Fine-grained Delaunay triangulation in a simulation of tumor spheroid growth

Alessio Del Fabbro; Edoardo Milotti; Roberto Chignola

REVIEW

Fine-grained Delaunay triangulation in a simulation of tumor spheroid growth

Alessio Del Fabbro, Roberto Chignola, Edoardo Milotti

Biological Physics · Submitted 28 Feb 2006 · 11 references

Pith's verdict: A plausible but unvalidated methods note: the CGAL integration is straightforward, and the Voronoi-face contact area is a real modeling inconsistency that the paper never confronts.

T0 · AI review

How this review was made: An AI referee (deepseek-v4-flash) read the paper on August 28, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 4 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv IDphysics/0603001 (version 1)

Pith IDLPUBKNCE, Pith's permanent record of this paper and its review

CitationAlessio Del Fabbro, Edoardo Milotti, Roberto Chignola. "Fine-grained Delaunay triangulation in a simulation of tumor spheroid growth." arXiv:physics/0603001 (2006). https://pith.science/paper/LPUBKNCE

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

This is a preliminary methods note, not a finished result. The useful piece is the claim that a Delaunay triangulation of cell centers gives both the neighbor list for short-range mechanical interactions and the Voronoi face areas needed for a discrete diffusion calculation. The complexity argument is correct: for short-range interactions you can get O(N log N) if you have an efficient way to identify neighbors, and the Delaunay triangulation is a natural candidate. Eq. (3) is a sensible diffusion balance on a graph, and the idea of adding environmental cells to impose boundary conditions is clean. The 2D figures show the approach can produce qualitatively reasonable oxygen gradients.

The soft spot is the one the stress-test note flags, and it is real. In Section 4 the cells are called simple spheres with different radii. In Section 3 the Voronoi face area A_{n,j} is used as the exchange area in Eq. (3), with the claim that the faces approximate the contact surfaces. For spheres, the actual contact area is zero unless the spheres overlap, and it then depends on the two radii and the center separation, not on the positions of surrounding points. Delaunay adjacency does not imply that two spheres touch, so Eq. (3) can allow mass transfer between cells that do not contact each other. Section 6's remark that the Voronoi tessellation gives a better approximation of actual cell shapes makes the problem worse: either the cells are deformable space-filling polyhedra, in which case the sphere-based mechanical model is inconsistent, or they are spheres, in which case the diffusion areas are wrong. This is load-bearing and the paper does not test it or justify it. No code, no data, no validation against experimental spheroid growth, and no performance numbers. The conclusion that CGAL has 'proved invaluable' is a value judgment, not a result.

For a methods description, the writing is clear and the geometric computations are standard. The paper might be useful to someone building a similar cell simulator, though they would have to make their own decision about how to handle cell shapes. It deserves a serious referee, because the modeling tension I described is exactly what a referee should force the authors to resolve: either drop the sphere language and treat cells as Voronoi polyhedra for all purposes, or replace A_{n,j} with something that depends on actual pair geometry. Even after that, the absence of any experimental comparison would keep it from being a strong paper. As it stands, I would not cite it as a validated method, but I would send it to peer review with a clear request for confrontation of the geometric approximation and at least minimal validation.

Referee report

Significance

If the central geometric identification is correct, the approach would be an attractive way to simulate large cell clusters at O(N log N) cost, and the use of Voronoi face areas as exchange surfaces is a natural discretization. The complexity reduction from Delaunay neighbor lists is standard and sound, and the environmental-cell treatment of boundary conditions is a reasonable device. The strengths of the paper are its clear statement of the computational pipeline and the concrete use of CGAL's iterators and circulators to construct the Voronoi dual. However, the load-bearing approximation that Voronoi face areas represent actual cell-cell contact areas is not validated and conflicts with the spherical-cell mechanical model, while the lack of any benchmark or experimental comparison leaves the accuracy claims unsupported. The paper would be a useful technical contribution if that geometric point were reconciled and tested; as it stands, the central claim is not established.

