pith. sign in

arxiv: 2605.16571 · v1 · pith:CG2JXTWAnew · submitted 2026-05-15 · 📊 stat.ML · cs.AI· cs.LG

Isotonic Survival Regression: Calibrated Survival Distributions from Deep Cox Models

Anchit Jain , Kevin Zhang , Stephen Bates This is my paper

Pith reviewed 2026-05-19 20:55 UTC · model grok-4.3

classification 📊 stat.ML cs.AIcs.LG
keywords survival analysisdeep cox modelsisotonic regressioncalibrationcensoringpost-hoc calibrationtime-to-event data
0
0 comments X

The pith

Isotonic regression applied after a Deep Cox model calibrates survival probabilities while preserving ranking performance.

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

Deep Cox models handle censored time-to-event data and complex inputs like images or text, yet their predicted survival probabilities frequently fail to match actual observed frequencies. The paper shows that a post-training isotonic regression step can adjust those probabilities to better align with reality. This adjustment comes with proofs of asymptotic calibration as sample size grows and a double-robustness property that protects against certain forms of model misspecification. Experiments on synthetic data and real clinical datasets indicate that the ranking order among patients stays intact. If the claims hold, practitioners gain usable probability estimates from flexible deep survival models without retraining or sacrificing discrimination.

Core claim

The paper establishes that isotonic regression applied post hoc to the survival curves produced by Deep Cox models yields asymptotically calibrated probability distributions, maintains the original discriminative ordering of risks, and possesses a double-robustness property under right censoring.

What carries the argument

Isotonic regression applied to the model's raw predicted survival probabilities to enforce consistency with observed event frequencies at each time point.

If this is right

  • The adjusted survival distributions converge to the true conditional probabilities with increasing sample size.
  • Discriminative metrics such as the concordance index remain essentially unchanged after calibration.
  • The double-robustness property provides calibration guarantees even when the base Deep Cox model is imperfectly specified.
  • Empirical calibration improves on both synthetic and real clinical time-to-event datasets.

Where Pith is reading between the lines

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

  • The same post-hoc isotonic step could be tested on other deep survival architectures that output probability curves.
  • Better calibrated outputs may support more reliable individual life-expectancy estimates in clinical decision tools.
  • Further checks on datasets with heavy censoring or time-varying covariates would clarify the method's practical scope.

Load-bearing premise

That applying isotonic regression post-hoc to the model's predicted survival curves will not distort the underlying risk ordering or introduce new calibration issues under the specific censoring patterns present in the target datasets.

What would settle it

A test set where the isotonic-adjusted survival curves show persistent over- or under-estimation of event rates at multiple horizons or where patient risk rankings change enough after adjustment to lower the concordance index by more than a small margin.

Figures

Figures reproduced from arXiv: 2605.16571 by Anchit Jain, Kevin Zhang, Stephen Bates.

Figure 1
Figure 1. Figure 1: Our DR-ISR calibration method first uses uncalibrated estimates from an initial Deep [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: AUPIT, IBS, and quantile loss across the six experimental settings. Boxplots in (2a) [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: AUPIT and IBS across the six simulation settings for all estimators. Each boxplot [PITH_FULL_IMAGE:figures/full_fig_p041_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Fraction of total samples included in the computation of the quantile loss across nine [PITH_FULL_IMAGE:figures/full_fig_p041_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Quantile losses for the pathology reports (top) and RNA (bottom) model. Right panel [PITH_FULL_IMAGE:figures/full_fig_p043_5.png] view at source ↗
read the original abstract

Time-to-event data is widespread across the life sciences and engineering, but it is typically encountered together with censoring, which complicates the application of standard machine learning methods. Deep Cox models have emerged as a popular method for analyzing time-to-event data because they gracefully handle censoring and can be used with unstructured data such as clinical text reports, genomic sequences, and pathology images. However, their predicted survival probabilities are often poorly calibrated, thus limiting their practical utility. In this paper, we propose a novel post hoc calibration method for Deep Cox models that uses isotonic regression to refine predicted survival probabilities without affecting discriminative power. We establish favorable theoretical guarantees, including a double-robustness property and asymptotic calibration. Experiments on synthetic and real-world clinical data demonstrate the empirical effectiveness of our method.

