Search

Controlling the bias is central to estimating semiparametric models. Many methods have been developed to control bias in estimating conditional expectations while maintaining a desirable variance order. However, these methods typically do not perform well at moderate sample sizes. Moreover, and perhaps related to their performance, nonoptimal windows are selected with undersmoothing needed to ensure the appropriate bias order. In this paper, we propose a recursive differencing estimator for conditional expectations. When this method is combined with a bias control targeting the derivative of the semiparametric expectation, we are able to obtain asymptotic normality under optimal windows. As suggested by the structure of the recursion, in a wide variety of triple index designs, the proposed bias control performs much better at moderate sample sizes than regular or higher-order kernels and local polynomials.

Introduction Many states have mandated breast density notification and insurance coverage for additional screening; yet, the association between such legislation and stage of diagnosis for breast cancer is unclear. This study investigates this association and examines the differential impacts among different age and race/ethnicity subgroups. Methods The Surveillance, Epidemiology, and End Results database was queried to identify patients with breast cancer aged 40–74 years diagnosed between 2005 and 2016. Using a difference-in-differences multinomial logistic model, the odds of being diagnosed at different stages of cancer relative to the localized stage depending on legislation and individual characteristics were examined. Analyses were conducted in 2020–2021. Results The study included 689,641 cases. Overall, the impact of notification legislation was not significant, whereas insurance coverage legislation was associated with 6% lower odds (OR=0.94, 95% CI=0.91, 0.96) of being diagnosed at the regional stage. The association between insurance coverage legislation and stage of diagnosis was even stronger among women aged 40–49 years, with 11% lower odds (OR=0.89, 95% CI=0.82, 0.96) of being diagnosed at the regional stage and 12% lower odds (OR=0.88, 95% CI=0.81, 0.96) of being diagnosed at the distant stage. Hispanic women benefited from notification laws, with 11% lower odds (OR=0.89, 95% CI=0.82, 0.97) of being diagnosed at distant stage. Neither notification nor supplemental screening insurance coverage legislation showed a substantial impact on Black women. Conclusions The findings imply that improving insurance coverage is more important than being notified overall. Raising awareness is important among Hispanic women; improving communication about dense breasts and access to screening might be more important than legislation among Black women.

Background Contemporary to the rapidly evolving landscape of cancer immunotherapy is the equally changing understanding of immune tumor microenvironments (TMEs) which is crucial to the success of these therapies. Their reliance on a robust host immune response necessitates clinical grade measurements of immune TMEs at diagnosis. In this study, we describe a stable tumor immunogenic profile describing immune TMEs in multiple tumor types with ability to predict clinical benefit from immune checkpoint inhibitors (ICIs). Methods A tumor immunogenic signature (TIGS) was derived from targeted RNA-sequencing (RNA-seq) and gene expression analysis of 1323 clinical solid tumor cases spanning 35 histologies using unsupervised analysis. TIGS correlation with ICI response and survival was assessed in a retrospective cohort of NSCLC, melanoma and RCC tumor blocks, alone and combined with TMB, PD-L1 IHC and cell proliferation biomarkers. Results Unsupervised clustering of RNA-seq profiles uncovered a 161 gene signature where T cell and B cell activation, IFNg, chemokine, cytokine and interleukin pathways are over-represented. Mean expression of these genes produced three distinct TIGS score categories: strong (n = 384/1323; 29.02%), moderate (n = 354/1323; 26.76%), and weak (n = 585/1323; 44.22%). Strong TIGS tumors presented an improved ICI response rate of 37% (30/81); with highest response rate advantage occurring in NSCLC (ORR = 36.6%; 16/44; p = 0.051). Similarly, overall survival for strong TIGS tumors trended upward (median = 25 months; p = 0.19). Integrating the TIGS score categories with neoplastic influence quantified via cell proliferation showed highly proliferative and strong TIGS tumors correlate with significantly higher ICI ORR than poorly proliferative and weak TIGS tumors [14.28%; p = 0.0006]. Importantly, we noted that strong TIGS and highly [median = not achieved; p = 0.025] or moderately [median = 16.2 months; p = 0.025] proliferative tumors had significantly better survival compared to weak TIGS, highly proliferative tumors [median = 7.03 months]. Importantly, TIGS discriminates subpopulations of potential ICI responders that were considered negative for response by TMB and PD-L1. Conclusions TIGS is a comprehensive and informative measurement of immune TME that effectively characterizes host immune response to ICIs in multiple tumors. The results indicate that when combined with PD-L1, TMB and cell proliferation, TIGS provides greater context of both immune and neoplastic influences on the TME for implementation into clinical practice.

