Utilizing 3D Fast Spin Echo Anatomical Imaging to Reduce the Number of Contrast Preparations in $T_{1\rho}$ Quantification of Knee Cartilage Using Learning-Based Methods

Chaoxing Huang; Fan Xiao; James F. Griffith; Junru Zhong; Ki-Wai Kevin Ho; Queenie Chan; Siyue Li; Tim-Yun Michael Ong; Weitian Chen; Ziqiang Yu

REVIEW 5 major objections 5 minor 34 references

Utilizing 3D Fast Spin Echo Anatomical Imaging to Reduce the Number of Contrast Preparations in $T_{1\rho}$ Quantification of Knee Cartilage Using Learning-Based Methods

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A deep network can estimate T1ρ maps of knee cartilage from a proton-density-weighted anatomical image and one T1ρ-weighted image, holding regional error below 5%.

desk verdict A clean feasibility study showing a U-Net can map PD-FSE plus one T1rho-weighted image to regional T1rho maps, but the clinical-accuracy claim overreaches and the scan-time savings depend on whether the PD is already in the protocol. read the letter →

arxiv 2502.08973 v1 pith:QRYB6BTD submitted 2025-02-13 eess.IV

Junru Zhong , Chaoxing Huang , Ziqiang Yu , Fan Xiao , Siyue Li , Tim-Yun Michael Ong , Ki-Wai Kevin Ho , Queenie Chan

show 2 more authors

James F. Griffith Weitian Chen

This is my paper · ORCID

classification eess.IV

keywords T1ρmappingkneecartilageosteoarthritisdeeplearningU-Netfastspinechoprotondensity-weightedquantitativeMRI

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that T1ρ relaxation maps of knee cartilage—a quantitative MRI biomarker for osteoarthritis—can be estimated from just two images: a proton-density-weighted anatomical fast spin-echo image plus a single T1ρ-weighted image, instead of the usual four T1ρ-weighted contrasts. The authors train a 2D U-Net and an MLP on 40 participants, using nonlinear least-squares fits from four T1ρ-weighted volumes as ground truth. The best deep-learning model in every tested input combination keeps regional percentage error below 5%, the clinical reproducibility target; for PD-plus-TSL=10 inputs the U-Net achieves 4.12 ± 3.01%, and for PD-plus-TSL=50 inputs 4.03 ± 2.63%. If this generalizes, it shortens knee cartilage T1ρ exams and lowers RF energy deposition, making quantitative cartilage assessment easier to fold into routine clinical MRI.

What carries the argument

The load-bearing object is a 2D U-Net with an output range limiter that estimates T1ρ voxel by voxel from a 64×64 patch pair of registered, Gaussian-smoothed PD-weighted and T1ρ-weighted FSE images. The limiter clamps predictions to [10, 100] ms using ReLU and a min-max gate, encoding prior knowledge that cartilage T1ρ lies in this range. Unlike an MLP that sees only voxel intensities, the U-Net pulls spatial context from both images; the paper's ROI-masking experiment shows that context outside the cartilage matters. The fitting target is the two-parameter mono-exponential equation $I_k = I_0 e^{-TSL_k/T_{1\rho}}$, with the PD image standing in for $I_0$.

What would settle it

Run the trained 2D U-Net on a held-out cohort scanned with a different scanner vendor or field strength; if the PD-weighted surrogate assumption is truly learned, the regional percentage error should stay below 5%, but if the network has overfit the 40-subject sampling, errors will rise sharply, mirroring the 52% RPE that the same PD-plus-TSL=10 input produces under NLLS fitting without learning.

Watch

Extended reading notes

Core claim

The central claim is that a proton-density-weighted FSE image can replace the TSL=0 image in the standard two-parameter mono-exponential T1ρ model, provided a deep network performs the fit. The network is treated as a noise-robust universal approximator that learns to correct the contrast mismatch between PD-weighted and true TSL=0 contrast. Using only a PD-weighted FSE volume and one T1ρ-weighted volume acquired at TSL=10 or 50 ms, the best 2D U-Net produces regional percentage errors of roughly 4%, comparable to its performance with two true T1ρ-weighted images, while the two-point NLLS baseline with the same PD-based inputs fails at 52% and 33% RPE. The authors therefore claim that the method reduces the number of T1ρ contrast preparations to one, cuts scan time, and permits shorter spin-lock times that ease hardware and SAR constraints.

