Firth Correction Research Articles

Cox proportional hazards regression in small studies of predictive biomarkers

Predictive biomarkers are essential for personalized medicine since they select the best treatment for a specific patient. However, of all biomarkers that are evaluated, only few are eventually used in clinical practice. Many promising biomarkers may be erroneously abandoned because they are investigated in small studies using standard statistical techniques which can cause small sample bias or lack of power. The standard technique for failure time endpoints is Cox proportional hazards regression with a multiplicative interaction term between binary variables of biomarker and treatment. Properties of this model in small studies have not been evaluated so far, therefore we performed a simulation study to understand its small sample behavior. As a remedy, we applied a Firth correction to the score function of the Cox model and obtained confidence intervals (CI) using a profile likelihood (PL) approach. These methods are generally recommended for small studies of different design. Our results show that a Cox model estimates the biomarker-treatment interaction term and the treatment effect in one of the biomarker subgroups with bias, and overestimates their standard errors. Bias is however reduced and power is increased with Firth correction and PL CIs. Hence, the modified Cox model and PL CI should be used instead of a standard Cox model with Wald based CI in small studies of predictive biomarkers.

Preoperative COVID-19 and Postoperative Mortality in Cancer Surgery: A South Korean Nationwide Study

BackgroundWe evaluated the impact of preoperative COVID-19 on early postoperative mortality in patients undergoing time-sensitive cancer surgery.MethodsThis retrospective, nationwide cohort study included adult patients who underwent various cancer (thyroid, breast, stomach, colorectal, hepatobiliary, genitourinary, lung, and multiple cancer) surgeries under general anesthesia in South Korea in 2022. Patients were grouped according to the duration from the date of COVID-19 confirmation to the date of surgery (0–2 weeks, 3–4 weeks, 5–6 weeks, and ≥7 weeks). Patients without preoperative COVID-19 also were included. Multivariable logistic regression analysis with Firth correction was performed to investigate the association between preoperative COVID-19 and 30-day and 90-day postoperative mortality. The covariates encompassed sociodemographic factors, the type of surgery, and vaccination status in addition to the aforementioned groups.ResultsOf the 99,555 patients analyzed, 30,933 (31.1%) were preoperatively diagnosed with COVID-19. Thirty-day mortality was increased in those who underwent surgery within 0–2 weeks after diagnosis of COVID-19 (adjusted odds ratio [OR], 1.47; 95% confidence interval [CI], 1.02–2.12; P = 0.038); beyond 2 weeks, there was no significant increase in mortality. A similar pattern was observed for 90-day mortality. Full vaccination against COVID-19 was associated with reduced 30-day (OR 0.38; 95% CI 0.29–0.50; P < 0.001) and 90-day (OR 0.39; 95% CI 0.33–0.46; P < 0.001) mortality.ConclusionsCancer surgery within 2 weeks of COVID-19 diagnosis was associated with increased early postoperative mortality. These findings support current guidelines that recommend postponing elective surgery for at least 2 weeks after the diagnosis of COVID-19.

Exploring depressive symptom trajectories in COVID-19 patients with clinically mild condition in South Korea using remote patient monitoring: longitudinal data analysis.

During the height of the COVID-19 pandemic, the Korean government temporarily allowed full scale telehealth care for safety and usability. However, limited studies have evaluated the impact of telehealth by analyzing the physical and/or mental health data of patients with COVID-19 diagnosis collected through telehealth targeting Korean population. This study aimed to identify subgroup of depressive symptom trajectories in patients with clinically mild COVID-19 using collected longitudinal data from a telehealth-based contactless clinical trial. A total of 199 patients with COVID-19 were accrued for contactless clinical trial using telehealth from March 23 to July 20, 2022. Depressive symptoms were measured using the patient health questionnaire-9 on the start day of quarantine, on the final day of quarantine, and 1 month after release from quarantine. Additionally, acute COVID-19 symptoms were assessed every day during quarantine. This study used a latent class mixed model to differentiate subgroups of depressive symptom trajectories and a logistic regression model with Firth's correction to identify associations between acute COVID-19 symptoms and the subgroups. Two latent classes were identified: class 1 with declining linearity at a slow rate and class 2 with increasing linearity. Among COVID-19 symptoms, fever, chest pain, and brain fog 1 month after release from quarantine showed strong associations with class 2 (fever: OR, 19.43, 95% CI, 2.30-165.42; chest pain: OR, 6.55, 95% CI, 1.15-34.61; brain fog: OR, 7.03, 95% CI 2.57-20.95). Sleeping difficulty and gastrointestinal symptoms were also associated with class 2 (gastrointestinal symptoms: OR, 4.76, 95% CI, 1.71-14.21; sleeping difficulty: OR, 3.12, 95% CI, 1.71-14.21). These findings emphasize the need for the early detection of depressive symptoms in patients in the acute phase of COVID-19 using telemedicine. Active intervention, including digital therapeutics, may help patients with aggravated depressive symptoms.

Finite-Sample Bias of the Linear Excess Relative Risk in Cohort Studies of Computed Tomography-Related Radiation Exposure and Cancer.

