A Probabilistic Model of Bilateral Lymphatic Spread in Head and Neck Cancer

Jan Unkelbach; Panagiotis Balermpas; Roman Ludwig; Sergi Benavente; Yoel Perez Haas

REVIEW 3 major objections 4 minor 24 references

A Probabilistic Model of Bilateral Lymphatic Spread in Head and Neck Cancer

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A probabilistic model maps which throat cancer patients can safely skip contralateral neck radiation.

desk verdict A solid, interpretable bilateral HMM with open code and a real internal contradiction between its de-escalation rule and its own data. read the letter →

arxiv 2501.16910 v1 pith:MU3HHBJD submitted 2025-01-28 physics.med-ph

Roman Ludwig , Yoel Perez Haas , Sergi Benavente , Panagiotis Balermpas , Jan Unkelbach This is my paper · ORCID

classification physics.med-ph

keywords oropharyngealsquamouscellcarcinomalymphnodelevelsoccultmetastasiscontralateralneckelectiveclinicaltargetvolumehiddenMarkovmodelmidlineextensionCTV-Nde-escalation

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 the risk of undetected ('occult') cancer spread to the lymph nodes on the side of the neck opposite an oropharyngeal tumor can be predicted per patient from clinical facts known before treatment: tumor position relative to the midline, T-category, and which lymph node levels are already involved on the same side. If this is right, many patients with lateralized tumors and a clinically negative contralateral neck can safely have the contralateral side left out of the elective radiation target volume, and patients with midline-crossing tumors can have that volume reduced to lymph node level II. The argument is carried by a bilateral hidden Markov model of lymphatic progression trained on 833 patients across four institutions, which reproduces observed bilateral involvement patterns with a compact set of interpretable parameters. The model identifies midline extension as the dominant risk factor for contralateral spread, with advanced T-stage and heavier ipsilateral involvement adding further risk.

What carries the argument

The engine is a hidden Markov model over the six lymph node levels I, II, III, IV, V, and VII on each side, where the hidden state is healthy or involved and observed through noisy imaging with fixed sensitivity and specificity. The bilateral extension keeps the ipsilateral transition graph, adds a contralateral graph with shared inter-level spread rates, and couples the sides through a time-prior that differs for early versus advanced T-category; the conditional independence $P(X^i, X^c \mid t) = P(X^i \mid t)\,P(X^c \mid t)$ encodes the assumption that contralateral disease spreads only from the primary tumor. Midline extension is a binary random variable $\epsilon$ that switches the contralateral tumor-spread rates from $b^{c,\epsilon=\mathrm{False}}_v$ to a linear mix $\alpha\,b^i_v + (1-\alpha)\,b^{c,\epsilon=\mathrm{False}}_v$, with $\alpha \approx 0.34$ in the fitted model. Parameters are learned by Markov chain Monte Carlo from 833 patients, and posterior risks for occult disease in any level are computed by marginalizing the joint distribution over hidden states.

What would settle it

Measure occult contralateral disease in a prospective cohort of lateralized T1-T2 oropharyngeal tumors with clinically negative contralateral necks who undergo elective contralateral neck dissection, and check whether contralateral level II involvement significantly exceeds the model's predicted risk; a proportion above roughly 5% would invalidate the de-escalation recommendation.

Watch

Extended reading notes

Core claim

The central claim is that bilateral lymphatic progression in oropharyngeal squamous cell carcinoma is explained by a single shared time axis: the two sides of the neck evolve independently, each governed by the same spread dynamics between lymph node levels, and are coupled only through the diagnosis time and through tumor spread rates that differ ipsilaterally versus contralaterally. Under this construction, contralateral involvement is driven primarily by whether the primary tumor crosses the midline (encoded as a random switch with probability $p_\epsilon$ per time step), with advanced T-stage and ipsilateral burden acting as secondary risk factors through later diagnosis times. For a lateralized tumor with a clinically negative contralateral neck, the model puts the occult contralateral level II risk near or below a 5% threshold across T-categories, and predicts that contralateral level III involvement is unlikely without level II involvement and level IV involvement is rare without level III, leading to the paper's volume de-escalation recommendations.

Load-bearing premise

The model assumes the two sides of the neck never drain into each other and that contralateral spread happens only through the primary tumor; if lymphatic rerouting or cross-midline connections occur, the predicted contralateral risks are too low.