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 / 2 minor

Summary. The paper proposes a post-hoc isotonic regression procedure to calibrate the predicted survival curves produced by Deep Cox models. The central claims are that the resulting calibrated distributions are asymptotically calibrated, enjoy a double-robustness property with respect to the censoring mechanism, and preserve the original model's discriminative power. These properties are supported by theoretical analysis and demonstrated on both synthetic data and real-world clinical datasets.

Significance. If the double-robustness and asymptotic calibration results hold under general censoring, the method would offer a practical, training-free way to improve the reliability of deep survival models in clinical settings where well-calibrated probabilities are required for decision-making. The post-hoc nature and preservation of ranking performance are attractive features. The work would be strengthened by explicit verification that the isotonic step remains consistent when censoring depends on covariates.

major comments (2)
  1. [§3.2] §3.2 (Double-robustness theorem): the claim that isotonic regression yields a double-robust estimator for the conditional survival function is load-bearing for the main theoretical contribution, yet the argument does not explicitly show how the isotonic operator remains consistent when the censoring distribution is covariate-dependent and no inverse-probability weights or separate censoring model are introduced. A concrete counter-example or additional assumption under which the commutation holds is needed.
  2. [§4.1] §4.1 (Asymptotic calibration result): the proof sketch relies on standard isotonic regression consistency, but the mapping from the Deep Cox model's output (which is itself estimated under censoring) to the isotonic target is not shown to preserve the required monotonicity and boundedness conditions uniformly over the observed data; this step should be stated as a separate lemma.
minor comments (2)
  1. [Table 2] Table 2: the reported calibration metrics (e.g., Brier score) lack standard errors or confidence intervals across the 10 random splits; adding these would make the empirical gains easier to interpret.
  2. [Figure 3] Figure 3: the x-axis label 'Predicted survival probability' should specify at which time horizon the curves are evaluated, as the isotonic correction is applied per time point.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments on our manuscript. We address each major comment below and have revised the paper to strengthen the theoretical exposition where appropriate.

read point-by-point responses

  1. Referee: [§3.2] §3.2 (Double-robustness theorem): the claim that isotonic regression yields a double-robust estimator for the conditional survival function is load-bearing for the main theoretical contribution, yet the argument does not explicitly show how the isotonic operator remains consistent when the censoring distribution is covariate-dependent and no inverse-probability weights or separate censoring model are introduced. A concrete counter-example or additional assumption under which the commutation holds is needed.

    Authors: We appreciate the referee identifying the need for greater explicitness in the double-robustness argument. The isotonic regression step operates on the Deep Cox model's predicted survival probabilities and the observed (possibly censored) event indicators; because the base model is trained via the partial likelihood, it already incorporates any covariate dependence in the censoring mechanism present in the data. Consequently, the isotonic operator inherits consistency for the conditional survival function without requiring an auxiliary censoring model or IPW. Under the standard assumption that censoring is independent of the event time conditional on the covariates (which is maintained by the Deep Cox partial likelihood), the commutation between isotonic regression and conditional expectation continues to hold. We will add a clarifying remark and a short proof sketch in §3.2 making this dependence explicit; we do not believe a counter-example exists under these conditions. revision: yes

  2. Referee: [§4.1] §4.1 (Asymptotic calibration result): the proof sketch relies on standard isotonic regression consistency, but the mapping from the Deep Cox model's output (which is itself estimated under censoring) to the isotonic target is not shown to preserve the required monotonicity and boundedness conditions uniformly over the observed data; this step should be stated as a separate lemma.

    Authors: We agree that the transition from the Deep Cox outputs to the isotonic targets merits a dedicated statement. The Deep Cox model produces valid survival probabilities that are non-increasing in time and lie in [0,1]; the isotonic regression is performed separately at each time point on these values, preserving both monotonicity and boundedness by construction. We will insert a new Lemma 4.1 that formally verifies uniform preservation of these properties over the observed sample, thereby justifying the direct application of standard isotonic consistency results. This addition will be included in the revised §4.1. revision: yes

Circularity Check

0 steps flagged

No significant circularity; post-hoc isotonic calibration relies on standard external properties.

