HCC Parallel Results Panel: six models, no combined stage
The Liver & HCC Parallel Results Panel reduces repeated data entry while keeping six models separate:
- Child-Turcotte-Pugh, using the standard-bilirubin INR variant in this compact panel
- Original ALBI grade
- OPTN Calculated MELD 3.0 under Policy 9.1.D
- Corrected BCLC 2026 stage
- Original HKLC-9 (2014) stage
- Original Milan criteria
The panel reports each component's own status, source version, result, applicability, and trace. It does not calculate a consensus stage, agreement or discordance, maximum severity, treatment choice, prognosis, transplant candidacy, or allocation priority.
The public form now asks for 18 clinical inputs plus one scope confirmation instead of 32 separate questions. ECOG and its HCC attribution are combined; liver status and LT context are combined; and the three distinct adult MELD listing pathways remain explicit.
Why outputs may differ
These models answer different source questions and use different inputs. BCLC 2026 does not use Child-Pugh as a stage determinant, while original HKLC-9 requires an interpreted Child-Pugh class. Milan uses tumor burden, vascular invasion, and spread but is not a transplant-candidacy decision. Calculated MELD 3.0 is an OPTN policy equation and is not HCC staging.
Differences are therefore described as model-scope differences, not as agreement or disagreement.
Component-specific availability
The panel does not require every component to be interpretable before showing the others.
- Missing sodium suppresses only MELD.
- Missing ascites or encephalopathy suppresses Child-Pugh and HKLC.
- Missing tumor-burden data suppresses only the dependent staging results.
- Diffuse disease can remain interpretable in HKLC while count-dependent BCLC and Milan remain unavailable.
Calculation requires one concise scope confirmation covering adult established chronic liver disease, confirmed HCC at initial complete staging, contemporaneous unconfounded liver values, and calculated-MELD context. No favorable clinical determinant is silently defaulted.
Vascular definitions
Users enter intrahepatic venous invasion, extrahepatic vascular invasion, and extrahepatic spread once. The panel derives each model's source-specific fields:
- BCLC macrovascular invasion: intrahepatic or extrahepatic vascular invasion
- HKLC IVI: intrahepatic venous invasion only
- HKLC EVM: extrahepatic vascular invasion or extrahepatic spread
- Milan evident vascular invasion: either vascular-invasion flag
- Milan regional/distant spread: extrahepatic spread
Verification boundary
The compact public contract and unchanged engine passed 12,292 deterministic implementation assertions on 2026-07-30. This verifies software orchestration against independent source-locked oracles. It is not prospective or independent clinical validation and does not replace the actual source documents or individualized assessment.