Citations

Background: Deep learning algorithms derived in homogeneous populations may be poorly generalizable and have the potential to reflect, perpetuate, and even exacerbate racial/ethnic disparities in health and health care. In this study, we aimed to (1) assess whether the performance of a deep learning algorithm designed to detect low left ventricular ejection fraction using the 12-lead ECG varies by race/ethnicity and to (2) determine whether its performance is determined by the derivation population or by racial variation in the ECG. Methods: We performed a retrospective cohort analysis that included 97 829 patients with paired ECGs and echocardiograms. We tested the model performance by race/ethnicity for convolutional neural network designed to identify patients with a left ventricular ejection fraction ≤35% from the 12-lead ECG. Results: The convolutional neural network that was previously derived in a homogeneous population (derivation cohort, n=44 959; 96.2% non-Hispanic white) demonstrated consistent performance to detect low left ventricular ejection fraction across a range of racial/ethnic subgroups in a separate testing cohort (n=52 870): non-Hispanic white (n=44 524; area under the curve [AUC], 0.931), Asian (n=557; AUC, 0.961), black/African American (n=651; AUC, 0.937), Hispanic/Latino (n=331; AUC, 0.937), and American Indian/Native Alaskan (n=223; AUC, 0.938). In secondary analyses, a separate neural network was able to discern racial subgroup category (black/African American [AUC, 0.84], and white, non-Hispanic [AUC, 0.76] in a 5-class classifier), and a network trained only in non-Hispanic whites from the original derivation cohort performed similarly well across a range of racial/ethnic subgroups in the testing cohort with an AUC of at least 0.930 in all racial/ethnic subgroups. Conclusions: Our study demonstrates that while ECG characteristics vary by race, this did not impact the ability of a convolutional neural network to predict low left ventricular ejection fraction from the ECG. We recommend reporting of performance among diverse ethnic, racial, age, and sex groups for all new artificial intelligence tools to ensure responsible use of artificial intelligence in medicine.

Abstract Background Nurse leaders who are mothers are at significant risk for experiencing stress, burnout, and occupational fatigue. Authentic Connections (AC) Groups is an intervention shown to be effective for fostering resilience among at‐risk moms, including physicians; however, it has not previously been tested with nurse leaders. Aims Our aims were to test the feasibility and acceptability of the AC Groups intervention with nurse leader mothers and examine its effects on participant resilience, as measured by increased self‐compassion and decreased distress, depression, perceived stress, and burnout. Methods A randomized controlled trial design was employed for this pilot study, with 36 nurse leaders at Mayo Clinic. AC participants attended group sessions for an hour per week for 12 weeks. Control group members were provided 1 hr per week of free time over 12 weeks. Multiple self‐report psychological measures were completed at baseline, post‐intervention, and 3‐month follow‐up. Results The AC Groups intervention was feasible and well‐received by nurse leaders. Session attendance rates averaged 92%. Despite the small n ’s, repeated measures of Analysis of Variance showed significantly greater improvements ( p < .05) for participants in the AC Groups than control condition for depression, self‐compassion, and perceived stress, with large effect sizes ( 0.18–0.22). In addition, effect sizes for anxiety and feeling loved approximated the moderate range ( 0.05 and .07). Linking Evidence to Action The AC intervention shows promise as a feasible intervention for mitigating nurse leader mothers’ stress by positively impacting indices of well‐being, including depression, self‐compassion, and perceived stress. Given, the prevalence of stress and burnout among nurse leaders, the effectiveness of the AC intervention in fostering resilience in this population has significant implications for research and practice. Further research is warranted with larger numbers from multiple sites, longer follow‐up periods, and biomarker measures of stress.

Abstract Background Serum free light chain (FLC) assays are used clinically to measure the concentration of κ and λ FLC in patients with suspected or diagnosed plasma cell proliferative disorders. Previous studies have demonstrated a loss of linearity in low concentration ranges of these assays. We hypothesized that this result could be caused by a matrix effect. Methods Recovery studies were performed for κ and λ FLC in both serum and saline using the Freelite assay (Binding Site) on a Cobas c502 system (Roche). Samples were analyzed either at the recommended dilution or undiluted. Follow-up studies were performed in varying matrices ranging from 0% to 100% saline. Retrospective patient data were analyzed to assess the impact on reported κ FLC, λ FLC, and κ/λ ratio. Results FLC in a serum matrix demonstrated underrecovery relative to samples diluted in saline for both κ and λ FLC. Of 255 patient samples with λ FLC measured undiluted (λ FLC &lt;6.0 mg/L), an unexpected gap was observed in patient results between 2.0 and 6.0 mg/L. In addition, 23 patients measured serially with λ FLC between 2.0 and 6.0 mg/L demonstrated dramatic changes in κ/λ ratio, with no changes in κ FLC, likely because of the matrix effect. Conclusions The κ and λ Freelite assays exhibit a matrix effect when samples are tested undiluted, which has the potential to affect the κ/λ ratio. Consequently, our laboratory has stopped reporting λ FLC &lt;6.0 mg/L.