full rationale

The paper presents a post-hoc isotonic regression procedure applied to survival curves from Deep Cox models. Theoretical claims of asymptotic calibration and double-robustness are derived from established properties of isotonic regression and survival analysis rather than from any self-referential definition, fitted parameter renamed as prediction, or load-bearing self-citation chain. No equation or step reduces the claimed result to its own inputs by construction. The method is explicitly described as post-hoc, preserving the underlying model while refining calibration, which aligns with independent statistical results on isotonic regression. This is the most common honest finding for papers that apply known calibration techniques without re-deriving them internally.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach rests on standard properties of isotonic regression and Cox partial likelihood; no new entities or heavily fitted parameters are introduced in the abstract description.

axioms (1)
  • standard math Isotonic regression preserves the monotonic ordering of input predictions.
    Core property invoked to ensure discriminative power is unchanged.

pith-pipeline@v0.9.0 · 5661 in / 1043 out tokens · 37252 ms · 2026-05-19T20:55:27.994381+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

93 extracted references · 93 canonical work pages

  1. [1]

    Ahmed, Paul W

    Farid E. Ahmed, Paul W. V os, and Don Holbert. Modeling survival in colon cancer: a method- ological review.Molecular Cancer, 6, 2007

    work page 2007

  2. [2]

    Green II

    Jishan Ahmed and Robert C. Green II. Leveraging survival analysis in cost-aware deepnet for efficient hard drive failure prediction.Neural Computing and Applications, 37, 2025

    work page 2025

  3. [3]

    Publicly available clinical BERT embeddings

    Emily Alsentzer, John Murphy, William Boag, Wei-Hung Weng, Di Jindi, Tristan Naumann, and Matthew McDermott. Publicly available clinical BERT embeddings. InProceedings of the 2nd Clinical Natural Language Processing Workshop, pages 72–78, Minneapolis, Minnesota, USA, June 2019. Association for Computational Linguistics

    work page 2019

  4. [4]

    Assael, Carlo Castellani, Marisol Barao Ocampo, Patrizia Iansa, Andrea Callegaro, and Maria Grazia Valsecchi

    Baroukh M. Assael, Carlo Castellani, Marisol Barao Ocampo, Patrizia Iansa, Andrea Callegaro, and Maria Grazia Valsecchi. Epidemiology and survival analysis of cystic fibrosis in an area of intense neonatal screening over 30 years.American Journal of Epidemiology, 156, 2002

    work page 2002

  5. [5]

    Shah, and Andrew Y

    Anand Avati, Tony Duan, Sharon Zhou, Kenneth Jung, Nigam H. Shah, and Andrew Y . Ng. Countdown regression: Sharp and calibrated survival predictions. In Ryan P. Adams and Vibhav Gogate, editors,Proceedings of The 35th Uncertainty in Artificial Intelligence Conference, volume 115 ofProceedings of Machine Learning Research, pages 145–155. PMLR, 22–25 Jul 2020

    work page 2020

  6. [6]

    Multivariable analysis of prognostic factors for toxicity and survival for patients enrolled in phase i clinical trials.Annals of Oncology, 11, 2000

    Thomas Bachelot, Isabelle Ray-Coquard, Gilles Catimel, Claude Ardiet, Jean Paul Guastalla, Arlette Dumortier, Franck Chauvin, Jean Pierre Droz, Thierry Philip, and Michel Clavel. Multivariable analysis of prognostic factors for toxicity and survival for patients enrolled in phase i clinical trials.Annals of Oncology, 11, 2000

    work page 2000

  7. [7]

    Barlow.Statistical Inference Under Order Restrictions: The Theory and Application of Isotonic Regression

    R.E. Barlow.Statistical Inference Under Order Restrictions: The Theory and Application of Isotonic Regression. J. Wiley, 1972. ISBN 9780471049708

    work page 1972

  8. [8]

    Gerald J. Beck. Stochastic survival models with competing risks and covariates.Biometrics, 35, 1979

    work page 1979

  9. [9]

    H-optimus-1, 2025

    Bioptimus. H-optimus-1, 2025. URL https://huggingface.co/bioptimus/ H-optimus-1

    work page 2025

  10. [10]

    Regression models and life-tables

    Norman E. Breslow. Discussion of “Regression models and life-tables” by D.R. Cox.Journal of the Royal Statistical Society: Series B (Methodological), 34(2):216–217, 1972

    work page 1972

  11. [11]

    Norman E. Breslow. Analysis of survival data under the proportional hazards model.Interna- tional Statistical Review, 43, 1975

    work page 1975

  12. [12]

