pith. sign in

arxiv: 2606.07195 · v1 · pith:UXQRDMBZnew · submitted 2026-06-05 · 🧮 math.NA · cs.NA

Adjoint-based Perfusion Estimation from Dynamic Contrast-Enhanced Ultrasound: Advection-Diffusion and Two-Compartment Models

Sophie Externbrink , Ahmed El Kaffas , Dimitre Hristov , Sebastian G\"otschel This is my paper

Pith reviewed 2026-06-27 21:10 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords perfusion estimationdynamic contrast-enhanced ultrasoundadjoint methodsTikhonov regularizationadvection-diffusiontwo-compartment modelparameter identificationinverse problems
0
0 comments X

The pith

Continuous adjoint equations enable efficient recovery of spatially varying blood flow velocities and perfusion parameters from dynamic contrast-enhanced ultrasound data.

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

The paper sets out a method for estimating spatially varying perfusion parameters and blood flow velocities from time series of contrast agent concentrations measured by ultrasound. It formulates the task as an inverse problem for both a standard advection-diffusion equation and a coupled two-compartment system of hyperbolic advection-reaction equations, the latter intended to be physiologically more faithful. Tikhonov regularization is introduced to stabilize the ill-posed recovery, and continuous adjoint equations are derived so that gradients for the minimization can be obtained at essentially the cost of one forward solve. The approach is tested on synthetic data and on actual in-vivo measurements. A reader would care because reliable perfusion maps could help assess how tumors respond to therapy.

Core claim

The central claim is that the parameter identification problem for perfusion velocities and exchange rates can be solved by Tikhonov-regularized least-squares minimization whose gradient is supplied by continuous adjoint equations for either the parabolic advection-diffusion model or the hyperbolic two-compartment model; the resulting reconstruction algorithms recover plausible parameter fields from both synthetic and in-vivo dynamic contrast-enhanced ultrasound data.

What carries the argument

Continuous adjoint equations that furnish the gradient of the Tikhonov-regularized misfit functional for the inverse perfusion problem.

If this is right

  • Gradient computation becomes feasible at the cost of one additional PDE solve rather than many finite-difference perturbations.
  • The two-compartment formulation can be treated with the same adjoint machinery as the simpler advection-diffusion model.
  • Regularized reconstructions remain stable enough to produce usable maps from noisy in-vivo ultrasound time series.
  • The same numerical discretization and optimization framework applies to both models, allowing direct comparison of their outputs.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method could be adapted to other time-resolved imaging modalities that supply concentration time courses.
  • If the models prove adequate, the recovered maps might be used to predict drug delivery or oxygen transport inside tumors.
  • Extending the framework to time-varying parameters or to three-dimensional domains would be a direct next numerical step.

Load-bearing premise

The chosen advection-diffusion and two-compartment models correctly describe how the contrast agent moves through the tissue.

What would settle it

A controlled experiment in which recovered velocity and perfusion maps are compared against independent ground-truth measurements obtained by another modality on the same tissue; systematic mismatch would falsify the claim that the recovered fields are physiologically meaningful.

Figures

Figures reproduced from arXiv: 2606.07195 by Ahmed El Kaffas, Dimitre Hristov, Sebastian G\"otschel, Sophie Externbrink.