Editorial extensions

If this is right

For lateralized tumors with no clinical contralateral involvement, the contralateral neck can be excluded from the elective CTV-N regardless of T-category or ipsilateral involvement, according to the 5% occult-risk threshold.
For tumors crossing the midline with a clinically negative contralateral neck, elective contralateral irradiation can be limited to lymph node level II.
Contralateral level III should be irradiated only when level II is clinically involved, and contralateral level IV only when level III involvement is confirmed.
Contralateral levels I, V, and VII generally do not require elective irradiation unless they are clinically involved.
The paper notes that the model's predictions are already guiding a clinical trial on volume de-escalation.

Reading between the lines

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

If confirmed prospectively, the same two-sided conditional-independence structure could be transferred to other head and neck subsites such as the oral cavity, hypopharynx, and larynx, where bilateral elective irradiation is standard.
The 5% occult-risk threshold is a clinical policy choice rather than a model output; a reader who prefers a stricter threshold would shrink or expand the recommended volumes accordingly.
A direct test of the no-cross-midline-drainage assumption could come from lymphoscintigraphy, or from comparing model predictions to pathological findings in elective contralateral neck dissections, which would reveal whether rerouted drainage after bulky ipsilateral disease is clinically significant.
Because the model outputs continuous per-level risk estimates, it could be connected to treatment planning to quantify expected reductions in xerostomia and dysphagia from unilateral or level-II-only irradiation.

Share X Bluesky LinkedIn Reddit HN

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,

The core contribution is real: a bilateral extension of the ipsilateral HMM with a midline-extension latent variable, trained on 833 OPSCC patients with MCMC, and fully reproducible code/data. The model is interpretable, the parameter estimates are sensible, and the paper is honest about several limitations (consensus-as-truth approximation, midline-extension prevalence mismatch by T-category). For a modeling paper, this is well above average.

But there's a load-bearing internal inconsistency that the authors don't address. Section 8.2 recommends unilateral RT for lateralized tumors with no contralateral clinical involvement, 'regardless of T-category or ipsilateral involvement,' based on a 5% acceptable risk threshold. Section 2.3.4 reports that among advanced T-category lateralized tumors with ipsilateral LNLs II+III involved, 12/54 (22.2%) had contralateral LNL II involvement. Using the paper's own imaging sensitivity (81%) and specificity (76%) from Section 7, a clinically negative contralateral II in that subgroup has posterior occult risk ~6.7% — above 5%. So either the model's posterior for this subgroup is >5%, in which case the blanket recommendation contradicts the model's own criterion, or the model's posterior is <5%, in which case the model fails to reproduce the prevalence it was trained on. The limitations section (8.3) doesn't mention this safety-relevant subgroup.

The other soft spots are milder. The evaluation is entirely in-sample; there is no held-out test set or cross-validation. The 'consensus' as truth is an approximation, acknowledged but potentially biasing both training and evaluation. The midline-extension prevalence mismatch is acknowledged and partly mitigated by conditioning, but it still affects the joint distribution. These are not fatal, but they reinforce that the clinical safety claims need external validation before they guide de-escalation.

Overall: this deserves peer review. The modeling is serious, the data/code are reproducible, and the clinical question is important. But the authors need to fix the threshold inconsistency — either refine the recommendation for the high-risk subgroup or show that the model's posterior for that subgroup is actually below 5% and explain why it deviates from the 22.2% prevalence. I'd send it to a journal with a request for that revision and for out-of-sample validation.

Referee Report

3 major / 4 minor

Summary. This paper extends a previously published hidden Markov model for ipsilateral lymphatic spread in oropharyngeal squamous cell carcinoma (OPSCC) to the contralateral neck. The bilateral model represents lymph node levels (LNLs) I, II, III, IV, V, and VII on both sides as binary hidden variables, evolves them synchronously over time, and couples the two sides only through the shared time to diagnosis and the primary tumor's midline extension. Parameters are inferred via MCMC from a pooled multi-institutional dataset of 833 patients. The authors report that the model reproduces observed contralateral involvement patterns, estimates risks of occult contralateral disease, and, based on a 5% risk threshold, recommends unilateral elective irradiation for lateralized tumors and limitation to LNL II for midline-crossing tumors with a clinically negative contralateral neck.