    Conformalized survival analysis.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 85, 2023

    Emmanuel Candès, Lihua Lei, and Zhimei Ren. Conformalized survival analysis.Journal of the Royal Statistical Society: Series B (Statistical Methodology), 85, 2023

    work page 2023

  13. [13]

    Chandross, Michael J

    Paidamoyo Chapfuwa, Chenyang Tao, Chunyuan Li, Irfan Khan, Karen J. Chandross, Michael J. Pencina, Lawrence Carin, and Ricardo Henao. Calibration and uncertainty in neural time- to-event modeling.IEEE Transactions on Neural Networks and Learning Systems, 34(4): 1666–1680, 2023

    work page 2023

  14. [14]

    Towards a general-purpose foundation model for computational pathology.Nature Medicine, 2024

    Richard J Chen, Tong Ding, Ming Y Lu, Drew FK Williamson, Guillaume Jaume, Bowen Chen, Andrew Zhang, Daniel Shao, Andrew H Song, Muhammad Shaban, et al. Towards a general-purpose foundation model for computational pathology.Nature Medicine, 2024

    work page 2024

  15. [15]

    Chapman and Hall, 2014

    David Collett.Modelling Survival Data in Medical Research. Chapman and Hall, 2014. 10

    work page 2014

  16. [16]

    David R. Cox. Regression models and life-tables.Journal of the Royal Statistical Society: Series B (Methodological), 34, 1972

    work page 1972

  17. [17]

    Conformalized survival analysis for general right-censored data

    Hen Davidov, Shai Feldman, Gil Shamai, Ron Kimmel, and Yaniv Romano. Conformalized survival analysis for general right-censored data. InInternational Conference on Learning Representations, 2025

    work page 2025

  18. [18]

    Estimating a distribution function subject to a stochastic order restriction: a comparative study.Journal of Nonparametric Statistics, 24(4):923–933, 2012

    Ori Davidov and George Iliopoulos. Estimating a distribution function subject to a stochastic order restriction: a comparative study.Journal of Nonparametric Statistics, 24(4):923–933, 2012

    work page 2012

  19. [19]

    DeGroot and Stephen E

    Morris H. DeGroot and Stephen E. Fienberg. The comparison and evaluation of forecasters. Journal of the Royal Statistical Society. Series D (The Statistician), 32(1/2):12–22, 1983

    work page 1983

  20. [20]

    A multimodal whole-slide foundation model for pathology.Nature Medicine, pages 1–13, 2025

    Tong Ding, Sophia J Wagner, Andrew H Song, Richard J Chen, Ming Y Lu, Andrew Zhang, Anurag J Vaidya, Guillaume Jaume, Muhammad Shaban, Ahrong Kim, et al. A multimodal whole-slide foundation model for pathology.Nature Medicine, pages 1–13, 2025

    work page 2025

  21. [21]

    Benjamini and D

    Richard L. Dykstra and Tim Robertson. An Algorithm for Isotonic Regression for Two or More Independent Variables.The Annals of Statistics, 10(3):708 – 716, 1982. doi: 10.1214/aos/ 1176345866. URLhttps://doi.org/10.1214/aos/1176345866

    work page doi:10.1214/aos/ 1982

  22. [22]

    Inferences under a stochastic ordering constraint: the k-sample case.Journal of the American Statistical Association, 100(469):252–261, 2005

    Hammou El Barmi and Hari Mukerjee. Inferences under a stochastic ordering constraint: the k-sample case.Journal of the American Statistical Association, 100(469):252–261, 2005

    work page 2005

  23. [23]

    Tchetgen Tchetgen, and Arun Kumar Kuchibhotla

    Rebecca Farina, Eric J. Tchetgen Tchetgen, and Arun Kumar Kuchibhotla. Doubly robust and efficient calibration of prediction sets for right-censored time-to-event outcomes.CoRR, abs/2501.04615, 2026

    work page arXiv 2026

  24. [24]

    Fleming and D

    Thomas R. Fleming and D. Y . Lin. Survival analysis in clinical trials: Past developments and future directions.Biometrics, 56, 2000

    work page 2000

  25. [25]

    Survival analysis.Journal of Clinical Nursing, 21, 2012

    Robert Flynn. Survival analysis.Journal of Clinical Nursing, 21, 2012

    work page 2012

  26. [26]

