pith. sign in

arxiv: 2604.20625 · v1 · submitted 2026-04-22 · 📊 stat.ME · stat.AP

Dynamic Prediction of the Target Survival Time in Metastatic Solid Tumor Cancer Clinical Trials

Sidi Wang , Kelley Kidwell , Bo Huang , Satrajit Roychoudhury This is my paper

Pith reviewed 2026-05-09 23:47 UTC · model grok-4.3

classification 📊 stat.ME stat.AP
keywords dynamic predictionoverall survivalprogression-free survivaljoint modelingcancer clinical trialsrenal cell carcinomasurvival maturitymultivariate models
0
0 comments X

The pith

Multivariate joint models using disease progression data can predict when overall survival will mature in metastatic cancer trials.

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

The paper develops statistical models to forecast the specific time when overall survival data in metastatic cancer randomized trials will reach sufficient maturity for reliable analysis. It achieves this by creating multivariate joint models that link components of disease progression to survival outcomes and adapts three standard association models for predictive use. A reader would care because trials often rely on faster-to-measure progression endpoints while regulators require mature survival data, creating uncertainty in planning follow-up duration and approval timelines. The approach is tested on real data from an advanced renal cell carcinoma trial, showing how different amounts of progression information can inform these forecasts.

Core claim

We propose a multivariate joint modeling approach considering components of progression and OS and extend three models commonly used for association to be used for OS prediction. To the best of our knowledge, this is the first comprehensive statistical study exploring the prediction of OS using different levels of information on disease progression and illustrating these models using a real, complex dataset from an advanced renal cell carcinoma clinical trial.

What carries the argument

Multivariate joint modeling of progression events and overall survival to dynamically forecast the target maturity time for survival data.

If this is right

  • Trial planners can use progression data to set more accurate follow-up durations and reduce unnecessary extensions.
  • Regulatory submissions for cancer drugs can rely on earlier, better-informed projections of when survival data will support inference.
  • Different levels of progression information, from basic to detailed, can be leveraged to refine the maturity predictions.
  • The adapted association models provide a flexible framework applicable across multiple metastatic solid tumor settings.

Where Pith is reading between the lines

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

  • The models could support interim monitoring systems that update maturity forecasts as new progression events arrive during an ongoing trial.
  • Similar joint modeling structures might address endpoint prediction challenges in non-cancer areas that track both surrogate and definitive outcomes.

Load-bearing premise

The patterns linking progression events to survival times seen in the training data will hold sufficiently in new trials to yield accurate maturity forecasts.

What would settle it

An independent dataset from a similar metastatic cancer trial in which the predicted target survival maturity times deviate substantially from the actual observed times when enough survival events have occurred.

Figures

Figures reproduced from arXiv: 2604.20625 by Bo Huang, Kelley Kidwell, Satrajit Roychoudhury, Sidi Wang.

Figure 1
Figure 1. Figure 1: Evolution of progression and death data during trial [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Multivariate joint modeling structure. where Yi(t) and σϵ are the target lesion measurement for subject i at time t and variability associated with it. Moreover, the lesion measurement process µ(t) is modeled by the linear model µi(t) = Xi(t) ′βµ + Wi(t), Wi(t) = b1i + b2it. Here, Xi(t) refers to a time-dependent covariate matrix for subject i’s lesion measurement. Finally, Wi(t) is a Gaussian process with… view at source ↗
Figure 3
Figure 3. Figure 3: (a) The scatterplot between time to progression (TTP) and overall survival (OS), (b) [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The bias (a) and root mean squared error (rMSE) (b) of the last (400th) death date [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: (a) Boxplots of the predicted number of deaths at primary analysis when 50 or 100 deaths are observed in the trial, with the true number of deaths being “146".(b) Boxplots of the posterior distributions of the last (341st) death date by baveJM, SPJM, the copula model between TTP and OS, the marginal Weibull baseline hazard model of OS, and the multi-state model. The date “2020-04-19" is the true last (341s… view at source ↗
Figure 6
Figure 6. Figure 6: Kaplan-Meier plots of the observed and predicted overall survival by baveJM (a), [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
read the original abstract