The linear excess relative risk (ERR) is the most commonly reported measure of association in radiation epidemiological studies, when individual dose estimates are available. While the asymptotic properties of the ERR estimator are well understood, there is evidence of small sample bias in case-control studies of treatment-related radiation exposure and second cancer risk. Cohort studies of cancer risk after exposure to low doses of radiation from diagnostic procedures, e.g., computed tomography (CT) examinations, typically have small numbers of cases and risks are small. Therefore, understanding the properties of the estimated ERR is essential for interpretation and analysis of such studies. We present results of a simulation study that evaluates the finite-sample bias of the ERR estimated by time-to-event analyses and its confidence interval using simulated data, resembling a retrospective cohort study of radiation-related leukemia risk after CT examinations in childhood and adolescence. Furthermore, we evaluate how the Firth-corrected estimator reduces the finite-sample bias of the classical estimator. We show that the ERR is overestimated by about 30% for a cohort of about 150,000 individuals, with 42 leukemia cases observed on average. The bias is reduced for higher baseline incidence rates and for higher values of the true ERR. As the number of cases increases, the ERR is approximately unbiased. The Firth correction reduces the bias for all cohort sizes to generally around or under 5%. Epidemiological studies showing an association between radiation exposure from pediatric CT and cancer risk, unless very large, may overestimate the magnitude of the relationship, while there is no evidence of an increased chance for false-positive results. Conducting large studies, perhaps by pooling individual studies to increase the number of cases, should be a priority. If this is not possible, Firth correction should be applied to reduce small-sample bias.

Bias reduction for semi-competing risks frailty model with rare events: application to a chronic kidney disease cohort study in South Korea.

In a semi-competing risks model in which a terminal event censors a non-terminal event but not vice versa, the conventional method can predict clinical outcomes by maximizing likelihood estimation. However, this method can produce unreliable or biased estimators when the number of events in the datasets is small. Specifically, parameter estimates may converge to infinity, or their standard errors can be very large. Moreover, terminal and non-terminal event times may be correlated, which can account for the frailty term. Here, we adapt the penalized likelihood with Firth's correction method for gamma frailty models with semi-competing risks data to reduce the bias caused by rare events. The proposed method is evaluated in terms of relative bias, mean squared error, standard error, and standard deviation compared to the conventional methods through simulation studies. The results of the proposed method are stable and robust even when data contain only a few events with the misspecification of the baseline hazard function. We also illustrate a real example with a multi-centre, patient-based cohort study to identify risk factors for chronic kidney disease progression or adverse clinical outcomes. This study will provide a better understanding of semi-competing risk data in which the number of specific diseases or events of interest is rare.

Treatment with Daratumumab Plus Bortezomib, Cyclophosphamide, and Dexamethasone May Result in Both Hematologic and Metabolic Complete Response to Achieve Long-Term Progression Free Survival Among Patients Presenting with Extra-Medullary Disease: A European Myeloma Network Study (EMN19)