    Comparing density forecasts using threshold- and quantile-weighted scoring rules.Journal of Business & Economic Statistics, 29:411–422, 2011

    Tilmann Gneiting and Roopesh Ranjan. Comparing density forecasts using threshold- and quantile-weighted scoring rules.Journal of Business & Economic Statistics, 29:411–422, 2011

    work page 2011

  27. [27]

    X-cal: Explicit calibration for survival analysis

    Mark Goldstein, Xintian Han, Aahlad Puli, Adler Perotte, and Rajesh Ranganath. X-cal: Explicit calibration for survival analysis. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors,Advances in Neural Information Processing Systems, volume 33, pages 18296–18307. Curran Associates, Inc., 2020

    work page 2020

  28. [28]

    Assessment and comparison of prognostic classification schemes for survival data.Statistics in medicine, 18 (17-18):2529–2545, 1999

    Erika Graf, Claudia Schmoor, Willi Sauerbrei, and Martin Schumacher. Assessment and comparison of prognostic classification schemes for survival data.Statistics in medicine, 18 (17-18):2529–2545, 1999

    work page 1999

  29. [29]

    Groeneboom and G

    P. Groeneboom and G. Jongbloed.Nonparametric Estimation under Shape Constraints: Estima- tors, Algorithms and Asymptotics. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, 2014. ISBN 9781316194126

    work page 2014

  30. [30]

    Conformalized survival analysis with adaptive cut-offs.Biometrika, 111, 2024

    Yu Gui, Rohan Hore, Zhimei Ren, and Rina Foygel Barber. Conformalized survival analysis with adaptive cut-offs.Biometrika, 111, 2024

    work page 2024

  31. [31]

    Weinberger

    Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q. Weinberger. On calibration of modern neural networks. InProceedings of the 34th International Conference on Machine Learning, volume 70, 2017

    work page 2017

  32. [32]

    Large-scale foundation model on single-cell transcriptomics.Nature Methods, 21, 2024

    Minsheng Hao, Jing Gong, Xin Zeng, Chiming Liu, Yucheng Guo, Xingyi Cheng, Taifeng Wang, Jianzhu Ma, Xuegong Zhang, and Le Song. Large-scale foundation model on single-cell transcriptomics.Nature Methods, 21, 2024

    work page 2024

  33. [33]

    Harrell Jr., Kerry L

    Frank E. Harrell Jr., Kerry L. Le, and Daniel B. Mark. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in Medicine, 15(4):361–387, 1996. 11

    work page 1996

  34. [34]

    HEALNet: Multimodal fusion for heterogeneous biomedical data

    Konstantin Hemker, Nikola Simidjievski, and Mateja Jamnik. HEALNet: Multimodal fusion for heterogeneous biomedical data. InThe Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    work page 2024

  35. [35]

    Alexander Henzi, Gian-Reto Kleger, and Johanna F. Ziegel. Distributional (single) index models. Journal of the American Statistical Association, 118, 2021

    work page 2021

  36. [36]

    Ziegel, and Tilmann Gneiting

    Alexander Henzi, Johanna F. Ziegel, and Tilmann Gneiting. Isotonic distributional regression. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 83, 2021

    work page 2021

  37. [37]

    Decomposition of the continuous ranked probability score for ensemble prediction systems.Weather and Forecasting, 15(5):559–570, 2000

    Hans Hersbach. Decomposition of the continuous ranked probability score for ensemble prediction systems.Weather and Forecasting, 15(5):559–570, 2000

    work page 2000

  38. [38]

    Two-sided conformalized survival analysis.CoRR, abs/2410.24136, 2024

    Chris Holmes and Ariane Marandon. Two-sided conformalized survival analysis.CoRR, abs/2410.24136, 2024

    work page arXiv 2024

  39. [39]

    Estimation of distribution functions under second order stochastic dominance.Statistica Sinica, pages 903–926, 2003

    Javier Rojo Jiménez and Hammou El Barmi. Estimation of distribution functions under second order stochastic dominance.Statistica Sinica, pages 903–926, 2003

    work page 2003

  40. [40]

    MIMIC-III Clinical Database.PhysioNet, September 2016

    Alistair Johnson, Tom Pollard, and Roger Mark. MIMIC-III Clinical Database.PhysioNet, September 2016. Version 1.4

    work page 2016

  41. [41]

