Evaluation of the Use of Cancer Registry Data for Comparative Effectiveness Research

Evaluating generalizability of oncology trial results to real-world patients using machine learning-based trial emulations

Randomized controlled trials (RCTs) evaluating anti-cancer agents often lack generalizability to real-world oncology patients. Although restrictive eligibility criteria contribute to this issue, the role of selection bias related to prognostic risk remains unclear. In this study, we developed TrialTranslator, a framework designed to systematically evaluate the generalizability of RCTs for oncology therapies. Using a nationwide database of electronic health records from Flatiron Health, this framework emulates RCTs across three prognostic phenotypes identified through machine learning models. We applied this approach to 11 landmark RCTs that investigated anti-cancer regimens considered standard of care for the four most prevalent advanced solid malignancies. Our analyses reveal that patients in low-risk and medium-risk phenotypes exhibit survival times and treatment-associated survival benefits similar to those observed in RCTs. In contrast, high-risk phenotypes show significantly lower survival times and treatment-associated survival benefits compared to RCTs. Our results were corroborated by a comprehensive robustness assessment, including examinations of specific patient subgroups, holdout validation and semi-synthetic data simulation. These findings suggest that the prognostic heterogeneity among real-world oncology patients plays a substantial role in the limited generalizability of RCT results. Machine learning frameworks may facilitate individual patient-level decision support and estimation of real-world treatment benefits to guide trial design.

Comparison of incident breast cancer cases in the largest national US tumor registries

BACKGROUND

This study compared incident breast cancer cases in the National Cancer Database (NCDB) and Surveillance, Epidemiology, End Results Program (SEER) to a national population cancer registry.

METHODS

Patients with malignant or in situ breast cancer (BC) 2010-2019 in the NCDB and SEER were compared to the US Cancer Statistics (USCS). Case coverage was estimated as the number of patients in the NCDB/SEER as a proportion of USCS cases.

RESULTS

The USCS reported 3,047,509 patients; 77.5% patients were included in the NCDB and 46.0% in SEER. Case ascertainment varied significantly by patient sex (both registries, p < .001). For males, 84.1% were captured in the NCDB, whereas only 77.5% of females were included. Case coverage in SEER was better for females than males (46.1% vs. 43.5%). Registries varied significantly by race/ethnicity (both p < .001). Case coverage in the NCDB was highest for non-Hispanic White (78.2%), non-Hispanic Black (77.7%), and non-Hispanic Asian or Pacific Islander (72.5%) BC patients, and lowest for Hispanic (56.4%) and non-Hispanic American Indian/Alaska Native (41.1%) patients. In SEER, case coverage was highest for non-Hispanic Asian or Pacific Islander (78.1%) and Hispanic (69.6%) patients and it was significantly lower for all other subgroups (non-Hispanic Black, 44.8%; non-Hispanic White, 42.4%; and non-Hispanic American Indian/Alaska Native, 36.6%).

CONCLUSIONS

National US tumor registries provide data for a large sampling of breast cancer patients, yet significant differences in case coverage were observed based on age, sex, and race/ethnicity. These findings suggest that analyses using these data sets and interpretation of findings should account for these meaningful variances.

Real-world data: bridging the gap between clinical trials and practice

Real-world data (RWD) are rapidly emerging sources of information for patients, clinicians and regulators. While randomized controlled trials (RCTs) reduce bias and confounding through the randomization process and provide the highest quality of evidence regarding drug efficacy, RCTs may be impractical or unfeasible for rare diseases or disease subsets. And yet, studies attempting to replicate clinical trial results using observational datasets have failed. Given the inherent differences between observational data and clinical trial results, this discordance is not surprising. However, RWD may still have independent value as complementary tools to trial results. In this viewpoint, we explore the challenges of RWD and discuss key questions that clinicians, patients, and regulators will need to consider when faced with positive efficacy data from clinical trials, and negative effectiveness data from real world studies. Finally, we explore novel trial designs that might help bridge the gap from RCTs to RWD.

Assessing the real-world effectiveness of 8 major metastatic breast cancer drugs using target trial emulation

BACKGROUND

Demonstration of trial emulation ability to benchmark randomised controlled trials (RCTs) from real-world data (RWD) is required to increase confidence in the use of routinely collected data for decision making in oncology.

METHODS

To assess the frequency with which emulation findings align with RCTs regarding effect size on overall survival (OS) in metastatic breast cancer (MBC), 8 of 13 pre-selected pivotal RCTs in MBC were emulated using data from 32,598 patients enrolled in the French ESME-MBC cohort between January 1, 2008 and December 31, 2021. Adjustment methods and confounders were selected a priori for each emulation; stabilized weight was the reference method to mitigate confounding. Concordance in OS hazard ratios with associated 95 % confidence intervals between RCTs and emulations were assessed used predefined metrics based on statistical significance, estimates, and standardized differences.

RESULTS

The effect sizes were consistent with RCT results in 7 out of the 8 emulations; 4 emulations achieved full statistical significance agreement; 5 emulations had a point estimate included in the RCT CI (estimate agreement); 6 emulations reported no significant differences between RCT and emulation (standardized difference agreement). Discrepancies related to residual confounders and significant shifts in prescription practices post-drug approval may arise in some cases.

CONCLUSION

Target trial emulation from RWD combined with appropriate adjustment can provide conclusions similar to RCTs in MBC. In oncology, this methodology offers opportunities for confirming the impact on long-term survival, for expanding indications in patients excluded from RCTs and for comparative effectiveness in single-arm trials using external control arms.

Patients with stage IV colorectal carcinoma selected for palliative primary tumor resection and systemic therapy survive longer compared with systemic therapy alone: a retrospective comparative cohort study

OBJECTIVE

To compare the survival of palliative stage IV colorectal cancer patients selected for primary tumor resection and systemic treatment (PTR+SYST) to patients with systemic treatment only (SYST).

BACKGROUND

About 20-25% of all colorectal cancer patients are diagnosed with stage IV disease. The benefit of primary tumor resection in the palliative situation is therefore of high concern. However, empirical evidence from randomized and observational studies is inconsistent.

METHODS

Mortality after PTR+SYST was compared to systemic treatment alone in a retrospective observational cohort of patients diagnosed 2012-2020 from the cancer registry in the federal state of Brandenburg (Germany), excluding patients with rectal cancer of the lower two-thirds, emergency procedures, unknown ECOG status, ECOG greater than 2, unknown metastatic status or unclear grading.

RESULTS

Of 480 patients, 416 died during an average follow-up of 23 months in mean. Twelve-month survival was 75% after PTR+SYST compared with 49% after SYST [hazard ratio (HR)=0.39, 95% CI 0.29-0.53, P <0.001]. The difference persisted to 36 months (28% vs. 13%, HR=0.53, 95% CI 0.43-0.66, P <0.001). Results were similar after multivariate adjustment, propensity score matching and delayed entry.

CONCLUSION AND RELEVANCE

Patients with stage IV colorectal carcinoma who are selected for primary tumor resection in combination with systemic therapy and who receive such treatment survive longer compared with patients who receive only systemic treatment. Whether the difference is due to the selection of patients or PTR remains unclear. At present, the current practice of selecting patients for PTR appears to do no harm.