Figure 1
Figure 1. Figure 1: Left: Schematic representation of a vascular system, vessel color indi￾cating presence of oxygenated (red) and de-oxygenated (blue) blood. Right: The finite resolution of the imaging system (ultrasound) eliminates small vessel details. 2.1. Advection-diffusion model. When modeling tracer in blood flow of an organ we need to be able to describe an in- and outflow via the main artery, respectively vein. Furt… view at source ↗
Figure 2
Figure 2. Figure 2: Sketch of transportation of tracer in blood flow through an organ using advection-diffusion (left) and the two-compartment model (right). Vessel Type Mean Flow Velocity Aorta 30 cm/s Arteries 30 cm/s - 20 mm/s Arterioles 20 mm/s Capillaries 0.3 mm/s Venules 3 mm/s Veins 3 mm/s - 10 mm/s Vena Cava 10 mm/s [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Cost functional and its contributions from the data mismatch and reg￾ularization term (left) as well as gradient norm of velocities (right; only showing velocity iterations) for the reconstruction of problem WTD with 5% noise and 4 optimization rounds, showing the progress of the optimization. The final cost functional values for all problems can be found in [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Reconstructed concentrations (top), the mismatch of reconstructed con￾centrations and artificial data (middle) and the relative error in percent (bottom) of problem WTD with 5% noise after 4 optimization rounds plotted at three time steps. velocities look similar to the previous ones. Despite this, the transfer coefficient (on a fixed conversion domain), is quite well identified, and is only 0.49 off the e… view at source ↗
Figure 5
Figure 5. Figure 5: Reconstructed concentrations (top), mismatch between reconstructed concentrations and artificial data (middle) and the relative error in percent (bottom) of problem NTD after 24 optimization rounds. squares. As a result, the data is smoothed and its precision enhanced, while preserving the underlying signal trends without distortion [27]. The difference before and after smoothing the concentration of trace… view at source ↗
Figure 6
Figure 6. Figure 6: Reconstructed velocities of problem WTD with 5% noise and 4 opti￾mization rounds. V 1 x V 2 x V 1 y V 2 y [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Reconstructed velocities of problem NTD after 24 optimization rounds [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Reconstructed concentrations (top), the mismatch of reconstructed con￾centrations and artificial data (middle) and the relative error in percent (bottom) of problem WTD with space-dependent κ after 3 optimization rounds plotted at three time steps. V 1 x V 2 x V 1 y V 2 y κ [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Reconstructed velocities and transfer coefficient function of problem WTD-s after 3 optimization rounds. though the final result is still far from a stationary point. To evaluate the reconstruction, simulated concentrations using the identified parameter fields are shown in [PITH_FULL_IMAGE:figures/full_fig_p012_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Visualisation of 3D ultrasound data at time t = 12s before (left) and after (right) smoothing [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Percentage of concentration over time for the whole ultrasound mea￾surement. Between measurement frame 13 and 46, indicated in the plot, 80% of the total concentration are inside the computational domain. model V 1 = (V 1 x V 1 y ) V 2 = (V 2 x V 2 y ) κ ∨ D {λi , i = 1, 2, 3} tol1 tol2 tolα Two-Compartment  0.1 −0.1  0.1 0.1  12 {10−4 , 10−4 , 10−5} 10−5 10−5 5 · 10−7 Advection-Diffusion  0.1 0.1  … view at source ↗
Figure 12
Figure 12. Figure 12: Two-compartment model: reconstructed concentrations (top row) with space-dependent κ and relative error (in percent) between the final simulated and measured concentrations (bottom row) [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Two-compartment model: reconstructed arterial (left) and venous (right) velocities. Arrows (one per cell) indicate the direction, coloring the mag￾nitude. Nonzero velocities are reconstructed only in areas with contrast agent con￾centrations [PITH_FULL_IMAGE:figures/full_fig_p014_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Two-compartment model: reconstructed space-dependent transfer co￾efficient κ. t1 = 1.47 t2 = 7.06 t3 = 14.11 t1 = 1.47 t2 = 7.06 t3 = 14.11 [PITH_FULL_IMAGE:figures/full_fig_p015_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Advection-diffusion model: reconstructed concentrations (top row) with space-dependent diffusivity and relative error (in percent) between the final simulated and measured concentrations (bottom row) [PITH_FULL_IMAGE:figures/full_fig_p015_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Advection-diffusion model: reconstructed velocity (left) and diffusivity (right) [PITH_FULL_IMAGE:figures/full_fig_p015_16.png] view at source ↗
read the original abstract

Tumor perfusion and vascular properties are important determinants of a cancer's response to therapy. In this paper, we discuss the estimation of spatially varying blood flow velocities and perfusion parameters from time-resolved contrast agent concentration data. We compare a standard parabolic advection-diffusion model against a two-compartment model governed by a coupled system of hyperbolic advection-reaction equations, which is physiologically more sound. To address the inherent ill-posedness of this parameter identification problem, we employ Tikhonov regularization and derive continuous adjoint equations necessary for efficient, gradient-based minimization. We discuss the numerical discretization of the state and adjoint systems using state-of-the-art schemes, and demonstrate the efficacy of the proposed reconstruction algorithms through numerical experiments on synthetic data and in vivo dynamic contrast-enhanced ultrasound measurements.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper develops adjoint-based reconstruction algorithms to estimate spatially varying blood flow velocities and perfusion parameters from time-resolved DCE-US concentration data. It compares a standard parabolic advection-diffusion model to a two-compartment model governed by coupled hyperbolic advection-reaction equations, employs Tikhonov regularization to handle ill-posedness, derives the corresponding continuous adjoint equations for gradient-based minimization, discusses state-of-the-art numerical discretizations, and demonstrates the methods on synthetic data generated from the forward models plus in vivo DCE-US measurements.