Introduction: Multiple myeloma (MM) with extramedullary disease (EMD) is an aggressive disease that requires innovative treatment strategies. Methods: The ongoing multinational, open-label phase 2 EMN19 trial (NCT04166565) enrolled patients (pts) with MM and EMD either as newly diagnosed MM (NDMM) or at first relapse (RMM). Pts received Daratumab (Dara) in combination with bortezomib (V), Cyclophosphamide (C), and Dexamethasone (D) (DaraVCD), either until disease progression or for a maximum of 36 months, along with autologous stem cell transplantation. Response to treatment was assessed using both the International Myeloma Working Group (IMWG) and the Impetus criteria for evaluating the EMD responses. Complete metabolic response (CMR) was defined as: Deauville scores 1, 2 or 3, with absence of FDG-avid bone marrow lesion(s), irrespective of a persistent mass on CT, or complete disappearance of EMD. Partial metabolic response as: Deauville score of 4 or 5 with decreased uptake compared to baseline and absence of structural progression development on CT or 50% reduction in EMD size. MRD assessment was performed locally at the time of CR or at an earlier time point, as per clinical practice. Pts without MRD assessment were included in the denominator when estimating the MRD(-) rate. Pts who had discontinued treatment by Cycle 4 were considered as non-responders (both in terms of hematologic and EMD response). Progression-Free Survival (PFS) was analyzed by the Kaplan-Meier method. Cox regression analysis with Firth's correction was used for the estimation of the hazard ratios (HRs) and 95% confidence intervals (CIs). Results: A cohort of 40 pts (73% NDMM; 28% RMM) was enrolled. The mean age was 58 years, with 22 (55%) of the pts being male. At baseline, 33 (83%) of the pts were in stage ≤II of the International Staging System and 6 (15%) were classified as high-risk according to FISH (t(4;14), t(14;16) and/or del17p). Among the pts, 22 (55%) had extramedullary plasmacytomas (PCTs), 14 (35%) had paraosseous involvement, and 4 (10%) had both. The median number of PCTs per pt was 2. At the cut-off date with a median follow-up (FU) of 23 months, 17 (43%) pts were still on treatment, while 23 (58%) had discontinued treatment mainly due to disease progression (13, 33%), inadequate response at the end of cycle 3 (5, 13%) or death (3, 8%). In the overall population, deepest responses were as follows: MRD(-) was achieved by 15 (38%) pts, ≥Complete response (CR) by 18 (45%) pts, ≥VGPR by 30 (75%) pts with an overall response rate (ORR) of 80% (32 pts). The median (range) time to MRD(-) was 19 months (4-33); 13 of the 15 MRD(-) pts were NDMM. Median PFS (24-months PFS rate) was 26 months (56%) in NDMM and 15 months (36%) in RMM. Overall, 28/40 (70%) pts had a concurrent hematologic and EMD response of partial response (PR) or better; the best hematologic responses achieved among those pts were: MRD(-) (15 pts), ≥CR (16 pts), VGPR (10 pts) and PR (2 pts). Additionally, 4/40 pts had a response of ≥PR (1 stringent CR, 1 CR and 2VGPR) without achieving an EMD response, while 8/40 pts had discontinued treatment by cycle 4 and did not have a hematologic or EMD response. The combinations of hematologic and EMD responses for 28 pts with concurrent hematologic and EMD response are shown in Figure 1. As seen in Figure 2, pts with concurrent hematologic CR (hemCR) and CMR had the best outcomes in terms of PFS, while having only one type of response (hematologic or EMD) was associated with inferior outcomes compared to having a response in and outside of marrow (p&amp;lt;0.001 overall and in NDMM). Overall, the respective HRs (95% CI) were: no response vs both types: 20.5 (3.5-119.8), p&amp;lt;0.001; one type vs both types: 8.3 (1.2-58.2), p=0.034. The median (range) time to first concurrent hemCR and CMR was 6 months (3-15); for pts achieving only one type of response (hemCR or CMR) the median (range) time to first response was 5 months (1-11). Conclusions: After a median FU of 23 months of DaraVCD treatment, 33% of the pts have achieved both deep hematological responses with MRD(-) and CMR resulting in a median PFS of 20 months. However, having only one type of response (hemCR or CMR) is associated with an inferior outcome. These results, in terms of response, are comparable to those found in the LYRA study in which NDMM EMD pts are treated using a similar DaraVCD protocol. Even in this high-risk population of unmet need, long-term deep responses can be achieved with DaraVCD in and outside of marrow.

Abstract 5591: Circulating tumor DNA for recurrence prediction and efficacy analysis in the ICONIC trial of peri-operative FLOT and avelumab (PD-L1) in localized esophago-gastric adenocarcinoma

Abstract Background: The utility of molecular residual disease (MRD) detection by circulating tumor (ct)DNA in early-stage (pT2+ or N+, M0) esophagogastric adenocarcinoma treated with peri-operative systemic therapy has not been assessed in prospective trials. Methods: This exploratory analysis of the phase 2 ICONIC trial (NCT03399071), assessed whether ctDNA can predict recurrence and determine the efficacy of 4xFLOT+avelumab (FLOT-A) before and after surgery. Exome sequencing of pre-treatment biopsies was successful in 24/26 patients (pts) (92.3%) who had received FLOT-A and had undergone surgery at the time of analysis. All pts had R0 resections. Tumor-informed ctDNA assays (SignateraTM) were designed for these 24 pts and 220 serial plasma samples were analyzed. Pathologic response was assessed using Mandard tumor regression grading (TRG1 complete, 2 excellent, 3 good, 4 poor and 5 no response). The median follow-up was 17.0m from surgery. Progression free survival (PFS) was calculated from surgery to radiological recurrence or death. Results: ctDNA was detected in 23/24 pts (95.8%) prior to treatment. PFS was significantly shorter for pts with a mean number of tumor DNA molecules per ml plasma in the middle &amp; highest tertile (p=0.049, HR=8.33, 95% CI: 1.01-1083, Firth Correction for Cox regression). 6 pts remained ctDNA-positive post neoadjuvant chemotherapy (NAC). None of these pts had a TRG1/2, 3 (50.0%) had TRG3 and 3 (50%) TRG4/5 in the resection specimen. Of 18 ctDNA-negative pts 5 (27.8%) had TRG1/2, 10 (55.6%) TRG3 and 3 (16.7%) TRG4/5. Post-surgery and prior to adjuvant therapy, 6/24 pts were ctDNA-positive. ctDNA positivity at this time point was associated with significantly shorter median PFS (12.9m) compared to ctDNA-negative status (PFS not reached, p&amp;lt;0.0001, HR=27, 95%CI: 3.0-241). Nodal status and Mandard TRG are routinely used as clinical predictors. Only the former was significantly associated with poor median PFS (pN+: 13.5m, pN-: not reached, p=0.027, HR=11, 95%CI: 1.3-98). Of 6 pts who remained ctDNA positive post-operatively, none achieved ctDNA clearance despite adjuvant FLOT-A. The median lead-time from ctDNA positivity after surgery to recurrence was 11.4 months. Conclusions: Post-NAC/pre-surgical ctDNA positivity correlated with worse pathological response. Persistent ctDNA after NAC &amp; surgery was a stronger predictor of recurrence than nodal status or TRG in the resection specimen. Post-operative adjuvant therapy failed to clear ctDNA in any of the pts who were ctDNA positive after surgery, indicating that administering more of the same treatment is ineffective. This provides an opportunity to test new adjuvant therapies in pts who remain ctDNA positive after surgery. Whether post-operative therapy can be omitted in pts who are ctDNA negative after surgery should be assessed in future trials. Citation Format: Marco Gerlinger, Anderley Gordon, Louise J. Barber, Georgios Laliotis, Avani Athauda, Benjamin Challoner, Andrew Woolston, Sonia Mansukhani, Matt Dunstan, Nikoletta Petrou, Komel Khabra, Retchel Lazaro-Alcausi, Richard Crux, Victoria Borja, Ruwaida Begum, Isma Rana, Charuta Palsuledesai, Meenakshi Malhotra, Minetta Liu, Adham Jurdi, Shruti Sharma, Sheela Rao, Sacheen Kumar, David Cunningham, Ian Chau, Naureen Starling, M Asif Chaudry. Circulating tumor DNA for recurrence prediction and efficacy analysis in the ICONIC trial of peri-operative FLOT and avelumab (PD-L1) in localized esophago-gastric adenocarcinoma. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5591.