Publication of observational studies making claims of causation over time

To examine methodology characteristics over time and investigate research impact before and after the start of the COVID-19 era, we analyzed original articles published in The New England Journal of Medicine between October 26, 2017 and August 27, 2022. April 1, 2020 was used as the defining date dividing before and after the COVID-19 era. Out of 1051 original articles, 515 (49 %) were before and 536 (51 %) were after the COVID-19 era. Two independent reviewers categorized and reconciled methodology into groups: "randomized trial" (715 articles), "uncontrolled experimental study" (128), "descriptive observational study" (168), and "observational study making a causal claim" (40). We extracted subsequent citations and Altmetric data for each article to assess impact. The median number of social media shares was 2272 (IQR: 743-7821) for observational studies making a causal conclusion, compared to 306 (IQR: 70-606) for randomized trials (p-value=<0.001). The median Altmetric score for randomized COVID-19 trials (2421, IQR: 1063-3920) was not significantly different than that of COVID-19 observational studies making a causal claim (2583, IQR: 1513-6197, p-value = 0.42), but it was significantly lower than descriptive observational COVID-19 studies (4093, IQR: 2545-6823, p-value = 0.04). We conclude that there has been a steady increase in the number and percentage of observational studies that make causal conclusions about the efficacy of an intervention. Research concerning COVID-19, regardless of methodology, has seen a sharp rise in dissemination as measured through Altmetric's social media score and subsequent citations.

Ethical Problems of Observational Studies and Big Data Compared to Randomized Trials

The temptation to use prospective observational studies (POS) instead of conducting difficult trials (RCTs) has always existed, but with the advent of powerful computers and large databases, it can become almost irresistible. We examine the potential consequences, were this to occur, by comparing two hypothetical studies of a new treatment: one RCT, and one POS. The POS inevitably submits more patients to inferior research methodology. In RCTs, patients are clearly informed of the research context, and 1:1 randomized allocation between experimental and validated treatment balances risks for each patient. In POS, for each patient, the risks of receiving inferior treatment are impossible to estimate. The research context and the uncertainty are down-played, and patients and clinicians are at risk of becoming passive research subjects in studies performed from an outsider's view, which potentially has extraneous objectives, and is conducted without their explicit, autonomous, and voluntary involvement and consent.

Adjuvant Ovarian Function Suppression in Premenopausal Hormone Receptor-Positive Breast Cancer

Importance

Few oncology studies have assessed the effectiveness of adjuvant ovarian function suppression (OFS) in observational settings for premenopausal hormone receptor-positive breast cancer. Target trial emulation is increasingly used for estimating treatment outcomes in observational cohorts.

Objectives

To describe hormone therapy and OFS treatment patterns (aim 1), examine the association between adding OFS to tamoxifen (TAM) or aromatase inhibitor (AI) and survival (aim 2), and examine the association between duration of hormone treatment (TAM or AI) plus OFS (H-OFS) and survival (aim 3).

Design, Setting, and Participants

This population-based cohort study included all premenopausal, early-stage breast cancer diagnoses between 2010 and 2020 in Alberta, Canada. Target trial emulation was conducted. Eligibility criteria were directly modeled after the Suppression of Ovarian Function Trial (SOFT) and Tamoxifen and Exemestane Trial (TEXT). Participants were followed up for a maximum of 5 years. Data were analyzed from July 2022 through March 2023.

Exposures

For aim 2, exposures were receiving the following baseline treatments for 2 years: AI + OFS (AI-OFS), TAM + OFS (T-OFS), and TAM alone. For aim 3, exposures were a 2-year or greater and a less than 2-year duration of H-OFS.

Main Outcomes and Measures

Recurrence-free survival was the primary outcome of interest. Marginal structural Cox models with inverse probability treatment and censoring weights were used to estimate hazard ratios (HRs), adjusted for baseline and time-varying confounding variables.

Results

Among 3434 female patients with premenopausal, early-stage breast cancer diagnoses (median [IQR] age, 45 [40-48] years), 2647 individuals satisfied SOFT and TEXT eligibility criteria. There were 2260 patients who initiated TAM, 232 patients who initiated T-OFS, and 155 patients who initiated AI-OFS; 192 patients received H-OFS for 2 or more years, and 195 patients received H-OFS for less than 2 years. The 5-year recurrence risks were not significantly lower in AI-OFS vs TAM (HR, 0.76; 95% CI, 0.38-1.33) or T-OFS vs TAM (HR, 0.87; 95% CI, 0.50-1.45) groups. Patients receiving H-OFS for 2 or more years had significantly better 5-year recurrence-free survival compared with those receiving H-OFS for less than 2 years (HR, 0.69; 95% CI, 0.54-0.90).

Conclusions and Relevance

This study found no significant reductions in recurrence risk for AI-OFS and T-OFS compared with TAM alone. H-OFS duration for at least 2 years was associated with significantly improved recurrence-free survival.

Design differences and variation in results between randomised trials and non-randomised emulations: meta-analysis of RCT-DUPLICATE data

Objective

To explore how design emulation and population differences relate to variation in results between randomised controlled trials (RCT) and non-randomised real world evidence (RWE) studies, based on the RCT-DUPLICATE initiative (Randomised, Controlled Trials Duplicated Using Prospective Longitudinal Insurance Claims: Applying Techniques of Epidemiology).

Design

Meta-analysis of RCT-DUPLICATE data.

Data sources

Trials included in RCT-DUPLICATE, a demonstration project that emulated 32 randomised controlled trials using three real world data sources: Optum Clinformatics Data Mart, 2004-19; IBM MarketScan, 2003-17; and subsets of Medicare parts A, B, and D, 2009-17.

Eligibility criteria for selecting studies

Trials where the primary analysis resulted in a hazard ratio; 29 RCT-RWE study pairs from RCT-DUPLICATE.

Results

Differences and variation in effect sizes between the results from randomised controlled trials and real world evidence studies were investigated. Most of the heterogeneity in effect estimates between the RCT-RWE study pairs in this sample could be explained by three emulation differences in the meta-regression model: treatment started in hospital (which does not appear in health insurance claims data), discontinuation of some baseline treatments at randomisation (which would have been an unusual care decision in clinical practice), and delayed onset of drug effects (which would be under-reported in real world clinical practice because of the relatively short persistence of the treatment). Adding the three emulation differences to the meta-regression reduced heterogeneity from 1.9 to almost 1 (absence of heterogeneity).

Conclusions

This analysis suggests that a substantial proportion of the observed variation between results from randomised controlled trials and real world evidence studies can be attributed to differences in design emulation.

Practice Patterns for Organ Preservation in US Patients With Rectal Cancer, 2006-2020

Importance

In March 2023, the National Comprehensive Cancer Network endorsed watch and wait for those with complete clinical response to total neoadjuvant therapy. Neoadjuvant therapy is highly efficacious, so this recommendation may have broad implications, but the current trends in organ preservation in the US are unknown.

Objective

To describe organ preservation trends among patients with rectal cancer in the US from 2006 to 2020.

Design, Setting, and Participants

This retrospective, observational case series included adults (aged ≥18 years) with rectal adenocarcinoma managed with curative intent from 2006 to 2020 in the National Cancer Database.

Exposure

The year of treatment was the primary exposure. The type of therapy was chemotherapy, radiation, or surgery (proctectomy, transanal local excision, no tumor resection). The timing of therapy was classified as neoadjuvant or adjuvant.

Main Outcomes and Measures

The primary outcome was the absolute annual proportion of organ preservation after radical treatment, defined as chemotherapy and/or radiation without tumor resection, proctectomy, or transanal local excision. A secondary analysis examined complete pathologic responses among eligible patients.

Results

Of the 175 545 patients included, the mean (SD) age was 63 (13) years, 39.7% were female, 17.4% had clinical stage I disease, 24.7% had stage IIA to IIC disease, 32.1% had stage IIIA to IIIC disease, and 25.7% had unknown stage. The absolute annual proportion of organ preservation increased by 9.8 percentage points (from 18.4% in 2006 to 28.2% in 2020; P < .001). From 2006 to 2020, the absolute rate of organ preservation increased by 13.0 percentage points for patients with stage IIA to IIC disease (19.5% to 32.5%), 12.9 percentage points for patients with stage IIIA to IIC disease (16.2% to 29.1%), and 10.1 percentage points for unknown stages (16.5% to 26.6%; all P < .001). Conversely, patients with stage I disease experienced a 6.1-percentage point absolute decline in organ preservation (from 26.4% in 2006 to 20.3% in 2020; P < .001). The annual rate of transanal local excisions decreased for all stages. In the subgroup of 80 607 eligible patients, the proportion of complete pathologic responses increased from 6.5% in 2006 to 18.8% in 2020 (P < .001).

