Hyperpolarized 13C Metabolic MRI L1-506

Medical ImagingReal-time metabolic kinetic imaging via dynamic nuclear polarization (multi-physics joint inverse)δ=5 · advancedL_DAG = 10.4📋 Stub — not mineable

📋

Unclaimed Principle — open for contribution

This Principle is declared in the catalog but has no reference solver, no pinned dataset, and is not registered on-chain. There is no reward pool. Submitting a cert against this Principle today will record the cert for reproducibility but pay zero PWM.

To claim it as a Bounty #7 contribution: open a PR adding (1) a reference solver, (2) ≥1 dataset pinned to IPFS, (3) updates to the L3 manifest with dataset CIDs. After verifier-agent triple-review, the founders' 3-of-5 multisig signs PWMRegistry.register() and the Principle becomes mineable.

Read the contribution guide →Open a claim issue →

Forward model E

I ni t ia l co n d i t i o n s : M_{p} y r (t = 0) = M_{0} \cdot P_{D} N P \cdot e x p (- T_{i} nj ec t / T 1_{p} y r_{p} o l a r i z er); M_{l} a c (0) = M_{a} l a (0) = M_{b} i c (0) = 0; O D E sy s t e m : d M_{p} y r / d t = - k_{P} L \cdot M_{p} y r - k_{P} A \cdot M_{p} y r - k_{P} B \cdot M_{p} y r - M_{p} y r / T 1_{p} y r + R_{i} nj ec t (t); d M_{l} a c / d t = + k_{P} L \cdot M_{p} y r - M_{l} a c / T 1_{l} a c; d M_{a} l a / d t = + k_{P} A \cdot M_{p} y r - M_{a} l a / T 1_{a} l a; d M_{b} i c / d t = + k_{P} B \cdot M_{p} y r - M_{b} i c / T 1_{b} i c; c h e mi c a l - s hi f t - r eso l v e d a c q u i s i t i o n : S_{m} (k, t) = in t e g r a l M_{m} (r, t) \cdot e x p (- i \cdot 2 \cdot p i \cdot k \cdot r) \cdot e x p (- i \cdot o m e g a_{m} \cdot T E) d r + n (k, t); j o in tl y in v er t S_{m} (k, t) \to (k_{P} L, k_{P} A, k_{P} B, T 1_{p} e r_{p} oo l) (r) g i v e ninj ec t i o nb o l u s, M_{0} \cdot P_{D} N P, o m e g a_{m} c h e mi c a l s hi f t s

Hyperpolarized 13C Metabolic MRI: joint multi-physics forward couples (i) DNP polarization transfer where 13C nuclear-spin polarization P(t) is enhanced by >10000x in a polarizer (3 T, 1.4 K, 100 mW microwave irradiation transferring electron polarization to nuclei) then dissolved and injected, with subsequent T1 decay (~30-60 s in vivo); (ii) chemical-exchange ODE kinetics where injected 13C-pyruvate is enzymatically converted to 13C-lactate (via LDH, rate k_PL), 13C-alanine (via ALT, rate k_PA), and 13C-bicarbonate (via PDH, rate k_PB), with each pool experiencing characteristic T1 decay; (iii) chemical-shift-resolved MR readout where spectrally-distinct metabolites (pyruvate ~171 ppm, lactate ~183 ppm, alanine ~178 ppm, bicarbonate ~163 ppm at 13C Larmor frequency) are encoded via spectral-spatial RF pulses, EPI/spiral readouts, or balanced SSFP. The forward DAG has 9 primitives with three coupling constraints (n_c = 3): (i) DNP polarization decay -> available metabolite signal magnitude (T1-weighted); (ii) chemical-exchange ODE -> multi-pool MR signal magnitude (rate-constant-weighted); (iii) T1 relaxation -> time-decay of all metabolite pools (per-pool T1). Recovery is posed as the joint inverse problem that recovers (k_PL, k_PA, k_PB, T1_pyr, T1_lac, T1_ala, T1_bic)(r) from time-resolved chemical-shift-resolved k-space data {S_pyr(k, t), S_lac(k, t), S_ala(k, t), S_bic(k, t)}. Difficulty tier delta = 5 with raw condition number kappa ~ 280 (limited by hyperpolarization decay window ~60-120s and SNR per metabolite) and effective kappa_eff ~ 50 after rate-constant-aware spatiotemporal regularization. Mismatch parameters: dnp_polarization_loss, b1_inhomogeneity, t1_uncertainty, injection_bolus_uncertainty, partial_volume_effect, susceptibility_artifact. Additive Gaussian thermal noise sets the data-fidelity floor; T1 decay sets the irreversible signal floor. See forward_model field for the closed-form joint imaging equation.

L-DAG

L.dnp_polarization -> L.injection_bolus -> L.chemical_exchange_ode -> L.bloch_dynamics -> L.rf_excitation -> L.spectral_spatial_encoding -> L.gradient_readout -> int.spectral -> int.temporal

L.dnp_polarizationL.injection_bolusL.chemical_exchange_odeL.bloch_dynamicsL.rf_excitationL.spectral_spatial_encodingL.gradient_readoutint.spectralint.temporal

Well-posedness W

Existence:: true
Uniqueness:: conditional
Stability:: conditional
κ:: 280

Existence of recovered metabolic rate-constant maps (k_PL, k_PA, k_PB)(r) and per-pool T1 maps is guaranteed within the declared Omega bounds. Uniqueness holds when at least 3 metabolite pools are sampled with sufficient temporal resolution (frame duration <= 5 s) and SNR per metabolite > 10 dB; degenerate cases (all rate constants near zero, or T1 << acquisition window) require regularization. Stability is moderately conditioned (kappa_eff ~ 50 after rate-constant-aware spatiotemporal regularization) — dnp_polarization_loss dominates absolute rate-constant scaling; injection_bolus_uncertainty dominates K_1-equivalent bias; t1_uncertainty contributes a per-pool scaling factor; partial_volume_effect dominates small-structure quantitation. Joint Hadamard well-posedness for the coupled DNP-chemical-exchange-MR forward is established by Bahrami et al. (2014), Larson et al. (2018), Maidens-Gordon-Arcak (2016 control-theoretic identifiability), and Brindle (2015 review on hyperpolarized MR imaging principles).

Solvability C

Solver class:: linear-operator + convex optimisation [least-squares ODE fitting + Tikhonov-regularized k-space inversion; HypMet two-step pipeline; joint reconstruction-kinetics ML] | nonlinear-least-squares [Levenberg-Marquardt for kinetics + IFFT for spatial] | linear-operator + deep neural prior [HypNet, KineticNet]
Convergence rate q:: 2
Complexity:: O(H * W * Z * N_frames * N_metabolites * log(H * W * Z)) per iteration via FFT-based spectral-spatial encoding + ODE forward pass; learned variants O(H W Z N_frames N_metabolites * F_theta_cost) per forward pass

Specs (0)

No L2 specs registered yet for this principle.