Significance. If the modeling assumptions are valid, the work provides an efficient computational framework for recovering physiologically relevant perfusion maps from DCE-US, which could support non-invasive evaluation of tumor response to therapy. The derivation of continuous adjoints and the direct comparison of the two transport models are technical strengths that enable scalable optimization and model selection. The inclusion of in vivo data adds practical relevance, though overall significance remains conditional on quantitative validation of reconstruction accuracy.

major comments (2)
  1. [Numerical Experiments] Numerical Experiments section: synthetic tests are generated from the identical forward models used in the inversion, confirming only that the optimizer recovers parameters under exact model match; this provides no evidence on robustness to model mismatch and therefore does not support the efficacy claim for the in vivo reconstructions.
  2. [Abstract and Results] Abstract and Results: no quantitative error metrics (e.g., relative L2 errors on recovered velocity or perfusion fields, convergence rates under mesh refinement, or regularization-parameter sensitivity) are supplied for either the synthetic or in vivo cases, leaving the central claim of algorithmic efficacy without measurable support.
minor comments (1)
  1. [Numerical Discretization] The discretization subsection would benefit from explicit statements of the CFL or stability conditions used for the hyperbolic two-compartment scheme.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below.

read point-by-point responses

  1. Referee: [Numerical Experiments] Numerical Experiments section: synthetic tests are generated from the identical forward models used in the inversion, confirming only that the optimizer recovers parameters under exact model match; this provides no evidence on robustness to model mismatch and therefore does not support the efficacy claim for the in vivo reconstructions.

    Authors: We agree that the synthetic experiments only verify the method under exact model match and therefore do not test robustness to mismatch. The in vivo experiments apply the method to real DCE-US data, where model mismatch is necessarily present, and produce physiologically plausible maps. To directly address the concern we will add a new subsection with synthetic mismatch experiments (data generated from one model, inverted with the other) in the revised manuscript. revision: yes

  2. Referee: [Abstract and Results] Abstract and Results: no quantitative error metrics (e.g., relative L2 errors on recovered velocity or perfusion fields, convergence rates under mesh refinement, or regularization-parameter sensitivity) are supplied for either the synthetic or in vivo cases, leaving the central claim of algorithmic efficacy without measurable support.

    Authors: We agree that quantitative metrics are needed to support the efficacy claims. In the revision we will add relative L2 errors on recovered fields for all synthetic cases, mesh-refinement convergence rates, and regularization-parameter sensitivity studies. For the in vivo data we will report data-misfit residuals and quantitative comparisons between the two models. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper derives continuous adjoint equations for Tikhonov-regularized gradient-based minimization of perfusion parameters in advection-diffusion and two-compartment models. Synthetic experiments test recovery when data is generated from the identical forward models, which is standard validation and does not reduce any output to an input by construction. In vivo results apply the same numerical scheme under the stated modeling assumptions without any self-definitional loops, fitted inputs renamed as predictions, or load-bearing self-citations. The derivation chain consists of independent PDE analysis and discretization steps that remain self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the physiological fidelity of the two transport models and on the effectiveness of Tikhonov regularization plus adjoint gradients for an ill-posed inverse problem; no free parameters, invented entities, or non-standard axioms are explicitly introduced in the abstract.

axioms (1)
  • domain assumption The advection-diffusion and two-compartment equations adequately describe contrast-agent transport in tissue
    Invoked when the models are presented as the basis for parameter estimation and when their relative physiological soundness is asserted.

pith-pipeline@v0.9.1-grok · 5674 in / 1288 out tokens · 17772 ms · 2026-06-27T21:10:29.409448+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

36 extracted references · 9 canonical work pages

  1. [1]

    Epidemiology of primary and secondary liver cancers

    A. Ananthakrishnan, V. Gogineni, and K. Saeian. “Epidemiology of primary and secondary liver cancers”. In:Seminars in Interventional Radiology23 (2006), pp. 47–63