BackgroundImmune checkpoint inhibitor-based therapies have achieved impressive success in the treatment of several cancer types. Predictive immune biomarkers, including PD-L1, MSI and TMB are well established as surrogate markers for immune evasion and tumor-specific neoantigens across many tumors. Positive detection across cancer types varies, but overall ~50% of patients test negative for these primary immune markers.1 In this study, we investigated the prevalence of secondary immune biomarkers outside of PD-L1, TMB and MSI.MethodsComprehensive genomic and immune profiling, including PD-L1 IHC, TMB, MSI and gene expression of 395 immune related genes was performed on 6078 FFPE tumors representing 34 cancer types, predominantly composed of lung cancer (36.7%), colorectal cancer (11.9%) and breast cancer (8.5%). Expression levels by RNA-seq of 36 genes targeted by immunotherapies in solid tumor clinical trials, identified as secondary immune biomarkers, were ranked against a reference population. Genes with a rank value ≥75th percentile were considered high and values were associated with PD-L1 (positive ≥1%), MSI (MSI-H or MSS) and TMB (high ≥10 Mut/Mb) status. Additionally, secondary immune biomarker status was segmented by tumor type and cancer immune cycle roles.ResultsIn total, 41.0% of cases were PD-L1+, 6.4% TMB+, and 0.1% MSI-H. 12.6% of cases were positive for >2 of these markers while 39.9% were triple negative (PD-L1-/TMB-/MSS). Of the PD-L1-/TMB-/MSS cases, 89.1% were high for at least one secondary immune biomarker, with 69.3% having ≥3 markers. PD-L1-/TMB-/MSS tumor types with ≥50% prevalence of high secondary immune biomarkers included brain, prostate, kidney, sarcoma, gallbladder, breast, colorectal, and liver cancer. High expression of cancer testis antigen secondary immune biomarkers (e.g., NY-ESO-1, LAGE-1A, MAGE-A4) was most commonly observed in bladder, ovarian, sarcoma, liver, and prostate cancer (≥15%). Tumors demonstrating T-cell priming (e.g., CD40, OX40, CD137), trafficking (e.g., TGFB1, TLR9, TNF) and/or recognition (e.g., CTLA4, LAG3, TIGIT) secondary immune biomarkers were most represented by kidney, gallbladder, and sarcoma (≥40%), with melanoma, esophageal, head & neck, cervical, stomach, and lung cancer least represented (≥15%).ConclusionsOur studies show comprehensive tumor profiling that includes gene expression can detect secondary immune biomarkers targeted by investigational therapies in ~90% of PD-L1-/TMB-/MSS cases. While genomic profiling could also provide therapeutic choices for a percentage of these patients, detection of secondary immune biomarkers by RNA-seq provides additional options for patients without a clear therapeutic path as determined by PD-L1 testing and genomic profiling alone.ReferenceHuang R S P, Haberberger J, Severson E, et al. A pan-cancer analysis of PD-L1 immunohistochemistry and gene amplification, tumor mutation burden and microsatellite instability in 48,782 cases. Mod Pathol 2021;34: 252–263.

BACKGROUND Non-small cell lung cancer (NSCLC) accounts for ~50% of brain metastases. We present the genomic and immune biomarker landscape for a cohort with metastatic NSCLC to the brain. METHODS We analyzed brain metastases FFPE tissue (n=137; ages 40-85y (mean 65y), 52% female, 48% male) vs. primary sites (n=5533; ages 24-100+y (mean 71y), 51% female, 49% male) for advanced or metastatic NSCLC patients with comprehensive genomic and immune biomarker profiling, including PD-L1 IHC, tumor mutational burden (TMB), Tumor Immunogenic Signature (TIGS), Cell Proliferation (CP), and Cancer Testis Antigen Burden (CTAB). RESULTS Genomic alteration (GA) frequency for NSCLC brain metastasis versus primary sites were similar among the most frequently mutated genes except for KRAS which was significantly higher among brain metastases (39.9% vs 25.5%, p< 0.0005). No significant differences were observed for TP53 (50.7% vs 50.2%), STK11 (11.5% vs 10.9%), CDKN2A (10.1% vs 7%), or EGFR (8.0% vs 11.5%). PD-L1 expression for all cases by IHC was not significantly different (mean TPS 29.9% vs 26.3%); however, brain metastases were more likely to be PD-L1 negative (TPS< 1%) (46.3% vs 33.3%, p< 0.005). Conversely, TMB was higher in brain metastases versus primary sites (12.9 vs 10.2 muts/MB, p< 8x10-10), with more TMB-high cases (10 muts/Mb) (57.0% vs 33.5%, p< 3x10-8). Similarly, brain metastases had a higher proportion of CTAB-high cases (68.6% vs 57.6%, p=0.01). The immune response biomarker score, TIGS, was more often weak in brain metastases (52.6% vs 34.0%, p< 9x10-6). CONCLUSION In NSCLC, metastatic brain lesions have a larger antigen burden, with increased TMB and CTAB, likely due to the immune privileged nature of the brain, which is reflected in the lower TIGS scores and PD-L1 positivity. Despite lower PD-L1 positivity, NSCLC brain metastases with negative PD-L1 IHC may potentially benefit from immunotherapy strategies given the high TMB and CTAB.

Timely and accurate identification of molecular alterations in solid tumors is essential for proper management of patients with advanced cancers. This has created a need for rapid, scalable comprehensive genomic profiling (CGP) systems that detect an increasing number of therapeutically-relevant variant types and molecular signatures. In this study, we assessed the analytical performance of the TruSight Oncology 500 High-Throughput assay for detection of somatic alterations from formalin-fixed paraffin-embedded tissue specimens. In parallel, we developed supporting software and automated sample preparation systems designed to process up to 70 clinical samples in a single NovaSeq 6000TM sequencing run with a turnaround time of 99% overall accuracy and precision. Our results demonstrate that the high-throughput CGP assay is a reliable method for accurate detection of molecular alterations in support of precision therapeutics in oncology. The supporting systems and scalable workflow allow for efficient interpretation and prompt reporting of hundreds of patient cancer genomes per week with excellent analytical performance.