Conclusions and Relevance

This case series shows that rectal cancer is increasingly being managed medically, especially among patients whose treatment historically relied on proctectomy. Given the National Comprehensive Cancer Network endorsement of watch and wait, the increasing trends in organ preservation, and the nearly 3-fold increase in complete pathologic responses, international professional societies should urgently develop multidisciplinary core outcome sets and care quality indicators to ensure high-quality rectal cancer research and care delivery accounting for organ preservation.

Real-World Data: Applications and Relevance to Cancer Clinical Trials

Randomized controlled trials (RCTs) are the gold standard for comparative-effectiveness research (CER). Since the 1980s, there has been a rise in the creation and utilization of large national cancer databases to provide readily accessible "real-world data" (RWD). This review article discusses the role of RCTs in oncology, and the role of RWD from the national cancer database in CER. RCTs remain the preferred study type for CER because they minimize confounding and bias. RCTs have challenges to conduct, including extensive time and resources, but these factors do not impact the internal validity of the result. Generalizability and external validity are potential limitations of RCTs. RWD is ideal for studying cancer epidemiology, patterns of care, disparities in care delivery, quality-of-care evaluation, and applicability of RCT data in specific populations excluded from RCTs. However, retrospective databases with RWD have limitations in CER due to unmeasured confounders and are often suboptimal in identifying causal treatment effects.

Gaps in the evidence underpinning high-risk medical devices in Europe at market entry, and potential solutions

AIM

To determine the level of evidence for innovative high-risk medical devices at market entry.

METHODS

We reviewed all Belgian healthcare payer (RIZIV-INAMI) assessor reports on novel implants or invasive medical devices (n = 18, Class IIb-III) available between 2018 to mid-2019 on applications submitted for inclusion on their reimbursement list. We also conducted a review of the literature on evidence gaps and an analysis of relevant legal and ethical frameworks within the European context.

FINDINGS

Conformity assessment of medical devices is based on performance, safety, and an acceptable risk-benefit balance. Information submitted for obtaining CE marking is confidential and legally protected, limiting access to clinical evidence. Seven out of the 18 RIZIV-INAMI assessor reports (39%) included a randomized controlled trial (RCT) using the novel device, whilst 2 applications (11%) referred to an RCT that used a different device. The population included was inappropriate or unclear for 3 devices (17%). Only half of the applications presented evidence on quality of life or functioning and 2 (11%) presented overall survival data. Four applications (22%) included no data beyond twelve months. The findings from the literature demonstrated similar problems with the study design and the clinical evidence.

DISCUSSION AND CONCLUSIONS

CE marking does not indicate that a device is effective, only that it complies with the law. The lack of transparency hampers evidence-based decision making. Despite greater emphasis on clinical benefit for the patient, the provisions of the European Medical Device Regulation (MDR) are not yet fully aligned with international ethical standards for clinical research. The MDR fails to address key issues, such as the lack of access to data submitted for CE marking and a failure to require evidence of clinical effectiveness. Indeed, a first report shows no improvement in the clinical evidence for implantable devices generated under the MDR. Thus, patients may continue to be exposed to ineffective or unsafe novel devices. The Health Technology Assessment Regulation plans for Joint Scientific Consultations for specific high-risk devices before companies begin their pivotal clinical investigations. The demanded comparative evidence should facilitate payer decisions. Nevertheless, there is also a need for legislation requiring comparative RCTs assessing patient-relevant outcomes for high-risk devices to ensure implementation, including development and implementation of common specifications for study designs.

DESIGN DIFFERENCES EXPLAIN VARIATION IN RESULTS BETWEEN RANDOMIZED TRIALS AND THEIR NON-RANDOMIZED EMULATIONS

Objectives

While randomized controlled trials (RCTs) are considered a standard for evidence on the efficacy of medical treatments, non-randomized real-world evidence (RWE) studies using data from health insurance claims or electronic health records can provide important complementary evidence. The use of RWE to inform decision-making has been questioned because of concerns regarding confounding in non-randomized studies and the use of secondary data. RCT-DUPLICATE was a demonstration project that emulated the design of 32 RCTs with non-randomized RWE studies. We sought to explore how emulation differences relate to variation in results between the RCT-RWE study pairs.

Methods

We include all RCT-RWE study pairs from RCT-DUPLICATE where the measure of effect was a hazard ratio and use exploratory meta-regression methods to explain differences and variation in the effect sizes between the results from the RCT and the RWE study. The considered explanatory variables are related to design and population differences.

Results

Most of the observed variation in effect estimates between RCT-RWE study pairs in this sample could be explained by three emulation differences in the meta-regression model: (i) in-hospital start of treatment (not observed in claims data), (ii) discontinuation of certain baseline therapies at randomization (not part of clinical practice), (iii) delayed onset of drug effects (missed by short medication persistence in clinical practice).

Conclusions

This analysis suggests that a substantial proportion of the observed variation between results from RCTs and RWE studies can be attributed to design emulation differences. (238 words).

Role of radiation in oligometastases and oligoprogression in metastatic non-small cell lung cancer: consensus and controversy

INTRODUCTION

The oligometastatic state in non-small cell lung cancer (NSCLC) has recently become well-established. However, the specific definition of oligometastases remains unclear. Several smaller randomized studies have investigated the safety and efficacy of radiation as metastasis-directed therapy (MDT) in oligometastatic NSCLC, which have led the way to larger studies currently accruing patients globally.

AREAS COVERED

This review covers the definitions of 'oligometastases' and explains why the oligometastatic state is becoming increasingly relevant in metastatic NSCLC. This includes the rationale for MDT in oligometastatic NSCLC, specifically reviewing stereotactic body radiation therapy (SBRT) as a treatment strategy. This review details many randomized trials that support radiation as MDT and introduces trials that are currently accruing patients. Finally, it explores some of the controversies that warrant further investigation.

EXPERT OPINION

Radiation treatment, specifically SBRT, has been shown to be safe, convenient, and cost-effective as MDT. As systemic therapy, including targeted agents and immunotherapy, continues to improve, the precise role(s) and timing of radiation therapy may evolve. However, radiation therapy as MDT will continue to be an integral part of treatment in patients with oligometastatic NSCLC.

Accommodating time-varying heterogeneity in risk estimation under the Cox model: a transfer learning approach

Transfer learning has attracted increasing attention in recent years for adaptively borrowing information across different data cohorts in various settings. Cancer registries have been widely used in clinical research because of their easy accessibility and large sample size. Our method is motivated by the question of how to utilize cancer registry data as a complement to improve the estimation precision of individual risks of death for inflammatory breast cancer (IBC) patients at The University of Texas MD Anderson Cancer Center. When transferring information for risk estimation based on the cancer registries (i.e., source cohort) to a single cancer center (i.e., target cohort), time-varying population heterogeneity needs to be appropriately acknowledged. However, there is no literature on how to adaptively transfer knowledge on risk estimation with time-to-event data from the source cohort to the target cohort while adjusting for time-varying differences in event risks between the two sources. Our goal is to address this statistical challenge by developing a transfer learning approach under the Cox proportional hazards model. To allow data-adaptive levels of information borrowing, we impose Lasso penalties on the discrepancies in regression coefficients and baseline hazard functions between the two cohorts, which are jointly solved in the proposed transfer learning algorithm. As shown in the extensive simulation studies, the proposed method yields more precise individualized risk estimation than using the target cohort alone. Meanwhile, our method demonstrates satisfactory robustness against cohort differences compared with the method that directly combines the target and source data in the Cox model. We develop a more accurate risk estimation model for the MD Anderson IBC cohort given various treatment and baseline covariates, while adaptively borrowing information from the National Cancer Database to improve risk assessment.

