Clinical Question

Can multimodal scoring refine postoperative risk stratification and treatment-association framing in ccRCC?

This macOS app presents the manuscript workflow as a compact video abstract: from DX whole-slide image intake to MRSS 2.0 subgroup assignment and postoperative interpretation.

DX WSI • KIPAN AIclinicalScore • Six-SNP score • MRSS 2.0

Low-resolution slide context remains visible during WSI scoring. The interface emphasizes the current case, not just the final score, to better match supplementary video expectations in scientific reporting.

Model Inputs

Three evidence streams are fused into one manuscript-aligned score.

The app derives a WSI score from the selected DX slide, combines it with KIPAN AIclinicalScore and six-SNP score, and computes MRSS 2.0 using the prespecified formula and cutoff.

DX Whole-Slide Image

Local path or file upload, restricted to TCGA DX slides for the current workflow.

WSI Score

Derived from native slide inference or communication-efficient 5x upload.

KIPAN AIclinicalScore

Generated from structured clinicopathologic variables through the existing model endpoint.

Six-SNP Score

Composed on device, then merged with the other two modalities into MRSS 2.0.

Primary Output

MRSS 2.0 returns subgroup assignment together with the underlying evidence trail.

During inference, the selected slide remains visible as a low-resolution preview. The final output shows score, subgroup, contribution structure, and the study-aligned postoperative interpretation layer.

Low-resolution H&E whole-slide image preview from a DX slide

Real low-resolution H&E WSI preview shown while the current DX slide is being processed

MRSS 2.0 formula MRSS 2.0 = 3.5424 × WSI_score_z + 1.1255 × six-SNP_score_z + 4.4860 × KIPAN AIclinicalScore_z − 1.543661814634

WSI score0.664

AIclinicalScore0.106

MRSS 2.00.842

High-risk subgroup Potential association with postoperative Sunitinib benefit in the study context.

Manuscript-aligned evidence box Cutoff-based subgroup assignment, model provenance, source pathway, and inference timestamp are shown alongside the result.

Validation and Limitation

Kaplan-Meier separation is displayed with manuscript endpoint wording and subgroup context.

The validation panel uses recurrence-free interval terminology for surgery-centered prognosis and keeps the treatment interpretation as subgroup association rather than standalone prescribing advice.

Research-use framing

High-risk subgroup

Potentially associated with postoperative Sunitinib benefit in the manuscript context.

Decision support only

The result screen emphasizes evidence provenance: model names, WSI source, AI-clinical source, selected slide identifier, formula, cutoff, and inference timestamp.

Kaplan-Meier subgroup separation

The Kaplan-Meier figure is embedded from the manuscript PDF for sharper display in the web iframe.

Limitation: the app should be interpreted as a research or decision-support interface. Final treatment choice still requires clinician review of stage, competing therapies, contraindications, and toxicity profile.

Example Patient

Postoperative ccRCC

ccRCC Post-op

Renal Cancer Radical nephrectomy

Scan

DX Slide

Input

Clinical

Assay

6-SNP

Model

MRSS 2.0

High Risk MRSS 2.0

Sunitinib

Low Risk MRSS 2.0

Follow-up