Gate-Before-Generate: A Dual-Layer Architecture for Output-Presence Routing in Chest X-ray Report Generation
BAI, T.-C.; YEH, S.-C.
Show abstract
CXR report generation may require a vision-language model (VLM) to produce both textual findings and spatial bounding boxes. Generative 4B-7B VLMs can emit non-empty outputs on normal images and empty outputs on abnormal images, motivating explicit structural routing. To evaluate whether a hard inference-time gate before a probabilistic VLM changes output-presence performance and to identify the mechanisms underlying paired STRUCT outcomes. We evaluated CXRxVLM v2, combining a frozen microsoft/rad-dino ViT-B/14 encoder with a 768[->]1 logistic probe (threshold 0.0557) and google/medgemma-4b-it with the pamessina/medgemma-4b-it-cure LoRA adapter. A seed=42 stratified cohort of 500 VinDr-CXR train-pool images (250 NORMAL, 250 ABNORMAL) was compared with Lingshu-7B A_baseline and D_fewshot configurations. Exact paired McNemar tests and stratum-level output-presence analyses were prespecified for the primary configurations; MedGemma 1.5 SigLIP was exploratory. CURE achieved STRUCT = 78.0% (390/500; Wilson 95% CI 74.2-81.4), versus 73.8% for Lingshu A_baseline and 74.2% for D_fewshot. Pairwise p-values were 0.0778, 0.1042, and 0.8642. The paired decomposition showed CURE ABNORMAL non-empty-output advantage of +13.6 percentage points versus Lingshu A (p = 0.0012; +14.0 points versus D, p = 0.0007), while Lingshu had higher NORMAL empty-output rates (+5.2 to +6.4 points; p = 0.0106 and p = 0.0004). The full pipeline used 8.87 GB VRAM and 4.92 s/image mean latency; 53% of records used a 25.7 ms warm gate-negative path after model loading. Equivalent overall STRUCT scores concealed two mechanistically different output regimes: CURE favored ABNORMAL non-empty outputs, whereas Lingshu favored NORMAL empty outputs. This paired decomposition, rather than the aggregate score alone, characterizes how hard-gated and probabilistic systems route output presence.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Using Adversarial Images to Assess the Stability of Deep Learning Models Trained on Diagnostic Images in Oncology 90%
- Actionability of Synthetic Data in a Heterogeneous and Rare Healthcare Demographic; Adolescents and Young Adults (AYAs) with Cancer 89%
- Histology-based Prediction of Therapy Response to Neoadjuvant Chemotherapy for Esophageal and Esophagogastric Junction Adenocarcinomas Using Deep Learning 89%
Similar papers in this journal
- Generative AI Enables Medical Image Segmentation in Ultra Low-Data Regimes 91%
- Segmenting functional tissue units across human organs using community-driven development of generalizable machine learning algorithms 89%
- Features fusion or not: harnessing multiple pathological foundation models using Meta-Encoder for downstream tasks fine-tuning 89%
Similar papers in this journal
- Tracking And Predicting COVID-19 Radiological Trajectory Using Deep Learning On Chest X-Rays: Initial Accuracy Testing 92%
- Dual Adversarial Deconfounding Autoencoder for joint batch-effects removal from multi-center and multi-scanner radiomics data 91%
- Toward Understanding COVID-19 Pneumonia: A Deep-learning-based Approach for Severity Analysis and Monitoring the Disease 91%
Similar papers in this journal
- Modular Clinical Decision Support Networks (MoDN)—Updatable, Interpretable, and Portable Predictions for Evolving Clinical Environments 92%
- Development and Validation of a Deep Learning Model for Detecting Signs of Tuberculosis on Chest Radiographs among US-bound Immigrants and Refugees 91%
- Designing a computer-assisted diagnosis system for cardiomegaly detection and radiology report generation 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.