Load-bearing premise

The result rests on the assumption that a PD-weighted FSE image can stand in for the TSL=0 T1ρ-weighted image and that the network's learned correction for that contrast mismatch keeps working on new patients; the paper's own NLLS baseline shows that without the learned correction the substitution fails (RPE ≈ 52%).

Editorial extensions

If this is right

Knee cartilage T1ρ maps can be acquired with one T1ρ-weighted volume plus the PD-weighted anatomical volume already used in many knee protocols, rather than four T1ρ-weighted volumes.
Shorter spin-lock times, such as TSL=10 ms, remain within the under-5% regional error target, which relaxes RF amplifier and SAR constraints.
Deep-learning fitting from PD-plus-single-TSL inputs substantially outperforms two-point NLLS fitting from the same inputs, so the claimed gain is not simply the signal equation.
The 2D U-Net benefits from anatomical context outside the cartilage ROI; masking out those voxels degrades performance.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Whether the PD-for-TSL=0 substitution is safe across scanner manufacturers, field strengths, and fat-suppression variants is untested; a cross-site study would decide whether the learned correction is a physical mapping or a cohort-specific shortcut.
If the PD-to-TSL=0 mapping is learnable, the same two-image pattern may extend to T2 mapping or other joints, since many protocols already include proton-density-weighted anatomical images.
The paper points toward deriving T1ρ without dedicated spin-lock preparation, but its experiments stop short of that; testing this would require replacing the T1ρ-weighted input with a CPMG-derived image.
The U-Net's advantage over NLLS in low-SNR conditions suggests it regularizes noise; quantifying that bias-variance trade-off against the theoretical precision limit for unbiased estimators is a natural next step.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

Quick take: this is a competent retrospective feasibility study, not a clinical breakthrough. The genuinely new bit is substituting a PD-weighted FSE image for the TSL=0 T1rho-weighted image and letting a U-Net correct the contrast mismatch. In this 40-subject, single-scanner dataset, the best models get regional percentage errors around 4% against the four-image NLLS reference, which is good evidence that the mapping can be learned.

What I like: the experiments are well organized. Five-fold cross-validation with a fixed split, two architectures, an ablation on ROI masking, and per-subject standard deviations reported. The finding that masking hurts is non-obvious and worth reporting. The limitations section is honest, and the paper doesn't oversell the novelty — it explicitly builds on prior deep learning T1rho work.

The soft spots are real but not fatal. The PD image is not I0; the paper's own two-point NLLS with PD as I0 gives 52% RPE. So the U-Net is learning a contrast correction that is specific to this scanner, this PD protocol, and this training distribution. With 40 subjects and no external validation, sub-5% RPE is an in-distribution result. The QIBA repeatability coefficient cited for the 5% threshold is about test-retest reproducibility, not agreement with a numerical reference, so claiming 'clinical standards' is a category error. And the scan-time argument is shaky: the PD FSE takes 7:20 min while the T1rho series takes 4:02 min. Unless the PD is already part of the standard exam, you are adding time, not saving it. The savings only materialize in a protocol that already includes PD FSE.

None of this sinks the core feasibility claim. The method does appear to estimate regional T1rho values in this dataset. Who should read it: anyone working on accelerated quantitative MRI, especially T1rho/T2 mapping. It deserves a serious referee; I'd send it to review, keep the technical claims, and push the authors to temper the clinical language and, ideally, add a hold-out site or at least a different PD protocol.

For my own work: wouldn't cite it in the next year, but I'd bring it to reading group as an example of a clean feasibility study with a clear limitation in generalizability.

Referee Report

5 major / 5 minor

Summary. The manuscript proposes an accelerated T1ρ quantification method for knee cartilage that uses one PD-weighted 3D FSE anatomical image and one T1ρ-weighted FSE image as inputs to deep learning models (a 2D U-Net and a 1D MLP), replacing the conventional acquisition of four T1ρ-weighted images. T1ρ maps estimated by these models are compared with ground truth maps from an NLLS fit of four T1ρ-weighted images (TSL=0,10,30,50 ms). The study evaluates four input combinations (PD-w or T1ρ-w TSL=0 with TSL=10 or 50 ms), three model variants, and masked versus unmasked U-Net inputs, using 5-fold cross-validation on 40 subjects from a single 3T Philips scanner. The authors report mean regional percentage error (RPE) below 5% for the best-performing models in all scenarios, including PD-based inputs, and concluded that the approach reduces scan time and maintains clinical standards.