Bayesian solution to the monotone likelihood in the standard mixture cure model

An advantage of the standard mixture cure model over an usual survival model is how it accounts for the population heterogeneity. It allows a joint estimation for the distribution related to the susceptible and non‐susceptible subjects. The estimation algorithm may provide coefficients when the likelihood cannot be maximized. This phenomenon is known as Monotone Likelihood (ML), common in survival and logistic regressions. The ML tends to appear in situations with small sample size, many censored times, many binary or unbalanced covariates. Particularly, it occurs when all uncensored cases correspond to one level of a binary covariate. The existing frequentist solution is an adaptation of the Firth correction, originally proposed to reduce bias of maximum likelihood estimates. It prevents estimates by penalizing the likelihood, with the penalty interpreted as the Bayesian Jeffreys prior. In this paper, the penalized likelihood of the standard mixture cure model is considered with different penalties (Bayesian priors). A Monte Carlo simulation study indicates good inference results, especially for balanced data sets. Finally, a real application involving a melanoma data illustrates the approach.

Quality of Surveillance Impacts the Colitis-Associated Advanced Neoplasia Risk: A Multicenter Case-Control Study

Although colorectal cancer (CRC) surveillance is embedded in clinical inflammatory bowel disease (IBD) practice, a subset of patients still develops advanced neoplasia (AN) (high-grade dysplasia [HGD] and/or CRC). We aimed to assess the impact of surveillance quality on AN risk in IBD. In this multicenter case-control study, we searched the Dutch nationwide pathology databank to identify IBD cases with AN and controls with indefinite or low-grade dysplasia. The surveillance colonoscopy preceding the index lesion (first indefinite for dysplasia [IND]/low-grade dysplasia [LGD] or AN) was used to assess the impact of surveillance quality. We assessed intervals, bowel preparation, cecal intubation, and absence of inflammation as primary quality indicators. In addition, we assessed chromoendoscopy, endoscopist expertise, hospital setting, and biopsy strategy. Associations of quality indicators with AN risk were determined with multivariable logistic regression analyses with Firth's correction. We included 137 cases and 138 controls. Delayed intervals (58.2% vs 39.6%) and active inflammation (65.3% vs 41.8%) were frequently present in cases and controls and were associated with AN (delayed interval: adjusted odds ratio [aOR], 2.00; 95% confidence interval [CI], 1.07-3.81; P= .03; active inflammation: aOR, 2.46; 95% CI, 1.33-4.61; P < .01). Surveillance compliant with primary quality indicators was associated with a reduced AN risk (aOR, 0.43; 95% CI, 0.22-0.91; P= .03), similar to chromoendoscopy (OR, 0.11; 95% CI, 0.01-0.89; P= .01). Other indicators were not significantly associated with AN. Surveillance compliant with primary quality indicators is associated with a reduced colitis-associated AN risk. Delayed surveillance intervals and active inflammation were associated with an increased AN risk. This underlines the importance of procedural quality, including endoscopic remission to optimize the effectiveness of endoscopic surveillance.

Association of CYP2D6 genotype predicted phenotypes with oxycodone requirements and side effects in children undergoing surgery