AbstractBackgroundSeveral registry-based analyses suggested a survival advantage for married versus single patients with pancreatic cancer. The mechanisms underlying the association of marital status and survival are likely multiple and complex and, therefore, may be obscured in analyses generated from large population-based databases. The goal of this research was to characterize this potential association of marital status with outcomes in patients with resected pancreatic cancer who underwent combined modality adjuvant therapy on a prospective clinical trial.Materials and MethodsThis is an ancillary analysis of 367 patients with known marital status treated on NRG Oncology/RTOG 97-04. Survival analysis was performed using the Kaplan-Meier method and compared using the log-rank test. Multivariate analysis was performed using the Cox proportional hazards regression model.ResultsOf 367 patients, 271 (74%) were married or partnered and 96 (26%) were single. Married or partnered patients were more likely to be male. There was no association between marital status and overall survival (OS) or disease-free survival (DFS) on univariate (hazard ratio [HR], 1.09 and 1.01, respectively) or multivariate analyses (HR, 1.05 and 0.98, respectively). Married or partnered male patients did not have improved survival compared with female or single patients.ConclusionAncillary analysis of data from NRG Oncology/RTOG 97-04 demonstrated no association between marital and/or partner status and OS or DFS in patients with resected pancreatic cancer who received adjuvant postoperative chemotherapy followed by concurrent external beam radiation therapy and chemotherapy. Clinical trial identification number. NCT00003216.

AimsChemotherapy‐induced peripheral neuropathy (PN) is a treatment limiting toxicity of paclitaxel. We evaluated if EPHA genetic variation (EPHA4, EPHA5, EPHA6, and EPHA8) is associated with PN sensitivity by accounting for variability in systemic paclitaxel exposure (time above threshold).MethodsGermline DNA from 60 patients with breast cancer was sequenced. PN was measured using the 8‐item sensory subscale (CIPN8) of the patient‐reported CIPN20. Associations for 3 genetic models were tested by incorporating genetics into previously published PN prediction models integrating measured paclitaxel exposure and cumulative treatment. Significant associations were then tested for association with PN‐related treatment disruption.ResultsEPHA5 rs7349683 (minor allele frequency = 0.32) was associated with increased PN sensitivity (β‐coefficient = 0.39, 95% confidence interval 0.11–0.67, p = 0.007). Setting a maximum tolerable threshold of CIPN8 = 30, optimal paclitaxel exposure target is shorter for rs7349683 homozygous (11.6 h) than heterozygous (12.6 h) or wild‐type (13.6 h) patients. Total number of missense variants (median = 0, range 0–2) was associated with decreased PN sensitivity (β‐coefficient: −0.42, 95% confidence interval −0.72 to −0.12, P = .006). No association with treatment disruption was detected for the total number of missense variants or rs7349683.ConclusionIsolating toxicity sensitivity by accounting for exposure is a novel approach, and rs7349683 represents a promising marker for PN sensitivity that may be used to individualize paclitaxel treatment.