Novel Prognostic Staging System for Patients With De Novo Metastatic Breast Cancer

PURPOSE

Given the heterogeneity and improvement in outcomes for metastatic breast cancer (MBC), we developed a staging system that refines prognostic estimates for patients with metastatic cancer at the time of initial diagnosis, de novo MBC (dnMBC), on the basis of survival outcomes and disease-related variables.

METHODS

Patients with dnMBC (2010-2016) were selected from the National Cancer Database (NCDB). Recursive partitioning analysis (RPA) was used to group patients with similar overall survival (OS) on the basis of clinical T category, grade, estrogen receptor (ER), progesterone receptor, human epidermal growth factor receptor 2, histology, organ system site of metastases (bone-only, brain-only, visceral), and number of organ systems involved. Three-year OS rates were used to assign a final stage: IVA: >70%, IVB: 50%-70%, IVC: 25 to <50%, and IVD: <25%. Bootstrapping was applied with 1,000 iterations, and final stage assignments were made based on the most commonly occurring assignment. Unadjusted OS was estimated. Validation analyses were conducted using SEER and NCDB.

RESULTS

At a median follow-up of 52.9 months, the median OS of the original cohort (N = 42,467) was 35.4 months (95% CI, 34.8 to 35.9). RPA stratified patients into 53 groups with 3-year OS rates ranging from 73.5% to 5.7%; these groups were amalgamated into four stage groups: 3-year OS, A = 73.2%, B = 61.9%, C = 40.1%, and D = 17% (log-rank P < .001). After bootstrapping, the survival outcomes for the four stages remained significantly different (log-rank P < .001). This staging system was then validated using SEER data (N = 20,469) and a separate cohort from the NCDB (N = 7,645) (both log-rank P < .001).

CONCLUSION

Our findings regarding the heterogeneity in outcomes for patients with dnMBC could guide future revisions of the current American Joint Committee on Cancer staging guidelines for patients with newly diagnosed stage IV disease. Our findings should be independently confirmed.

Commercializing Personal Health Information: A Critical Qualitative Content Analysis of Documents Describing Proprietary Primary Care Databases in Canada

BACKGROUND

Commercial data brokers have amassed large collections of primary care patient data in proprietary databases. Our study objective was to critically analyze how entities involved in the collection and use of these records construct the value of these proprietary databases. We also discuss the implications of the collection and use of these databases.

METHODS

We conducted a critical qualitative content analysis using publicly available documents describing the creation and use of proprietary databases containing Canadian primary care patient data. We identified relevant commercial data brokers, as well as entities involved in collecting data or in using data from these databases. We sampled documents associated with these entities that described any aspect of the collection, processing, and use of the proprietary databases. We extracted data from each document using a structured data tool. We conducted an interpretive thematic content analysis by inductively coding documents and the extracted data.

RESULTS

We analyzed 25 documents produced between 2013 and 2021. These documents were largely directed at the pharmaceutical industry, as well as shareholders, academics, and governments. The documents constructed the value of the proprietary databases by describing extensive, intimate, detailed patient-level data holdings. They provided examples of how the databases could be used by pharmaceutical companies for regulatory approval, marketing and understanding physician behaviour. The documents constructed the value of these data more broadly by claiming to improve health for patients, while also addressing risks to privacy. Some documents referred to the trade-offs between patient privacy and data utility, which suggests these considerations may be in tension.

CONCLUSION

Documents in our analysis positioned the proprietary databases as socially legitimate and valuable, particularly to pharmaceutical companies. The databases, however, may pose risks to patient privacy and contribute to problematic drug promotion. Solutions include expanding public data repositories with appropriate governance and external regulatory oversight.

Reexamining Time From Breast Cancer Diagnosis to Primary Breast Surgery

Importance

Although longer times from breast cancer diagnosis to primary surgery have been associated with worse survival outcomes, the specific time point after which it is disadvantageous to have surgery is unknown. Identifying an acceptable time to surgery would help inform patients, clinicians, and the health care system.

Objective

To examine the association between time from breast cancer diagnosis to surgery (in weeks) and overall survival and to describe factors associated with surgical delay. The hypothesis that there is an association between time to surgery and overall survival was tested.

Design, Setting, and Participants

This was a case series study that used National Cancer Database (NCDB) data from female individuals diagnosed with breast cancer from 2010 to 2014 (with 5-year follow-up to 2019). The NCDB uses hospital registry data from greater than 1500 Commission on Cancer-accredited facilities, accounting for 70% of all cancers diagnosed in the US. Included participants were females 18 years or older with stage I to III ductal or lobular breast cancer who underwent surgery as the first course of treatment. Patients with prior breast cancer, missing receptor information, neoadjuvant or experimental therapy, or who were diagnosed with breast cancer on the date of their primary surgery were excluded. Multivariable Cox regression was used to evaluate factors associated with overall survival. Patients were censored at death or last follow-up. Covariates included age and tumor characteristics. Multinomial regression was performed to identify factors associated with longer time to surgery, using surgery 30 days or less from diagnosis as the reference group. Data were analyzed from March 15 to July 7, 2022.

Exposures

Time to receipt of primary breast surgery.

Measures

The primary outcome measure was overall survival.

Results

The final cohort included 373 334 patients (median [IQR] age, 61 [51-70] years). On multivariable Cox regression analysis, time to surgery 9 weeks (57-63 days) or later after diagnosis was associated with worse overall survival (hazard ratio, 1.15; 95% CI, 1.08-1.23; P < .001) compared with surgery between 0 to 4 weeks (1-28 days). By multinomial regression, factors associated with longer times to surgery (using surgery 1-30 days from diagnosis as a reference) included the following: (1) younger age, eg, the adjusted odds ratio (OR) for patients 45 years or younger undergoing surgery 31 to 60 days from diagnosis was 1.32 (95% CI, 1.28-1.38); 61 to 74 days, 1.64 (95% CI, 1.52-1.78); and greater than 74 days, 1.58 (95% CI, 1.46-1.71); (2) uninsured or Medicaid status, eg, the adjusted OR for patients with Medicaid undergoing surgery 31 to 60 days from diagnosis was 1.35 (95% CI, 1.30-1.39); 61 to 74 days, 2.13 (95% CI, 2.01-2.26); and greater than 74 days, 3.42 (95% CI, 3.25-3.61); and (3) lower neighborhood household income, eg, the adjusted OR for patients with household income less than $38,000 undergoing surgery 31 to 60 days from diagnosis was 1.35 (95% CI, 1.02-1.07); 61 to 74 days, 1.21 (95% CI, 1.15-1.27); and greater than 74 days, 1.53 (95% CI, 1.46-1.61).

Conclusions and Relevance

Findings of this case series study suggest the use of 8 weeks or less as a quality metric for time to surgery. Time to surgery of greater than 8 weeks may partly be associated with disadvantageous social determinants of health.

Risk Stratification for Management of Solitary Fibrous Tumor/Hemangiopericytoma of the Central Nervous System

INTRODUCTION

Solitary fibrous tumor/hemangiopericytoma (SFT/HPC) of the central nervous system (CNS) is a rare meningeal tumor. Given the absence of prospective or randomized data, there are no standard indications for radiotherapy. Recently, the NRG Oncology and EORTC cooperative groups successfully accrued and completed the first prospective trials evaluating risk-adapted adjuvant radiotherapy strategies for meningiomas. Using a similar framework, we sought to develop prognostic risk categories that may predict the survival benefit associated with radiotherapy, using two large national datasets.