Significance. If the risk estimates are reliable, the model could support personalized CTV-N de-escalation in a substantial patient group, with meaningful toxicity reduction. The paper's strengths include a compact and interpretable parameterization, a multi-institutional dataset, a reproducible implementation with publicly available code and data, and explicit MCMC convergence diagnostics. The clinical significance is, however, conditional on the model's ability to generalize beyond the training cohort and on the safety of the proposed risk threshold for the highest-risk subgroups; the current manuscript does not yet establish those conditions.

major comments (3)

[§5.1, §6.2] The model evaluation in §6.2 is entirely in-sample: the predicted prevalences in Figures 7–9 are compared with the observed prevalences in the same 833 patients used for training (§5.1). Because the model parameters are fit to these data, the agreement shown in those figures is expected and does not by itself demonstrate that the model will provide accurate risk estimates for new patients. Given that the paper makes a clinical recommendation (unilateral CTV-N exclusion for lateralized tumors), the authors should provide at least an internal validation, for example by cross-validation or by training on three of the four institutions in Table 1 and testing on the fourth. Without such validation, the claim that the model 'accurately and precisely describes observed patterns' is overstated.
[§8.2 with §2.3.4 and §7.1] The blanket recommendation in §8.2 that for lateralized tumors with no contralateral clinical involvement 'unilateral radiotherapy is sufficient, regardless of T-category or ipsilateral involvement' is not reconciled with the paper's own data. Section 2.3.4 reports a 22.2% prevalence (12 of 54 patients) of contralateral level II involvement for advanced T-category lateralized tumors with ipsilateral levels II and III involved. For a patient in this subgroup with a clinically negative contralateral level II, the imaging characteristics assumed in §7 (sensitivity 81%, specificity 76%) yield a posterior occult-disease risk of about 6.7%, above the 5% threshold used throughout the paper. Conditioning on additional contralateral levels being clinically negative could lower this posterior, but the paper does not report the model's predicted risk for this exact subgroup in §7.1 or Figure 10. Since this is the subgroup for which the 'regardless' recommendation is most consequential, the authors should either report the model's posterior for this scenario and demonstrate that it falls below 5%, or qualify the recommendation. As written, the central de-escalation claim is not directly supported by the presented risk estimates.
[§5.1, §B, §8.3.2] The training procedure treats the consensus involvement states as the true hidden states X (§5.1), while the posterior risk predictions in §7 apply an observation model with sensitivity/specificity to the same LNLs. For non-surgical patients the consensus is itself derived from imaging using the same sensitivity/specificity values (§B, Table 3), so the training targets already incorporate diagnostic uncertainty that is later re-applied as if it were the clinical observation process. The paper acknowledges this approximation in §8.3.2, but it does not assess how sensitive the occult-risk estimates are to it. A sensitivity analysis—for example, training the model only on pathologically confirmed cases and validating on clinically diagnosed cases, or treating the consensus as a latent variable—would strengthen confidence in the reported risk estimates.

minor comments (4)

[§3, §5.3, §4.3] There are several typos: 'diagnositc' in Section 3 after Eq. (3), 'enumator' in §5.3, 'matrx' in §4.3, and 'probabiltiy' in §3.1; these should be corrected.
[Figures 7–9] The y-axis is omitted in Figures 7–9; the authors state that the numerical value is not intuitively interpretable, but this makes it difficult to assess whether the predicted histograms match the width of the observed beta posteriors. Consider adding a labeled axis or a scale.
[§4.2] In the recursive formula for P(Xc, ϵ=True|τ+1), the variable τ is used without explicit definition; please clarify that τ indexes time steps from 0 to tmax.
[§7] The sensitivity and specificity values are introduced for 'imaging' but the numbers correspond to CT in Table 3; please specify the modality (e.g., CT) to avoid ambiguity for readers.

Circularity Check

1 steps flagged · score 3.0 of 10

Prevalence 'predictions' are in-sample fits from the same 833-patient cohort; central occult-risk claims are model extrapolations rather than independent predictions.

fitted input called prediction [Section 5.1 and 5.3; Eq. 7; Figs. 7-9]
"We trained the model using the dataset of 833 patients described in section 2. The consensus decision on lymphatic involvement is assumed to correspond to the true hidden state of involvement X. ... We evaluate the model’s ability to describe the observed frequencies of lymphatic involvement patterns. We compare the prevalence of selected involvement patterns in the data to the model’s predicted prevalence, given patient scenarios."

The MCMC inference maximizes the cohort log-likelihood (Eq. 7) over exactly these 833 patients' involvement patterns. The 'predicted prevalence' shown in Figs. 7-9 is the same fitted joint distribution marginalized to each scenario, while the 'observed' beta posteriors are the empirical frequencies that entered the likelihood. Agreement is therefore an in-sample goodness-of-fit rather than an independent prediction. The model is compact, so the match is not an exact identity, but the comparison is partly forced by the fitting procedure. The central occult-risk estimates for clinically negative LNLs remain generative Bayes outputs and are not directly fitted to an occult-disease endpoint, which is why the circularity is only partial.

