BEACON-HCC
ForHCC across all stages, North American practice context
TL;DRNew North American HCC allocation framework; expert decisions 96.6% concordant with BEACON vs 72.4% with BCLC 2025 across 29 cases.
First major allocation framework to place EBRT/SBRT alongside TARE as a first-line locoregional option rather than a BCLC afterthought, extending to Class 1D (Vp1-2) and Class 3B (Vp3/4). The panel's own citation base for Class 3B is RTOG1112, which showed only a trend (p=0.06) for SBRT vs sorafenib, so the elevation runs ahead of randomised evidence.
For a cirrhotic patient with unifocal HCC not eligible for resection or transplant, this framework treats SBRT as a legitimate first-line ablative option rather than a fallback; it does not apply to Class 4A, where the panel explicitly withholds EBRT outside symptom palliation.
SBRT enters the allocation boxes as a first-line locoregional option across Classes 1A through 3A, not a salvage line, which changes referral expectations at tumor board. No dose or fractionation is specified anywhere in the framework, and the Class 3B citation is RTOG1112 at p=0.06, so the elevation runs ahead of randomised evidence.
The framework pushes back on reflexive systemic therapy for macrovascular invasion: Class 1D (Vp1-2) is routed to local or surgical options, and adjuvant therapy after complete response is explicitly not recommended. Systemic therapy stays the cornerstone only for Class 3B, 4A, and 4B.
Resection is preferred without CSPH, transplant with it, and both extend beyond early stage after downstaging response. Class 3B carries a cited RCT showing resection after neoadjuvant radiation beat up-front resection, and Vp1-2 patients with deep durable responses become transplant candidates rather than automatic exclusions.
9 details 4 trials watching
Modified Delphi consensus, 20 North American multidisciplinary experts, iterative discussion and voting over two rounds. Followed by a 29-case pilot concordance exercise against external experts.
Framework spans the full HCC spectrum, from unifocal ≤3 cm (Class 1A) to distant metastases (Class 4B). Class assignment keys on tumor burden, Vp level of portal invasion, and poor prognostic features (>8 cm, AFP >1000 ng/mL, poor differentiation).
EBRT including SBRT is named a first-line locoregional option in Classes 1A through 3A, on par with TARE, RFA/MWA, and TACE. No dose or fractionation is specified anywhere in the framework: the stated rationale is ablative dose delivered selectively to limit radiation-induced liver injury. In Class 4A, EBRT is not routinely indicated outside palliation of symptomatic disease.
Expert decisions matched BEACON in 28 of 29 cases (96.6%) vs 72.4% for BCLC 2025. The single discordant case was combination systemic plus locoregional therapy for vascular invasion reaching the right atrium.
| BEACON class | BCLC 2025 | AJCC v8 | Adaptation from BCLC |
|---|---|---|---|
| Class 1A | Very early (0) or early (A) | Stage 1A or 1B | Aligned; framework elevates SBRT and TARE |
| Class 1B | Early (A) | Stage 1B | Aligned |
| Class 1C | Early (A) | Stage 2 | Aligned |
| Class 1D | Advanced (C) | Stage 2 | Macrovascular invasion may suit local or surgical therapy: TARE, SBRT, resection |
| Class 2 | Early (A) | Stage 1A or 1B | Aligned; framework elevates SBRT and TARE |
| Class 3A | Intermediate (B) or advanced (C) | Stage 3A | Aligned; framework elevates SBRT and TARE |
| Class 3B | Advanced (C) | Stage 3B or 4A | Macrovascular invasion may be amenable to local therapy in selected cases |
| Class 4A | Intermediate (B) | Stage 3A | Extensive intrahepatic spread treated as systemic process |
| Class 4B | Advanced (C) | Stage 4B | Stage and treatment framework aligned |
BCLC keeps radiation outside its core allocation algorithm; BEACON moves it in. The supporting evidence the panel cites is uneven: DOSISPHERE (ORR 71 vs 36%, OS 26.6 vs 10.7 mo) for personalized-dosimetry TARE is randomised, while the SBRT case in Class 3B rests on RTOG1112, which reached only p=0.06 for OS vs sorafenib.