METHODS

We queried the National Cancer Database (NCDB) and the Surveillance, Epidemiology, and End Results (SEER) databases for all newly diagnosed cases of SFT/HPC within the CNS. Risk categories were created, as follows: low risk-grade 1, with any extent of resection (EOR) and grade 2, with gross-total resection; intermediate risk-grade 2, with biopsy/subtotal resection; high risk-grade 3 with any EOR. The Kaplan-Meier method and Cox proportional hazards regressions were used to determine the association of risk categories with overall and cause-specific survival. We then determined the association of radiotherapy with overall survival in the NCDB, stratified by risk group.

RESULTS

We identified 866 and 683 patients from the NCDB and SEER databases who were evaluated, respectively. In the NCDB, the 75% survival times for low- (n = 312), intermediate- (n = 239), and high-risk (n = 315) patients were not reached, 86 months (HR 1.60 (95% CI 1.01-2.55)), and 55 months (HR 2.56 (95% CI 1.68-3.89)), respectively. Our risk categories were validated for overall and cause-specific survival in the SEER dataset. Radiotherapy was associated with improved survival in the high- (HR 0.46 (0.29-0.74)) and intermediate-risk groups (HR 0.52 (0.27-0.99)) but not in the low-risk group (HR 1.26 (0.60-2.65)). The association of radiotherapy with overall survival remained significant in the multivariable analysis for the high-risk group (HR 0.55 (0.34-0.89)) but not for the intermediate-risk group (HR 0.74 (0.38-1.47)). Similar results were observed in a time-dependent landmark sensitivity analysis.

CONCLUSION

Risk stratification based on grade and EOR is prognostic of overall and cause-specific survival for SFT/HPCs of the CNS and performs better than any individual clinical factor. These risk categories appear to predict the survival benefit from radiotherapy, which is limited to the high-risk group and, potentially, the intermediate-risk group. These data may serve as the basis for a prospective study evaluating the management of meningeal SFT/HPCs.

Prognostic Utility of Tumor Stage versus American Thyroid Association Risk Class in Thyroid Cancer

OBJECTIVE

To evaluate the prognostic strengths of American Joint Committee on Cancer (AJCC) staging and American Thyroid Association (ATA) risk classification in well-differentiated thyroid cancer (DTC), and their implications in guiding medical decision-making and epidemiological study designs.

METHODS

The 2004-2017 National Cancer Database was queried for DTC patients. Cox proportional hazards (CPH) and Kaplan-Meier analyses modeled patient mortality and overall survival, respectively. Each CPH model was evaluated by its concordance index, measure of explained randomness (MER), Akaike information criterion (AIC), and area under receiver operating characteristic curve (AUC).

RESULTS

Overall, 134,226 patients were analyzed, with an average age of 48.1 ± 15.1 years (76.9% female). Univariate CPH models using AJCC staging demonstrated higher concordance indices, MERs, and AUCs than those using ATA risk classification (all p < 0.001). Multivariable CPH models using AJCC staging demonstrated higher concordance indices (p = 0.049), MERs (p = 0.046), and AUCs (p = 0.002) than those using ATA risk classification. The AICs of multivariable AJCC staging and ATA risk models were 7.564 × 104 and 7.603 × 104 , respectively. AJCC stage I tumors were associated with greater overall survival than those classified as ATA low risk, whereas AJCC stages II-III and stage IV tumors demonstrated worse survival than ATA intermediate- and high-risk tumors, respectively (all p < 0.001).

CONCLUSION

AJCC staging may be a more predictive system for patient survival than ATA risk. The prognostic utility of these two systems converges when additional demographic and clinical factors are considered. AJCC staging was found to classify patients across a wider range of survival patterns than the ATA risk stratification system.

LEVEL OF EVIDENCE

4 Laryngoscope, 133:205-211, 2023.

Impact of academic medical center access on outcomes in multiple myeloma

Treatment at academic cancer centers (ACs) is associated with improved survival across hematologic malignancies, though the benefit in multiple myeloma (MM) has not been examined. This study aims to evaluate survival outcomes at Commission on Cancer accredited ACs compared to non-academic centers (NACs) for patients receiving MM-directed therapy. The National Cancer Database (NCDB) was used to identify demographics and overall survival (OS) of MM patients diagnosed from 2004 to 2017 and to compare outcomes by facility type. Survival analysis was repeated in a propensity score matched cohort, with NACs matched 1:1 to ACs by age, race, comorbidity score, insurance, year of diagnosis, distance traveled, and income. Of 163 375 MM patients, 44.5% were treated at ACs. Patients at ACs were more likely to receive MM-directed therapy compared to NACs (81% vs. 73%, p < .001). For patients receiving treatment, median OS at ACs was 71.3 months versus 41.2 months at NACs (p < .001). When adjusted for baseline demographics, patients treated at ACs had reduced mortality; hazard ratio (HR) 0.79 (95% CI 0.78-0.81, p < .001). The propensity score matched cohort maintained this survival benefit with a median OS of 59.9 months at ACs versus 37.0 months at NACs (p < .001), HR of 0.66 (95% CI 0.64-0.67, p < .001). ACs treated younger patients with fewer comorbidities and were more likely to treat racial minorities and patients with Medicaid or private insurance, and the uninsured. In this analysis, MM patients treated at ACs have significantly improved survival. While potentially related to access to specialized care, socioeconomic factors that drive facility selection may also contribute.

Improved Postoperative Outcomes after Prehabilitation for Colorectal Cancer Surgery in Older Patients: An Emulated Target Trial

BACKGROUND

The aim of this study was to assess the effect of a multimodal prehabilitation program on perioperative outcomes in colorectal cancer patients with a higher postoperative complication risk, using an emulated target trial (ETT) design.

PATIENTS AND METHODS

An ETT design including overlap weighting based on propensity score was performed. The study consisted of all patients with newly diagnosed colorectal cancer (2016-2021), in a large nonacademic training hospital, who were candidate to elective colorectal cancer surgery and had a higher risk for postoperative complications defined by: age ≥ 65 years and or American Society of Anesthesiologists score III/IV. Intention-to-treat (ITT) and per-protocol analyses were performed to evaluate the effect of prehabilitation compared with usual care on perioperative complications and length of stay (LOS).

RESULTS

Two hundred fifty-one patients were included: 128 in the usual care group and 123 patients in the prehabilitation group. In the ITT analysis, the number needed to treat to reduce one or more complications in one person was 4.2 (95% CI 2.6-10). Compared with patients in the usual care group, patients undergoing prehabilitation had a 55% lower comprehensive complication score (95% CI -71 to -32%). There was a 33% reduction (95% CI -44 to -18%) in LOS from 7 to 5 days.

CONCLUSIONS

This study showed a clinically relevant reduction of complications and LOS after multimodal prehabilitation in patients undergoing colorectal cancer surgery with a higher postoperative complication risk. The study methodology used may serve as an example for further larger multicenter comparative effectiveness research on prehabilitation.

Outcomes of Local Excision Compared to Radical Excision of Rectal Gastrointestinal Stromal Tumors: A Propensity-Score Matched Analysis of the NCDB

BACKGROUND

Gastrointestinal stromal tumors (GISTs) account for <1% of gastrointestinal cancers. The present study aimed to assess the outcomes of local and radical excision of non-metastatic rectal GISTs.

METHODS

This study was a retrospective cohort analysis of patients with non-metastatic rectal GISTs in the NCDB. Patients were divided according to the surgical approach into local and radical excision and were compared in regard to the baseline characteristics and outcomes. A propensity-score matched analysis was performed to match the two groups for baseline confounders. The main outcomes were 5-year overall survival (OS), surgical margins, hospital stay, short-term mortality, and readmission.