We propose instrumental variables (IV) estimators for averaged conditional treatment effects and the parameters upon which they depend in the context of a semiparametric outcome model with endogenous discrete treatment variables. For this model, the treatment impacts are unknown functions of a vector of indices that depend on a finite dimensional parameter vector. We develop the theory for an estimator of these impacts when they are averaged over regions of interest. We prove identification, consistency and N-asymptotic normality of the estimators. We also show that they are efficient under correct model specification. Further, we show that they are robust to misspecification of the propensity score model. In the Monte Carlo study, the estimators perform well over a wide variety of designs covering both correct and incorrect propensity score model specification.

Purpose of Study: Cancer testis antigens (CTA) are highly immunogenic genes with the ability to cause cancer-specific immune responses when expressed. Their tumor cell-specific expression makes them a key target of natural T cell response, cancer vaccines, immune checkpoint blockade (ICB), and cell-based immunotherapies in a wide range of tumor types. In this study, we assess the pan-cancer distribution and ICB survival association of CTA burden (CTAB) in real-world solid tumors. Procedure: Three tumor sample cohorts were studied: 1) a pan-cancer discovery cohort to develop a low- and high-CTAB cutoff (n=5450, 39 tumor types), 2) a TCGA cohort (n=19923, 32 tumor types) used to validate the classifier based on CTAB distribution and serve as a non-ICB-treated population, and 3) an ICB-treated retrospective cohort to validate the classification on overall survival (OS) (n=242, 3 tumor types). The expression levels of 17 CTA were measured using targeted RNA-Seq of FFPE tumor samples and then ranked against a pan-cancer reference population. CTAB was calculated for each sample, cohort and tumor type as the sum of the 17 CTA gene expression ranks. The discovery cohort median CTAB of 171 was used to classify all three cohorts into high- and low-CTAB groups. OS analysis was performed on the TCGA and ICB-treated cohorts using a CoxPH regression model to determine the Hazard Ratio (HR). Results: The three cohorts demonstrated overlapping single-peak, left-skewed CTAB distribution curves centered at CTAB values between 170 (discovery cohort) and 256 (retrospective cohort). When grouping by tumor types and ordering by median CTAB, the CTAB distributions for tumor types within all three cohorts were comparable. CoxPH regression analysis revealed an association between the CTAB threshold classifier and OS in both the ICB-treated retrospective and non-ICB TCGA cohorts. However, the direction of this association differed between the two cohorts, with high-CTAB samples having better survival (HR=0.936, p=0.076) in the ICB-treated retrospective cohort and worse survival (HR: 1.007, p=0.084) in the non-ICB-treated cohort. Conclusion: Our studies show that the CTAB distribution was maintained across the discovery and TCGA cohorts and a wide range of tumor types, supporting that the CTAB classifier is valid and histology agnostic. Additionally, when evaluating the ICB and non-ICB-treated cohorts, CTAB demonstrated the ability to predict OS, pointing to the utility of ICB in supporting CTA-specific natural immune response. However, further studies are necessary to verify these mechanisms of response to ICB as well as cancer vaccines and cell-based immunotherapies. Additional validation is needed to establish the predictive utility of CTAB alone and in combination with other immune oncology biomarkers for resistance or response. Citation Format: R.J. Seager, Erik Van Roey, Shuang Gao, Blake Burgher, Paul DePietro, Mary Nesline, Roger Klein, Shengle Zhang, Jeffrey M. Conroy, Sarabjot Pabla. Cancer testis antigen burden: Pan-cancer distribution and survival implications [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5137.

PURPOSE: The immunohistochemistry companion diagnostic test for pembrolizumab (IHC 22C3 pharmDx) lacks sensitivity, challenging immunotherapy selection for NSCLC patients with lower levels of expression. Unlike IHC 22C3, which restricts assessment of PD-L1 expression to viable tumor cells as a tumor proportion score (% TPS), mRNA next generation sequencing (RNA-seq) measures PD-L1 expression in the tumor microenvironment for both tumor and inflammatory background cells. RNA-seq previously demonstrated concordance with IHC and may be a robust alternative testing method for multiple tumor types. Here, we sought to optimize PD-L1 RNA-seq cutoff values in NSCLC to improve clinical sensitivity. PROCEDURE: NSCLC patients included in the study (n=3,465) were tested for PD-L1 expression by IHC 22C3 and clinically validated RNA-seq, measured as % rank (0-100) relative to a reference population based on normalized reads per million (nRPM). Patients were divided into an RNA-seq cut-off discovery cohort (n=3,168), and a test cohort pembrolizumab treated patients. Principal components analysis (PCA) was used to classify patients based on test results and explore cut-off values in the discovery cohort. Kaplan Meier curves and a Cox proportional hazards regression models assessed overall survival (OS) hazard ratios (HR) for RNA-seq versus standard of care IHC cut-offs in the test cohort. RESULTS: Unsupervised PCA clustering identified three distinct PD-L1 groups separated by combinations of significant over- and under-representation of RNA-seq and IHC result measures from prior testing. The groups were labeled as “low” (rank ≤40), “moderate” (rank 41-73), and “high” (rank ≥74), based on the median RNA-seq rank for each group (+/- 1SD for low and high). Both the low and moderate groups were overrepresented by patients in the PD-L1 IHC low and negative groups. The moderate group was overrepresented by patients with moderately high PD-L1 RNA-seq ranks (median=70), while the low group was overrepresented by patients that were not PD-L1 high by RNA-seq. The high group was overrepresented by patients high for PD-L1 by both IHC and RNA-seq. OS HRs were better for RNA-seq high versus moderate (HR=0.05, CI 0.00-0.63, p=.02), and RNA-seq high versus low (HR=0.16, CI 0.03-0.86, p=.03) groups compared to standard of care IHC 22C3 high versus low groups, (HR=0.21, CI 0.04-1.07, p=.06). Findings were non-significant for the RNA-seq moderate versus low groups, likely due to the limited and disproportionately high number of patients with poor performance status in these groups. CONCLUSIONS: PD-L1 expression by RNA-seq demonstrated improved clinical sensitivity in predicting OS versus standard of care PD-LI IHC in a pembrolizumab treated NSCLC patient cohort. Additional studies are needed to further define cut-offs in the context of performance status, and better understand immune escape mechanisms in the moderate group. Citation Format: Mary K. Nesline, Sarabjot Pabla, Yong Hee Lee, Paul DePietro, Amy Early, Roger Klein, Shengle Zhang, Jeffrey Conroy. PD-L1 expression by RNA-sequencing and survival from pembrolizumab in non-small cell lung cancer (NSCLC) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1259.