Significance. If the claims hold, the approach could reduce the number of T1ρ contrast preparations needed for cartilage T1ρ mapping, potentially easing integration into clinical workflows. The experimental design is clear: 5-fold cross-validation, standard metrics (MAE, MAPE, RE, RPE), and ablations over inputs, architectures, and masking. The inclusion of an NLLS reference for each input combination is a useful benchmark. The central in-sample result—that a U-Net can map a PD image plus a single T1ρ-weighted image to a T1ρ map with mean regional error below 5%—is plausible and internally consistent. However, the clinical significance is currently overstated: the QIBA threshold is applied to a different statistical quantity, the scan-time reduction is not supported by Table 1, and no external validation is provided, so the sub-5% RPE remains an in-distribution result on a single-scanner, 40-subject dataset.

major comments (5)

[Section 2.4.5, Eq. (6)] The translation of the QIBA within-subject coefficient of variation (4–5%) into a target RPE of 5% is not justified: the QIBA CV describes test–retest repeatability of repeated T1ρ measurements, not agreement between a learned estimate and an NLLS reference. Moreover, Table 3 reports mean RPE ± SD; for PD-w TSL=10 the mean RPE is 4.12 ± 3.01%, so a substantial fraction of subjects have RPE exceeding 5%. The abstract's 'below 5%' is a mean, not a per-subject guarantee, and the paper should report the percentage of subjects meeting the threshold.
[Section 2.1, Eq. (1), Table 3] The hypothesis that a PD-weighted FSE image has contrast comparable to the TSL=0 T1ρ-weighted image contradicts the definition in Eq. (1), where I0 is the TSL=0 image from the same magnetization-prepared acquisition. The paper's own NLLS reference using PD as I0 gives RPE 52.35 ± 6.10% (Table 3), quantifying the contrast mismatch. The deep network therefore does not simply approximate Eq. (1) for PD-based inputs; it learns a dataset-specific mapping from PD to the T1ρ distribution. With 40 subjects from one 3T Philips system and a fixed PD protocol, the sub-5% RPE is an in-distribution result, and no external validation (different sites, field strengths, vendors, or PD normalization) is provided. The claim that the method 'maintains clinical standards' requires such validation.
[Table 1, Abstract] The claimed reduction in scan time is not supported by the acquisition parameters in Table 1. The PD-weighted FSE acquisition takes 7:20 min and the T1ρ-weighted acquisition takes 4:02 min; using both as inputs requires 11:22 min, which is longer than the four T1ρ-weighted images used for the ground truth (4:02 min). The conclusion that the method 'reduces scan time' is only valid if the PD volume is already acquired clinically as part of a standard knee MRI protocol, but the manuscript does not establish this condition. Please provide such evidence or revise the conclusion.
[Section 2.4.4, Table 3] The reference NLLS method uses the same two inputs (e.g., PD and one T1ρ-weighted image) and incorrectly assumes that Eq. (1) holds with PD as I0. This is a physically invalid assumption, so the comparison in Table 3 does not demonstrate that the deep learning method 'outperforms' a legitimate NLLS approach; it demonstrates that a trained network can compensate for the contrast mismatch. A fairer NLLS baseline would fit with an unknown scaling factor between PD and I0, or the comparison should be framed as a learned mapping versus a physics-based fit under a misspecified model.
[Section 3, Table 3] For PD-based inputs, the voxel-level errors remain large (e.g., unmasked U-Net with PD-w TSL=50: MAE 6.53 ± 1.04 ms, MAPE 14.73 ± 2.10%) while the RPE is 4.03 ± 2.63%. The low regional error results from cancellation of positive and negative voxel errors in regional averaging. The paper should report the full distribution of subject-level RPE (the boxplots in Figure 5 partially address this) and discuss whether the large voxel-level errors are clinically acceptable for applications that use voxel-wise T1ρ values, since the clinical claim is based on regional metrics.

minor comments (5)