Overall survival (OS) is the gold standard for assessing patient benefit and cost-effectiveness of new cancer drugs. However, it is often difficult to use OS as the primary endpoint in randomized clinical trials (RCTs) for patients with metastatic cancer due to multiple reasons. In recent years, progression-free survival (PFS) has increasingly been used as the primary endpoint in metastatic cancer RCTs to accelerate development. However, regulatory authorities often seek mature OS data for approval. Therefore, it is critical to determine the target time when OS data are expected to be mature for reliable statistical inference. Motivated by an advanced renal cell carcinoma (RCC) clinical trial, we develop and investigate different prediction models leveraging information from disease progression to improve target OS prediction times. We propose a multivariate joint modeling approach considering components of progression and OS and extend three models commonly used for association to be used for OS prediction. To the best of our knowledge, this is the first comprehensive statistical study exploring the prediction of OS using different levels of information on disease progression and illustrating these models using a real, complex dataset. Our findings have significant implications for OS prediction.

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 manuscript proposes a multivariate joint modeling approach to dynamically predict the target time at which overall survival (OS) data reach maturity in metastatic solid tumor cancer trials. It extends three standard association models (shared random effects, current value, and slope) to incorporate components of progression-free survival (PFS) and other progression information, illustrates the framework on data from a single advanced renal cell carcinoma (RCC) trial, and claims this is the first comprehensive statistical study of OS prediction using varying levels of progression data.

Significance. If the models prove accurate and transportable, the work could help trialists estimate OS maturity timing more reliably, supporting faster yet statistically sound regulatory decisions in oncology. The joint-modeling extension for dynamic prediction is a methodological contribution, and the use of a real, complex clinical dataset is a positive feature.

major comments (2)
  1. [§3] §3 (Results): No quantitative prediction performance metrics (e.g., mean absolute error, calibration, or time-dependent AUC for the predicted target OS maturity time), cross-validation results, or direct comparisons against simpler baselines (PFS-only or covariate-only predictors) are reported, preventing assessment of whether the joint-model extensions improve accuracy.
  2. [§4] §4 (Discussion): All model fitting, association estimation, and (internal) evaluation occur within one advanced RCC trial. The central claim that the learned progression-OS associations can forecast OS maturity in future trials therefore rests on an untested generalizability assumption; no external validation on independent trials or other solid-tumor populations is presented.
minor comments (2)
  1. Notation for the extended association structures and the precise mapping from joint-model parameters to the predicted target OS time could be made more explicit, perhaps with a dedicated equation or algorithmic box.
  2. A few references to recent dynamic-prediction literature in joint models appear to be missing; adding them would better situate the extensions.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each of the major comments below.

read point-by-point responses

  1. Referee: [§3] §3 (Results): No quantitative prediction performance metrics (e.g., mean absolute error, calibration, or time-dependent AUC for the predicted target OS maturity time), cross-validation results, or direct comparisons against simpler baselines (PFS-only or covariate-only predictors) are reported, preventing assessment of whether the joint-model extensions improve accuracy.

    Authors: We agree that including quantitative performance metrics would allow for a more rigorous assessment of the models' predictive accuracy. In the revised version of the manuscript, we will add evaluations using mean absolute error, calibration measures, and time-dependent AUC for the predicted target OS maturity time. We will also include cross-validation results and direct comparisons to simpler baselines, such as PFS-only or covariate-only predictors, to quantify the benefit of incorporating the joint modeling of progression and survival data. revision: yes

  2. Referee: [§4] §4 (Discussion): All model fitting, association estimation, and (internal) evaluation occur within one advanced RCC trial. The central claim that the learned progression-OS associations can forecast OS maturity in future trials therefore rests on an untested generalizability assumption; no external validation on independent trials or other solid-tumor populations is presented.

    Authors: The analysis is indeed conducted within a single trial, as the manuscript uses this as a case study to illustrate the proposed multivariate joint modeling framework. We do not claim that the specific associations learned are directly transportable without further validation; rather, the contribution lies in the methodological extension and the exploration of different association structures. We will update the Discussion section to more clearly articulate this limitation and the assumptions involved in applying the models to future trials. External validation on independent datasets is not feasible with the current resources but represents an important direction for subsequent research. revision: partial

standing simulated objections not resolved
  • External validation using data from independent trials or other solid tumor populations, as such data are not available to the authors.