BackgroundCancer of unknown primary (CUP) is a rare tumor type accounting for 2% of solid cancers. In the subset of CUP cases where tumor of origin is posited and treated as such, no clear clinical benefit has been demonstrated. Furthermore, CUP patients treated by empiric platinum-based regimens have low response and survival rates of approximately 20%.1 2 Support of tissue-agnostic marker-directed immunotherapy is growing because it targets the immune system rather than the tumor, with some efficacy evidence emerging for CUP.3 Identifying new targets for immunotherapeutic opportunities in this heterogeneous and difficult to treat patient group is a critical unmet need.MethodsComprehensive genomic and immune marker profiling by NGS4 was performed on FFPE tissue for CUP tumors (n=298) as indicated by physicians’ test orders from >100 clinical practice sites. Histology was verified by a molecular pathologist as part of pre-analytic test quality control, with cases representing tumors with adenocarcinoma (58%), carcinoma (26%), squamous (10%), and neuroendocrine (6%) histologic features. RNA-expression levels of immune genes that are current targets in non-CUP immunotherapy clinical trials (n=36) were ranked against a reference population (≥75th percentile=high), and described by histologic type, along with PD-L1 IHC (22C3) expression, tumor mutational burden (TMB) and genomic variants.Results90% of all CUP tumors had at least 1 highly expressed immune gene target in active immunotherapy trials, with the most frequent being TGFB1 (47%) and CCL2 (39%). 55% of CUP tumors were PD-L1 IHC 22C3 positive (>=1% TPS), and 21% had high TMB (>=10 mut/Mb) in CUP tumors with neuroendocrine (32%), carcinoma (30%), squamous cell (21%), and adenocarcinoma (17%) histologic features. Overall, 26% of CUP patient tumors, mostly adenocarcinomas (28%) and carcinomas (27%), harbored genomic variants (n=77) with FDA approved targeted therapies in other tumor types. The most frequently immunogenic CUP tumors were carcinomas, showing high RNA-seq expression of 26/36 genes in at least 20% of patients, most represented by CD20, CD27, TLR8, and PD-L1. High expression of CD40, CSF1R, TIM3, and VISTA was most common in adenocarcinomas. Squamous cell carcinomas were relatively immunogenic, with frequent high expression of 17/36 immune genes, uniquely including MAGEA4. Neuroendocrine tumors were the least immunogenic, with frequent high expression in only 4/36 genes, including ADORA2A (42%) and MAGEA1 (37%).ConclusionsCUP tumors diversely express both standard marker and novel immunotherapeutic targets based on histology and may benefit from selective access to clinical trials for these therapies.ReferencesNational Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®) Occult Primary (Cancer of Unknown Primary [CUP]), Version 2.2021. Fort Washington, Pennsylvania: National Comprehensive Cancer Network; 2021. https://www.nccn.org/professionals/physician_gls/pdf/occult.pdf.Laprovitera N, Riefolo M, Ambrosini E, Klec C, Pichler M, Ferracin M. Cancer of unknown primary: Challenges and progress in clinical management. Cancers (Basel) 2021;13(3):1–30. doi:10.3390/cancers130304513.Naing A, Meric-Bernstam F, Stephen B, et al. Phase 2 study of pembrolizumab in patients with advanced rare cancers. J Immunother Cancer 2020;8(1):e000347. doi:10.1136/jitc-2019-0003474.Conroy JM, Pabla S, Glenn ST, et al. Analytical validation of a next-generation sequencing assay to monitor immune responses in solid tumors. J Mol Diagnostics 2018;20(1):95–109. doi:10.1016/j.jmoldx.2017.10.001

Background. Robotic surgical systems are expensive to own and operate, and the purchase of such technology is an important decision for hospital administrators. Most prior literature focuses on the comparison of clinical outcomes between robotic surgery and other laparoscopic or open surgery. There is a knowledge gap about what drives hospitals’ decisions to purchase robotic systems. Objective. To identify factors associated with a hospital’s acquisition of advanced surgical systems. Method. We used 2002 to 2011 data from the State of California Office of Statewide Health Planning and Development to examine robotic surgical system purchase decisions of 476 hospitals. We used a probit estimation allowing heteroscedasticity in the error term including a set of two equations: one binary response equation and one heteroscedasticity equation. Results. During the study timeframe, there were 78 robotic surgical systems purchased by hospitals in the sample. Controlling for hospital characteristics such as number of available beds, teaching status, nonprofit status, and patient mix, the probit estimation showed that market-level directly relevant surgery volume in the previous year (excluding the hospital’s own volume) had the largest impact. More specifically, hospitals in high volume (>50,000 surgeries v. 0) markets were 12 percentage points more likely to purchase robotic systems. We also found that hospitals in less competitive markets (i.e., Herfindahl index above 2500) were 2 percentage points more likely to purchase robotic systems. Limitations. This study has limitations common to observational database studies. Certain characteristics such as cultural factors cannot be accurately quantified. Conclusions. Our findings imply that potential market demand is a strong driver for hospital purchase of robotic surgical systems. Market competition does not significantly increase the adoption of new expensive surgical technologies.