    2006

  2. [2]

    Anderson.Compartmental Modeling and Tracer Kinetics

    David H. Anderson.Compartmental Modeling and Tracer Kinetics. Springer, 1983

    1983

  3. [3]

    Minimization of Functions Having Lipschitz Continuous First Partial Deriva- tives

    Larry Armijo. “Minimization of Functions Having Lipschitz Continuous First Partial Deriva- tives”. In:Pacific Journal of Mathematics16.1 (1966), pp. 1–3

    1966

  4. [4]

    An optimal control approach to optical flow computation

    A. Borz` ı, K. Ito, and K. Kunisch. “An optimal control approach to optical flow computation”. In:International Journal for Numerical Methods in Fluids40.1-2 (2002), pp. 231–240.doi: https://doi.org/10.1002/fld.273

    work page doi:10.1002/fld.273 2002

  5. [5]

    Numerical solution of the Navier-Stokes equations

    Alexandre Joel Chorin. “Numerical solution of the Navier-Stokes equations”. In:Math. Comp. 22 (1968), pp. 745–762.doi:10.1090/S0025-5718-1968-0242392-2

    work page doi:10.1090/s0025-5718-1968-0242392-2 1968

  6. [6]

    Parameter Estimation in the 1-D Transport Equation with Advection

    C. Coles. “Parameter Estimation in the 1-D Transport Equation with Advection”. In:PERG- AMON Computers and Mathematics with Applications44 (2002), pp. 1493–1502

    2002

  7. [7]

    Reconstruction of a Velocity Field for a 3D Advection Diffusion Equation

    Yi-Xin Dou and Bo Han. “Reconstruction of a Velocity Field for a 3D Advection Diffusion Equation”. In:International Journal of Thermal Sciences50 (2011), pp. 2340–2354

    2011

  8. [8]

    Spatial Characterization of Tumor Perfusion Properties from 3D DCE- US Perfusion Maps Are Early Predictors of Cancer Treatment Response

    Ahmed El Kaffas et al. “Spatial Characterization of Tumor Perfusion Properties from 3D DCE- US Perfusion Maps Are Early Predictors of Cancer Treatment Response”. In:Scientific Reports 10.1 (Apr. 24, 2020), p. 6996.issn: 2045-2322.doi:10.1038/s41598-020-63810-1

    work page doi:10.1038/s41598-020-63810-1 2020

  9. [9]

    European Cancer Information System (ECIS).Fact Sheet Liver Cancer in the EU.https:// ecis.jrc.ec.europa.eu/sites/default/files/2024-10/2024-10-03_Factsheet_Liver- Cancer.pdf. 2024

    2024

  10. [10]

    2026.url: https://doi.org/10.5281/zenodo.19063183(visited on 03/17/2026)

    Sophie Externbrink.Parameter Identification for a Tracer Model (BaseVersion). 2026.url: https://doi.org/10.5281/zenodo.19063183(visited on 03/17/2026)

    work page doi:10.5281/zenodo.19063183(visited 2026

  11. [11]

    Implicit–explicit (IMEX) Runge–Kutta methods for non-hydrostatic atmospheric models

    D. J. Gardner, J. E. Guerra, F. P. Hamon, D. R. Reynolds, P. A. Ullrich, and C. S. Woodward. “Implicit–explicit (IMEX) Runge–Kutta methods for non-hydrostatic atmospheric models”. In:Geoscientific Model Development11.4 (2018), pp. 1497–1515

    2018

  12. [12]

    Goodman.Speckle phenomena in optics: theory and applications

    Joseph W. Goodman.Speckle phenomena in optics: theory and applications. SPIE Digital Library, 2020.doi:10.1117/3.2548484

    work page doi:10.1117/3.2548484 2020

  13. [13]

    Guyton and J

    A. Guyton and J. Hall.Textbook of Medical Physiology. W.B. Saunders, 1996

    1996

  14. [14]

    A survey of nonlinear conjugate gradient methods

    William W. Hager and Hongchao Zhang. “A survey of nonlinear conjugate gradient methods”. In:Pacific Journal of Optimization2 (2006), pp. 35–58. REFERENCES 17

    2006

  15. [15]

    Identification of the advection- diffusion equation for predicting green water propagation