full rationale

The only concrete reduction of a 'prediction' to model inputs is the prevalence comparison performed on the training cohort: parameters learned from the 833 patients are used to compute 'predicted' prevalences for the same patients, so agreement reflects fitting as well as model structure. The central clinical claim—occult contralateral risk and the 5%-threshold de-escalation recommendation—is a generative model output rather than a direct fit to the clinical endpoint, and the model structure is grounded in previously published, code-reproduced work on the ipsilateral HMM; no load-bearing self-citation chain or imported uniqueness theorem was found. The no-cross-drainage factorization (Eq. 9) is a transparent anatomical assumption, not a conclusion derived from itself. The apparent tension between the 22.2% observed contralateral-II prevalence in advanced lateralized tumors with ipsilateral II+III involvement and the blanket unilateral-RT recommendation in Section 8.2 is a clinical-consistency or safety concern, not a circularity, and therefore does not increase the circularity score.

Assumptions & free parameters 22 free parameters · 6 assumptions · 0 invented entities

The model uses 19 fitted spread and timing parameters plus hand-fixed p_early and tmax, and the clinical recommendation depends on a 5% risk threshold. The DAG structure is inherited from the authors' prior model selection. No new physical entities are introduced.

free parameters (22)

b_i_1 = 2.80% ± 0.26%
Ipsilateral direct spread from primary tumor to LNL I; learned via MCMC.
b_i_2 = 34.89% ± 1.40%
Ipsilateral direct spread from primary tumor to LNL II; learned via MCMC.
b_i_3 = 5.45% ± 0.66%
Ipsilateral direct spread from primary tumor to LNL III; learned via MCMC.
b_i_4 = 0.94% ± 0.18%
Ipsilateral direct spread from primary tumor to LNL IV; learned via MCMC.
b_i_5 = 1.83% ± 0.22%
Ipsilateral direct spread from primary tumor to LNL V; learned via MCMC.
b_i_7 = 2.32% ± 0.26%
Ipsilateral direct spread from primary tumor to LNL VII; learned via MCMC.
b_c_1 = 0.29% ± 0.09%
Contralateral direct spread from primary tumor to LNL I for lateralized tumors; learned via MCMC.
b_c_2 = 2.46% ± 0.29%
Contralateral direct spread from primary tumor to LNL II for lateralized tumors; learned via MCMC.
b_c_3 = 0.14% ± 0.07%
Contralateral direct spread from primary tumor to LNL III for lateralized tumors; learned via MCMC.
b_c_4 = 0.19% ± 0.08%
Contralateral direct spread from primary tumor to LNL IV for lateralized tumors; learned via MCMC.
b_c_5 = 0.05% ± 0.04%
Contralateral direct spread from primary tumor to LNL V for lateralized tumors; learned via MCMC.
b_c_7 = 0.50% ± 0.17%
Contralateral direct spread from primary tumor to LNL VII for lateralized tumors; learned via MCMC.
alpha = 33.87% ± 4.32%
Mixing parameter that raises contralateral tumor spread rates when the tumor crosses the midline (eq. 12).
t12 = 62.50% ± 16.69%
LNL-to-LNL spread probability from LNL I to II; learned via MCMC.
t23 = 14.23% ± 1.64%
LNL-to-LNL spread probability from LNL II to III; learned via MCMC.
t34 = 15.86% ± 1.93%
LNL-to-LNL spread probability from LNL III to IV; learned via MCMC.
t45 = 14.58% ± 3.76%
LNL-to-LNL spread probability from LNL IV to V; learned via MCMC.
p_adv = 44.98% ± 1.99%
Binomial time-prior parameter for advanced T-category tumors, controlling expected time to diagnosis.
p_epsilon = 8.16% ± 0.48%
Per-time-step probability that the primary tumor crosses the midline.
p_early = 0.3 (fixed)
Time-prior parameter for early T-category, fixed by hand rather than inferred.
tmax = 10 (fixed)
Maximum number of time steps in the HMM, chosen by hand.
5% risk threshold = 0.05
Clinical decision threshold assumed for whether elective irradiation of an LNL is recommended; not a model parameter, but drives the recommendations.