This paper addresses the estimation of a semiparametric sample selection index model where both the selection rule and the outcome variable are binary. Since the marginal effects are often of primary interest and are difficult to recover in a semiparametric setting, we focus on developing an estimator for the marginal effects. This marginal effect estimator uses only observations where the selection probability is above a certain threshold. A key innovation is that this high probability set is adaptive to the data. We establish the large sample properties of the marginal effect estimator as well as those for an index estimator upon which it depends. Monte Carlo studies show that these estimators perform well in finite samples.

This article investigates the degree of intergenerational transmission of education for individuals from the National Longitudinal Survey of Youth 1979. Rather than identifying the causal effect of parental education via instrumental variables we exploit the feature of the transmission mechanism responsible for its endogeneity. More explicitly, we assume the intergenerational transfer of unobserved ability is invariant to the economic environment. This, combined with the heteroskedasticity resulting from the interaction of unobserved ability with socioeconomic factors, identifies the causal effect. We conclude that the observed intergenerational educational correlation reflects both a causal parental educational effect and a transfer of unobserved ability.

Semiparametric methods are widely employed in applied work where the ability to conduct inferences is important. To establish asymptotic normality for making inferences, bias control mechanisms are often used in implementing semiparametric estimators. The first contribution of this paper is to propose a mechanism that enables us to establish asymptotic normality with regular kernels. In so doing, we argue that the resulting estimator performs very well in finite samples.Semiparametric models are commonly estimated under a single index assumption. Because the consistency of the estimator critically depends on this assumption being correct, our second objective is to develop a test for it. To ensure that the test statistic has good size and power properties in finite samples, we employ a bias control mechanism similar to that underlying the estimator. Furthermore, we structure the test so that its form adapts to the model under the alternative hypothesis. Monte Carlo results confirm that the bias control and the adaptive feature significantly improve the performance of the test statistic in finite samples.

This paper provides a control function estimator to adjust for endogeneity in the triangular simultaneous equations model where there are no available exclusion restrictions to generate suitable instruments. Our approach is to exploit the dependence of the errors on exogenous variables (e.g. heteroscedasticity) to adjust the conventional control function estimator. The form of the error dependence on the exogenous variables is subject to restrictions, but is not parametrically specified. In addition to providing the estimator and deriving its large-sample properties, we present simulation evidence which indicates the estimator works well.

SUMMARY This paper formulates a likelihood-based estimator for a double-index, semiparametric binary response equation. A novel feature of this estimator is that it is based on density estimation under local smoothing. While the proofs differ from those based on alternative density estimators, the finite sample performance of the estimator is significantly improved. As binary responses often appear as endogenous regressors in continuous outcome equations, we also develop an optimal instrumental variables estimator in this context. For this purpose, we specialize the double-index model for binary response to one with heteroscedasticity that depends on an index different from that underlying the ‘mean response’. We show that such (multiplicative) heteroscedasticity, whose form is not parametrically specified, effectively induces exclusion restrictions on the outcomes equation. The estimator developed exploits such identifying information. We provide simulation evidence on the favorable performance of the estimators and illustrate their use through an empirical application on the determinants, and affect, of attendance at a government-financed school. Copyright  2009 John Wiley & Sons, Ltd.

This paper employs conditional second moments to identify the impact of education in wage regressions where education is treated as endogenous. This approach avoids the use of instrumental variables in a setting where instruments are frequently not available. We employ this methodology to estimate the returns to schooling for a sample of Australian workers. We find that accounting for the endogeneity of education in this manner increases the estimated return to education from 6 percent to 10 percent.

We propose instrumental variable(IV) estimators for quantile marginal effects and the parameters upon which they depend in a semiparametric outcome model with endogenous discrete treatment variables. We prove identification, consistency, and asymptotic normality of the estimators. We also show that they are efficient under correct model specification. Further, we show that they are robust to misspecification of the treatment model in that consistency and asymptotic normality continue to hold in this case. In the Monte Carlo study, the estimators perform well over diverse designs covering both correct and incorrect treatment model specifications.