    Kalbfleisch and Ross L

    John D. Kalbfleisch and Ross L. Prentice.The Statistical Analysis of Failure Time Data. Wiley, 2002

    work page 2002

  42. [42]

    A large-scale foundation model for bulk transcriptomes.bioRxiv, 2025

    Boming Kang, Rui Fan, Meizheng Yi, Chunmei Cui, and Qinghua Cui. A large-scale foundation model for bulk transcriptomes.bioRxiv, 2025. doi: 10.1101/2025.06.11.659222

    work page doi:10.1101/2025.06.11.659222 2025

  43. [43]

    Kaplan and Paul Meier

    Edward L. Kaplan and Paul Meier. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association, 53, 1958

    work page 1958

  44. [44]

    Katzman, Uri Shaham, Jonathan Bates, Tingting Jiang, and Yuval Kluger

    Jared L. Katzman, Uri Shaham, Jonathan Bates, Tingting Jiang, and Yuval Kluger. Deepsurv: personalized treatment recommender system using a cox proportional hazards deep neural network.BMC Medical Research Methodology, 18, 2018

    work page 2018

  45. [45]

    Tcga-reports: A machine-readable pathology report resource for benchmarking text-based ai models.Patterns, 5(3):100933, 2024

    Jenna Kefeli and Nicholas Tatonetti. Tcga-reports: A machine-readable pathology report resource for benchmarking text-based ai models.Patterns, 5(3):100933, 2024

    work page 2024

  46. [46]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In Yoshua Bengio and Yann LeCun, editors,ICLR (Poster), 2015

    work page 2015

  47. [47]

    Klein and Melvin L

    John P. Klein and Melvin L. Moeschberger.Survival Analysis: Techniques for Censored and Truncated Data. Springer, 1997

    work page 1997

  48. [48]

    Knaus, Frank E

    William A. Knaus, Frank E. Harrell Jr, Charles J. Fisher Jr, Douglas P. Wagner, Steven M. Opal, Jerald C. Sadoff, Elizabeth A. Draper, Cynthia A. Walawander, Kathleen Conboy, and Thaddeus H. Grasela. The clinical evaluation of new drugs for sepsis: A prospective study design based on survival analysis.JAMA, 270, 1993

    work page 1993

  49. [49]

    Econometric Society Monographs

    Roger Koenker.Quantile Regression. Econometric Society Monographs. Cambridge University Press, 2005

    work page 2005

  50. [50]

    Time-to-event prediction with neural networks and cox regression.Journal of Machine Learning Research, 20, 2019

    Håvard Kvamme, Ørnulf Borgan, and Ida Scheel. Time-to-event prediction with neural networks and cox regression.Journal of Machine Learning Research, 20, 2019

    work page 2019

  51. [51]

    Stephen W. Lagakos. General right censoring and its impact on the analysis of survival data. Biometrics, 35(1), 1979

    work page 1979

  52. [52]

    Nicholas R. Latimer. Survival analysis for economic evaluations alongside clinical trials– extrapolation with patient-level data: Inconsistencies, limitations, and a practical guide.Medical Decision Making, 33(6):743–754, 2013

    work page 2013

  53. [53]

    K. M. Leung, Robert M. Elashoff, and Abdelmonem A. Afifi. Censoring issues in survival analysis.Annual Review of Public Health, 1997. 12

    work page 1997

  54. [54]

    Hoadley, Laila M

    Jianfang Liu, Tara Lichtenberg, Katherine A. Hoadley, Laila M. Poisson, Alexander J. Lazar, and et al. An integrated tcga pan-cancer clinical data resource to drive high-quality survival outcome analytics.Cell, 173(2):400–416.e11, 2018

    work page 2018

  55. [55]

    Epithelioid hemangioendothelioma: Incidence, mortality, prognostic factors, and survival analysis using the surveillance, epidemiology, and end results database

    Zhen Liu and Shuting He. Epithelioid hemangioendothelioma: Incidence, mortality, prognostic factors, and survival analysis using the surveillance, epidemiology, and end results database. Journal of Oncology, 2022(1), 2022

    work page 2022

  56. [56]

    Data-efficient and weakly supervised computational pathology on whole-slide images.Nature Biomedical Engineering, 5(6):555–570, 2021