assumptions (6)

domain assumption No direct lymphatic drainage between ipsilateral and contralateral LNLs; contralateral spread occurs only via the primary tumor
Anatomical claim used to factorize the joint prior in equation 9; if wrong for some patients, contralateral risks would be underestimated.
domain assumption Spread rates between LNLs are symmetric across the two sides of the neck (eq. 11)
Assumed to share parameters and reduce dimensionality; plausible but not empirically verified per side.
ad hoc to paper Midline extension is a binary random variable that switches the contralateral transition matrix via linear mixing (eq. 12)
A chosen functional form for how midline crossing increases contralateral spread; alpha is fitted to data.
ad hoc to paper Time to diagnosis follows a binomial distribution with one parameter per T-category group
Modeling convenience for marginalizing over unknown diagnosis time; the form is not derived from biology.
domain assumption Consensus diagnostic status equals the true hidden involvement state during training
Simplification stated in section 5.1 and discussed as a limitation in section 8.3.2; pathological confirmation is not available for all patients.
domain assumption The lymphatic DAG in figure 2, including which LNL-to-LNL arcs exist, is the structure selected in the authors' prior model-selection work
The graph structure is inherited from reference 16, where it was chosen by maximizing model evidence on overlapping data.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Probabilistic Model of Bilateral Lymphatic Spread in Head and Neck Cancer." pith.science (2026). https://pith.science/paper/MU3HHBJD

@misc{pith2026250116910,
  author       = {Pith},
  title        = {Pith review of: A Probabilistic Model of Bilateral Lymphatic Spread in Head and Neck Cancer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MU3HHBJD}},
  note         = {Machine review of arXiv:2501.16910}
}

read the original abstract

Current guidelines for elective nodal irradiation in oropharyngeal squamous cell carcinoma (OPSCC) recommend including large portions of the contralateral lymph system in the clinical target volume (CTV-N), even for lateralized tumors with no clinical lymph node involvement in the contralateral neck. This study introduces a probabilistic model of bilateral lymphatic tumor progression in OPSCC to estimate personalized risks of occult disease in specific lymph node levels (LNLs) based on clinical involvement, T-stage, and tumor lateralization. Building on a previously developed hidden Markov model for ipsilateral spread, we extend the approach to the contralateral neck. The model represents LNLs I, II, III, IV, V, and VII on both sides of the neck as binary hidden variables (healthy/involved), connected via arcs representing spread probabilities. These probabilities are learned using Markov chain Monte Carlo (MCMC) sampling from a dataset of 833 OPSCC patients, enabling the model to reflect the underlying lymphatic progression dynamics. The model accurately and precisely describes observed patterns of involvement with a compact set of interpretable parameters. Midline extension of the primary tumor is identified as the primary risk factor for contralateral involvement, with advanced T-stage and extensive ipsilateral involvement further increasing risk. Occult disease in contralateral LNL III is highly unlikely if upstream LNL II is clinically negative, and in contralateral LNL IV, occult disease is exceedingly rare without LNL III involvement. For lateralized tumors not crossing the midline, the model suggests the contralateral neck may safely be excluded from the CTV-N. For tumors extending across the midline but with a clinically negative contralateral neck, the CTV-N could be limited to LNL II, reducing unnecessary exposure of normal tissue while maintaining regional tumor control.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 19 canonical work pages

[1]

Radiotherapy and Oncology 69(3), 227–236 (2003) https: //doi.org/10.1016/j.radonc.2003.09.011

Grégoire, V., Levendag, P., Ang, K.K., Bernier, J., Braaksma, M., Budach, V., Chao, C., Coche, E., Cooper, J.S., Cosnard, G., Eisbruch, A., El-Sayed, S., Emami, B., Grau, C., Hamoir, M., Lee, N., Maingon, P., Muller, K., Rey- chler, H.: CT-based delineation of lymph node levels and related CTVs in the node-negative neck: DAHANCA, EORTC, GORTEC, NCIC,RTOG ...

work page doi:10.1016/j.radonc.2003.09.011 2003
[2]

DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG consensus guidelines

Grégoire, V., Ang, K., Budach, W., Grau, C., Hamoir, M., Langendijk, J.A., Lee, A., Le, Q.-T., Maingon, P., Nutting, C., O’Sullivan, B., Porceddu, S.V., Lengele, B.: Delineation of the neck node levels for head and neck tumors: A 2013 update. DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG consensus guidelines. Radiotherapy and Oncology110(1), 172–181...

work page doi:10.1016/j.radonc.2013.10.010 2014
[3]

Radiotherapy and Oncology126(1), 3–24 (2018) https://doi.org/10.1016/j.radonc.2017.10.016

Grégoire, V., Evans, M., Le, Q.-T., Bourhis, J., Budach, V., Chen, A., Eisbruch, A., Feng, M., Giralt, J., Gupta, T., Hamoir, M., Helito, J.K., Hu, C., Hunter, K., Johansen, J., Kaanders, J., Laskar, S.G., Lee, A., Maingon, P., Mäkitie, A., Micciche’, F., Nicolai, P., O’Sullivan, B., Poitevin, A., Porceddu, S., Składowski, K., Tribius, S., Waldron, J., We...

