EviDent-CBCT: Evidence-Bottlenecked Report Generation from Dental CBCT under Non-Exhaustive Report Supervision

EviDent-CBCT：非详尽报告监督下牙科 CBCT 的证据瓶颈报告生成

Abstract: Dento-maxillofacial cone-beam CT (CBCT) reports may contain dozens of tooth-specific, anatomical, and spatial findings from a single 3D scan. Learning to generate such reports from limited clinical data is challenging because routine reports may not exhaustively document image findings, and a non-mention may reflect either absence or non-reporting.

摘要： 牙颌面锥形束 CT (CBCT) 报告可能包含单次 3D 扫描中数十项针对牙齿、解剖结构和空间位置的发现。由于常规报告可能不会详尽记录所有影像发现，且“未提及”可能意味着“不存在”或“未报告”，因此从有限的临床数据中学习生成此类报告极具挑战性。

We present EviDent-CBCT, an evidence-bottlenecked framework designed for this incomplete supervision. An anatomy-aware network maps each CBCT scan to a discrete record of tooth-level, global, and tooth-IAC evidence. A dental-logic consistency projection reconciles incompatible evidence before a deterministic renderer and an image-blind local language model generate the report using only this record.

我们提出了 EviDent-CBCT，这是一个专为这种不完整监督设计的证据瓶颈框架。一个具备解剖学感知能力的神经网络将每次 CBCT 扫描映射为包含牙齿级、全局级和牙齿-下牙槽神经管 (tooth-IAC) 证据的离散记录。在确定性渲染器和图像盲局部语言模型仅利用该记录生成报告之前，牙科逻辑一致性投影会先对不兼容的证据进行调和。

For tooth-level evidence, reliability-aware training uses eligible non-mentions as reduced-weight negatives, while unreported global and tooth-IAC labels remain unknown. A metal-sensitive input channel preserves intensity cues from dental materials.

对于牙齿级证据，可靠性感知训练将符合条件的“未提及”项作为权重降低的负样本，而未报告的全局和牙齿-IAC 标签则保持未知。一个金属敏感输入通道保留了来自牙科材料的强度线索。

Across three validation runs, EviDent-CBCT achieves $0.666\pm0.006$ merged evidence set-F1 and $0.402\pm0.003$ RadFact-Lite-Dental logical-F1, versus $0.371\pm0.018$ for the strongest controlled direct baseline. In the ODIN 2026 challenge, it ranked second in automated evaluation and third in blinded clinical Arena comparison on the hidden test set. These results support the discrete evidence record as an effective and auditable interface for CBCT report generation.

在三次验证运行中，EviDent-CBCT 的合并证据集 F1 分数为 $0.666\pm0.006$，RadFact-Lite-Dental 逻辑 F1 分数为 $0.402\pm0.003$，而最强的受控直接基准模型仅为 $0.371\pm0.018$。在 ODIN 2026 挑战赛中，它在自动化评估中排名第二，在隐藏测试集的盲法临床竞技场对比中排名第三。这些结果证明，离散证据记录是 CBCT 报告生成中一种有效且可审计的接口。