Organ Dysfunction in Patients with Advanced Melanoma Treated with Immune Checkpoint Inhibitors.

BACKGROUND: Real-world data enables evaluation of immune checkpoint inhibitor (ICI) use in advanced melanoma management. We examined characteristics and outcomes of ICI-treated patients with advanced melanoma and organ dysfunction (baseline and emergent). MATERIALS AND METHODS: This retrospective observational study used electronic health records derived from a nationwide data set to examine advanced melanoma patients treated with first-line ICIs (2011-2018). Clinical characteristics, real-world time to treatment discontinuation (rwTTD), and overall survival (OS) were analyzed for patients with normal organ function and those with organ dysfunction prior to ICI initiation. Patients with emergent dysfunction in the 90 days following ICI initiation were identified, and potentially associated characteristics were explored. RESULTS: Of 2,407 patients included, 1,884 and 1,717 had evaluable renal and hepatic laboratory values, respectively. Patients with baseline renal dysfunction (2.4%) were older and more frequently male, and less frequently treated with ICI combinations, than patients with normal renal function. Patients with baseline hepatic dysfunction (2.8%) were similar to patients with normal hepatic function regarding demographics and treatments received. Patients with baseline organ dysfunction displayed shorter rwTTD and OS. Among patients with normal baseline organ function, 4.6% and 7.4% developed renal and hepatic dysfunction within 90 days of ICI initiation, respectively; this was associated with combination ICI treatment. CONCLUSION: Patients with advanced melanoma and baseline organ dysfunction frequently receive ICI treatment but have poorer clinical outcomes than patients with normal organ function. Among patients with normal renal and hepatic function at ICI initiation, emergent organ dysfunction rates in this real-world cohort are similar to those reported in clinical trials. IMPLICATIONS FOR PRACTICE: Real-world data provide an opportunity to understand treatment patterns, toxicity, and clinical outcomes among patients treated outside of clinical trials. This study confirms that patients with advanced melanoma and baseline renal or hepatic dysfunction are being treated with ICI therapy more frequently as monotherapy than in combination therapy. For those real-world patients with normal baseline organ function, emergent renal and hepatic dysfunction are both more common in patients treated with combination versus ICI monotherapy.

Using Propensity Score Analysis of Survey Data to Estimate Population Average Treatment Effects: A Case Study Comparing Different Methods.

BACKGROUND: Many studies in psychological and educational research aim to estimate population average treatment effects (PATE) using data from large complex survey samples, and many of these studies use propensity score methods. Recent advances have investigated how to incorporate survey weights with propensity score methods. However, to this point, that work had not been well summarized, and it was not clear how much difference the different PATE estimation methods would make empirically. PURPOSE: The purpose of this study is to systematically summarize the appropriate use of survey weights in propensity score analysis of complex survey data and use a case study to empirically compare the PATE estimates using multiple analysis methods that include ordinary least squares regression, weighted least squares regression, and various propensity score applications. METHODS: We first summarize various propensity score methods that handle survey weights. We then demonstrate the performance of various analysis methods using a nationally representative data set, the Early Childhood Longitudinal Study-Kindergarten to estimate the effects of preschool on children's academic achievement. The correspondence of the results was evaluated using multiple criteria. RESULTS AND CONCLUSIONS: It is important for researchers to think carefully about their estimand of interest and use methods appropriate for that estimand. If interest is in drawing inferences to the survey target population, it is important to take the survey weights into account, particularly in the outcome analysis stage for estimating the PATE. The case study shows, however, not much difference among various analysis methods in one applied example.

Primary tumor location and survival in colorectal cancer: A retrospective cohort study.