    J.V. Hern´ andez-Fontes, L. Torres, E. Mendoza, and M. Escudero. “Identification of the advection- diffusion equation for predicting green water propagation”. In:Elsevier Ocean Engineering214 (2020)

    2020

  16. [16]

    Dynamic Contrast-Enhanced Ultrasound Modeling of an Analog to Pseudo- Diffusivity in Intravoxel Incoherent Motion Magnetic Resonance Imaging

    Dimitre Hristov, Lauri Mustonen, Rie von Eyben, Sebastian G¨ otschel, Michael Minion, and Ahmed El Kaffas. “Dynamic Contrast-Enhanced Ultrasound Modeling of an Analog to Pseudo- Diffusivity in Intravoxel Incoherent Motion Magnetic Resonance Imaging”. In:IEEE Transac- tions On Medical Imaging10 (2022)

    2022

  17. [17]

    Global epidemiology of NAFLD- related HCC: trends, predictions, risk factors and prevention

    Daniel Q. Huang, Hashem B. El-Serag, and Rohit Loomba. “Global epidemiology of NAFLD- related HCC: trends, predictions, risk factors and prevention”. eng. In:Nature Reviews. Gas- troenterology & Hepatology18.4 (2021), pp. 223–238.issn: 1759-5045.doi:10.1038/s41575- 020-00381-6

    work page doi:10.1038/s41575- 2021

  18. [18]

    Polarization of Tumor-Associated Macrophages: A Novel Strategy for Vascular Normalization and Antitumor Immunity

    Yuhui Huang, Matija Snuderl, and Rakesh K. Jain. “Polarization of Tumor-Associated Macrophages: A Novel Strategy for Vascular Normalization and Antitumor Immunity”. In:Cancer Cell19.1 (2011), pp. 1–2

    2011

  19. [19]

    Clinical Intravoxel Incoherent Motion and Diffusion MR Imaging: Past, Present, and Future

    M. Iima and D. Le Bihan. “Clinical Intravoxel Incoherent Motion and Diffusion MR Imaging: Past, Present, and Future”. In:Radiology278.1 (2016), pp. 13–32

    2016

  20. [20]

    Johnson.Lecture notes in Introduction To Numerical Methods

    Steven G. Johnson.Lecture notes in Introduction To Numerical Methods. Apr. 2019

    2019

  21. [21]

    P. R. Peebles Jr.Probability, Random Variables and Random Signal Principles. McGraw-Hill, Inc., 2001

    2001

  22. [22]

    Accuracy and efficiency of adjoint state based parameter identification for fractional advection diffusion equation with space-dependent coefficients

    Boris Maryshev, Alain Cartalade, Christelle Latrille, and Marie-Christine N´ eel. “Accuracy and efficiency of adjoint state based parameter identification for fractional advection diffusion equation with space-dependent coefficients”. In:IEEE Xplore(2016)

    2016

  23. [23]

    Robust estimation of ultrasound pulses using outlier-resistant de-noising

    O. Michailovich and D. Adam. “Robust estimation of ultrasound pulses using outlier-resistant de-noising”. In:IEEE Transactions on Medical Imaging22.3 (2003), pp. 368–381.doi:10. 1109/TMI.2003.809603

    arXiv 2003

  24. [24]

    The ”Fingerprint

    Sneha R. Raoa, Sarah E. Shelton, and Paul A. Dayton. “The ”Fingerprint” of Cancer Extends Beyond Solid Tumor Boundaries: Assessment With a Novel Ultrasound Imaging Approach”. In:IEEE Trans Biomedical Engineering63.5 (2015), pp. 1082–1086

    2015

  25. [25]

    Global burden of primary liver cancer in 2020 and predictions to 2040

    Harriet Rumgay, Melina Arnold, Jacques Ferlay, Olufunmilayo Lesi, Citadel J. Cabasag, J´ erˆ ome Vignat, Mathieu Laversanne, Katherine A. McGlynn, and Isabelle Soerjomataram. “Global burden of primary liver cancer in 2020 and predictions to 2040”. In:Journal of Hepatology 77.6 (2022), pp. 1598–1606.issn: 0168-8278.doi:10.1016/j.jhep.2022.08.021

    work page doi:10.1016/j.jhep.2022.08.021 2020

  26. [26]