work page 2018
[4]

Seminars in Radiation Oncology12(3), 238–249 (2002) https://doi.org/10.1053/srao.2002.32435

Eisbruch, A., Foote, R.L., O’Sullivan, B., Beitler, J.J., Vikram, B.: Intensity- modulated radiation therapy for head and neck cancer: Emphasis on the selection and delineation of the targets. Seminars in Radiation Oncology12(3), 238–249 (2002) https://doi.org/10.1053/srao.2002.32435

work page arXiv 2002
[5]

Radiotherapy and Oncology134, 1–9 (2019) https: //doi.org/10.1016/j.radonc.2019.01.018

Biau, J., Lapeyre, M., Troussier, I., Budach, W., Giralt, J., Grau, C., Kazmierska, J., Langendijk, J.A., Ozsahin, M., O’Sullivan, B., Bourhis, J., Grégoire, V.: Selection of lymph node target volumes for definitive head and neck radiation therapy: A 2019 Update. Radiotherapy and Oncology134, 1–9 (2019) https: //doi.org/10.1016/j.radonc.2019.01.018

work page doi:10.1016/j.radonc.2019.01.018 2019
[6]

30 International Journal of Radiation Oncology*Biology*Physics53(5), 1174–1184 (2002) https://doi.org/10.1016/S0360-3016(02)02881-X

Chao, K.S.C., Wippold, F.J., Ozyigit, G., Tran, B.N., Dempsey, J.F.: Determi- nation and delineation of nodal target volumes for head-and-neck cancer based on patterns of failure in patients receiving definitive and postoperative IMRT. 30 International Journal of Radiation Oncology*Biology*Physics53(5), 1174–1184 (2002) https://doi.org/10.1016/S0360-3016(...

work page doi:10.1016/s0360-3016(02)02881-x 2002
[7]

Radiation Oncology 6(1), 97 (2011) https://doi.org/10.1186/1748-717X-6-97

Vorwerk, H., Hess, C.F.: Guidelines for delineation of lymphatic clinical target vol- umes for high conformal radiotherapy: Head and neck region. Radiation Oncology 6(1), 97 (2011) https://doi.org/10.1186/1748-717X-6-97

work page doi:10.1186/1748-717x-6-97 2011
[8]

British Journal of Oral and Maxillofacial Surgery47(1), 5–9 (2009) https://doi

Ferlito, A., Silver, C.E., Rinaldo, A.: Elective management of the neck in oral cavity squamous carcinoma: Current concepts supported by prospective studies. British Journal of Oral and Maxillofacial Surgery47(1), 5–9 (2009) https://doi. org/10.1016/j.bjoms.2008.06.001

work page doi:10.1016/j.bjoms.2008.06.001 2009

Show all 24 references

[9]

Scientific Reports 11(1), 12261 (2021) https://doi.org/10.1038/s41598-021-91544-1

Ludwig, R., Pouymayou, B., Balermpas, P., Unkelbach, J.: A hidden Markov model for lymphatic tumor progression in the head and neck. Scientific Reports 11(1), 12261 (2021) https://doi.org/10.1038/s41598-021-91544-1

2021 doi
[10]

arXiv (2023)

Ludwig, R., Schubert, A., Barbatei, D., Bauwens, L., Hoffmann, J.-M., Werlen, S., Elicin, O., Dettmer, M., Zrounba, P., Pouymayou, B., Balermpas, P., Grégoire, V., Giger, R., Unkelbach, J.: Modelling the Lymphatic Metastatic Progression Pathways of OPSCC from Multi-Institution...

2023 doi arXiv
[11]

Physics in Medicine & Biology64(16), 165003 (2019) https://doi.org/10.1088/1361-6560/ab2a18

Pouymayou, B., Balermpas, P., Riesterer, O., Guckenberger, M., Unkelbach, J.: A Bayesian network model of lymphatic tumor progression for personalized elective CTV definition in head and neck cancers. Physics in Medicine & Biology64(16), 165003 (2019) https://doi.org/10.1088/1...

2019 doi
[12]

Data in Brief43, 108345 (2022) https://doi.org/10.1016/j.dib.2022.108345

Ludwig, R., Hoffmann, J.-M., Pouymayou, B., Morand, G., Däppen, M.B., Guckenberger,M.,Grégoire,V.,Balermpas,P.,Unkelbach,J.:Adatasetonpatient- individual lymph node involvement in oropharyngeal squamous cell carcinoma. Data in Brief43, 108345 (2022) https://doi.org/10.1016/j.d...