BACKGROUND: Hepatocellular carcinoma is one of the most common malignant tumors worldwide. Imaging techniques, specially computed tomography and ultrasound, are among the most useful diagnostic tools, although the accuracy of these methods may have a significant variability. AIMS: To determine the prevalence of hepatocellular carcinoma in cirrhotic patients undergoing orthotopic liver transplantation at "Santa Casa de Misericórdia" of Porto Alegre, RS, Brazil; to estimate the sensitivity of computed tomography and ultrasound in pretransplantation detection of hepatocellular carcinoma in this population; to correlate the radiological characteristics with anatomopathological findings. MATERIALS AND METHODS: Retrospective prevalence study. Population: adult, cirrhotic patients undergoing orthotopic liver transplantation from January 1990 to July 2003. Among the 292 transplanted patients, 31 cases of hepatocellular carcinoma were diagnosed, of which 29 were included in the study. Tumor characteristics in both ultrasound and computed tomography were compared to those observed in anatomopathological examination. RESULTS: Prevalence of hepatitis C virus infection among patients with diagnosis of hepatocellular carcinoma was 93.5%, and the prevalence of hepatocellular carcinoma among transplanted patients was 10.6%. The overall sensitivity of the imaging techniques was 70.3% for computed tomography and 72% for ultrasound. CONCLUSION: The prevalence of hepatocellular carcinoma at our institution, as well as the sensitivity of both ultrasound and computed tomography to detect such tumors at pretransplantation screening were similar to those found by other authors, while the prevalence of hepatitis C virus infection, the most common etiological agent for liver disease in our patients, is one of the highest ever reported in literature. Factors influencing hepatocellular carcinoma detection rates were: time from examination to liver transplantation; acquisition of computed tomography images during arterial phase; lesion size. Arterial phase proved to be the most useful part of computed tomography examination in this study.RACIONAL: O carcinoma hepatocelular é um dos tumores malignos mais comuns em todo o mundo. Exames de imagens, especialmente tomografia computadorizada e ultra-sonografia, estão entre as principais técnicas diagnósticas, embora a acurácia destes métodos possa apresentar significativa variabilidade. OBJETIVOS: Determinar a prevalência de carcinoma hepatocelular em pacientes cirróticos submetidos a transplante hepático na Santa Casa de Misericórdia de Porto Alegre, RS; estimar a sensibilidade da tomografia computadorizada e da ultra-sonografia na detecção pré-transplante de carcinoma hepatocelular nesse grupo de pacientes; correlacionar características radiológicas com achados anatomopatológicos. MATERIAIS E MÉTODOS: Estudo de prevalência retrospectivo. População: pacientes adultos, cirróticos, submetidos a transplante hepático de janeiro de 1990 a julho de 2003. Entre os 292 pacientes transplantados, foi diagnosticado 31 casos de carcinoma hepatocelular, dos quais 29 foram incluídos no estudo. As características tomográficas e ecográficas dos tumores diagnosticados pré-transplante foram comparadas com as observadas em exame anatomopatológico. RESULTADOS: A prevalência da infecção pelo vírus da hepatite C nos pacientes com diagnóstico de carcinoma hepatocelular foi de 93,5% e a prevalência deste entre os pacientes transplantados foi de 10,6%. A sensibilidade dos métodos de imagem na detecção de casos de carcinoma hepatocelular foi de 70,3% para tomografia computadorizada e de 72% para ecografia. CONCLUSÃO: A prevalência de carcinoma hepatocelular na instituição onde foi desenvolvido o estudo, bem como a sensibilidade da ultra-sonografia e da tomografia computadorizada para detecção dessa neoplasia na avaliação pré-transplante foi semelhante à relatada na literatura. Em contrapartida, a prevalência de infecção pelo vírus da hepatite C, fator etiológico de hepatopatia mais freqüente nos pacientes desta série, é das maiores já relatadas. Os fatores que influenciaram as taxas de detecção de carcinoma hepatocelular foram: tempo decorrido entre realização do exame e transplante; realização de tomografia computadorizada com fase arterial; tamanho da lesão. A fase arterial provou ser a mais importante no diagnóstico de carcinoma hepatocelular neste estudo.

We develop a √n-consistent and asymptotically normal estimator of the parameters (regression coefficients and threshold points) of a semiparametric ordered response model under the assumption of independence of errors and regressors. The independence assumption implies shift restrictions allowing identification of threshold points up to location and scale. The estimator is useful in various applications, particularly in new product demand forecasting from survey data subject to systematic misreporting. We apply the estimator to assess exaggeration bias in survey data on demand for a new telecommunications service.

This paper addresses the estimation of a semiparametric sample selection index model where both the selection rule and the outcome variable are binary. Since the marginal effects are often of primary interest and are difficult to recover in a semiparametric setting, we develop estimators for both the marginal effects and the underlying model parameters. The marginal effect estimator uses only observations where the selection probability is above a certain threshold. A key innovation is that this high probability set is adaptive to the data. The model parameter estimator is a quasi-likelihood estimator based on regular kernels with bias corrections. We establish their large sample properties and provide simulation evidence confirming that these estimators perform well in finite samples.

Market researchers often conduct surveys asking respondents to estimate their future demand for new products. However, projected demand may exhibit systematic bias. For example, the more respondents like a product, the more they may exaggerate their demand. We found evidence of such exaggeration in a recent survey of demand for a potential new video product. In this paper, we develop a computationally tractable procedure that corrects for a general form of systematic bias in demand projections. This general form is characterized by a monotonictransformation of projected demand, and covers exaggeration bias as a special case.

Os autores relatam o caso de um paciente branco, de quatro anos de idade, com comprometimento neurológico progressivo. Tomografia computadorizada do crânio e ressonância magnética evidenciaram lesão expansiva no tronco cerebral. Subseqüentemente, foi feito diagnóstico histopatológico de gliossarcoma. Trata-se de um tumor raro do sistema nervoso central que, na grande maioria dos casos, acomete pacientes acima dos 40 anos de idade e tem localização supratentorial. Há poucos casos relatados de gliossarcomas em crianças, sobretudo na região infratentorial. Dados epidemiológicos, assim como achados mais freqüentes na tomografia computadorizada e ressonância magnética, são discutidos.