    How to Identify the Right Patients for the Right Treatment in Metastatic Colorectal Cancer (mCRC)

    Z. Saridaki, N. Asimakopoulou, I. Boukovinas, and J. Souglakos. “How to Identify the Right Patients for the Right Treatment in Metastatic Colorectal Cancer (mCRC)”. In:Curr Colorec- tal Cancer Rep.11.4 (2015), pp. 151–159

    2015

  27. [27]

    Smoothing and Differentiation of Data by Simplified Least Squares Procedures

    Abraham Savitzky and M. J. E. Golay. “Smoothing and Differentiation of Data by Simplified Least Squares Procedures”. In:Analytical Chemistry36.8 (1964), pp. 1627–1639

    1964

  28. [28]

    Essentially Non-oscillatory and Weighted Essentially Non-oscillatory Schemes for Hyperbolic Conservation Laws

    Chi-Wang Shu. “Essentially Non-oscillatory and Weighted Essentially Non-oscillatory Schemes for Hyperbolic Conservation Laws”. In:ICASE1.97–65 (1997)

    1997

  29. [29]

    Tracer kinetic modelling in MRI: estimating perfusion and capillary permeability

    S.P. Sourbron and D.L. Buckley. “Tracer kinetic modelling in MRI: estimating perfusion and capillary permeability”. In:Physics in Medicine and Biology57.2 (2011), pp. 1–33

    2011

  30. [30]

    A Tracer-Kinetic Field Theory for Medical Imaging

    Steven Sourbron. “A Tracer-Kinetic Field Theory for Medical Imaging”. In:IEEE Transactions On Medical Imaging33.4 (2014)

    2014

  31. [31]

    Update on the Role of Imaging in Management of Metastatic Colorectal Cancer 1

    S. Harsha Tirumani, K. Won Kim, M. Nishino, S.A. Howard, K.M. Krajewski, J.P. Jagan- nathan, J.M. Cleary, N.H. Ramaiya, and A.B. Shinagare. “Update on the Role of Imaging in Management of Metastatic Colorectal Cancer 1”. In:Radio Graphics34.1 (2014), pp. 1908– 1928

    2014

  32. [32]

    Cassola, M

    Thodsawit Tiyarattanachai, Simona Turco, John R. Eisenbrey, Corinne E. Wessner, Alexan- dra Medellin-Kowalewski, Stephanie Wilson, Andrej Lyshchik, Aya Kamaya, and Ahmed El Kaffas. “A Comprehensive Motion Compensation Method for In-Plane and Out-of-Plane Mo- tion in Dynamic Contrast-Enhanced Ultrasound of Focal Liver Lesions”. In:Ultrasound in Medicine an...

    work page doi:10.1016/j 2022

  33. [33]

    Fredi Tr¨ oltzsch.Optimal Control of Partial Differential Equations: Theory, Methods and Ap- plications. Vol. 112. Graduate Studies in Mathematics. AMS, 2010.doi:10.1090/gsm/112

    work page doi:10.1090/gsm/112 2010

  34. [34]

    Convective-Dispersion Modeling in 3D Contrast- Ultrasound Imaging for the Localization of Prostate Cancer

    R. R. Wildeboer, R. J. G. van Sloun, S. G. Schalk, C. K. Mannaerts, J. C. van der Linden, P. Huang, H. Wijkstra, and M. Mischi. “Convective-Dispersion Modeling in 3D Contrast- Ultrasound Imaging for the Localization of Prostate Cancer”. In:IEEE Transactions on Medical Imaging37.12 (2018), pp. 2593–2602. 18 REFERENCES

    2018

  35. [35]

    Convergence Conditions for Ascent Methods

    Philip Wolfe. “Convergence Conditions for Ascent Methods”. In:SIAM Review11.2 (1969), pp. 226–235

    1969

  36. [36]

    Cooperative Filtering and Parameter Identification for Advection–Diffusion Processes Using a Mobile Sensor Network

    Jie You, Ziqiao Zhang, Fumin Zhang, and Wencen Wu. “Cooperative Filtering and Parameter Identification for Advection–Diffusion Processes Using a Mobile Sensor Network”. In:IEEE Transactions on Control Systems Technology31.2 (2023), pp. 527–542. AppendixA.Derivation of Adjoint System and Gradient for the Advection-Diffusion Model We follow the formal Lagra...

    2023