2022
[13]

Data in Brief52, 110020 (2024) https://doi.org/10.1016/j.dib.2023.110020

Ludwig, R., Schubert, A., Barbatei, D., Bauwens, L., Werlen, S., Elicin, O., Dettmer, M., Zrounba, P., Balermpas, P., Pouymayou, B., Grégoire, V., Giger, R., Unkelbach, J.: A multi-centric dataset on patient-individual pathological lymph node involvement in head and neck squam...

2024
[14]

European Journal of Radiology64(2), 266–272 (2007) https://doi.org/ 10.1016/j.ejrad.2007.02.037

De Bondt, R.B.J., Nelemans, P.J., Hofman, P.A.M., Casselman, J.W., Kremer, B., Van Engelshoven, J.M.A., Beets-Tan, R.G.H.: Detection of lymph node metastases in head and neck cancer: A meta-analysis comparing US, USgFNAC, CT and MR imaging. European Journal of Radiology64(2), ...

2007 doi
[15]

JNCI: Journal of the National Cancer Institute100(10), 712–720 (2008) https://doi.org/10.1093/jnci/djn125

Kyzas, P.A., Evangelou, E., Denaxa-Kyza, D., Ioannidis, J.P.A.: 18F- fluorodeoxyglucose positron emission tomography to evaluate cervical node 31 metastases in patients with head and neck squamous cell carcinoma: A meta- analysis. JNCI: Journal of the National Cancer Institute...

2008 doi
[16]

Scientific Reports14(1), 15750 (2024) https://doi.org/10.1038/s41598-024-66012-1

Ludwig, R., Schubert, A.D., Barbatei, D., Bauwens, L., Hoffmann, J.-M., Werlen, S., Elicin, O., Dettmer, M., Zrounba, P., Pouymayou, B., Balermpas, P., Grégoire, V., Giger, R., Unkelbach, J.: Modelling the lymphatic metastatic progression pathways of OPSCC from multi-instituti...

2024 doi
[17]

Foreman-Mackey, D., Hogg, D.W., Lang, D., Goodman, J.: Emcee: The MCMC Hammer.PublicationsoftheAstronomicalSocietyofthePacific 125(925),306–312 (2013) https://doi.org/10.1086/670067

2013 doi
[18]

Statistics and Computing18(4), 435–446 (2008) https: //doi.org/10.1007/s11222-008-9104-9

Ter Braak, C.J.F., Vrugt, J.A.: Differential Evolution Markov Chain with snooker updater and fewer chains. Statistics and Computing18(4), 435–446 (2008) https: //doi.org/10.1007/s11222-008-9104-9

2008 doi
[19]

The Astrophysical Journal Supplement Series210(1), 11 (2013) https://doi.org/10.1088/0067-0049/210/1/11

Nelson, B., Ford, E.B., Payne, M.J.: RUN DMC: AN EFFICIENT, PARALLEL CODE FOR ANALYZING RADIAL VELOCITY OBSERVATIONS USING N -BODY INTEGRATIONS AND DIFFERENTIAL EVOLUTION MARKOV CHAIN MONTE CARLO. The Astrophysical Journal Supplement Series210(1), 11 (2013) https://doi.org/10....

2013 doi
[20]

Benson, N., Whipple, M., Kalet, I.J.: A Markov model approach to predicting regional tumor spread in the lymphatic system of the head and neck. AMIA ... Annual Symposium proceedings. AMIA Symposium, 31–35 (2006)

2006
[21]

AMIA Annual Symposium Proceedings 2016, 1832–1839 (2017)

Jung, H., Law, A., Grunblatt, E., Wang, L.L., Kusano, A., Mejino, J.L.V., Whipple, M.E.: Development of a Novel Markov Chain Model for the Prediction of Head and Neck Squamous Cell Carcinoma Dissemination. AMIA Annual Symposium Proceedings 2016, 1832–1839 (2017)

2017
[22]

Clinical Trial Registration NCT06563362, clinicaltrials.gov (August 2024)

University of Zurich: Personalized Volume-deescalated Elective Nodal Irradiation in Oropharyngeal Head and Neck. Clinical Trial Registration NCT06563362, clinicaltrials.gov (August 2024)

2024
[23]

Sarrut, D., Rit, S.: Proceedings of the XX-th International Conference on the Use of Computers in Radiation Therapy (ICCR), Lyon, France (2024)

2024
[24]

PhD thesis, University of Zurich, Zurich (2023)

Ludwig, R.: Modelling Lymphatic Metastatic Progression in Head and Neck Cancer. PhD thesis, University of Zurich, Zurich (2023). https://doi.org/10.5167/ uzh-231470 32