Major objections (3)

[§3, Eq. (3), §4] The diffusion calculation in Eq. (3) uses the Voronoi face area A_n,j as the area through which chemicals are exchanged between neighboring cells n and j, but Section 4 states that cells are approximated by simple spheres with different radii. For non-overlapping spheres the physical contact area is zero, and the Voronoi face is a polygon on the perpendicular bisector plane whose area is determined by the global point configuration rather than by the two cell radii. Delaunay adjacency, defined by the empty circumsphere property, does not imply that two spheres touch. Consequently Eq. (3) permits diffusive exchange between non-contacting cells and cannot represent the membrane transport described by Eq. (2) without additional justification. The authors must either replace the spherical cell model with a space-filling Voronoi cell model (which would then make the geometry consistent but would require a different mechanical interaction model) or provide quantitative evidence that, for the packings produced by their simulator, the Voronoi face area is a good approximation to the actual exchange area.
[§5, §6, Figs. 2–6] The paper contains no quantitative validation of the proposed diffusion and mechanical schemes. Figures 2–6 are illustrative, and the text does not report measured oxygen gradients, spheroid growth curves, comparisons against known analytic solutions of diffusion into a sphere, or convergence studies of the discrete Voronoi-face discretization. The claim that the approach yields physically accurate spheroid behavior is therefore unsupported. I request benchmarks against at least a standard diffusion problem with analytic solution (for example, steady-state diffusion into a sphere with constant boundary concentration) and, if available, comparison with experimental spheroid data, including error estimates.
[§3, §6] Section 6 states that the Voronoi tessellation 'yields a better approximation of the actual cell shapes, as a byproduct,' which exposes a central tension: cells in the simulator are modeled as spheres for mechanical purposes, but their biochemical exchange areas are taken from Voronoi polyhedra. These two geometric descriptions are inconsistent. The paper must address this inconsistency explicitly and demonstrate that the hybrid model does not introduce systematic errors into the diffusion calculation, or modify the cell representation so that the geometry used for mechanics and the geometry used for diffusion are the same.

Minor objections (4)

[Abstract] The abstract contains a typo: 'enviroment' should be 'environment'.
[§2, §4] The complexity claim is stated inconsistently: Section 2's bullet list says the triangulation reduces computational complexity to O(N), while Section 4 correctly states O(N log N) for the triangulation and force calculation. This should be reconciled.
[§6] There are grammatical errors in the CGAL description: 'we can to compute the dual diagram' should read 'we can compute the dual diagram,' and the sentence beginning 'One should have to pay attention to the degenerate cases' is awkward and should be rewritten.
[§6] The discussion of degenerate cases and the infinite vertex is compressed to the point of being hard to follow; a short explanation of how the infinite-vertex facets on the convex hull are handled in the diffusion calculation would improve clarity.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

This paper shows that a Delaunay triangulation of cell centers can serve as the computational backbone of a tumor-spheroid simulator, reducing the cost of mechanical interactions from O(N^2) to O(N log N) while supplying the face areas…

The reading

The paper argues that the Delaunay triangulation of cell centers should be the core data structure in a simulator of tumor spheroid growth. It provides the neighbor list for mechanical forces, cutting the calculation from O($N^{2}$) to O(N log N), and its dual Voronoi tessellation gives the face areas and volumes needed to solve diffusion of oxygen and nutrients at the level of individual cells. The authors implement the scheme in 3D using a computational-geometry library and test a 2D prototype. If the approach works, it becomes feasible to simulate spheroids containing up to about a million cells while preserving the saturating Michaelis-Menten transport kinetics that continuum models miss.

What carries the argument

The central object is the Delaunay triangulation of the cell centers—the triangulation in which the circumsphere of every tetrahedron contains no other center, a property called the empty-sphere property—together with its dual Voronoi tessellation, the partition of space into the regions closer to each center than to any other. The Delaunay triangulation supplies the adjacency list of touching cells, bounding the mechanical neighborhood to about twelve neighbors per cell in 3D and making the force computation linear after the triangulation is constructed. The dual Voronoi tessellation provides the geometric quantities in the discrete diffusion equation: the area of the face shared with each neighbor, the volume of the cell for densities, and the boundary via the convex hull. The machinery does the dual work of complexity reduction and biophysical bookkeeping; a computational-geometry library is used to build and update the triangulation, and vertex labels together with iterators and circulators are used to reconstruct the Voronoi faces.

Extended reading notes

Core claim

The central claim is that the same geometric object—the Delaunay triangulation of the centers of cells and of the surrounding environmental cells—simultaneously solves the two most expensive parts of a spheroid simulation. On the mechanical side, because cell interactions are short-ranged, only cells that are neighbors in the triangulation exert forces on each other; with about twelve neighbors per cell on average, the force calculation becomes O(N) once the O(N log N) triangulation is built. On the biochemical side, the Voronoi cells dual to the triangulation approximate the actual cell shapes, so the area of a shared Voronoi face is used as the contact surface in a discrete diffusion equation (Eq. 3) that balances influx, outflux, and absorption for each cell. The diffusion problem is then solved by a relaxation method on this disordered graph, with known concentrations on the convex hull supplying boundary conditions. The paper establishes that a fine-grained, cell-level simulation can be built on the Delaunay-Voronoi pair without ever explicitly treating cell pairs that do not touch.