RESULTS

228 patients (54.8% male) with rectal GISTs were included. Before matching, 127 (55.7%) patients underwent local excision and 101 (44.3%) had radical excision. Patients who underwent local excision had more cT1-T2 and low-grade GISTs whereas patients who had radical excision received more neoadjuvant systemic treatment. After matching for clinical T stage, tumor grade, and neoadjuvant systemic therapy, 52 patients were included in each group. Local excision had a significantly higher rate of positive resection margins (42.2% vs. 19.1%, p = 0.02) and a shorter hospital stay (0 vs. 3 days, p < 0.001) than radical excision. The two groups had similar mean OS (139.8 vs. 133.1 months, p = 0.52).

CONCLUSIONS

Local excision was associated with a significantly higher incidence of positive resection margins and shorter hospital stay, yet similar overall survival to radical excision.

Controversies in the Front-Line Treatment of Systemic Peripheral T Cell Lymphomas

Systemic peripheral T cell lymphomas (PTCL) are a rare and clinically and biologically heterogeneous group of disorders with scarce and generally low-quality evidence guiding their management. In this manuscript, we tackle the current controversies in the front-line treatment of systemic PTCL including (1) whether CNS prophylaxis should be administered; (2) whether CHOEP should be preferred over CHOP; (3) what role brentuximab vedotin should have; (4) whether stem cell transplant (SCT) consolidation should be used and whether autologous or allogeneic; (5) how should molecular subtypes (including DUSP22 or TP63-rearranged ALCL or GATA3 or TBX21 PTCL, NOS) impact therapeutic decisions; and (6) whether there is a role for targeted agents beyond brentuximab vedotin.

Understanding cancer genetic risk assessment motivations in a remote tailored risk communication and navigation intervention randomized controlled trial

Background

National guidelines recommend cancer genetic risk assessment (CGRA) (i.e. genetic counseling prior to genetic testing) for women at increased risk for hereditary breast and ovarian cancer (HBOC). Less than one-half of eligible women obtain CGRA, leaving thousands of women and their family members without access to potentially life-saving cancer prevention interventions.

Purpose

The Genetic Risk Assessment for Cancer Education and Empowerment Project (GRACE) addressed this translational gap, testing the efficacy of a tailored counseling and navigation (TCN) intervention vs. a targeted print brochure vs. usual care on CGRA intentions. Selected behavioral variables were theorized to mediate CGRA intentions.

Methods

Breast and ovarian cancer survivors meeting criteria for guideline-based CGRA were recruited from three state cancer registries (N = 654), completed a baseline survey, and were randomized. TCN and targeted print arms received the brochure; TCN also participated in a tailored, telephone-based decision coaching and navigation session grounded in the Extended Parallel Process Model and Ottawa Decision Support Framework. Participants completed a one-month assessment. Logistic regression was used to compare the rate of CGRA intentions. CGRA intentions and theorized mediator scores (continuous level variables) were calculated using mixed model analysis.

Results

CGRA intentions increased for TCN (53.2%) vs. targeted print (26.7%) (OR = 3.129; 95% CI: 2.028, 4.827, p < .0001) and TCN vs. usual care (23.1%) (OR = 3.778, CI: 2.422, 5.894, p < .0001). Perceived risk (p = 0.023) and self-efficacy (p = 0.035) mediated CGRA intentions in TCN.

Conclusions

Improvements in CGRA intentions and theorized mediators support the use of a tailored communication intervention among women at increased HBOC risk. (Clinicaltrials.gov: NCT03326713.)Trial registration: ClinicalTrials.gov identifier: NCT03326713.

Population-level and individual-level explainers for propensity score matching in observational studies

Precis

The exclusion of unmatched observations in propensity score matching has implications for the generalizability of causal effects. Machine learning methods can help to identify how the study population differs from the unmatched subpopulation.

Background

There has been widespread use of propensity scores in evaluating the effect of cancer treatments on survival, particularly in administrative databases and cancer registries. A byproduct of certain matching schemes is the exclusion of observations. Borrowing an analogy from clinical trials, one can view these exclusions as subjects that do not satisfy eligibility criteria.

Methods

Developing identification rules for these “data-driven eligibility criteria” in observational studies on both population and individual levels helps to ascertain the population on which causal effects are being made. This article presents a machine learning method to determine the representativeness of causal effects in two different datasets from the National Cancer Database.

Results

Decision trees reveal that groups with certain features have a higher probability of inclusion in the study population than older patients. In the first dataset, younger age categories had an inclusion probability of at least 0.90 in all models, while the probability for the older category ranged from 0.47 to 0.65. Most trees split once more on an even higher age at a lower node, suggesting that the oldest patients are the least likely to be matched. In the second set of data, both age and surgery status were associated with inclusion.

Conclusion

The methodology presented in this paper underscores the need to consider exclusions in propensity score matching procedures as well as complementing matching with other propensity score adjustments.

Impact of Preoperative Chemotherapy Features on Patient Outcomes after Hepatectomy for Initially Unresectable Colorectal Cancer Liver Metastases: A LiverMetSurvey Analysis

Background: Prognostic factors have been extensively reported after resection of colorectal liver metastases (CLM); however, specific analyses of the impact of preoperative systemic anticancer therapy (PO-SACT) features on outcomes is lacking. Methods: For this real-world evidence study, we used prospectively collected data within the international surgical LiverMetSurvey database from all patients with initially-irresectable CLM. The main outcome was Overall Survival (OS) after surgery. Disease-free (DFS) and hepatic-specific relapse-free survival (HS-RFS) were secondary outcomes. PO-SACT features included duration (cumulative number of cycles), choice of the cytotoxic backbone (oxaliplatin- or irinotecan-based), fluoropyrimidine (infusional or oral) and addition or not of targeted monoclonal antibodies (anti-EGFR or anti-VEGF). Results: A total of 2793 patients in the database had received PO-SACT for initially irresectable diseases. Short (<7 or <13 cycles in 1st or 2nd line) PO-SACT duration was independently associated with longer OS (HR: 0.85 p = 0.046), DFS (HR: 0.81; p = 0.016) and HS-RFS (HR: 0.80; p = 0.05). All other PO-SACT features yielded basically comparable results. Conclusions: In this international cohort, provided that PO-SACT allowed conversion to resectability in initially irresectable CLM, surgery performed as soon as technically feasible resulted in the best outcomes. When resection was achieved, our findings indicate that the choice of PO-SACT regimen had a marginal if any, impact on outcomes.

Clinical trial design in the era of precision medicine

Recent rapid biotechnological breakthroughs have led to the identification of complex and unique molecular features that drive malignancies. Precision medicine has exploited next-generation sequencing and matched targeted therapy/immunotherapy deployment to successfully transform the outlook for several fatal cancers. Tumor and liquid biopsy genomic profiling and transcriptomic, immunomic, and proteomic interrogation can now all be leveraged to optimize therapy. Multiple new trial designs, including basket and umbrella trials, master platform trials, and N-of-1 patient-centric studies, are beginning to supplant standard phase I, II, and III protocols, allowing for accelerated drug evaluation and approval and molecular-based individualized treatment. Furthermore, real-world data, as well as exploitation of digital apps and structured observational registries, and the utilization of machine learning and/or artificial intelligence, may further accelerate knowledge acquisition. Overall, clinical trials have evolved, shifting from tumor type-centered to gene-directed and histology-agnostic trials, with innovative adaptive designs and personalized combination treatment strategies tailored to individual biomarker profiles. Some, but not all, novel trials now demonstrate that matched therapy correlates with superior outcomes compared to non-matched therapy across tumor types and in specific cancers. To further improve the precision medicine paradigm, the strategy of matching drugs to patients based on molecular features should be implemented earlier in the disease course, and cancers should have comprehensive multi-omic (genomics, transcriptomics, proteomics, immunomic) tumor profiling. To overcome cancer complexity, moving from drug-centric to patient-centric individualized combination therapy is critical. This review focuses on the design, advantages, limitations, and challenges of a spectrum of clinical trial designs in the era of precision oncology.