[Section 4] In the Discussion, 'a shorter T SKk' should read 'a shorter TSLk'.
[Figure 6 caption] The caption contains a typo: 'osteoarthriths' should be 'osteoarthritis'.
[Equation (2)] The limiter expression 'ˆy = {ymin, ReLU(x) + ymin, ymax}' is nonstandard; it should be written as a clamping operation, e.g., 'ˆy = clamp(ReLU(x) + ymin, ymin, ymax)', to avoid ambiguity.
[Table 3] The column header 'PD-w2 TSL=10ms' is confusing because the superscript '2' appears to be a footnote marker but is not explained in the table caption. Please reformat the headers for clarity.
[Section 2.3.1] The statement 'underwent Gaussian smoothing with a radius of three' does not specify the units of the radius (voxels or millimeters) or the kernel size; please clarify.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the result is supervised regression to an external NLLS ground truth under cross-validation, not a derivation from fitted constants; self-citations are contextual and non-load-bearing.

full rationale

The paper is an empirical supervised-regression study rather than a derivation. T1rho maps are produced by training a 2D U-Net or 1D MLP on 40 subjects with five-fold cross-validation, with ground truth provided by NLLS fitting of four T1rho-weighted images (Section 2.4.4). Equation (1) is used to motivate the two-input design, and PD is hypothesized as a surrogate for I0 (Section 2.1), but the network output is not algebraically derived from Eq. (1); therefore the fact that PD is not truly I0 (quantified by the paper's own NLLS baseline, RPE 52.35 +/- 6.10% for PD-w + TSL=10ms in Table 3) is an empirical correctness and generalization issue, not a definitional circularity. The RPE below 5% is an evaluation outcome, not a quantity forced by construction; regional averaging can make RPE low even when voxel-wise MAE remains large, but that is a metric choice, not a circular step. The authors explicitly acknowledge the main limitations in the Discussion (retrospective design, unoptimized acquisition parameters, absence of longitudinal validation), and these limitations are weighed in the verdict but do not constitute circular reasoning. Self-citations to the authors' prior dataset (ref. 17) and pulse-sequence papers (refs. 13-15) provide data and sequence context; the methods are described in the text and the target maps come from an external NLLS reference, so the load-bearing result does not reduce to a self-citation chain. No uniqueness theorem or ansatz is smuggled in via citation. The score of 1 reflects only the presence of non-load-bearing self-citations for the dataset and sequence; no specific circular step was identified.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central method rests on several domain assumptions that are reasonable in quantitative MRI but unvalidated externally. The PD-as-I0 surrogate is the most consequential: without it, the acquisition strategy loses its premise; the DL model compensates but only within the training distribution. No new physical entities are introduced.

free parameters (4)

Output range limiter bounds (ymin, ymax) = ymin=10 ms, ymax=100 ms
Chosen from prior experience (Section 2.3.2), clamps every prediction and can bias estimates at the extremes.
Gaussian smoothing radius = 3 voxels
Applied to all inputs (Section 2.3.1); affects SNR and spatial resolution.
Training schedule and optimizer settings = lr=0.001, exponential decay 0.9, 1000 epochs; MLP: RMSProp lr=0.001, weight decay 0.0003, batch 512
Standard but dataset-specific choices; not derived from first principles.
Patch sampling bias toward cartilage ROI = Not quantified
Patches are randomly cropped with higher probability near the ROI (Section 2.3.2); the exact probability is not stated and affects the training distribution.

assumptions (6)