Circularity Check

0 steps flagged

No circularity in the statistical modeling derivation

full rationale

The paper develops multivariate joint models for predicting target OS maturity time by extending standard association models with progression components. The prediction task is defined externally based on trial design needs, and the models are constructed from established joint modeling techniques rather than being tautological with the fitted values. No self-definitional steps, fitted inputs called predictions, or load-bearing self-citations are present in the described approach. The central claim of improved prediction using different levels of progression information stands independently of the specific RCC dataset used for illustration.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard survival-analysis assumptions for joint modeling of time-to-event data; no new entities are postulated and the free parameters are the usual association and baseline hazard parameters estimated from data.

free parameters (1)
  • association parameters between PFS and OS components
    These parameters are estimated from the observed data to capture the dependence between progression and survival.
axioms (1)
  • domain assumption The joint distribution of progression and survival times can be adequately described by a multivariate model with a specified association structure.
    Invoked when extending standard association models to the prediction task.

pith-pipeline@v0.9.0 · 5504 in / 1236 out tokens · 71309 ms · 2026-05-09T23:47:39.681881+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages

  1. [1]

    European Journal of Cancer , volume=

    New Response Evaluation Criteria in Solid Tumours: Revised RECIST Guideline (Version 1.1) , author=. European Journal of Cancer , volume=. 2009 , publisher=

    work page 2009

  2. [2]

    Biometrika , volume=

    A Model for Association in Bivariate Life Tables and Its Application in Epidemiological Studies of Familial Tendency in Chronic Disease Incidence , author=. Biometrika , volume=. 1978 , publisher=

    work page 1978

  3. [3]

    Statistical Methodology , volume=

    Model Weights for Model Choice and Averaging , author=. Statistical Methodology , volume=. 2007 , publisher=

    work page 2007

  4. [4]

    New England Journal of Medicine , volume=

    Avelumab Plus Axitinib Versus Sunitinib for Advanced Renal-Cell Carcinoma , author=. New England Journal of Medicine , volume=. 2019 , publisher=

    work page 2019

  5. [5]

    Journal of Clinical Oncology , volume=

    Surrogate End Points for Median Overall Survival in Metastatic Colorectal Cancer: Literature-Based Analysis From 39 Randomized Controlled Trials of First-Line Chemotherapy , author=. Journal of Clinical Oncology , volume=. 2007 , publisher=

    work page 2007

  6. [6]

    The Cancer Journal , volume=

    Overall Survival: Still the Gold Standard: Why Overall Survival Remains the Definitive End Point in Cancer Clinical Trials , author=. The Cancer Journal , volume=. 2009 , publisher=

    work page 2009

  7. [7]

    BMC Cancer , volume=

    Surrogate Endpoints for Overall Survival in Digestive Oncology Trials: Which Candidates? A Questionnaires Survey Among Clinicians and Methodologists , author=. BMC Cancer , volume=. 2010 , publisher=

    work page 2010

  8. [8]

    Journal of Oncology Practice , volume=

    Impact of Accuracy of Survival Predictions on Quality of End-Of-Life Care Among Patients With Metastatic Cancer Who Receive Radiation Therapy , author=. Journal of Oncology Practice , volume=. 2019 , publisher=

    work page 2019

  9. [9]

    Journal of Clinical Epidemiology , volume=

    Measuring the Accuracy of Prognostic Judgments in Oncology , author=. Journal of Clinical Epidemiology , volume=. 1997 , publisher=

    work page 1997

  10. [10]

    Statistics in Medicine , volume=

    Accuracy of Point Predictions in Survival Analysis , author=. Statistics in Medicine , volume=. 2001 , publisher=

    work page 2001

  11. [11]

    Statistics in Medicine , volume=

    A Statistical Model for the Dependence Between Progression-Free Survival and Overall Survival , author=. Statistics in Medicine , volume=. 2009 , publisher=

    work page 2009

  12. [12]

    JNCI: Journal of the National Cancer Institute , volume=

    Detecting an Overall Survival Benefit That Is Derived From Progression-Free Survival , author=. JNCI: Journal of the National Cancer Institute , volume=. 2009 , publisher=

    work page 2009

  13. [13]

    Journal of the National Cancer Institute , volume=

    Detecting Overall Survival Benefit Derived From Survival Postprogression Rather Than Progression-Free Survival , author=. Journal of the National Cancer Institute , volume=. 2015 , publisher=

    work page 2015

  14. [14]

