Accounting for Delayed Entry in Analyses of Overall Survival in Clinico-Genomic Databases

Backenroth, Daniel; Snider, Jeremy; Shen, Ronglai; Seshan, Venkatraman; Castellanos, Emily; McCusker, Margaret; Feuchtbaum, Dana; Gönen, Mithat; Sarkar, Somnath

doi:10.1158/1055-9965.epi-21-0876

Cited by 5 publications

(7 citation statements)

References 21 publications

(26 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These challenges are not unique to the GENIE BPC data set. 7 The results underscore the need for a comprehensive understanding of NGS testing patterns within the target populations to which researchers aim to generalize the results of clinicogenomic studies. A summary of these challenges and mitigation strategies is provided in Table 4.…”

Section: Discussionmentioning

confidence: 97%

“…In this context, risk set adjustment alone may not yield unbiased estimates of survival time and associations between biomarkers and survival. [5][6][7][8] When patients are added to the analytic cohort at a specific time because of increased mortality risk at that time, the core assumption of risk set adjustment that cohort entry and survival times are statistically quasi-independent 9 is violated. This phenomenon, often described as informative cohort entry, informative left truncation, or temporal selection bias, 6 can yield biased estimates of outcomes such as overall survival and biomarker hazard ratios (HRs) when compared with what would be observed if NGS were uncorrelated with clinical risk.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Biomarker Inference and the Timing of Next-Generation Sequencing in a Multi-Institutional, Cross-Cancer Clinicogenomic Data Set

Kehl,

Lavery,

Brown

et al. 2024

JCO Precis Oncol

View full text Add to dashboard Cite

PURPOSE Observational clinicogenomic data sets, consisting of tumor next-generation sequencing (NGS) data linked to clinical records, are commonly used for cancer research. However, in real-world practice, oncologists frequently request NGS in search of treatment options for progressive cancer. The extent and impact of this dynamic on analysis of clinicogenomic research data are not well understood. METHODS We analyzed clinicogenomic data for patients with non–small cell lung, colorectal, breast, prostate, pancreatic, or urothelial cancers in the American Association for Cancer Research Biopharmaceutical Consortium cohort. Associations between baseline and time-varying clinical characteristics and time from diagnosis to NGS were measured. To explore the impact of informative cohort entry on biomarker inference, statistical interactions between selected biomarkers and time to NGS with respect to overall survival were calculated. RESULTS Among 7,182 patients, time from diagnosis to NGS varied significantly by clinical factors, including cancer type, calendar year of sequencing, institution, and age and stage at diagnosis. NGS rates also varied significantly by dynamic clinical status variables; in an adjusted model, compared with patients with stable disease at any given time after diagnosis, patients with progressive disease by imaging or oncologist assessment had higher NGS rates (hazard ratio for NGS, 1.61 [95% CI, 1.45 to 1.78] and 2.32 [95% CI, 2.01 to 2.67], respectively). Statistical interactions between selected biomarkers and time to NGS with respect to survival, potentially indicating biased biomarker inference results, were explored. CONCLUSION To evaluate the appropriateness of a data set for a particular research question, it is crucial to measure associations between dynamic cancer status and the timing of NGS, as well as to evaluate interactions involving biomarkers of interest and NGS timing with respect to survival outcomes.

show abstract

Section: Discussionmentioning

confidence: 97%

Section: Introductionmentioning

confidence: 99%

Biomarker Inference and the Timing of Next-Generation Sequencing in a Multi-Institutional, Cross-Cancer Clinicogenomic Data Set

Kehl,

Lavery,

Brown

et al. 2024

JCO Precis Oncol

View full text Add to dashboard Cite

show abstract

“…This model incorporated the demographic, clinical, and genomic covariates used in the time-to-metastasis models (age, sex, race, ethnicity, smoking history, stage at diagnosis, and 10 total mutation/copy number alteration variables). Risk set adjustment 21 was not performed, since informative cohort entry has previously been demonstrated in clinico-genomic datasets 12 , 13 , and risk set adjustment could still yield biased results in the event of informative entry. Since this analysis was designed to specifically assess the effect of TP53 mutations on patient survival, no correction for multiple hypotheses was performed.…”