Oxycodone is a commonly used oral opioid in children for treating postoperative pain. Highly polymorphic gene CYP2D6 metabolizes oxycodone into its more potent metabolite, oxymorphone. We hypothesized that altered activity due to CYP2D6 polymorphisms will influence oxycodone requirements {relative oxycodone use [oxycodone morphine equivalents (MEq)/total MEq] to maintain analgesia} (primary outcome) and risk for oxycodone induced side-effects such as respiratory depression (RD) and emesis (secondary outcomes). We also explored the influence of genotype availability and provider guidance on oral opioid prescription patterns. Patients who underwent Nuss procedure and spine fusion with CYP2D6 genotyping results available preoperatively were included. Data on demographics, genotypes, oral opioids, pain scores, RD and emesis were collected. Univariate and multivariable regression for comparison of CYP2D6 genotype predicted poor, ultrarapid, intermediate metabolizers (PM, UM and IM) phenotype with normal metabolizers (NM) for outcomes were performed. Stratified logistic regression was conducted in low (oxycodone/total MEq <0.5) and high (and oxycodone/total MEq >0.5) oxycodone use groups for RD and emesis, with application of firth correction due to quasi-complete separations. Breslow-Day test was used to evaluate odds ratios for prescribing genotype directed opioid between control group (2012-15) (where providers were alerted to genotyping results availability but not directed to use them while prescribing) and genotype directed groups (2016-18) (where providers were directed to use the genotyping results available to them while prescribing oxycodone after surgery). Of 193 subjects (age 15.9±0.25 years, 28.5% female, 93.78% White; 101 NM, 76 IM, 10 PM and 6 UM), 77.72% underwent pectus surgery. CYP2D6 phenotype was associated with oxycodone MEq/total MEq requirements (P<0.001). Both PM and UM phenotypes had lower oxycodone requirements compared to NM [-0.316 (SE 0.098), P=0.005 and -0.432 (SE 0.113), P<0.001 respectively]. CYP2D6 phenotype was associated with RD in high use oxycodone group (P=0.018) but not low use oxycodone groups (P=0.634). No phenotype association was found for emesis. Oxycodone was prescribed to 91.24% of NM/IM vs. 66.67% of PM/UM (P=0.129) in control group and 94.64% of NM/IM vs. 28.57% of PM/UM (P<0.001) in the genotype-directed group. PM/UM phenotypes in genotype directed group had a lower chance of being prescribed oxycodone (effect size =-2.775; SE 1.566; P=0.076). Our findings suggest CYP2D6 genotypes are associated with oxycodone requirements for analgesia and may influence risk for RD. Genotype availability and guidance likely influence oral opioid prescription pattern after surgery. Our findings are limited by small sample size for UM/PM groups.

Effects of Prenatal Phthalate Exposure on Early Childhood Expressive Language Development Trajectories

BACKGROUND AND AIM: Phthalates are common endocrine disruptors that can impair neurodevelopment. We investigated potential links between prenatal exposure to phthalates and child language development trajectories given prior inconsistent evidence across individual phthalates. METHODS: We investigated fourteen urinary phthalate metabolites, measured in maternal urine during pregnancy, and child expressive language scores from the Mullen Scales of Early Learning (MSEL) at 6, 12, 24, and 36 months in 231 children from an enhanced autism-risk cohort (19% with autism): Early Autism Risk Longitudinal Investigation (EARLI), born 2009 - 2013 from California, Maryland, and Pennsylvania. We used multiple imputation to handle missingness in MSEL (~37%) and phthalates (~9%), and assigned children to six distinct expressive language trajectories after selecting the latent class growth analysis model with best goodness-of-fit measures. We estimated relative risk ratios (RRR) of being in different language trajectories (&amp;#x22;high growth” as reference) associated with two-fold change in creatinine adjusted phthalates using multinomial logistic regression with Firth correction, adjusting for child sex, birth season, income, and maternal age, race, and education. RESULTS: Some low molecular weight phthalates were associated with sub-optimal language trajectories, e.g. RRRs of 1.33 or higher for mono-2-hydroxy-iso-butyl phthalate, peaking at 1.51 (1.02, 2.24) for the delayed growth low trajectory. In contrast, some high molecular weight phthalates exhibited protective effects for delayed growth trajectories but adverse effects for plateau and tracking trajectories: e.g. mono-isononyl phthalate with RRR 0.72 (0.59, 0.89) with delayed growth low but 1.29 (1.08, 1.53) with the tracking low trajectory. Some phthalates exhibited null patterns. CONCLUSIONS: Prenatal phthalate exposure may impact the trajectory of expressive language development in early childhood, with differences by individual phthalates, perhaps by molecular weight group. We observed adverse and protective influences of phthalates that may be influenced by small numbers in each trajectory class. KEYWORDS: phthalates, language development, trajectory analysis

Longitudinal reoperation risk after apical suspension procedures in female pelvic reconstructive surgery