Differences in receipt of multimodality therapy by race, insurance status, and socioeconomic disadvantage in patients with resected pancreatic cancer

BACKGROUND AND METHODS: Racial and socioeconomic disparities in receipt of adjuvant chemotherapy affect patients with pancreatic cancer. However, differences in receipt of neoadjuvant chemotherapy among patients undergoing resection are not well-understood. A retrospective cross-sectional cohort of patients with resected AJCC Stage I/II pancreatic ductal adenocarcinoma was identified from the National Cancer Database (2014-2017). Outcomes included receipt of neoadjuvant versus adjuvant chemotherapy, or receipt of either, defined as multimodality therapy and were assessed by univariate and multivariate analysis.

RESULTS

Of 19 588 patients, 5098 (26%) received neoadjuvant chemotherapy, 9624 (49.1%) received adjuvant chemotherapy only, and 4757 (24.3%) received no chemotherapy. On multivariable analysis, Black patients had lower odds of neoadjuvant chemotherapy compared to White patients (OR: 0.80, 95% CI: 0.67-0.97) but no differences in receipt of multimodality therapy (OR: 0.89, 95% CI: 0.77-1.03). Patients with Medicaid or no insurance, low educational attainment, or low median income had significantly lower odds of receiving neoadjuvant chemotherapy or multimodality therapy.

CONCLUSIONS

Racial and socioeconomic disparities persist in receipt of neoadjuvant and multimodality therapy in patients with resected pancreatic adenocarcinoma.

DISCUSSION

Policy and interventional implementations are needed to bridge the continued socioeconomic and racial disparity gap in pancreatic cancer care.

Bridging the gap between oncology clinical trials and real-world data: evidence on replicability of efficacy results using German claims data

Aims: Using German claims, the authors replicated the CHAARTED trial in metastatic hormone-sensitive prostate cancer. Methods: The authors identified metastatic hormone-sensitive prostate cancer patients replicating the inclusion/exclusion criteria of CHAARTED. Patients treated with docetaxel in combination with androgen deprivation therapy (ADT) at first line (docetaxel group) were compared with patients treated with ADT monotherapy (ADT mono group). After propensity score matching, overall survival was compared between the matched cohorts. Results: The authors included 441 patients. After propensity score matching, two equally sized matched cohorts of 74 patients each were compared in terms of overall survival. The hazard ratio (HR) was 0.71 (95% CI: 0.42-1.19), comparable to the HR in CHAARTED (HR: 0.72; 95% CI: 0.59-0.89). Conclusions: Using early comparative evidence from real-world data for regulatory and health technology assessment decisions is useful.

Novel Use of Patient-Specific Covariates From Oncology Studies in the Era of Biomedical Data Science: A Review of Latest Methodologies

In this article, we review different applications of how to incorporate individual patient variables into clinical research within oncology. These methodologies range from the more traditional use of baseline covariates from randomized clinical trials, as well as observational studies, to using covariates to generalize the results of randomized clinical trials to other populations. Individual patient variables also allow for the consideration of heterogeneity in treatment effects and individualized treatment rules. We primarily consider two treatment groups and mostly focus on time-to-event outcomes where such methodologies have been well established and widely applied. We also discuss more conceptually newer statistical research that has not been widely applied in clinical oncology, but is likely to make an impact in future oncology research. With the increasing amount of biomedical data available for analysis, it is inevitable that more methods are developed to make best use of information, to advance oncology research.

Population-based estimates of overtreatment with adjuvant systemic therapy in early breast cancer patients with data from the Netherlands and the USA

Purpose

Although adjuvant systemic therapy (AST) helps increase breast cancer-specific survival (BCSS), there is a growing concern for overtreatment. By estimating the expected BCSS of AST using PREDICT, this study aims to quantify the number of patients treated with AST without benefit to provide estimates of overtreatment.

Methods

Data of all non-metastatic unilateral breast cancer patients diagnosed in 2015 were retrieved from cancer registries from The Netherlands and the USA. The PREDICT tool was used to estimate AST survival benefit. Overtreatment was defined as the proportion of patients that would have survived regardless of or died despite AST within 10 years. Three scenarios were evaluated: actual treatment, and recommendations by the Dutch or USA guidelines.

Results

59.5% of Dutch patients were treated with AST. 6.4% (interquartile interval [IQI] = 2.5, 8.2%) was expected to survive at least 10 years due to AST, leaving 93.6% (IQI = 91.8, 97.5%) without AST benefit (overtreatment). The lowest expected amount of overtreatment was in the targeted and chemotherapy subgroup, with 86.5% (IQI = 83.4, 89.6%) overtreatment, and highest in the only endocrine treatment subgroup, with 96.7% (IQI = 96.0, 98.1%) overtreatment. Similar results were obtained using data from the USA, and guideline recommendations.

Conclusion

Based on PREDICT, AST prevents 10-year breast cancer death in 6.4% of the patients treated with AST. Consequently, AST yields no survival benefit to many treated patients. Especially improved personalization of endocrine therapy is relevant, as this therapy is widely used and is associated with the highest amount of overtreatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10549-022-06550-2.

Bias Reduction through Analysis of Competing Events (BRACE) Correction to Address Cancer Treatment Selection Bias in Observational Data

BACKGROUND

Cancer treatments can paradoxically appear to reduce the risk of non-cancer mortality in observational studies, due to residual confounding. Here we introduce a method, Bias Reduction through Analysis of Competing Events (BRACE), to reduce bias in the presence of residual confounding.

METHODS

BRACE is a novel method for adjusting for bias from residual confounding in proportional hazards models. Using standard simulation methods, we compared BRACE vs. Cox proportional hazards regression in the presence of an unmeasured confounder. We examined estimator distributions, bias, mean squared error (MSE), and coverage probability. We then estimated treatment effects of high vs. low intensity treatments in 36,630 prostate cancer, 4,069 lung cancer, and 7,117 head/neck cancer patients, using the Veterans Affairs database. We analyzed treatment effects on cancer-specific mortality (CSM), non-cancer mortality (NCM), and overall survival (OS), using conventional multivariable Cox and propensity score (adjusted using inverse probability weighting) models, vs. BRACE-adjusted estimates.

RESULTS

In simulations with residual confounding, BRACE uniformly reduced both bias and MSE. In the absence of bias, BRACE introduced bias toward the null, albeit with lower MSE. BRACE markedly improved coverage probability, but with a tendency toward overcorrection for effective but non-toxic treatments. For each clinical cohort, more intensive treatments were associated with significantly reduced hazards for CSM, NCM, and OS. BRACE attenuated OS estimates, yielding results more consistent with findings from randomized trials and meta-analyses.

CONCLUSIONS

BRACE reduces bias and MSE when residual confounding is present and represents a novel approach to improve treatment effect estimation in non-randomized studies.

Retrospective comparative effectiveness research: will changing the analytical methods change the results?