BACKGROUND: Primary tumor location is a prognostic factor for metastatic colorectal cancer (mCRC). Post hoc analyses of mCRC clinical trials, including FIRE-3, CALGB/SWOG 80405, suggest that primary tumor location is also predictive of survival benefit with cetuximab or bevacizumab in combination with 5-fluorouracil-based chemotherapy. AIM: Evaluate prognostic/predictive roles of primary tumor location in real-world mCRC patients treated with cetuximab or bevacizumab plus 5-fluorouracil-based chemotherapy. METHODS: This retrospective cohort study selected patients with KRAS wild-type mCRC who initiated first-line therapy with cetuximab or bevacizumab in combination with 5-fluorouracil/leucovorin/irinotecan (FOLFIRI) or 5-fluorouracil/ leucovorin/oxaliplatin (FOLFOX) between January 2013 and April 2017 from the Flatiron Health electronic health record-derived database of de-identified patient-level data in the United States. Primary tumor location was abstracted from patients' charts. Left-sided primary tumor location (LPTL) was defined as tumors that originated in the splenic flexure, descending colon, sigmoid colon, or rectum; right-sided primary tumor location (RPTL) was defined as tumors that originated from the appendix, cecum, ascending colon, hepatic flexure, or transverse colon. Propensity score matching was used to balance the baseline demographic and clinical characteristics between patients treated with cetuximab and patients treated with bevacizumab. Kaplan-Meier and Cox regression methods were used for survival analyses. RESULTS: A total of 1312 patients met the selection criteria. Of 248 cetuximab plus FOLFIRI or FOLFOX patients, 164 had LPTL and 84 had RPTL; of 1064 bevacizumab plus FOLFIRI or FOLFOX patients, 679 had LPTL and 385 had RPTL. Cetuximab LPTL and RPTL patients were more likely to receive FOLFIRI vs bevacizumab patients (LPTL: 64.0% vs 24.3%; RPTL: 76.2% vs 24.9%, P < 0.001). Stage at initial diagnosis was different between cetuximab RPTL vs bevacizumab RPTL patients (P < 0.001); cetuximab RPTL patients were more likely to have stage III disease (44.0% vs 22.6%), while bevacizumab RPTL patients were more likely to have stage IV disease (65.7% vs 48.8%). Cetuximab RPTL patients were more likely to have a documented history of adjuvant chemotherapy vs bevacizumab RPTL patients (47.6% vs 22.3%, P < 0.001). In the propensity score-matched sample, median overall survival (OS) was 29.7 mo (95%CI: 26.9-35.2) for LPTL patients vs 18.3 mo (95%CI: 15.8-21.3) for RPTL patients (P < 0.001). Median OS was 29.7 mo (95%CI: 27.4-NA) for cetuximab LPTL patients vs 29.1 mo (95%CI: 26.6-35.6) for bevacizumab LPTL patients (HR = 0.87; 95%CI: 0.63-1.19; P = 0.378) and 17.0 mo (95%CI: 12.0-32.6) for cetuximab RPTL patients vs 18.8 mo (95%CI: 15.8-22.3) for bevacizumab RPTL patients (HR = 1.00; 95%CI: 0.68-1.46; P = 0.996). The interaction of treatment and primary tumor location was not significant in the Cox regression. CONCLUSION: In this real-world mCRC cohort, the prognostic role of primary tumor location was substantiated, but not the predictive role for treatment with cetuximab vs bevacizumab in combination with 5-fluorouracil-based chemotherapy.

It's all about balance: propensity score matching in the context of complex survey data.

Many research studies aim to draw causal inferences using data from large, nationally representative survey samples, and many of these studies use propensity score matching to make those causal inferences as rigorous as possible given the non-experimental nature of the data. However, very few applied studies are careful about incorporating the survey design with the propensity score analysis, which may mean that the results do not generate population inferences. This may be because few methodological studies examine how to best combine these methods. Furthermore, even fewer of them investigate different non-response mechanisms. This study examines methods for handling survey weights in propensity score matching analyses of survey data under different non-response mechanisms. Our main conclusions are: (i) whether the survey weights are incorporated in the estimation of the propensity score does not impact estimation of the population treatment effect, as long as good population treated-comparison balance is achieved on confounders, (ii) survey weights must be used in the outcome analysis, and (iii) the transferring of survey weights (i.e., assigning the weights of the treated units to the comparison units matched to them) can be beneficial under certain non-response mechanisms.

Measuring Model Misspecification: Application to Propensity Score Methods with Complex Survey Data.