Load-bearing premise

The diffusion calculation assumes that the shared face area between two Voronoi cells equals the actual contact surface through which chemicals pass between two real cells; real cells are not space-filling polyhedra, so the accuracy of the simulated oxygen and nutrient profiles depends on this geometric proxy being close enough.

Editorial extensions

If this is right

Simulations of spheroids with around a million cells become computationally feasible, because the total time complexity drops from O(N^2) to O(N log N).
Oxygen and nutrient diffusion is computed cell by cell on an irregular graph, preserving the saturating Michaelis-Menten kinetics that averaged continuum descriptions cannot represent.
Because the Voronoi tessellation approximates cell shapes, the same data structure supplies contact areas and volumes for other purposes, not just neighbor lists.
Dynamic updates of the triangulation can track cell division, death, and slow mechanical relaxation without rebuilding the whole structure from scratch each time step.
The method includes the surrounding aqueous medium as fictitious cells, so container walls and the air-medium interface become natural boundary conditions for the diffusion equation.

Reading between the lines

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

A testable consequence follows from the Voronoi-face approximation: if real cells do not pack as space-filling convex polyhedra, the computed exchange areas will be systematically wrong, and this should show up as a bias in simulated oxygen profiles near the spheroid rim that grows as the cell packing becomes less convex.
The same Delaunay backbone could be reused for other short-range, geometry-dependent processes—such as cell-cell adhesion forces, juxtacrine signaling, or electrical coupling—by substituting the appropriate face-area weights into the adjacency sum, all at the same O(N log N) cost.
Because mitotic events locally invalidate the triangulation, the method couples geometry to biology; one could use the change in the diffusion solution across a single cell division as a convergence check for the spatial discretization.
The relaxation method on a disordered graph should converge to the continuum diffusion solution in the limit of fine cell spacing; comparing the two in a simple spherical geometry would provide a quantitative benchmark for the discretization error.

Original abstract

The simulation of many-particle systems often requires the detailed knowledge of proximity relations to reduce computational complexity and to provide a basis for specific calculations. Here we describe the basic scheme of a simulator of tumor spheroid growth: the calculation of mechanical interactions between cells and of the concentrations of diffusing chemicals requires a backbone provided by the Delaunay triangulation and the volumes of the associated Voronoi regions. Thus the Delaunay triangulation provides both the proximity relations needed to reduce the computational complexity and the basic structures that are needed to carry out the calculation of the biochemical interactions between cells and with the enviroment. A 3D version of the simulator uses the CGAL library as an essential component for the efficient computation of the Delaunay triangulation and of the Voronoi regions.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 3,736 input tokens.

What the abstract itself states, as probabilities: claims first 0.03 · states limitation 0.07 · human participants 0.03 · controlled comparison 0.02 · randomized assignment 0.01 · states quantitative result 0.03 · compares with prior results 0.04 · mentions code or data availability 0.04.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

The simulation of many-particle systems often requires the detailed knowledge of proximity relations to reduce computational complexity and to provide a basis for specific calculations. Here we describe the basic scheme of a simulator of tumor spheroid growth: the calculation of mechanical interactions between cells and of the concentrations of diffusing chemicals requires a backbone provided by the Delaunay triangulation and the volumes of the associated Voronoi regions. Thus the Delaunay triangulation provides both the proximity relations needed to reduce the computational complexity and the basic structures that are needed to carry out the calculation of the biochemical interactions between cells and with the enviroment. A 3D version of the simulator uses the CGAL library as an essential component for the efficient computation of the Delaunay triangulation and of the Voronoi regions.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	The simulation of many-particle systems often requires the detailed knowledge of proximity relations to reduce computational complexity and to provide a basis for specific calculations.	background 1.00	0.05	0.04	asserted 0.89
s1	Here we describe the basic scheme of a simulator of tumor spheroid growth: the calculation of mechanical interactions between cells and of the concentrations of diffusing chemicals requires a backbone provided by the Delaunay triangulation and the volumes of the associated Voronoi regions.	method 0.79	0.79	0.13	asserted 1.00
s2	Thus the Delaunay triangulation provides both the proximity relations needed to reduce the computational complexity and the basic structures that are needed to carry out the calculation of the biochemical interactions between cells and with the enviroment.	interpretation 0.64	0.43	0.35	asserted 1.00
s3	A 3D version of the simulator uses the CGAL library as an essential component for the efficient computation of the Delaunay triangulation and of the Voronoi regions.	method 1.00	0.67	0.07	asserted 1.00

What kind of paper the classifier reports this is: method or tool 0.79 · evidence from simulation or computation 0.96.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** Conceptual scheme of the simulator. An initializati

**Figure 2.** Early phase in the calculation of mechanical relaxat

**Figure 3.** Intermediate step in the calculation of the mechanic

**Figure 4.** This figure shows the relaxed configuration of the syst

**Figure 5.** This figure shows the Delaunay triangulation (backgr

**Figure 6.** This figure shows the result obtained in a test run with

Figures reproduced from arXiv:physics/0603001 by the authors.

