasion); all high-grade
substage T1m / T1e T1m: ≤ 0.5mm invasion; T1e: > 0.5mm invasion
grade G2 / G3 G2: moderately differentiated; G3: poorly differentiated
reTUR Yes / No Re-transurethral resection (TUR) performed before BCG induction
LVI Yes / No Lymphovascular invasion observed on H&E slide
variant UCC / UCC + Variant Urothelial carcinoma alone or with variant histology
EORTC High risk / Highest risk European Organization for Research and Treatment of Cancer (EORTC) risk classification
no_instillations Integer Total number of BCG instillations. "-1" indicates missing data.
BRS BRS1 / BRS2 / BRS3 Biomarker-derived BCG response subtype from RNA-seq
Refrence Standard
time_to_HG_recur_or_FUend Float (months) Time to high-grade recurrence or end of follow-up in months
Additional information (not available in evaluation/test)
progression 0 / 1 Progression to advanced disease (1-true/0-false)
HG_recur_BCG_failure 0 / 1 BCG failure (1-true/0-false)
time_to_prog_or_FUend Float (months) Time to progression or end of follow-up in months
time_to_FUend Float (months) Time to end of follow-up in months

Evaluation Metric

Model performance is evaluated using the censored concordance index (C-index). This metric measures the proportion of all comparable patient pairs where the model correctly predicts the ordering of outcomes.

Two patients are considered comparable if:

  • Both experienced the event (e.g., recurrence) at different times, or
  • One experienced the event, and the other was event-free, but with a longer observed follow-up time

A pair is not comparable if both patients experienced the event at the same time.

A pair is considered concordant if the patient with the higher predicted risk score has a shorter actual survival time. In other words, the model correctly orders the two patients in terms of risk.

The C-index ranges from: 0.5 → random predictions 1.0 → perfect concordance

The complete evaluation pipeline, including code for computing the censored concordance index, will be made publicly available to ensure transparency and reproducibility.

Download Training Data

  1. Install AWS CLI https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html
  2. Bucket name: s3://chimera-challenge/task3/
  3. Command line: aws s3 sync --no-sign-request s3://chimera-challenge/task3/ <destination_path>