Model misspecification is a potential problem for any parametric-model based analysis. However, the measurement and consequences of model misspecification have not been well formalized in the context of causal inference. A measure of model misspecification is proposed, and the consequences of model misspecification in non-experimental causal inference methods are investigated. The metric is then used to explore which estimators are more sensitive to misspecification of the outcome and/or treatment assignment model. Three frequently used estimators of the treatment effect are considered, all of which rely on the propensity score: (1) full matching, (2) 1:1 nearest neighbor matching, and (3) weighting. The performance of these estimators is evaluated under two different sampling designs: (1) simple random sampling (SRS) and (2) a two-stage stratified survey. As the degree of misspecification of either the propensity score or outcome model increases, so does the bias and the root mean square error, while the coverage decreases. Results are similar for the simple random sample and a complex survey design.

An imputation-based solution to using mismeasured covariates in propensity score analysis.

Although covariate measurement error is likely the norm rather than the exception, methods for handling covariate measurement error in propensity score methods have not been widely investigated. We consider a multiple imputation-based approach that uses an external calibration sample with information on the true and mismeasured covariates, multiple imputation for external calibration, to correct for the measurement error, and investigate its performance using simulation studies. As expected, using the covariate measured with error leads to bias in the treatment effect estimate. In contrast, the multiple imputation for external calibration method can eliminate almost all the bias. We confirm that the outcome must be used in the imputation process to obtain good results, a finding related to the idea of congenial imputation and analysis in the broader multiple imputation literature. We illustrate the multiple imputation for external calibration approach using a motivating example estimating the effects of living in a disadvantaged neighborhood on mental health and substance use outcomes among adolescents. These results show that estimating the propensity score using covariates measured with error leads to biased estimates of treatment effects, but when a calibration data set is available, multiple imputation for external calibration can be used to help correct for such bias.

Advanced-stage hepatocellular carcinoma with portal vein thrombosis: conventional versus drug-eluting beads transcatheter arterial chemoembolization.

OBJECTIVES: Our study sought to compare the overall survival in patients with hepatocellular carcinoma (HCC) and portal venous thrombosis (PVT), treated with either conventional trans-arterial chemoembolization (cTACE) or drug-eluting beads (DEB) TACE. METHODS: This retrospective analysis included a total of 133 patients, treated without cross-over and compared head-to-head by means or propensity score weighting. Mortality was compared using survival analysis upon propensity score weighting. Adverse events and liver toxicity grade ≥3 were recorded and reported for each TACE. In order to compare with historical sorafenib studies, a sub-group analysis was performed and included patients who fulfilled the SHARP inclusion criteria. RESULTS: The median overall survival (MOS) of the entire cohort was 4.53 months (95 % CI, 3.63-6.03). MOS was similar across treatment arms, no significant difference between cTACE (N = 95) and DEB-TACE (N = 38) was observed (MOS of 5.0 vs. 3.33 months, respectively; p = 0.157). The most common adverse events after cTACE and DEB- TACE, respectively, were as follows: post-embolization syndrome [N = 57 (30.0 %) and N = 38 (61.3 %)], diarrhea [N = 3 (1.6 %) and N = 3 (4.8 %)], and encephalopathy [N = 11 (5.8 %) and N = 2 (3.2 %)]. CONCLUSION: Our retrospective study could not reveal a difference in toxicity and efficiency between cTACE and DEB-TACE for treatment of advanced stage HCC with PVT. KEY POINTS: • Conventional TACE (cTACE) and drug-eluting-beads TACE (DEB-TACE) demonstrated equal safety profiles. • Survival rates after TACE are similar to patients treated with sorafenib. • Child-Pugh class and tumor burden are reliable predictors of survival.

A doubly robust estimator for the average treatment effect in the context of a mean-reverting measurement error.

One of the main limitations of causal inference methods is that they rely on the assumption that all variables are measured without error. A popular approach for handling measurement error is simulation-extrapolation (SIMEX). However, its use for estimating causal effects have been examined only in the context of an additive, non-differential, and homoscedastic classical measurement error structure. In this article we extend the SIMEX methodology, in the context of a mean reverting measurement error structure, to a doubly robust estimator of the average treatment effect when a single covariate is measured with error but the outcome and treatment and treatment indicator are not. Throughout this article we assume that an independent validation sample is available. Simulation studies suggest that our method performs better than a naive approach that simply uses the covariate measured with error.