    Ming Y Lu, Drew FK Williamson, Tiffany Y Chen, Richard J Chen, Matteo Barbieri, and Faisal Mahmood. Data-efficient and weakly supervised computational pathology on whole-slide images.Nature Biomedical Engineering, 5(6):555–570, 2021

    work page 2021

  57. [57]

    Lu, Bowen Chen, Andrew Zhang, Drew F

    Ming Y . Lu, Bowen Chen, Andrew Zhang, Drew F. K. Williamson, Richard J. Chen, Tong Ding, Long Phi Le, Yung-Sung Chuang, and Faisal Mahmood. Visual language pretrained multiple instance zero-shot transfer for histopathology images. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19764–19775, June 2023

    work page 2023

  58. [58]

    Lu, Bowen Chen, Drew F

    Ming Y . Lu, Bowen Chen, Drew F. K. Williamson, Richard J. Chen, Ivy Liang, Tong Ding, Guillaume Jaume, Igor Odintsov, Long Phi Le, Georg Gerber, Anil V . Parwani, Andrew Zhang, and Faisal Mahmood. A visual-language foundation model for computational pathology.Nature Medicine, 30, 2024

    work page 2024

  59. [59]

    A visual-language foundation model for computational pathology.Nature Medicine, 30:863–874, 2024

    Ming Y Lu, Bowen Chen, Drew FK Williamson, Richard J Chen, Ivy Liang, Tong Ding, Guil- laume Jaume, Igor Odintsov, Long Phi Le, Georg Gerber, et al. A visual-language foundation model for computational pathology.Nature Medicine, 30:863–874, 2024

    work page 2024

  60. [60]

    Huiyan Luo, Jin Lu, Yuxian Bai, and et al. Effect of camrelizumab vs placebo added to chemotherapy on survival and progression-free survival in patients with advanced or metastatic esophageal squamous cell carcinoma: The escort-1st randomized clinical trial.JAMA, 326(10), 2021

    work page 2021

  61. [61]

    Federated cox proportional hazards model with multicentric privacy-preserving lasso feature selection for survival analysis from the perspective of personalized medicine

    Carlotta Masciocchi, Benedetta Gottardelli, Mariachiara Savino, Luca Boldrini, Antonella Martino, Ciro Mazzarella, Mariangela Massaccesi, Vincenzo Valentini, and Andrea Damiani. Federated cox proportional hazards model with multicentric privacy-preserving lasso feature selection for survival analysis from the perspective of personalized medicine. In2022 I...

    work page 2022

  62. [62]

    Revisiting the calibration of modern neural networks

    Matthias Minderer, Josip Djolonga, Rob Romijnders, Frances Hubis, Xiaohua Zhai, Neil Houlsby, Dustin Tran, and Mario Lucic. Revisiting the calibration of modern neural networks. InAdvances in Neural Information Processing Systems, volume 34, 2021

    work page 2021

  63. [63]

    Monotone least squares and isotonic quantiles

    Alexandre Mösching and Lutz Dümbgen. Monotone least squares and isotonic quantiles. Electronic Journal of Statistics, 14(1):24 – 49, 2020

    work page 2020

  64. [64]

    Susan Murphy, Anthony Rossini, and A. W. van der Vaart. Maximum likelihood estimation in the proportional odds model.Journal of the American Statistical Association, 1997

    work page 1997

  65. [65]

    Dusenberry, Linchuan Zhang, Ghassen Jerfel, and Dustin Tran

    Jeremy Nixon, Michael W. Dusenberry, Linchuan Zhang, Ghassen Jerfel, and Dustin Tran. Measuring calibration in deep learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2019

    work page 2019

  66. [66]

    Cronin, Sean F

    Anne-Michelle Noone, Kathleen A. Cronin, Sean F. Altekruse, Nadia Howlader, Denise R. Lewis, Valentina I. Petkov, and Lynne Penberthy. Cancer incidence and survival trends by subtype using data from the surveillance epidemiology and end results program, 1992–2013. Cancer Epidemiology, Biomarkers & Prevention, 26, 2017

    work page 1992

  67. [67]

    Machine failure prediction using survival analysis.Future Internet, 15(5), 2023

    Dimitris Papathanasiou, Konstantinos Demertzis, and Nikos Tziritas. Machine failure prediction using survival analysis.Future Internet, 15(5), 2023

    work page 2023

  68. [68]