This paper addresses the estimation of a semiparametric sample selection index model where both the selection rule and the outcome variable are binary. Since the marginal effects are often of primary interest and are difficult to recover in a semiparametric setting, we develop estimators for both the marginal effects and the underlying model parameters. The marginal effect estimator only uses observations which are members of a high probability set in which the selection problem is not present. A key innovation is that this high probability set is data dependent. The model parameter estimator is a quasi-likelihood estimator based on regular kernels with bias corrections. We establish their large sample properties and provide simulation evidence confirming that these estimators perform well in finite samples.

In testing choice models, there are contexts in which it is important to focus on choice probabilities as opposed to the parameters of choice models. This paper examines the former by formulating a distance measure between parametric choice probabilities under the null hypothesis and either a nonparametric or a semiparametric alternative. To access where the parametric model is ‘unsatisfactory’, we apply the distance measure to ordered data subsets (‘index quantiles’). We establish the large-sample distribution of the test statistics, examine their asymptotic local power properties, and apply the tests to a model of labor force participation.

Os autores relatam o caso de uma paciente do sexo feminino, 51 anos de idade, branca, com diagnóstico estabelecido de lúpus eritematoso sistêmico associado a evidência radiológica de fibrose pulmonar grave. A epidemiologia das manifestações pleuropulmonares do lúpus eritematoso sistêmico, dando ênfase à fibrose pulmonar, é discutida.The authors report the case of a 51-year-old white female with proven diagnosis of systemic lupus erythematosus and radiological evidence of severe pulmonary fibrosis. The epidemiology of pleuropulmonary manifestations of systemic lupus erythematosus emphasizing pulmonary fibrosis is discussed.

Os autores relatam o caso de uma paciente do sexo feminino, 51 anos de idade, branca, com diagnóstico estabelecido de lúpus eritematoso sistêmico associado a evidência radiológica de fibrose pulmonar grave. A epidemiologia das manifestações pleuropulmonares do lúpus eritematoso sistêmico, dando ênfase à fibrose pulmonar, é discutida.

An innovation which bypasses the need for instruments when estimating endogenous treatment effects is identification via conditional second moments. The most general of these approaches is Klein and Vella (J Econom 154:154-164, 2010), which models the conditional variances semiparametrically. While this is attractive, as identification is not reliant on parametric assumptions for variances, the nonparametric aspect of the estimation may discourage practitioners from its use. This paper outlines how the estimator can be implemented parametrically. The use of parametric assumptions is accompanied by a large reduction in computational and programming demands. We illustrate the approach by estimating the return to education using a sample drawn from the National Longitudinal Survey of Youth 1979. Accounting for endogeneity increases the estimate of the return to education from 6. 8 to 11. 2%. © 2010 Springer-Verlag., Farré acknowledges the support of the Barcelona GSE Research Network and the Government of Catalonia, and the financial support from the Spanish Ministry of Science (Grant ECO2008-05721). This work was done while Farré held a CSIC-JaeDoc contract, cofinanced by the European Social Fund

We use panel data from the US Health and Retirement Study 1992-2002 to estimate the effect of self-assessed health limitations on active labor market participation of men around retirement age. Self-assessments of health and functioning typically introduce an endogeneity bias when studying the effects of health on labor market participation. This results from justification bias, reflecting an individual’s tendency to provide answers which "justify" his labor market activity, and individual-specific heterogeneity in providing subjective evaluations. We address both concerns. We propose a semiparametric binary choice procedure which incorporates potentially nonadditive correlated individual-specific effects. Our estimation strategy identifies and estimates the average partial effects of health and functioning on labor market participation. The results indicate that poor health and functioning play a major role in the labor market exit decisions of older men.

We use panel data from the US Health and Retirement Study 1992-2002 to estimate the effect of self-assessed health limitations on active labor market participation of men around retirement age. Self-assessments of health and functioning typically introduce an endogeneity bias when studying the effects of health on labor market participation. This results from justification bias, reflecting an individual’s tendency to provide answers which "justify" his labor market activity, and individual-specific heterogeneity in providing subjective evaluations. We address both concerns. We propose a semiparametric binary choice procedure which incorporates potentially nonadditive correlated individual-specific effects. Our estimation strategy identifies and estimates the average partial effects of health and functioning on labor market participation. The results indicate that poor health and functioning play a major role in the labor market exit decisions of older men.

This paper provides a control function estimator to adjust for endogeneity in the triangular simultaneous equations model where there are no available exclusion restrictions to generate suitable instruments. Our approach is to exploit the dependence of the errors on exogenous variables (e.g. heteroscedasticity) to adjust the conventional control function estimator. The form of the error dependence on the exogenous variables is subject to restrictions, but is not parametrically specified. In addition to providing the estimator and deriving its large-sample properties, we present simulation evidence which indicates the estimator works well.

Os autores relatam o caso de um paciente branco, de quatro anos de idade, com comprometimento neurológico progressivo. Tomografia computadorizada do crânio e ressonância magnética evidenciaram lesão expansiva no tronco cerebral. Subseqüentemente, foi feito diagnóstico histopatológico de gliossarcoma. Trata-se de um tumor raro do sistema nervoso central que, na grande maioria dos casos, acomete pacientes acima dos 40 anos de idade e tem localização supratentorial. Há poucos casos relatados de gliossarcomas em crianças, sobretudo na região infratentorial. Dados epidemiológicos, assim como achados mais freqüentes na tomografia computadorizada e ressonância magnética, são discutidos. The authors report a case of a white four-year-old boy with progressive neurological symptoms due to a brain stem mass lesion identified by computed tomography and magnetic resonance imaging scans. The diagnosis of gliosarcoma was confirmed by histological examination. This rare type of central nervous system tumor affects mainly patients over 40 years of age and is most commonly supratentorial. Only a few cases of gliosarcoma in children have been reported, particularly in the infratentorial region. The epidemiological data as well as the most frequent computed tomography and magnetic resonance imaging findings are discussed.

