Why does circuit depth reduction for current quantum hardware limit the complexity of medical imaging tasks?

Looking at this question and the archive, I need to be direct: these preprints don't address medical imaging specifically, nor do they explain why circuit depth reduction would limit (rather than enable) imaging tasks.

That said, I can sketch the general principle from what's here.

In quantum-classical hybrid algorithms for medical imaging, the quantum processor typically encodes image data and performs some computational task — feature extraction, classification, or generation — before measurement [AG-2024.05-1722]. The number of gates you apply (circuit depth) directly determines two things:

1. Noise accumulation: Each gate introduces errors. Shallower circuits suffer less total noise [AG-2024.02-2049], which is why reducing depth improves performance.

2. Information capacity: Paradoxically, *very shallow* circuits may not have enough expressive power to capture complex image features. Parameterized circuits need sufficient depth to encode the complexity of the problem [AG-2024.05-2574].

So the tension is this: medical imaging tasks are *complex* — they require discriminating subtle anatomical features. A circuit constrained to minimal depth for noise resilience might lack the expressivity to learn or represent those features effectively. The data encoding strategy itself becomes crucial; different ways of mapping image pixels into quantum states have very different information-theoretic costs [AG-2024.05-1722].

The archive doesn't contain a paper that directly analyzes this trade-off for medical imaging or proves a specific depth limit.