domain assumption The T1ρ signal follows the mono-exponential model Ik = I0 exp(-TSL_k / T1ρ) (Eq. 1).
Standard model used for ground truth NLLS fitting and as the premise that two images encode T1ρ.
ad hoc to paper A PD-weighted FSE image has contrast comparable to a TSL=0 T1ρ-weighted image.
Hypothesized in Section 2.1; the paper's NLLS baseline with PD as I0 gives RPE around 52%, so the validity of this premise is only rescued by the deep network's learned correction.
domain assumption NLLS fitting from four T1ρ-weighted images provides ground truth T1ρ maps.
No phantom, histology, or other independent reference is used; all training targets and evaluation metrics depend on this reference (Section 2.4.4).
ad hoc to paper The QIBA within-subject coefficient of variation of 4-5% for test-retest T1ρ maps can be translated into an RPE threshold of 5% against this NLLS ground truth.
Section 2.4.5 equates two different statistical quantities without demonstration.
domain assumption Rigid, affine, and symmetric deformable registration accurately aligns PD and T1ρ volumes.
Relies on a validated method (ref 18), but registration errors propagate into inputs, training targets, and evaluation (Section 2.3.1).
domain assumption The five-fold cross-validation split is subject-level, preventing leakage of slices from the same subject across train and test.
The paper says 'same five-fold split' (Section 2.3.2) but does not explicitly state that the split is by participant; if not, reported errors would be optimistic.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Utilizing 3D Fast Spin Echo Anatomical Imaging to Reduce the Number of Contrast Preparations in $T_{1\rho}$ Quantification of Knee Cartilage Using Learning-Based Methods." pith.science (2026). https://pith.science/paper/QRYB6BTD

@misc{pith2026250208973,
  author       = {Pith},
  title        = {Pith review of: Utilizing 3D Fast Spin Echo Anatomical Imaging to Reduce the Number of Contrast Preparations in $T_1\rho$ Quantification of Knee Cartilage Using Learning-Based Methods},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QRYB6BTD}},
  note         = {Machine review of arXiv:2502.08973}
}

abstract

Purpose: To propose and evaluate an accelerated $T_{1\rho}$ quantification method that combines $T_{1\rho}$-weighted fast spin echo (FSE) images and proton density (PD)-weighted anatomical FSE images, leveraging deep learning models for $T_{1\rho}$ mapping. The goal is to reduce scan time and facilitate integration into routine clinical workflows for osteoarthritis (OA) assessment. Methods: This retrospective study utilized MRI data from 40 participants (30 OA patients and 10 healthy volunteers). A volume of PD-weighted anatomical FSE images and a volume of $T_{1\rho}$-weighted images acquired at a non-zero spin-lock time were used as input to train deep learning models, including a 2D U-Net and a multi-layer perceptron (MLP). $T_{1\rho}$ maps generated by these models were compared with ground truth maps derived from a traditional non-linear least squares (NLLS) fitting method using four $T_{1\rho}$-weighted images. Evaluation metrics included mean absolute error (MAE), mean absolute percentage error (MAPE), regional error (RE), and regional percentage error (RPE). Results: Deep learning models achieved RPEs below 5% across all evaluated scenarios, outperforming NLLS methods, especially in low signal-to-noise conditions. The best results were obtained using the 2D U-Net, which effectively leveraged spatial information for accurate $T_{1\rho}$ fitting. The proposed method demonstrated compatibility with shorter TSLs, alleviating RF hardware and specific absorption rate (SAR) limitations. Conclusion: The proposed approach enables efficient $T_{1\rho}$ mapping using PD-weighted anatomical images, reducing scan time while maintaining clinical standards. This method has the potential to facilitate the integration of quantitative MRI techniques into routine clinical practice, benefiting OA diagnosis and monitoring.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 31 canonical work pages

[1]

W., Guermazi A., Demehri S., Wirth W., Kijowski R

Roemer F. W., Guermazi A., Demehri S., Wirth W., Kijowski R.. Imaging in Osteoarthritis. Osteoarthritis and Cartilage. 2022;30(7):913–934

work page 2022
[2]

The QIBA Proﬁle for MRI-based Compositional Imaging of Knee Cartilage

Chalian Majid, Li Xiaojuan, Guermazi Ali, et al. The QIBA Proﬁle for MRI-based Compositional Imaging of Knee Cartilage. Radiology. 2021;301(2):423–432. 12 ZHONG et al

work page 2021
[3]

The Role of Relax- ation Times in Monitoring Proteoglycan Depletion in Articular Cartilage

Mlyn´ arik Vladim ´ ır, Trattnig Siegfried, Huber Monika, Zembsch Alexander, Imhof Herwig. The Role of Relax- ation Times in Monitoring Proteoglycan Depletion in Articular Cartilage. Journal of Magnetic Resonance Imaging. 1999;10(4):497–502

work page 1999
[4]