References

Works cited (11)

[1]

L. Norton, Kinetic concepts in the systemic drug therapy of breast cancer, Semin. Oncol. 26 (1999) 11

View on Pith 1999
[2]

R. M. Sutherland, Cell and Environment Interactions in T umor Microregions: The Multicell Spheroid Model, Science 240 (1988) 177

View on Pith 1988
[3]

R. Chignola et al., Forecasting the growth of multicell t umour spheroids: implications for the dynamic growth of solid tumours, Cell P rolif. 33 (2000) 219

View on Pith 2000
[4]

D. L. McElwain and G. J. Pettet, Cell migration in multice ll spheroids: swimming against the tide, Bull. Math Biol. 55 (1993) 655

View on Pith 1993
[5]

R. Chignola and E. Milotti, A phenomenological approach to the simulation of metabolism and proliferation of large tumour cell popula tions, Phys. Bio. 2 (2005) 8

View on Pith 2005
[6]

R. Chignola and E. Milotti, Numerical simulation of tumo r spheroid dynamics, Physica A338 (2004) 261

View on Pith 2004
[7]

J. Barnes and P. Hut, A hierarchical O(N log N ) force-calculation algorithm, Nature 324 (1986) 446

View on Pith 1986
[8]

C. B. Barber, D. P. Dobkin, and H. T. Huhdanpaa, The Quickh ull Algorithm for Convex Hulls, ACM Trans. Math. Software, 22 (1996), 469

View on Pith 1996
[9]

Full online documentation is available at http://www.cgal.org

View on Pith
[10]

J.-D. Boissonnat, O. Devillers, S. Pion, M. Teillaud, a nd M. Yvinec, Triangulations in CGAL, Computational Geometry 22 (2002) 5

View on Pith 2002
[11]

R. Chignola, C. Della Pellegrina, A. Del Fabbro, and E. M ilotti, Thresholds, long delays and stability from generalized allosteric eﬀect in protein networks, submitted (preprint: arXiv:q-bio/0601045). 11

View on Pith Pith review arXiv

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.

[1] [1]

L. Norton, Kinetic concepts in the systemic drug therapy of breast cancer, Semin. Oncol. 26 (1999) 11

View on Pith 1999

[2] [2]

R. M. Sutherland, Cell and Environment Interactions in T umor Microregions: The Multicell Spheroid Model, Science 240 (1988) 177

View on Pith 1988

[3] [3]

R. Chignola et al., Forecasting the growth of multicell t umour spheroids: implications for the dynamic growth of solid tumours, Cell P rolif. 33 (2000) 219

View on Pith 2000

[4] [4]

D. L. McElwain and G. J. Pettet, Cell migration in multice ll spheroids: swimming against the tide, Bull. Math Biol. 55 (1993) 655

View on Pith 1993

[5] [5]

R. Chignola and E. Milotti, A phenomenological approach to the simulation of metabolism and proliferation of large tumour cell popula tions, Phys. Bio. 2 (2005) 8

View on Pith 2005

[6] [6]

R. Chignola and E. Milotti, Numerical simulation of tumo r spheroid dynamics, Physica A338 (2004) 261

View on Pith 2004

[7] [7]

J. Barnes and P. Hut, A hierarchical O(N log N ) force-calculation algorithm, Nature 324 (1986) 446

View on Pith 1986

[8] [8]

C. B. Barber, D. P. Dobkin, and H. T. Huhdanpaa, The Quickh ull Algorithm for Convex Hulls, ACM Trans. Math. Software, 22 (1996), 469

View on Pith 1996

[9] [9]

Full online documentation is available at http://www.cgal.org

View on Pith

[10] [10]

J.-D. Boissonnat, O. Devillers, S. Pion, M. Teillaud, a nd M. Yvinec, Triangulations in CGAL, Computational Geometry 22 (2002) 5

View on Pith 2002

[11] [11]

R. Chignola, C. Della Pellegrina, A. Del Fabbro, and E. M ilotti, Thresholds, long delays and stability from generalized allosteric eﬀect in protein networks, submitted (preprint: arXiv:q-bio/0601045). 11

View on Pith Pith review arXiv