Abstract Background Schwannomas are rare tumours that pose a significant management challenge in the abdomen, retroperitoneum and pelvis. No data are available to inform management strategy. Methods A collaborative international cohort study, across specialist sarcoma units, was conducted to include adults presenting between 2000 and 2017 with histopathologically confirmed schwannomas within the abdomen, retroperitoneum or pelvis. Results Of 485 patients across 12 centres, 38 (7·8 per cent) were discharged without follow-up, 199 (41·0 per cent) underwent early resection and 248 (51·1 per cent) had radiological monitoring. Of these 248 patients, 96 (38·7 per cent) eventually had surgery, giving an overall resection rate of 60·8 per cent (295 of 485). At baseline, median tumour volume was 90·1 (i.q.r. 26·5–262·0) cm3. The estimated growth rate was 10·5 (95 per cent c.i. 9·4 to 11·6) per cent per year, and was consistent in the short term (within 2 years of diagnosis) and long term (beyond 2 years) (ρ = 0·405, P = 0·021). A decision to operate was more common in symptomatic patients (P &lt; 0·001) and for rapidly growing tumours (growth rate more than 20 per cent per year) (P = 0·025). R0/R1 resection was achieved in 91·6 per cent of patients (263 of 287). Kaplan–Meier long-term recurrence rates after R0/R1 resection were 2·3 and 6·7 per cent at 3 and 5 years respectively. Conclusion Specific recommendations include: indications for early surgery, prediction of growth from radiological monitoring, promotion of selective submacroscopic resection and cessation of postoperative imaging surveillance.

Abstract Background Adjuvant bisphosphonates, when given in a low-estrogen environment, can decrease breast cancer recurrence and death. Treatment guidelines include recommendations for adjuvant bisphosphonates in postmenopausal patients. SWOG/Alliance/Canadian Cancer Trials Group/ECOG-ACRIN/NRG Oncology study S0307 compared the efficacy of three bisphosphonates in early-stage breast cancer. Methods Patients with stage I–III breast cancer were randomly assigned to 3 years of intravenous zoledronic acid, oral clodronate, or oral ibandronate. The primary endpoint was disease-free survival (DFS) with overall survival as a secondary outcome. All statistical tests were two-sided. Results A total of 6097 patients enrolled. Median age was 52.7 years. Prior to being randomly assigned, 73.2% patients indicated preference for oral vs intravenous formulation. DFS did not differ across arms in a log-rank test (P = .49); 5-year DFS was 88.3% (zoledronic acid: 95% confidence interval [CI] = 86.9% to 89.6%), 87.6% (clodronate: 95% CI = 86.1% to 88.9%), and 87.4% (ibandronate: 95% CI = 85.6% to 88.9%). Additionally, 5-year overall survival did not differ between arms (log rank P = .50) and was 92.6% (zoledronic acid: 95% CI = 91.4% to 93.6%), 92.4% (clodronate: 95% CI = 91.2% to 93.5%), and 92.9% (ibandronate: 95% CI = 91.5% to 94.1%). Bone as first site of recurrence did not differ between arms (P = .93). Analyses based on age and tumor subtypes showed no treatment differences. Grade 3/4 toxicity was 8.8% (zoledronic acid), 8.3% (clodronate), and 10.5% (ibandronate). Osteonecrosis of the jaw was highest for zoledronic acid (1.26%) compared with clodronate (0.36%) and ibandronate (0.77%). Conclusions We found no evidence of differences in efficacy by type of bisphosphonate, either in overall analysis or subgroups. Despite an increased rate of osteonecrosis of the jaw with zoledronic acid, overall toxicity grade differed little across arms. Given that patients expressed preference for oral formulation, efforts to make oral agents available in the United States should be considered.

