Cleveland Clinic, IBM Build 46-Qubit Cancer AI That Beats Classical Models on 150 Samples
Updated
Updated · Technology Networks · Aug 3
Cleveland Clinic, IBM Build 46-Qubit Cancer AI That Beats Classical Models on 150 Samples
2 articles · Updated · Technology Networks · Aug 3
Summary
Q-CHIPP improved neoantigen prediction under data-limited conditions, giving Cleveland Clinic and IBM a quantum AI model that outperformed classical methods in Science Advances tests.
150 samples were enough for the quantum convolutional neural network to learn biologically meaningful patterns, addressing a key bottleneck in immunotherapy research where training data are scarce.
46 qubits powered a full-length peptide modeling milestone on quantum hardware, extending the approach beyond smaller proof-of-concept tasks.
196 candidate neoantigens from a recently published Cleveland Clinic atlas highlight the scale problem: tumors can contain thousands of possible targets, but only a few trigger immune responses.
The team said sharper neoantigen prediction could speed personalized cancer immunotherapies and next-generation cancer vaccines by identifying stronger therapeutic targets earlier.