In medicine, retrospective cohort studies are used to compare treatments to one another. We hypothesize that the outcomes of retrospective comparative effectiveness research studies can be heavily influenced by biostatistical analytic choices, thereby leading to inconsistent conclusions. We selected a clinical scenario currently under investigation: survival in metastatic prostate, breast, or lung cancer after systemic vs systemic + definitive local therapy. We ran >300 000 regression models (each representing a publishable study). Each model had various forms of analytic choices (to account for bias): propensity score matching, left truncation adjustment, landmark analysis, and covariate combinations. There were 72 549 lung, 14 904 prostate, and 13 857 breast cancer patients included. In the most basic analysis, which omitted propensity score matching, left truncation adjustment, and landmark analysis, all of the HRs were < 1 (generally, 0.60-0.95, favoring addition of local therapy), with all P-values < 0.001. Left truncation adjustment landmark analysis produced results with non-significant P-values. The combination of propensity score matching, left truncation adjustment, landmark analysis, and covariate combinations generally produced P-values that were > 0.05 and/or HRs that were > 1 (favoring systemic therapy alone). The use of more statistical methods to reduce the selection bias caused reported HR ranges to approach 1.0. By varying analytic choices in comparative effectiveness research, we generated contrary outcomes. Our results suggest that some retrospective observational studies may find a treatment improves outcomes for patients, while another similar study may find it does not, simply based on analytical choices. This article is protected by copyright. All rights reserved.

Evaluation of sociodemographic and baseline patient characteristic differences in cervical cancer patients treated with either external beam or brachytherapy boost

PURPOSE

Chemoradiation is considered the standard of care for locally advanced cervical cancer. While brachytherapy (BT) boost is associated with improved survival and less toxicity compared to external beam radiation therapy (EBRT) boost, it is unclear why many patients do not receive a BT boost. In this study, we compared sociodemographic and baseline patient characteristics between patients receiving EBRT boost versus BT boost.

METHODS

We analyzed patients in the National Cancer Database diagnosed between 2004 and 2016 with FIGO stage IIB-IVA cervical cancer treated with nonpalliative doses of chemoradiation. Logistic regression analysis was utilized to evaluate BT utilization over time and by other clinicopathological and sociodemographic factors.

RESULTS

Overall, 5764 patients were evaluated, of which 4937 (86%) underwent BT boost. Using multivariable logistic regression, higher FIGO stage was a significant predictor for utilization of EBRT versus BT boost, with odds ratio 2.92 (95% confidence interval [CI] 2.04-4.16; p < 0.001), 2.68 (95%CI 2.22-3.24; p < 0.001), and 4.51 (95%CI 3.05-6.67; p < 0.001) for IIIA, IIIB, and IVA, respectively, compared to IIB. Increased utilization of EBRT boost was also associated with community cancer facility types, lower income (based on zip code), earlier year of diagnosis, and higher comorbidity score.

CONCLUSIONS

In FIGO stage IIB-IVA cervical cancer patients treated with nonpalliative doses of chemoradiation, overall utilization of BT is 86%. Higher FIGO stage, community cancer facilities, lower income, earlier year of diagnosis, and higher comorbidity score were significant predictors of EBRT boost utilization. Future studies are needed to better understand reasons for this as higher FIGO stage patients are the mostly likely to benefit from a BT boost.

Commission on Cancer CP3R Compliance Rates for Treatment of Patients With Triple Negative and HER2+ Breast Cancer: A National Cancer Database Analysis

BACKGROUND

Cancer Program Practice Profile Reports (CP3R) metrics were released by the Commission on Cancer to provide standards for high-quality care. One metric is the recommendation of combination chemotherapy or chemo-immunotherapy (CIT) within 120 days of diagnosis for women under 70 with AJCC T1cN0M0 or Stage IB-III HER2+ or hormone receptor negative breast cancer ([Multi-agent chemotherapy] MAC). Our study assesses national concordance rates for MAC and CIT.

METHODS

The National Cancer Database was queried from 2004-2014.

RESULTS

122,045 patients met criteria, of whom treatment for 101,800 (83.4%) patients was concordant with MAC and CIT. Treatment concordance increased from 75.7% in 2004 to 89.5% in 2014. For HER2+ patients, use of CIT treatment downtrended with progression of pathological stage, from 70.1% (stage I) to 58.1% (stage III). Mean overall survival of patients whose treatment was concordant with MAC and CIT was longer than that of patients who were non-concordant (146.6 vs 143.8 months, P <.01). On Cox regression, there was a survival benefit for concordant patients who were treated at academic hospitals (HR .89, 95% CI 0.802-.976) and had private insurance (HR .76, 95% CI 0.65-.89).

CONCLUSION

Compliance with MAC and CIT has improved over the past decade and is associated with a significant improvement in overall survival.

Electronic Health Record Network Research in Infectious Diseases

With the marked increases in electronic health record (EHR) use for providing clinical care, there have been parallel efforts to leverage EHR data for research. EHR repositories offer the promise of vast amounts of clinical data not easily captured with traditional research methods and facilitate clinical epidemiology and comparative effectiveness research, including analyses to identify patients at higher risk for complications or who are better candidates for treatment. These types of studies have been relatively slow to penetrate the field of infectious diseases, but the need for rapid turnaround during the COVID-19 global pandemic has accelerated the uptake. This review discusses the rationale for her network projects, opportunities and challenges that such networks present, and some prior studies within the field of infectious diseases.

Associations of Medicaid Expansion With Insurance Coverage, Stage at Diagnosis, and Treatment Among Patients With Genitourinary Malignant Neoplasms

IMPORTANCE

Health insurance coverage is associated with improved outcomes in patients with cancer. However, it is unknown whether Medicaid expansion through the Patient Protection and Affordable Care Act (ACA) was associated with improvements in the diagnosis and treatment of patients with genitourinary cancer.

OBJECTIVE

To assess the association of Medicaid expansion with health insurance status, stage at diagnosis, and receipt of treatment among nonelderly patients with newly diagnosed kidney, bladder, or prostate cancer.

DESIGN, SETTING, AND PARTICIPANTS

This case-control study included adults aged 18 to 64 years with a new primary diagnosis of kidney, bladder, or prostate cancer, selected from the National Cancer Database from January 1, 2011, to December 31, 2016. Patients in states that expanded Medicaid were the case group, and patients in nonexpansion states were the control group. Data were analyzed from January 2020 to March 2021.

EXPOSURES

State Medicaid expansion status.

MAIN OUTCOMES AND MEASURES

Insurance status, stage at diagnosis, and receipt of cancer and stage-specific treatments. Cases and controls were compared with difference-in-difference analyses.

RESULTS

Among a total of 340 552 patients with newly diagnosed genitourinary cancers, 94 033 (27.6%) had kidney cancer, 25 770 (7.6%) had bladder cancer, and 220 749 (64.8%) had prostate cancer. Medicaid expansion was associated with a net decrease in uninsured rate of 1.1 (95% CI, -1.4 to -0.8) percentage points across all incomes and a net decrease in the low-income population of 4.4 (95% CI, -5.7 to -3.0) percentage points compared with nonexpansion states. Expansion was also associated with a significant shift toward early-stage diagnosis in kidney cancer across all income levels (difference-in-difference, 1.4 [95% CI, 0.1 to 2.6] percentage points) and among individuals with low income (difference-in-difference, 4.6 [95% CI, 0.3 to 9.0] percentage points) and in prostate cancer among individuals with low income (difference-in-difference, 3.0 [95% CI, 0.3 to 5.7] percentage points). Additionally, there was a net increase associated with expansion compared with nonexpansion in receipt of active surveillance for low-risk prostate cancer of 4.1 (95% CI, 2.9 to 5.3) percentage points across incomes and 4.5 (95% CI, 0 to 9.0) percentage points among patients in low-income areas.

CONCLUSIONS AND RELEVANCE

These findings suggest that Medicaid expansion was associated with decreases in uninsured status, increases in the proportion of kidney and prostate cancer diagnosed in an early stage, and higher rates of active surveillance in the appropriate, low-risk prostate cancer population. Associations were concentrated in population residing in low-income areas and reinforce the importance of improving access to care to all patients with cancer.