Section: Methodsmentioning

confidence: 99%

“…First, as our study retrospectively examined the effect of genomic alterations on patient outcome, differences in treatment or other factors associated with specific mutations (e.g., administration of targeted therapies to patients with EGFR mutations) made it difficult to isolate the effect of certain genomic changes. Additionally, our study is vulnerable to selection bias and to informative cohort entry 12 , 13 , since it only included patients who underwent primary tumor genomic sequencing, which is more likely to be performed in patients who later developed recurrent or progressive disease.…”

Section: Introductionmentioning

confidence: 99%

Genomic analysis of early-stage lung cancer reveals a role for TP53 mutations in distant metastasis

Egeren

Kohli²,

Catalano³

et al. 2022

Sci Rep

View full text Add to dashboard Cite

Patients with non-small cell lung cancer (NSCLC) who have distant metastases have a poor prognosis. To determine which genomic factors of the primary tumor are associated with metastasis, we analyzed data from 759 patients originally diagnosed with stage I–III NSCLC as part of the AACR Project GENIE Biopharma Collaborative consortium. We found that TP53 mutations were significantly associated with the development of new distant metastases. TP53 mutations were also more prevalent in patients with a history of smoking, suggesting that these patients may be at increased risk for distant metastasis. Our results suggest that additional investigation of the optimal management of patients with early-stage NSCLC harboring TP53 mutations at diagnosis is warranted in light of their higher likelihood of developing new distant metastases.

show abstract

“…Because the day of mortality was withheld, the day was imputed as the 15th of the same month or, if there was evidence of clinical activity after this, to the last calendar day of that month. To circumvent a left-truncation bias caused by eligibility potentially occurring after time zero, entry into the analysis was delayed until the day of earliest eligibility (the latest of their second visit or on their genetic report date) relative to the index date (risk-set adjustment) (19,20). For the weighted cohorts we used robust variance estimation.…”

Section: ### Generalizabilitymentioning

confidence: 99%

Extending Inferences from Sample to Target Populations: On the Generalizability of a Real-World Clinico-Genomic Database Non-Small Cell Lung Cancer Cohort

Thomas,

Collin,

Berrocal-Almanza

et al. 2023

Preprint

View full text Add to dashboard Cite

The representativeness of Real-World Data is generally assumed, but findings will rarely generalise to the target population when the potential outcomes under treatment are influenced by variables causative of selection into a study. Using a de-identified nationwide US Clinico-Genomic Database (CGDB) Non-Small Cell Lung Cancer (NSCLC) cohort as an example of collider bias, we assess its representativeness in relation to two target populations: a superset of all NSCLC patients in the Flatiron Health network and Surveillance, Epidemiology and End Results cancer registrations. Informed by causal Directed Acyclic Graphs, the CGDB cohort was weighted to be representative of the target populations and real-world overall survival (rwOS) was re-estimated. Despite Standardised Differences suggesting differences in individual covariates between sample and target populations, the conditional distributions of selection were alike and indices of generalizability were very high (≥ 0.96 on a proportional scale of 0–1). Estimates of rwOS in a population weighted to be representative did not differ from naive estimates in the unweighted cohort. The Tipton generalizability index provides a quantitative assessment of the generalizability of findings that can be used to determine the influence of selection biases.

show abstract

Accounting for Delayed Entry in Analyses of Overall Survival in Clinico-Genomic Databases

Abstract: Development and stock ownership in Johnson & Johnson. MM reports current employment by GRAIL, Inc. and stock ownership in Illumina. RS, VS, MG report no conflict of interests.

Cited by 5 publications

References 21 publications

Biomarker Inference and the Timing of Next-Generation Sequencing in a Multi-Institutional, Cross-Cancer Clinicogenomic Data Set

Biomarker Inference and the Timing of Next-Generation Sequencing in a Multi-Institutional, Cross-Cancer Clinicogenomic Data Set

Genomic analysis of early-stage lung cancer reveals a role for TP53 mutations in distant metastasis

Extending Inferences from Sample to Target Populations: On the Generalizability of a Real-World Clinico-Genomic Database Non-Small Cell Lung Cancer Cohort

Contact Info

Product

Resources

About