Proteoglycan-Induced Changes in T1ρ-relaxation of Articular Cartilage at 4T

Akella Sarma V.S., Reddy Regatte Ravinder, Gougoutas Alexander J., et al. Proteoglycan-Induced Changes in T1ρ-relaxation of Articular Cartilage at 4T. Magnetic Resonance in Medicine. 2001;46(3):419–423

work page 2001
[5]

W., Sharaﬁ Azadeh, Otazo Ricardo, Regatte Ravinder R

Zibetti Marcelo V. W., Sharaﬁ Azadeh, Otazo Ricardo, Regatte Ravinder R.. Accelerated Mono- and Biexponen- tial 3D-T1ρ Relaxation Mapping of Knee Cartilage Using Golden Angle Radial Acquisitions and Compressed Sens- ing. Magnetic Resonance in Medicine. 2020;83(4):1291– 1309

work page 2020
[6]

SuperMAP: Deep Ultrafast MR Relaxometry with Joint Spatiotemporal Undersampling

Li Hongyu, Yang Mingrui, Kim Jee Hun, et al. SuperMAP: Deep Ultrafast MR Relaxometry with Joint Spatiotemporal Undersampling. Magnetic Resonance in Medicine. 2023;89(1):64–76

work page 2023
[7]

Cram´ er–Rao Bound-Informed Training of Neural Net- works for Quantitative MRI

Zhang Xiaoxia, Duchemin Quentin, Liu* Kangning, et al. Cram´ er–Rao Bound-Informed Training of Neural Net- works for Quantitative MRI. Magnetic Resonance in Medicine. 2022;88(1):436–448

work page 2022
[8]

Bias-Reduced Neural Networks for Parameter Estimation in Quantitative MRI

Mao Andrew, Flassbeck Sebastian, Assl¨ ander Jakob. Bias-Reduced Neural Networks for Parameter Estimation in Quantitative MRI. Magnetic Resonance in Medicine. 2024;92(4)

work page 2024

Show all 34 references

[9]

Breathing Freely: Self-supervised Liver T1rho Mapping from A Sin- gle T1rho-weighted Image

Huang Chaoxing, Qian Yurui, Hou Jian, et al. Breathing Freely: Self-supervised Liver T1rho Mapping from A Sin- gle T1rho-weighted Image. In: International Conference on Medical Imaging with Deep Learning; 2022

2022
[10]

Uncertainty-Aware Self-Supervised Neural Network for Liver T1ρ Mapping with Relaxation Constraint

Huang Chaoxing, Qian Yurui, Yu Simon Chun-Ho, et al. Uncertainty-Aware Self-Supervised Neural Network for Liver T1ρ Mapping with Relaxation Constraint. Physics in Medicine & Biology. 2022;67(22):225019

2022
[11]

Mul- tilayer Feedforward Networks Are Universal Approxima- tors

Hornik Kurt, Stinchcombe Maxwell, White Halbert. Mul- tilayer Feedforward Networks Are Universal Approxima- tors. Neural Networks. 1989;2(5):359–366

1989
[12]

Variability of CubeQuant T1rho, Quan- titative DESS T2, and Cones Sodium MRI in Knee Car- tilage

Jordan Caroline D., Monu Uchechukwuka D., McWalter Emily J., et al. Variability of CubeQuant T1rho, Quan- titative DESS T2, and Cones Sodium MRI in Knee Car- tilage. Osteoarthritis and Cartilage. 2014;22(10):1559– 1567

2014
[13]

3D Quantita- tive Imaging of T1rho & T2

Chen Weitian, Takahashi Austin, Han Eric. 3D Quantita- tive Imaging of T1rho & T2. In: Proceeding of the Inter- national Society for Magnetic Resonance in Medicine; 2011; Montr´ eal, Qu´ ebec, Canada

2011
[14]

Errors in Quantitative T1rho Imaging and the Correction Methods

Chen Weitian. Errors in Quantitative T1rho Imaging and the Correction Methods. Quantitative Imaging in Medicine and Surgery. 2015;5(4):58391–58591

2015
[15]

Breath- Hold Black Blood Quantitative T1rho Imaging of Liver Using Single Shot Fast Spin Echo Acquisition

Chen Weitian, Chan Queenie, W´ ang Y ` ı-Xi´ ang J.. Breath- Hold Black Blood Quantitative T1rho Imaging of Liver Using Single Shot Fast Spin Echo Acquisition. Quantita- tive Imaging in Medicine and Surgery. 2016;6(2):16877– 16177