Os autores relatam o caso de um paciente do sexo masculino, com 55 anos de idade, branco, com diagnóstico radiológico e histopatológico pós-cirúrgico de cisto hidático pulmonar gigante. A epidemiologia, fisiopatologia e características radiológicas desta doença são discutidas. The authors report a case of a 55-year-old white male patient with radiological and postsurgical histopathological diagnosis of a giant lung hydatid cyst. The epidemiological, physiopathological and radiological findings of this disease are discussed.

Os autores relatam o caso de uma paciente do sexo feminino, 51 anos de idade, branca, com diagnóstico estabelecido de lúpus eritematoso sistêmico associado a evidência radiológica de fibrose pulmonar grave. A epidemiologia das manifestações pleuropulmonares do lúpus eritematoso sistêmico, dando ênfase à fibrose pulmonar, é discutida.

Os autores relatam o caso de uma paciente do sexo feminino, 51 anos de idade, branca, com diagnóstico estabelecido de lúpus eritematoso sistêmico associado a evidência radiológica de fibrose pulmonar grave. A epidemiologia das manifestações pleuropulmonares do lúpus eritematoso sistêmico, dando ênfase à fibrose pulmonar, é discutida. The authors report the case of a 51-year-old white female with proven diagnosis of systemic lupus erythematosus and radiological evidence of severe pulmonary fibrosis. The epidemiology of pleuropulmonary manifestations of systemic lupus erythematosus emphasizing pulmonary fibrosis is discussed.

Os autores relatam o caso de uma paciente do sexo feminino, 51 anos de idade, branca, com diagnóstico estabelecido de lúpus eritematoso sistêmico associado a evidência radiológica de fibrose pulmonar grave. A epidemiologia das manifestações pleuropulmonares do lúpus eritematoso sistêmico, dando ênfase à fibrose pulmonar, é discutida. The authors report the case of a 51-year-old white female with proven diagnosis of systemic lupus erythematosus and radiological evidence of severe pulmonary fibrosis. The epidemiology of pleuropulmonary manifestations of systemic lupus erythematosus emphasizing pulmonary fibrosis is discussed.

Os autores relatam o caso de um paciente do sexo masculino, com 55 anos de idade, branco, com diagnóstico radiológico e histopatológico pós-cirúrgico de cisto hidático pulmonar gigante. A epidemiologia, fisiopatologia e características radiológicas desta doença são discutidas.The authors report a case of a 55-year-old white male patient with radiological and postsurgical histopathological diagnosis of a giant lung hydatid cyst. The epidemiological, physiopathological and radiological findings of this disease are discussed.

This paper proposes an estimator for discrete choice models that makes no assumption concerning the functional form of the choice probability function, where this function can be characterized by an index. The estimator is shown to be consistent, asymptotically normally distributed, and to achieve the semiparametric efficiency bound. Monte-Carlo evidence indicates that there may be only modest efficiency losses relative to maximum likelihood estimation when the distribution of the disturbances is known, and that the small-sample behavior of the estimator in other cases is good.

We have used a proprietary data set of newly hired semi-skilled production workers at one location of a large unionized firm to investigate several issues in labor economics. This data set is unique in several respects: the workers in our sample faced the same wage schedules, had the same promotional opportunities, the same job tenure (zero), similar working conditions, and had jobs for which we were able to record their physical output. We analyze these data by formulating a simultaneous equation model to explain wages, output, education, and a worker's quit decision. The model is estimated by maximum likelihood and subjected to a variety of specification tests. Such tests include a comparison of the standard error estimates that form the basis for White's information test, and White's version of a Hausman specification test. Our principal findings are: 1. Individuals that choose more education than we would expect from their observed characteristics have lower than expected quit propensities. We argue that this low quit propensity is one of the unmeasured (and unobserved) attributes that sorting models posit are correlated with education and hence distort the usual estimates of rates of return to education. 2. The performance of non-whites in our sample was no lower than that of whites. However, on their previous jobs non-whites received lower wages than did whites. 3. The output per hour of males in our sample was higher than that of females; however, we were unable to conclude from our data whether these productivity differences could explain the higher wages received by men on their previous jobs. Moreover, this output difference may be transitory and may diminish with on-the-job learning. 4. The expected value of alternative wages had a positive (but not statistically very significant) effect on quit rates. Workers with better alternative opportunities were more likely to quit (all workers had the same opportunities on their current job). 5. Finally we found that workers with high output levels were more likely to quit than were workers with average output levels.

This paper determines the effect of estimation risk on optimal portfolio choice under uncertainty. In most realistic problems, the parameters of return distributions are unknown and are estimated using available economic data. Traditional analysis neglects estimation risk by treating the estimated parameters as if they were the true parameters to determine the optimal choice under uncertainty. We show that for normally distributed returns and ‘non-informative’ or ‘invariant’ priors, the admissible set of portfolios taking the estimation uncertainty into account is identical to that given by traditional analysis. However, as a result of estimation risk, the optimal portfolio choice differs from that obtained by traditional analysis. For other plausible priors, the admissible set, and consequently the optimal choice, is shown to differ from that in traditional analysis.