To analyze longitudinal reoperation risk for recurrent pelvic organ prolapse (POP) among the 4 apical suspension procedures over a period of 2 through 15 years. This multicenter, retrospective cohort study included adult women who underwent sacrocolpopexy (SCP), uterosacral ligament suspension (USLS), sacrospinous ligament fixation (SSLF), or colpocleisis (CC) from January 1, 2006 through December 31, 2018, with follow up through July 25, 2021. Women who underwent vaginal prolapse repair with mesh augmentation or concomitant rectopexy were excluded. Data were abstracted using procedural and diagnosis codes with manual review of 10% of each variable. The primary outcome was reoperation for POP in any compartment. Rates were compared using a X2 test with Bonferroni correction. Time to event distributions was contrasted with a log rank test. Multivariate analysis using a Cox proportional hazards model with Firth correction evaluated the following predictors: index surgery, patient characteristics (age, race/ethnicity, body mass index [BMI], smoking status), concomitant procedures (hysterectomy, compartment repairs, incontinence procedures), and year of index surgery. Censoring events included exit from the health maintenance organization and death. This cohort included 9,117 women with mean age of 60.6 (SD ±11.6) years and BMI of 28 kg/m2 (±4.9) at the time of index surgery. Most women were Hispanic (48.3%) or White (40.2%). The overall reoperation rate was 5.1%. Rates by procedure are shown in Table 1. Significant differences were found in the following pairwise comparisons: SCP (3.6%) versus USLS (6.1%; P <0.0006), SCP (3.6%) versus SSLF (9%; P <0.0006), SCP (3.6%) versus CC (0.8%; P <0.0006), USLS (6.1%) versus CC (0.8%; P <0.0006), and SSLF (9%) versus CC (0.8%; P <0.0006). Reoperation rate did not significantly differ between USLS (6.1%) and SSLF (9%; P =0.074) after correction for multiplicity. Similarly, pairwise comparisons of time to event distributions were significant (P =0.0003-0.0018), except for USLS versus SSLF (P =0.05). Index procedure was found to be a significant predictor of reoperation when compared to SCP (Table 1, P =0.0003-0.0024). No other covariates were shown to consistently predict reoperation upon either crude or adjusted analysis. While overall reoperation rate after apical suspension is low, index surgery is a significant predictor of reoperation. CC offers the most durable prolapse repair, followed by SCP. USLS and SSLF appear to be comparable to one another in durability.View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Modified score function for monotone likelihood in the semiparametric mixture cure model.

The cure fraction models are intended to analyze lifetime data from populations where some individuals are immune to the event under study, and allow a joint estimation of the distribution related to the cured and susceptible subjects, as opposed to the usual approach ignoring the cure rate. In situations involving small sample sizes with many censored times, the detection of nonfinite coefficients may arise via maximum likelihood. This phenomenon is commonly known as monotone likelihood (ML), occurring in the Cox and logistic regression models when many categorical and unbalanced covariates are present. An existing solution to prevent the issue is based on the Firth correction, originally developed to reduce the estimation bias. The method ensures finite estimates by penalizing the likelihood function. In the context of mixture cure models, the ML issue is rarely discussed in the literature; therefore, this topic can be seen as the first contribution of our paper. The second major contribution, not well addressed elsewhere, is the study of the ML issue in cure mixture modeling under the flexibility of a semiparametric framework to handle the baseline hazard. We derive the modified score function based on the Firth approach and explore finite sample size properties of the estimators via a Monte Carlo scheme. The simulation results indicate that the performance of coefficients related to the binary covariates are strongly affected to the imbalance degree. A real illustration, in the melanoma dataset, is discussed using a relatively novel data set collected in a Brazilian universityhospital.

Prediction of post-traumatic neuropathy following impacted mandibular third molar removal

ObjectivesThe extraction of impacted mandibular third molars is a common surgical procedure often associated with complications including post-traumatic neuropathy. Previous work has focused on identifying confounding factors, but a robust preoperative risk prediction model remains elusive. MethodsUsing a dataset of 648 patients and 812 impacted mandibular third molars, we used least absolute shrinkage and selection operator (LASSO) to fit prediction models based on risk factors assessed at both the tooth and patient levels. In addition, we fitted multivariable logistic regression models with the Firth correction for generalized estimating equations (GEE). ResultsThe LASSO model for post-traumatic neuropathy identified distoangular impaction of ≥ 45° (odds ratio [OR] = 2.9), proximity to the inferior alveolar nerve of ≤ 3 mm (OR = 1.9), disadvantageous curving (OR = 1.4), and psychiatric conditions (OR = 2.1) as predictors [area under the receiving operator characteristic curve (AUC) = 0.75]. Among other complications analyzed, the LASSO model for bleeding identified deep embedding or full impaction (OR = 1.8), psychiatric conditions (OR = 1.3), and age (OR = 0.9) as predictors (AUC = 0.64). These associations between predictors and postoperative complications were fundamentally reinforced by the corresponding GEE models. ConclusionsOur findings point to the predictability of post-traumatic neuropathy and bleeding based on tooth anatomy and patient characteristics, overall suggesting that preoperatively identifiable factors can predict the risk of adverse outcomes in the extraction of impacted mandibular third molars. Clinical significanceMandibular third molar extraction is both a routine procedure and a leading cause of trigeminal neuropathy. Prevention of post-traumatic neuropathy, aided by individualized preoperative risk prediction, is of high clinical relevance.

