Fixed-Budget Parameter-Efficient Training with Frozen Encoders Improves Multimodal Chest X-Ray Classification

Multimodal chest X-Ray analysis often fine-tunes large vision-language models, which is computationally costly. We study parameter-efficient training (PET) strategies including frozen encoders, BitFit, LoRA, and adapters for multi-label classification on the Indiana University Chest X-Ray dataset (3,851 image-report pairs; 579 test samples). To mitigate data leakage, we redact pathology terms from reports used as text inputs while retaining clinical context. Under a fixed parameter budget ($\mathbf{2. 3 7 M}$ parameters, $\mathbf{2. 5 1 \%}$ of total), all PET variants achieve AUROC between 0.892 and 0.908, outperforming full fine-tuning (0.770 AUROC) which uses 94.3 M trainable parameters, a $40 \times$ reduction. External validation on CheXpert (224,316 images, $58 \times$ larger) confirms scalability: all PET methods achieve $>0.69$ AUROC with $<9 \%$ trainable parameters, with Adapter achieving best performance (0.7214 AUROC). Budgetmatched comparisons reveal that vision-only models (0.653 AUROC, 1.06 M parameters) outperform budget-matched multimodal models (0.641 AUROC, 1.06 M parameters), indicating improvements arise primarily from parameter allocation rather than cross-modal synergy. While PET methods show degraded calibration (ECE: 0.29-0.34) compared to simpler models (ECE: 0.049), this represents a tractable limitation addressable through post-hoc calibration methods. These findings demonstrate that frozen encoder strategies provide superior discrimination at substantially reduced computational cost, though calibration correction is essential for clinical deployment.