2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Radiotherapy and Oncology 69(3), 227–236 (2003) https: //doi.org/10.1016/j.radonc.2003.09.011

Grégoire, V., Levendag, P., Ang, K.K., Bernier, J., Braaksma, M., Budach, V., Chao, C., Coche, E., Cooper, J.S., Cosnard, G., Eisbruch, A., El-Sayed, S., Emami, B., Grau, C., Hamoir, M., Lee, N., Maingon, P., Muller, K., Rey- chler, H.: CT-based delineation of lymph node levels and related CTVs in the node-negative neck: DAHANCA, EORTC, GORTEC, NCIC,RTOG ...

work page doi:10.1016/j.radonc.2003.09.011 2003

[2] [2]

DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG consensus guidelines

Grégoire, V., Ang, K., Budach, W., Grau, C., Hamoir, M., Langendijk, J.A., Lee, A., Le, Q.-T., Maingon, P., Nutting, C., O’Sullivan, B., Porceddu, S.V., Lengele, B.: Delineation of the neck node levels for head and neck tumors: A 2013 update. DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG consensus guidelines. Radiotherapy and Oncology110(1), 172–181...

work page doi:10.1016/j.radonc.2013.10.010 2014

[3] [3]

Radiotherapy and Oncology126(1), 3–24 (2018) https://doi.org/10.1016/j.radonc.2017.10.016

Grégoire, V., Evans, M., Le, Q.-T., Bourhis, J., Budach, V., Chen, A., Eisbruch, A., Feng, M., Giralt, J., Gupta, T., Hamoir, M., Helito, J.K., Hu, C., Hunter, K., Johansen, J., Kaanders, J., Laskar, S.G., Lee, A., Maingon, P., Mäkitie, A., Micciche’, F., Nicolai, P., O’Sullivan, B., Poitevin, A., Porceddu, S., Składowski, K., Tribius, S., Waldron, J., We...

work page 2018

[4] [4]

Seminars in Radiation Oncology12(3), 238–249 (2002) https://doi.org/10.1053/srao.2002.32435

Eisbruch, A., Foote, R.L., O’Sullivan, B., Beitler, J.J., Vikram, B.: Intensity- modulated radiation therapy for head and neck cancer: Emphasis on the selection and delineation of the targets. Seminars in Radiation Oncology12(3), 238–249 (2002) https://doi.org/10.1053/srao.2002.32435

work page arXiv 2002

[5] [5]

Radiotherapy and Oncology134, 1–9 (2019) https: //doi.org/10.1016/j.radonc.2019.01.018

Biau, J., Lapeyre, M., Troussier, I., Budach, W., Giralt, J., Grau, C., Kazmierska, J., Langendijk, J.A., Ozsahin, M., O’Sullivan, B., Bourhis, J., Grégoire, V.: Selection of lymph node target volumes for definitive head and neck radiation therapy: A 2019 Update. Radiotherapy and Oncology134, 1–9 (2019) https: //doi.org/10.1016/j.radonc.2019.01.018

work page doi:10.1016/j.radonc.2019.01.018 2019

[6] [6]

30 International Journal of Radiation Oncology*Biology*Physics53(5), 1174–1184 (2002) https://doi.org/10.1016/S0360-3016(02)02881-X

Chao, K.S.C., Wippold, F.J., Ozyigit, G., Tran, B.N., Dempsey, J.F.: Determi- nation and delineation of nodal target volumes for head-and-neck cancer based on patterns of failure in patients receiving definitive and postoperative IMRT. 30 International Journal of Radiation Oncology*Biology*Physics53(5), 1174–1184 (2002) https://doi.org/10.1016/S0360-3016(...

work page doi:10.1016/s0360-3016(02)02881-x 2002

[7] [7]

Radiation Oncology 6(1), 97 (2011) https://doi.org/10.1186/1748-717X-6-97

Vorwerk, H., Hess, C.F.: Guidelines for delineation of lymphatic clinical target vol- umes for high conformal radiotherapy: Head and neck region. Radiation Oncology 6(1), 97 (2011) https://doi.org/10.1186/1748-717X-6-97

work page doi:10.1186/1748-717x-6-97 2011

[8] [8]

British Journal of Oral and Maxillofacial Surgery47(1), 5–9 (2009) https://doi

Ferlito, A., Silver, C.E., Rinaldo, A.: Elective management of the neck in oral cavity squamous carcinoma: Current concepts supported by prospective studies. British Journal of Oral and Maxillofacial Surgery47(1), 5–9 (2009) https://doi. org/10.1016/j.bjoms.2008.06.001

work page doi:10.1016/j.bjoms.2008.06.001 2009