2016
[16]

T., Borthakur Arijitt, Elliott Mark A., et al

Witschey Walter R. T., Borthakur Arijitt, Elliott Mark A., et al. Artifacts in T1 ρ-weighted Imaging: Com- pensation for B1 and B0 Field Imperfections. Journal of Magnetic Resonance. 2007;186(1):75–85

2007
[17]

A Sys- tematic Post-Processing Approach for Quantitative T1 ρ Imaging of Knee Articular Cartilage

Zhong Junru, Yao Yongcheng, Xiao Fan, et al. A Sys- tematic Post-Processing Approach for Quantitative T1 ρ Imaging of Knee Articular Cartilage. arXiv:2409.12600. 2024

2024 arXiv
[18]

B., Epstein C

Avants B. B., Epstein C. L., Grossman M., Gee J. C.. Symmetric Diﬀeomorphic Image Registration with Cross- Correlation: Evaluating Automated Labeling of Elderly and Neurodegenerative Brain. Medical Image Analysis. 2008;12(1):26–41

2008
[19]

U-Net: Convolutional Networks for Biomedical Image Segmentation

Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer-Assisted Intervention, Pt Iii. 2015;9351:234–241

2015
[20]

Deep Learning Using Rectiﬁed Lin- ear Units (ReLU)

Agarap Abien Fred. Deep Learning Using Rectiﬁed Lin- ear Units (ReLU). arXiv:1803.08375. 2019

2019 arXiv
[21]

Adam: A Method for Stochastic Optimization

Kingma Diederik P., Ba Jimmy. Adam: A Method for Stochastic Optimization. arXiv:1412.6980. 2017

2017 arXiv
[22]

Generating Sequences With Recurrent Neu- ral Networks

Graves Alex. Generating Sequences With Recurrent Neu- ral Networks. arXiv:1308.0850. 2014

2014 arXiv
[23]

Python 3 Reference Manual

Van Rossum Guido, Drake Fred L.. Python 3 Reference Manual. Scotts Valley, CA: CreateSpace; 2009

2009
[24]

Scikit-Image: Image Processing in Python

Walt St´ efan, Sch¨ onberger Johannes L., Nunez-Iglesia s Juan, et al. Scikit-Image: Image Processing in Python. PeerJ. 2014;2:e453

2014
[25]

PyTorch: An Imperative Style, High-Performance Deep Learning Library

Paszke Adam, Gross Sam, Massa Francisco, et al. PyTorch: An Imperative Style, High-Performance Deep Learning Library. In: Advances in Neural Information Processing Systems; 2019; Vancouver, BC

2019
[26]

PyTorch Lightning

Falcon William, The PyTorch Lightning team . PyTorch Lightning. 2019

2019
[27]

MONAI: Medical Open Network for AI Zenodo2024

Consortium MONAI. MONAI: Medical Open Network for AI Zenodo2024
[28]

User-Guided 3D Active Contour Segmen- tation of Anatomical Structures: Signiﬁcantly Improved Eﬃciency and Reliability

Yushkevich Paul A., Piven Joseph, Cody Hazlett Heather, et al. User-Guided 3D Active Contour Segmen- tation of Anatomical Structures: Signiﬁcantly Improved Eﬃciency and Reliability. Neuroimage. 2006;31(3):1116– 1128

2006
[29]

MATLAB Version: 23.2.0 (R2023b) The MathWorks Inc.2023

The MathWorks Inc. . MATLAB Version: 23.2.0 (R2023b) The MathWorks Inc.2023

2023
[30]

Towards Quantiﬁcation of Pulsed Spinlock and CEST at Clinical MR Scanners: An Analytical Inter- leaved Saturation–Relaxation (ISAR) Approach

Roeloﬀs Volkert, Meyer Christian, Bachert Peter, Zaiss Moritz. Towards Quantiﬁcation of Pulsed Spinlock and CEST at Clinical MR Scanners: An Analytical Inter- leaved Saturation–Relaxation (ISAR) Approach. NMR in Biomedicine. 2015;28(1):40–53

2015
[31]