    Toward conditional distribution calibration in survival prediction

    Shi-Ang Qi, Yakun Yu, and Russell Greiner. Toward conditional distribution calibration in survival prediction. InAdvances in Neural Information Processing Systems, volume 37, 2024. 13

    work page 2024

  69. [69]

    Conformalized survival distributions: A generic post-process to increase calibration

    Shi-Ang Qi, Yakun Yu, and Russell Greiner. Conformalized survival distributions: A generic post-process to increase calibration. InProceedings of the 41st International Conference on Machine Learning, volume 235, 2024

    work page 2024

  70. [70]

    Conformal predictive intervals in survival analysis: A resampling approach.Biometrics, 81, 2025

    Jing Qin, Jin Piao, Jing Ning, and Yu Shen. Conformal predictive intervals in survival analysis: A resampling approach.Biometrics, 81, 2025

    work page 2025

  71. [71]

    Med-bert: Pretrained contextu- alized embeddings on large-scale structured electronic health records for disease prediction.npj Digital Medicine, 4, 2021

    Laila Rasmy, Yang Xiang, Ziqian Xie, Cui Tao, and Degui Zhi. Med-bert: Pretrained contextu- alized embeddings on large-scale structured electronic health records for disease prediction.npj Digital Medicine, 4, 2021

    work page 2021

  72. [72]

    Robins and Andrea Rotnitzky

    James M. Robins and Andrea Rotnitzky. Recovery of information and adjustment for dependent censoring using surrogate markers. In Nicholas P. Jewell, Klaus Dietz, and Vern T. Farewell, editors,AIDS Epidemiology. Birkhäuser, 1992

    work page 1992

  73. [73]

    Estimation of machine setup and changeover times by survival analysis.Computers & Industrial Engineering, 153, 2021

    Tamas Ruppert, Robert Csalodi, and Janos Abonyi. Estimation of machine setup and changeover times by survival analysis.Computers & Industrial Engineering, 153, 2021

    work page 2021

  74. [74]

    Patrick Schober and Thomas R. Vetter. Survival analysis and interpretation of time-to-event data: The tortoise and the hare.Anesthesia & Analgesia, 127, 2018

    work page 2018

  75. [75]

    Conformal survival bands for risk screening under right- censoring

    Matteo Sesia and Vladimir Svetnik. Conformal survival bands for risk screening under right- censoring. InProceedings of the Fourteenth Symposium on Conformal and Probabilistic Prediction with Applications, volume 266, 2025

    work page 2025

  76. [76]

    Training-set conditionally valid prediction sets with right- censored data.CoRR, abs/2505.02213, 2025

    Wenwen Si and Hongxiang Qiu. Training-set conditionally valid prediction sets with right- censored data.CoRR, abs/2505.02213, 2025

    work page arXiv 2025

  77. [77]

    Kochan, Julian Trollor, and Henry Brodaty

    Annette Spooner, Emily Chen, Arcot Sowmya, Perminder Sachdev, Nicole A. Kochan, Julian Trollor, and Henry Brodaty. A comparison of machine learning methods for survival analysis of high-dimensional clinical data for dementia prediction.Scientific Reports, 10, 2020

    work page 2020

  78. [78]

    Bilmes, Tanmoy Bhattacharya, and Sarah Michalak

    Sunil Thulasidasan, Gopinath Chennupati, Jeff A. Bilmes, Tanmoy Bhattacharya, and Sarah Michalak. On mixup training: Improved calibration and predictive uncertainty for deep neural networks. InAdvances in Neural Information Processing Systems, volume 32, 2019

    work page 2019

  79. [79]

    Tsiatis.Semiparametric Theory and Missing Data

    Anastasios A. Tsiatis.Semiparametric Theory and Missing Data. Springer, 2006

    work page 2006

  80. [80]

    Molecular-driven foundation model for oncologic pathology,

    Anurag Vaidya, Andrew Zhang, Guillaume Jaume, Andrew H Song, Tong Ding, Sophia J Wagner, Ming Y Lu, Paul Doucet, Harry Robertson, Cristina Almagro-Perez, et al. Molecular- driven foundation model for oncologic pathology.arXiv preprint arXiv:2501.16652, 2025

    work page arXiv 2025

Showing first 80 references.