The novel NSAID-derivative, MCI-030, prevents ovarian cancer in the egg-laying hen by increasing tumor cell apoptosis and decreasing beta-catenin and MAPK oncogenic signaling

Objectives: The egg-laying hen is an ideal model to study chemoprevention of epithelial ovarian cancer (EOC) as aging hens have a high rate of spontaneous EOC and similar clinicopathological characteristics of the disease to women. Our previous research revealed that the novel non-COX inhibitory NSAID-derivative, MCI-030, decreased ovarian cancer incidence in the hen. MCI-030 inhibits PDE10A, which was shown to be an effective therapeutic target in colon and lung cancers, reducing beta-catenin and MAPK signaling. Molecular analysis of MCI-030 treated hens was performed to study the mechanism by which PDE10A inhibition can reduce ovarian cancer risk. Methods: A total of 225 White leghorn hens aged 2.5 to 3 years old were treated through their feed with MCI-030 or its parent compound, sulindac, for 6- and 12-months and compared to a control group (Figure 1A). Necropsy was performed at each time point to formally assess for tumors with 60 hens evaluated at 6-months and 85 at 12-months. Logistic regression with Firth correction to estimate effect size for this pilot study was performed. Ovary and oviduct tissues were harvested and formalin-fixed or flash-frozen then paraffin-embedded or homogenized, respectively. Molecular analyses were performed using fluorescent immunohistochemistry, TUNEL assays, and western blotting. Results: A trend of decreased risk of malignancy was observed in the MCI-030 group at 6- and 12-months versus control with odds ratios (OR) of 0.56 and 0.75, respectively (95% CI 0.121-2.633, 0.22-2.58). In contrast, this trend was not observed for sulindac (6-month OR=1.56, 95% CI 0.405-6.06, 12-month OR=0.94, 95% CI 0.29-3.03). Assessment of apoptosis in hen ovaries by TUNEL assay demonstrated that MCI-030 significantly increased apoptosis both overall and in ovarian tumors (Figure 1B). Western blotting revealed that pERK was significantly decreased in ovaries of sulindac and MCI-030 hens (Figure 1C). Immunohistochemistry showed that MCI-030 significantly decreased nuclear translocation of beta-catenin and levels of active beta-catenin in ovarian tumors (Figure 1D). Download : Download high-res image (224KB) Download : Download full-size image Conclusions: MCI-030 appears to be a promising agent for the chemoprevention and treatment of ovarian cancer. Molecular analysis of its efficacy in hens revealed that MCI-030 increases tumor apoptosis and decreases pERK and beta-catenin oncogenic signaling. Further investigation of MCI-030 in additional model systems and in-human trials is warranted.

Survival of Left-to-Right Shunt Repair in Children with Pulmonary Arterial Hypertension at a Tertiary Hospital in a Low-to-Middle-Income Country.

Background:In low-to-middle-income countries, repair of the left-to-right shunts congenital heart disease (CHD) are often done with existing pulmonary arterial hypertension (PAH). Long-term outcomes data of this condition in either low-to-middle or high-income countries are limited. We conducted a study to evaluate the outcomes of children with PAH related to left-to-right shunt CHD who underwent surgical or transcatheter repair.Methods:All cases of PAH related to left-to-right shunt CHD repairs from 2015–2018 were retrospectively reviewed with additional new patients who underwent repair within our study period (2019–2020). Cases with complex congenital heart disease and incomplete hemodynamic echocardiography or catheterization measurements were excluded. Kaplan-Meier curves, log-rank test, Cox regression with Firth’s correction and restricted mean survival time were used for survival analysis.Results:Of the 118 patients, 103 patients were enrolled and 15 patients were excluded due to complex congenital heart disease and missing hemodynamic measurements prior to repair. Overall, median age at intervention was 44 months, mPAP mean was 43.17 ± 16.05 mmHg and Pulmonary Vascular Resistance index (PVRi) mean was 2.84 ± 2.09 (WU.m2). Nine patients died after repair. The survival rate for patients with PAH-CHD at 1 day, 30 days and 1400 days (4 years) was 96.1%, 92.1%, and 91.0% respectively. Patients with persisting PAH after correction had –476.1 days (95% confidence interval [CI]: –714.4, –237,8) shorter survival over 4 years of follow up compared to patients with reversed PAH. PVRi was found to be the influencing covariate of the difference of restricted mean survival time between these groups.Conclusion:In low-to-middle income settings, with accurate PAH reversibility assessment prior to intervention, repair of left-to-right shunt CHD with existing PAH in children has a favourable outcome. Inferior survival is found in patients with persistence of PAH. PVRi at baseline predicts between-group survival difference.

Omission of Cortical Renorrhaphy During Robotic Partial Nephrectomy: A Vattikuti Collective Quality Initiative Database Analysis

To analyze the outcomes of patients in whom cortical (outer) renorrhaphy (CR) was omitted during robotic partial nephrectomy (RPN). We analyzed 1453 patients undergoing RPN, from 2006 to 2018, within a large multi-institutional database. Patients having surgery for bilateral tumors (n=73) were excluded. CR and no-CR groups were compared in terms of operative and ischemia time, estimated blood loss (EBL), complications, surgical margins, hospital stay, change in estimated glomerular filtration rate (eGFR), and need of angioembolization. Inverse probability of treatment weighting with Firth correction for center code was performed to account for selection bias. CR was omitted in 120 patients (8.7%); 1260 (91.3%) patients underwent both inner layer and CR. There was no difference in intraoperative complications (7.4% CR; 8.9% no-CR group; P = .6), postoperative major complications (1% and 2.8% in CR and no-CR groups, respectively; P = .2), or median drop in eGFR (7.3 vs 10.4 mL/min/m2). The no-CR group had a higher incidence of minor complications (26.7% vs 5.5% in CR group; P < .001). EBL was 100 mL (IQR 50-200) in both groups (P = .6). Angioembolization was needed in 0.7% patients in CR vs 1.4% in no-CR group (P = .4). Additionally, there was no difference in median operative time (168 vs 162 min; P = .2) or ischemia time (18 vs 17 min; P = .7). In selected patients with renal masses, single layer renorrhaphy does not significantly improve operative time, ischemia time, or eGFR after RPN. There is a higher incidence of minor complications, but not major perioperative complications after no-CR technique.

