Search

In this paper, we present a practical approach for computing the sandwich variance estimator in 2-stage regression model settings. As a motivating example for 2-stage regression, we consider regression calibration, a popular approach for addressing covariate measurement error. The sandwich variance approach has rarely been applied in regression calibration, despite its requiring less computation time than popular resampling approaches for variance estimation, specifically the bootstrap. This is probably because it requires specialized statistical coding. Here we first outline the steps needed to compute the sandwich variance estimator. We then develop a convenient method of computation in R for sandwich variance estimation, which leverages standard regression model outputs and existing R functions and can be applied in the case of a simple random sample or complex survey design. We use a simulation study to compare the sandwich estimator to a resampling variance approach for both settings. Finally, we further compare these 2 variance estimation approaches in data examples from the Women's Health Initiative (1993–2005) and the Hispanic Community Health Study/Study of Latinos (2008–2011). In our simulations, the sandwich variance estimator typically had good numerical performance, but simple Wald bootstrap confidence intervals were unstable or overcovered in certain settings, particularly when there was high correlation between covariates or large measurement error. [ABSTRACT FROM AUTHOR]

HER2-positive (HER2+) breast cancer historically exhibited the most adverse local outcomes and, consequently, has had limited representation among trials of partial breast irradiation (PBI). We hypothesized that with contemporary HER2-targeted agents, patients may now exhibit excellent disease control outcomes after adjuvant PBI. Using a prospectively maintained institutional database, we identified patients with HER2+ breast cancer treated with breast conserving surgery (BCS) and PBI from 2000 to 2022. Salient clinicopathologic and treatment parameters were collected. All patients received external beam PBI to 40 Gy in 10 daily fractions. Cumulative incidence functions were calculated to estimate the incidence of local recurrence (LR) with the competing risk of death. We identified 1248 patients who underwent PBI at our institution, of whom the study cohort comprised 52 (4%) with HER2+ breast cancer (median age, 64 years; range, 44-87). Nearly all had T1 tumors (98%; median size, 12 mm [range, 1-21 mm]). Most had estrogen receptor positive disease (88%), and all patients had negative surgical margins. Nearly all underwent sentinel node biopsy (94%), with the remainder undergoing no surgical axillary evaluation. Forty-two (81%) received chemotherapy; 40 (77%), hormone therapy; and 42 (81%), HER2-directed therapy, most commonly trastuzumab. At 143.8 person-years of follow-up (range, 7-226 months for each patient), we observed 2 LR events at 14 and 26 months, respectively, yielding a 2-year LR rate of 3.8%. No regional or distant recurrences were observed, nor were any breast-specific mortality events. Two deaths were observed, both without evidence of disease. Among a cohort of patients with HER2+ early-stage breast cancer managed with BCS and PBI, we observed a 2-year LR rate of 3.8%, with no regional or distant recurrences and excellent overall survival. These findings require confirmation with longer follow-up among larger cohorts but appear consistent with the excellent results of contemporary randomized trials of PBI unselected for HER2 status. [ABSTRACT FROM AUTHOR]

Purpose: Newer-generation tyrosine kinase inhibitors (TKIs) for non-small cell lung cancer (NSCLC) with epidermal growth factor receptor ( EGFR ) mutations and anaplastic lymphoma kinase ( ALK ) rearrangements have demonstrated high CNS activity. The optimal use of up-front stereotactic radiosurgery (SRS) for brain metastases (BM) in patients eligible for CNS-penetrant TKIs is controversial, and data to guide patient management are limited., Materials and Methods: Data on TKI-naïve patients with EGFR- and ALK-driven NSCLC with BM treated with CNS-penetrant TKIs with and without up-front SRS were retrospectively collected from seven academic centers in the United States. Time-to-CNS progression and overall survival (OS) were analyzed, with multivariable adjustment in Fine & Gray and Cox proportional hazards models for clinically relevant factors., Results: From 2013 to 2022, 317 patients were identified (200 TKI-only and 117 TKI + SRS). Two hundred fifty (79%) and 61 (19%) patients received osimertinib and alectinib, respectively. Patients receiving TKI + SRS were more likely to have BM ≥1 cm ( P < .001) and neurologic symptoms ( P < .001) at presentation. Median OS was similar between the TKI and TKI + SRS groups (median 41 v 40 months, respectively; P = .5). On multivariable analysis, TKI + SRS was associated with a significant improvement in time-to-CNS progression (hazard ratio [HR], 0.63 [95% CI, 0.42 to 0.96]; P = .033). Local CNS control was significantly improved with TKI + SRS (HR, 0.30 [95% CI, 0.16 to 0.55]; P < .001), whereas no significant differences were observed in distant CNS control. Subgroup analyses demonstrated a greater benefit from TKI + SRS in patients with BM ≥1 cm in diameter for time-to-CNS progression and CNS progression-free survival., Conclusion: The addition of up-front SRS to CNS-penetrant TKI improved time-to-CNS progression and local CNS control, but not OS, in patients with BM from EGFR - and ALK-driven NSCLC. Patients with larger BM (≥1 cm) may benefit the most from up-front SRS.

Importance: The outcomes of patients with low-risk thyroid cancer who undergo surgery following a period of active surveillance (AS) are not well-defined., Objective: To evaluate surgical, pathologic, and oncologic outcomes among patients undergoing conversion surgery (CS) following AS for low-risk papillary thyroid carcinoma., Design, Setting, and Participants: In this cohort study, patients who underwent CS for disease progression were compared with patients who underwent CS without disease progression and with a propensity score-matched cohort of patients who underwent initial surgery (IS). The median (IQR) postsurgical follow-up time was 40.3 (18.0-59.0) months. Patients were treated at a quaternary cancer referral center in the United States., Exposures: Surgery., Main Outcomes and Measures: Surgical complications, pathologic characteristics, overall survival (OS), and recurrence-free survival (RFS)., Results: Of 550 patients who underwent AS, 55 (10.0%) had CS, of whom 39 (7.1%) had surgery due to suspected disease progression (median [IQR] age, 48 [39-56] years; 32 [82.1%] female). There were no clinically meaningful differences in rates of surgical sequalae between the progression CS group (12 of 39 [30.7%]) and the nonprogression CS group (7 of 16 [43.8%]) (Cramer V, 0.2; 95% CI, 0.01-0.5). The 5-year OS was 100% (95% CI, 100%-100%) in both the disease-progression CS cohort and the IS cohort. Although the cohort of patients undergoing CS after disease progression was by definition a subset with more aggressive tumor behavior, no clinically meaningful differences were observed in the rates of regional recurrence (2 of 39 [5.1%] vs 0 of 39 patients with IS), local recurrence (0 patients), distant metastasis (0 patients), or disease-specific mortality (0 patients) when compared with the matched IS group. Five-year RFS rates were similar: 100% in the IS group and 86% (95% CI, 70%-100%) in the CS group., Conclusions and Relevance: In this cohort study, CS for suspected disease progression was associated with surgical and oncologic outcomes similar to IS, supporting the feasibility and safety of AS for patients with low-risk papillary thyroid carcinoma.