    JAMA Network Open , volume=

    Assessment of Progression-Free Survival as a Surrogate End Point of Overall Survival in First-Line Treatment of Ovarian Cancer: A Systematic Review and Meta-Analysis , author=. JAMA Network Open , volume=. 2020 , publisher=

    work page 2020

  15. [15]

    Statistics in Medicine , volume=

    Joint Modeling of Progression-Free and Overall Survival and Computation of Correlation Measures , author=. Statistics in Medicine , volume=. 2019 , publisher=

    work page 2019

  16. [16]

    Pharmacoeconomics , volume=

    Surrogate Endpoints in Health Technology Assessment: An International Review of Methodological Guidelines , author=. Pharmacoeconomics , volume=. 2020 , publisher=

    work page 2020

  17. [17]

    The AAPS Journal , volume=

    Predicting Overall Survival and Progression-Free Survival Using Tumor Dynamics in Advanced Breast Cancer Patients , author=. The AAPS Journal , volume=. 2019 , publisher=

    work page 2019

  18. [18]

    Journal of Thoracic Oncology , volume=

    Relationship Between Overall Survival and Response or Progression-Free Survival in Advanced Non--Small Cell Lung Cancer Patients Treated With Anti--Pd-1/Pd-L1 Antibodies , author=. Journal of Thoracic Oncology , volume=. 2016 , publisher=

    work page 2016

  19. [19]

    Clinical Pharmacology & Therapeutics , volume=

    Evaluation of Tumor Size Response Metrics to Predict Survival in Oncology Clinical Trials , author=. Clinical Pharmacology & Therapeutics , volume=. 2014 , publisher=

    work page 2014

  20. [20]

    Clinical Pharmacology & Therapeutics , volume=

    Elucidation of Relationship Between Tumor Size and Survival in Non-small-Cell Lung Cancer Patients Can Aid Early Decision Making in Clinical Drug Development , author=. Clinical Pharmacology & Therapeutics , volume=. 2009 , publisher=

    work page 2009

  21. [21]

    Journal of Clinical Oncology , volume=

    Evaluation of Tumor-Size Response Metrics to Predict Overall Survival in Western and Chinese Patients With First-Line Metastatic Colorectal Cancer , author=. Journal of Clinical Oncology , volume=. 2013 , publisher=

    work page 2013

  22. [22]

    Journal of Clinical Oncology , volume=

    Model-Based Prediction of Phase III Overall Survival in Colorectal Cancer on the Basis of Phase II Tumor Dynamics , author=. Journal of Clinical Oncology , volume=. 2009 , publisher=

    work page 2009

  23. [23]

    Cpt: Pharmacometrics & Systems Pharmacology , volume=

    A New Method to Model and Predict Progression Free Survival Based on Tumor Growth Dynamics , author=. Cpt: Pharmacometrics & Systems Pharmacology , volume=. 2020 , publisher=

    work page 2020

  24. [24]

    British Journal of Clinical Pharmacology , volume=

    Models for Change in Tumour Size, Appearance of New Lesions and Survival Probability in Patients With Advanced Epithelial Ovarian Cancer , author=. British Journal of Clinical Pharmacology , volume=. 2016 , publisher=

    work page 2016

  25. [25]

    Clyde, David Draper and Ei George, and a Rejoinder by the Authors , author=

    Bayesian Model Averaging: A Tutorial (With Comments by M. Clyde, David Draper and Ei George, and a Rejoinder by the Authors , author=. Statistical Science , volume=. 1999 , publisher=

    work page 1999

  26. [26]

    2020 , publisher=

    Statistical Models to Capture the Association Between Progression-Free and Overall Survival in Oncology Trials , author=. 2020 , publisher=

    work page 2020

  27. [27]

    JAMA Oncology , volume=

    Interpretability of Cancer Clinical Trial Results Using Restricted Mean Survival Time as an Alternative to the Hazard Ratio , author=. JAMA Oncology , volume=. 2017 , publisher=

    work page 2017

  28. [28]

    2022 , note =

    survRM2: Comparing Restricted Mean Survival Time , author =. 2022 , note =

    work page 2022

  29. [29]

    Fonctions de r

    Sklar, M , booktitle=. Fonctions de r

    work page

  30. [30]

    Statistical Methods in Medical Research , volume=

    Multi-State Models for Event History Analysis , author=. Statistical Methods in Medical Research , volume=. 2002 , publisher=

    work page 2002

  31. [31]