A Fast and Accurate Method for Genome-Wide Time-to-Event Data Analysis and Its Application to UK Biobank

With increasing biobanking efforts connecting electronic health records and national registries to germline genetics, the time-to-event data analysis has attracted increasing attention in the genetics studies of human diseases. In time-to-event data analysis, the Cox proportional hazards (PH) regression model is one of the most used approaches. However, existing methods and tools are not scalable when analyzing a large biobank with hundreds of thousands of samples and endpoints, and they are not accurate when testing low-frequency and rare variants. Here, we propose a scalable and accurate method, SPACox (a saddlepoint approximation implementation based on the Cox PH regression model), that is applicable for genome-wide scale time-to-event data analysis. SPACox requires fitting a Cox PH regression model only once across the genome-wide analysis and then uses a saddlepoint approximation (SPA) to calibrate the test statistics. Simulation studies show that SPACox is 76-252 times faster than other existing alternatives, such as gwasurvivr, 185-511 times faster than the standard Wald test, and more than 6,000 times faster than the Firth correction and can control type I error rates at the genome-wide significance level regardless of minor allele frequencies. Through the analysis of UK Biobank inpatient data of 282,871 white British European ancestry samples, we show that SPACox can efficiently analyze large sample sizes and accurately control type I error rates. We identified 611 loci associated with time-to-event phenotypes of 12 common diseases, of which 38 loci would be missed within a logistic regression framework with a binary phenotype defined as event occurrence status during the follow-up period.

Ranking hospitals when performance and risk factors are correlated: A simulation-based comparison of risk adjustment approaches for binary outcomes.

BackgroundThe conceptualization of hospital quality indicators usually includes some form of risk adjustment to account for hospital differences in case mix. For binary outcome variables like in-hospital mortality, frequently utilized risk adjusted measures include the standardized mortality ratio (SMR), the risk standardized mortality rate (RSMR), and excess risk (ER). All of these measures require the estimation of expected hospital mortality, which is often based on logistic regression models. In this context, an issue that is often neglected is correlation between hospital performance (e.g. care quality) and patient-specific risk factors. The objective of this study was to investigate the impact of such correlation on the adequacy of hospital rankings based on different measures and methods.MethodsUsing Monte Carlo simulation, the impact of correlation between hospital care quality and patient-specific risk factors on the adequacy of hospital rankings was assessed for SMR/RSMR, and ER based on logistic regression and random effects logistic regression. As an alternative method, fixed effects logistic regression with Firth correction was considered. The adequacies of the resulting hospital rankings were assessed by the shares of hospitals correctly classified into quintiles according to their true (unobserved) care qualities.ResultsThe performance of risk adjustment approaches based on logistic regression and random effects logistic regression declined when correlation between care quality and a risk factor was induced. In contrast, fixed-effects-based estimations proved to be more robust. This was particularly true for fixed-effects-logistic-regression-based ER. In the absence of correlation between risk factors and care quality, all approaches showed similar performance.ConclusionsCorrelation between risk factors and hospital performance may severely bias hospital rankings based on logistic regression and random effects logistic regression. ER based on fixed effects logistic regression with Firth correction should be considered as an alternative approach to assess hospital performance.

"Third Window" and "Single Window" Effects Impede Surgical Success: Analysis of Retrofenestral Otosclerosis Involving the Internal Auditory Canal or Round Window.

Background and Objectives: We aimed to identify prognostic computed tomography (CT) findings in retrofenestral otosclerosis, with particular attention paid to the role of otosclerotic lesion area in predicting post-stapedotomy outcome. Materials and Methods: We included 17 subjects (23 ears) with retrofenestral otosclerosis who underwent stapedotomy. On preoperative CT, the presence of cavitating lesion and involvement of various subsites (cochlea, round window [RW], vestibule, and semicircular canal) were assessed. Pre- and post-stapedotomy audiometric results were compared according to the CT findings. The surgical outcomes were analyzed using logistic regression with Firth correction. Results: Cavitating lesions were present in 15 of 23 ears (65.2%). Involvement of the RW was the strongest predictor of unsuccessful surgical outcome, followed by involvement of the internal auditory canal (IAC) and the cochlea. Conclusions: RW and IAC involvement in retrofenestral otosclerosis were shown to predict unsuccessful outcomes. While a “third window” effect caused by extension of a cavitating lesion into the IAC may dissipate sound energy and thus serve as a barrier to desirable postoperative audiological outcome, a “single window” effect due to an extension of retrofenestral otosclerosis into the RW may preclude a good surgical outcome, even after successful stapedotomy, due to less compressible cochlear fluid and thus decreased linear movement of the piston.