The 72.4% BCLC comparator was scored by BEACON's own authors against cases the same group selected, and 29 cases cannot exercise nine classes evenly. Round-1 per-class agreement is unreadable from the source table because vote counts and class labels arrived unpaired.
The framework's real claim is that intrahepatic burden and Vp level, not BCLC's stage buckets, should drive allocation, and that radiation-based modalities are burden-appropriate rather than stage-restricted. Whether that improves outcomes is untested: validation against empiric clinical data is deferred to the HCC-Live Consortium.
Delphi consensus framework, not an efficacy study. Concordance figures measure opinion against opinion. Elevates EBRT/TARE to first-line locoregional status in a formal allocation system.
- Head-to-head SBRT vs TARE for liver-confined HCC n=146 · primary completion 2028-12 · randomised phase 2 TARE vs SBRT, =3 HCC lesionsrecruiting Comparison of SBRT and SIRT With Combination IO for Locally-advanced, Unresectable HCCs (BIIRTH) Phase 2/3n=106 · primary completion 2034-03 · TACE-SBRT vs Y90 SIRT, PFS primary
- Does BEACON allocation improve survival vs BCLC allocation
- Optimal sequencing of ICI with locoregional therapy active Durvalumab/Tremelimumab in Neoadjuvant and Adjuvant Setting in Patients With HCC Treated by by Percutaneous Ablation Procedure Phase 2n=30 · primary completion 2027-09 · durva/treme before and after percutaneous ablationrecruiting Using Radiotherapy and Immunotherapy to Treat Advanced Liver Cancer Before Transplant Phase 1/2n=48 · primary completion 2031-06 · Y90/SBRT plus atezo-bev, transplant conversion
📚 Sources · 📄 1 paper
Abstract
The longer read
The substance of BEACON-HCC is not the 96.6% concordance figure, which measures little more than whether a framework written by 20 experts agrees with the judgment of experts drawn from the same community. It is the structural decision to move external beam radiation and radioembolization from the margins of the allocation algorithm into its center. BCLC has long treated radiation-based therapy as something practiced but unallocated, present in the literature and absent from the box-and-arrow diagram that most tumor boards actually use. BEACON puts SBRT and TARE in the boxes, and it does so across a wide span of classes, from unifocal small tumors through segmental portal invasion.
That is a defensible reading of where the field has moved, but the evidence backing each placement is uneven in a way the framework's uniform presentation obscures. For TARE the panel can point to randomised data: DOSISPHERE showed that personalized dosimetry roughly doubled ORR (71 vs 36%) and more than doubled median OS (26.6 vs 10.7 months) against standard dosimetry. For SBRT in locally advanced disease the strongest citation is RTOG1112, which compared SBRT plus sorafenib to sorafenib and landed at p=0.06 for overall survival. A trend is a reasonable basis for offering a therapy; it is a thinner basis for elevating it inside a formal allocation system that tumor boards will treat as settled. The panel's own honesty helps here, since Table 3 states plainly that few high-quality data establish superiority of any one locoregional modality over another, which is an argument for modality-agnostic allocation rather than for any particular ranking.
The class definitions are the more durable contribution. Splitting unifocal disease by size and by poor prognostic features, and splitting vascular invasion by Vp level rather than treating any macrovascular invasion as uniformly advanced, tracks the biology better than BCLC's stage C catch-all. A patient with a Vp1 branch thrombus and a patient with main trunk invasion do not share a prognosis or a set of reasonable options, and BCLC's grouping of them has pushed a lot of treatable disease toward systemic therapy by default. BEACON's Class 1D and Class 3B split is the practical payoff of the whole exercise.
What should temper confidence is that the validation exercise cannot test what the framework claims. Twenty-nine cases across nine classes leaves some classes represented by one or two patients, and concordance with external experts measures cultural agreement within North American academic practice, not correctness. The 72.4% BCLC figure was generated by the framework's authors on cases they selected, which is the least persuasive form of comparison available. Notably, the classes where BEACON diverges most from BCLC are also where the panel's own internal agreement was weakest, which is consistent with genuine uncertainty rather than a settled position being formalized. The planned HCC-Live Consortium validation against empiric outcomes is the analysis that would matter, and until it reports this remains a well-reasoned proposal about how to allocate rather than evidence that allocating this way helps patients live longer.