    Statistics in Medicine , volume=

    Tutorial in Biostatistics: Competing Risks and Multi-State Models , author=. Statistics in Medicine , volume=. 2007 , publisher=

    work page 2007

  32. [32]

    1977 , publisher=

    The Theory of Stochastic Processes , author=. 1977 , publisher=

    work page 1977

  33. [33]

    Statistical Methods in Medical Research , volume=

    Multi-State Models for the Analysis of Time-To-Event Data , author=. Statistical Methods in Medical Research , volume=. 2009 , publisher=

    work page 2009

  34. [34]

    2) , author=

    Bayesian Data Analysis (Vol. 2) , author=. 2014 , publisher=

    work page 2014

  35. [35]

    0 User Manual [Computer Software Manual] , author=

    Jags Version 4.3. 0 User Manual [Computer Software Manual] , author=. Retrieved from Sourceforge. net/projects/mcmc-jags/files/Manuals/4. x , volume=

    work page

  36. [36]

    2022 , note =

    rjags: Bayesian Graphical Models using MCMC , author =. 2022 , note =

    work page 2022

  37. [37]

    Endpoints News , year =

    Zachary Brennan , title =. Endpoints News , year =

    work page

  38. [38]

    European Urology , volume=

    Survival Prediction in Everolimus-Treated Patients With Metastatic Renal Cell Carcinoma Incorporating Tumor Burden Response in the Record-1 Trial , author=. European Urology , volume=. 2013 , publisher=

    work page 2013

  39. [39]

    Statistics in Medicine , volume=

    Quantifying the Association Between Progression-Free Survival and Overall Survival in Oncology Trials Using Kendall’s , author=. Statistics in Medicine , volume=. 2019 , publisher=

    work page 2019

  40. [40]

    Statistics in Medicine , volume=

    Joint Modeling of Progression-Free Survival and Overall Survival by a Bayesian Normal Induced Copula Estimation Model , author=. Statistics in Medicine , volume=. 2013 , publisher=

    work page 2013

  41. [41]

    Cancer , volume=

    Progression-Free Survival as a Predictor of Overall Survival in Metastatic Renal Cell Carcinoma Treated With Contemporary Targeted Therapy , author=. Cancer , volume=. 2011 , publisher=

    work page 2011

  42. [42]

    Trials , volume=

    How Is Overall Survival Assessed in Randomised Clinical Trials in Cancer and Are Subsequent Treatment Lines Considered? A Systematic Review , author=. Trials , volume=. 2023 , publisher=

    work page 2023

  43. [43]

    2011 , publisher=

    NICE DSU Technical Support Document 2: A Generalised Linear Modelling Framework for Pairwise and Network Meta-Analysis of Randomised Controlled Trials , author=. 2011 , publisher=

    work page 2011

  44. [44]

    Report by the Decision Support Unit , year=

    NICE DSU Technical Support Document 14: Survival Analysis for Economic Evaluations Alongside Clinical Trials-Extrapolation With Patient-Level Data , author=. Report by the Decision Support Unit , year=

    work page

  45. [45]

    2006 , publisher=

    Guidelines for the Economic Evaluation of Health Technologies: Canada , author=. 2006 , publisher=

    work page 2006

  46. [46]

    Health Do, editor

    Guidelines for Preparing Submissions to the Pharmaceutical Benefits Advisory Committee (PBAC) , author=. Health Do, editor. Canberra2016 , year=

    work page

  47. [47]

    Journal of Comparative Effectiveness Research , volume=

    Novel and Existing Flexible Survival Methods for Network Meta-Analyses , author=. Journal of Comparative Effectiveness Research , volume=. 2022 , publisher=

    work page 2022

  48. [48]

    Statistics in medicine , volume=

    Shared parameter models for the joint analysis of longitudinal data and event times , author=. Statistics in medicine , volume=. 2006 , publisher=

    work page 2006

  49. [49]

    Report of the DIA Bayesian joint modeling working group , author=

    Joint modeling of survival and longitudinal non-survival data: current methods and issues. Report of the DIA Bayesian joint modeling working group , author=. Statistics in medicine , volume=. 2015 , publisher=

    work page 2015

  50. [50]

    Biometrics , volume=

    Joint modeling of longitudinal and survival data via a common frailty , author=. Biometrics , volume=. 2004 , publisher=

    work page 2004