ImageNet Classiﬁcation with Deep Convolutional Neural Networks

Krizhevsky Alex, Sutskever Ilya, Hinton Geoﬀrey E.. ImageNet Classiﬁcation with Deep Convolutional Neural Networks. Commun. ACM. 2017;60(6):84–90

2017
[32]

Dive into Deep Learning

Zhang Aston, Lipton Zachary C., Li Mu, Smola Alexan- der J.. Dive into Deep Learning . Cambridge University Press; 2023

2023
[33]

Mark, Bronskill Michael J

Santyr Giles E, Henkelman R. Mark, Bronskill Michael J. Variation in Measured Transverse Relaxation in Tissue Resulting from Spin Locking with the CPMG Sequence. Journal of Magnetic Resonance (1969). 1988;79(1):28– 44

1969
[34]

T1ρ Imaging of Articular Cartilage Using a CPMG Pulse Train for Spin Locking

Gold G, Han E, Hargreaves B, Wright G, Brittain J. T1ρ Imaging of Articular Cartilage Using a CPMG Pulse Train for Spin Locking. In: Proceeding of the Interna- tional Society for Magnetic Resonance in Medicine; 2006; Seattle, Washington, USA. ZHONG et al 13 How to cite this ar...

2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

W., Guermazi A., Demehri S., Wirth W., Kijowski R

Roemer F. W., Guermazi A., Demehri S., Wirth W., Kijowski R.. Imaging in Osteoarthritis. Osteoarthritis and Cartilage. 2022;30(7):913–934

work page 2022

[2] [2]

The QIBA Proﬁle for MRI-based Compositional Imaging of Knee Cartilage

Chalian Majid, Li Xiaojuan, Guermazi Ali, et al. The QIBA Proﬁle for MRI-based Compositional Imaging of Knee Cartilage. Radiology. 2021;301(2):423–432. 12 ZHONG et al

work page 2021

[3] [3]

The Role of Relax- ation Times in Monitoring Proteoglycan Depletion in Articular Cartilage

Mlyn´ arik Vladim ´ ır, Trattnig Siegfried, Huber Monika, Zembsch Alexander, Imhof Herwig. The Role of Relax- ation Times in Monitoring Proteoglycan Depletion in Articular Cartilage. Journal of Magnetic Resonance Imaging. 1999;10(4):497–502

work page 1999

[4] [4]

Proteoglycan-Induced Changes in T1ρ-relaxation of Articular Cartilage at 4T

Akella Sarma V.S., Reddy Regatte Ravinder, Gougoutas Alexander J., et al. Proteoglycan-Induced Changes in T1ρ-relaxation of Articular Cartilage at 4T. Magnetic Resonance in Medicine. 2001;46(3):419–423

work page 2001

[5] [5]

W., Sharaﬁ Azadeh, Otazo Ricardo, Regatte Ravinder R

Zibetti Marcelo V. W., Sharaﬁ Azadeh, Otazo Ricardo, Regatte Ravinder R.. Accelerated Mono- and Biexponen- tial 3D-T1ρ Relaxation Mapping of Knee Cartilage Using Golden Angle Radial Acquisitions and Compressed Sens- ing. Magnetic Resonance in Medicine. 2020;83(4):1291– 1309

work page 2020

[6] [6]

SuperMAP: Deep Ultrafast MR Relaxometry with Joint Spatiotemporal Undersampling

Li Hongyu, Yang Mingrui, Kim Jee Hun, et al. SuperMAP: Deep Ultrafast MR Relaxometry with Joint Spatiotemporal Undersampling. Magnetic Resonance in Medicine. 2023;89(1):64–76

work page 2023

[7] [7]

Cram´ er–Rao Bound-Informed Training of Neural Net- works for Quantitative MRI

Zhang Xiaoxia, Duchemin Quentin, Liu* Kangning, et al. Cram´ er–Rao Bound-Informed Training of Neural Net- works for Quantitative MRI. Magnetic Resonance in Medicine. 2022;88(1):436–448

work page 2022

[8] [8]

Bias-Reduced Neural Networks for Parameter Estimation in Quantitative MRI

Mao Andrew, Flassbeck Sebastian, Assl¨ ander Jakob. Bias-Reduced Neural Networks for Parameter Estimation in Quantitative MRI. Magnetic Resonance in Medicine. 2024;92(4)

work page 2024