Objective:

To develop a non-invasive framework for assessing the bone marrow immune microenvironment in multiple myeloma using routine diagnostic examinations, specifically Wright–Giemsa-stained bone marrow aspirate smears and whole-body 18F-FDG PET/CT.

Approach:

• Data Fusion: Employed a contrastive fusion module to align imaging modalities with flow cytometry reference panel, producing an Immune Dysfunction Index (IDI) to evaluate immune competence.

Key Findings:

• Cytomorphologic indices correlated with flow cytometric fractions (Spearman ρ values of 0.68–0.81, p<0.001).
• Unsupervised clustering identified immune-competent and immune-exhausted archetypes with differing progression-free survival and daratumumab response.
• Adding the IDI to conventional risk markers improved the concordance index from 0.58 to 0.75 (p<0.001) and increased the AUC for daratumumab response from 0.55 to 0.81 (p<0.001).

Interpretation:

ImmunoCast-MM provides a non-invasive profiling method for the bone marrow immune microenvironment.

Limitations:

• Retrospective study design may limit generalizability.
• Evaluation conducted in a single cohort of newly diagnosed patients.

Conclusion:

ImmunoCast-MM reframes standard diagnostic examinations as tools for non-invasive profiling of the bone marrow immune microenvironment in multiple myeloma.