Abstract Abstract 3110 High-dose therapy followed by autologous stem cell transplantation (ASCT) is standard therapy for patients with relapsed or refractory Hodgkin Lymphoma (HL), although the optimal conditioning regimen remains uncertain. The three drug alkylating agent regimen, BuMelTt, has been established as our institutional regimen. We performed a retrospective review of outcomes obtained with BuMelTt compared to those achieved with other standard conditioning regimens. 133 patients with relapsed/refractory HL who underwent ASCT between January 1990 and December 2009 were analyzed. 62 patients received BuMelTt and 71 patients received other standard preparative regimens (Standard) consisting of CBV (44), CyVp16TBI (20), BEAM (4), CyTBI (2), and Mel (1). The median follow-up (range) was 4.3 years (0.01–14.24) in BuMelTt and 3.9 years (0.07–20.21) in Standard. The two groups (BuMelTt vs Standard) were balanced for gender (65% male vs 65% male), median age (29 vs 33), median number of prior regimens (2 vs 2), and pre-transplant disease status. Disease status was categorized as PR2/CR2 (48% vs 42%), primary induction failure sensitive (16% vs 17%), and Other which included PIF-refractory, 1st relapse refractory, and greater than 1st relapse (35% vs 41%). The 5 year overall survival (OS) was superior for patient who received BuMelTt conditioning (74% vs 54%, p = 0.03). There was also a trend for improved 5 year event free survival (EFS) for BuMelTt at 56% vs 39% for Standard (p=0.07). Other risk factors for outcome after ASCT were also analyzed and included disease status, more than 2 chemotherapy regimens before ASCT, Karnofsky score < 90% at ASCT, time from diagnosis to ASCT < 12 months, prior extranodal involvement, and age > 40 years. By univariate analysis, >2 chemotherapy regimens before ASCT (HR 1.7, p=0.05) and disease status of Other (HR 2.8, p<0.01) were predictive for lower OS. There was a trend for influence on OS for BuMelTt (HR 0.6, p=0.07) and PIF-sensitive (HR 1.9, p=0.09). By multivariate analysis, disease status of PIF-sensitive or Other predicted worse outcomes for OS (HR 2.4 & 2.9, p 0.03 & <0.01, respectively), and there was still a trend for benefit with BuMelTt (HR 0.6, p=0.09), although >2 chemotherapy regimens lost significance (p=0.22). For EFS, disease status of PIF-sensitive or Other were prognostic in univariate analysis (HR 2.1 or 2.3, p=0.02 or <0.01, respectively), and there was a trend for improved EFS with BuMelTt (HR 0.66, p=0.08). In multivariate analysis, disease status of PIF-sensitive or Other remained prognostic (p=0.03 & 0.004, respectively), while BuMelTt was not significant (p=0.27). There was no significant difference in 100 day non-relapse related mortality for BuMelTt vs Standard conditioning (5% vs 3%, p=0.88). More patients developed NCI CTCAE grade 3/4 mucositis with BuMelTt than Standard (89% vs 61%, p<0.01), but fewer patients had bacteremia with BuMelTt than Standard (7% vs 23%, p=0.02). There was no difference in duration of hospitalization, episodes of febrile neutropenia, incidence of veno-occlusive disease, time to neutrophil engraftment, or time to platelet engraftment. Our results suggest that BuMelTt may improve OS and EFS compared to standard conditioning regimens with increased but manageable toxicity. A case matched control study is planned. Disclosures: No relevant conflicts of interest to declare.

Abstract Previous work has suggested a relationship between disease activity and in vivo birth rates of chronic lynphocytic leukemia (CLL) cells, as measured by incorporation of deuterium into DNA during heavy water (2H2O) consumption. By extension, CLL-cell kinetics may be a useful prognostic marker in early stage disease, with an increased birth rate predicting shorter time to disease progression and poor prognosis. We are conducting a multi-center study to evaluate the relevance of proliferation rates to both known prognostic markers and to disease progression during 2.5 years of clinical follow up in previously untreated patients with early stage disease. Patients are given a daily dose of 2H2O for 6 weeks and CLL-cell birth rates are measured. We examined the leukemia cells of 36 patients for known prognostic markers (ZAP 70 expression, CD38 expression, IgVH mutation status) and CLL birth rates. Subjects were Rai Stage 0 (n=15), 1 (n=17), or 2 (n=4); mean age 57 +/− 9 (range 41–78); mean time from diagnosis to enrollment 2.1 +/− 2.2 y (range 0.1–10 y). ZAP-70 expression (20% cutoff) was positive in 44%, CD38 expression (30% cutoff) was positive in 35%, and use of unmutated IgVH genes was found in 45% of the subjects. CLL-cell birth rates ranged from 0.09 to 0.97% per day. Utilizing the previously published cut-off value of 0.35% newly divided cells per day, 33% of subjects had a high proliferation rate. Data on all biomarkers were available for 33 subjects. Nine subjects had 4/4 favorable prognostic indicators and 5 subjects had 4/4 unfavorable prognostic indicators. Agreement between prognostic markers, as assessed by kappa coefficient, was strongest for the ZAP-70 v. IgVH mutational status comparison. Comparisons were also significant for ZAP-70 v. CD38 and IgVH mutational status v. CD38. Birth rate was not significantly correlated with any of the other prognostic markers, suggesting it may be a useful independent predictor of disease progression. Isolation of CLL-cells from the bone marrow at the end of the 6 wk labeling period in 5 subjects revealed matching proliferation rates in the marrow and blood in 3 subjects. In 2 subjects, marrow enrichment was lower, suggesting a more slowly proliferating pool in the marrow, the pathophysiologic significance of which is unclear. In 5 subjects, kinetic analysis revealed a very delayed entry of newly divided cells into the blood compartment, presumably from tissue sites of proliferation. This parameter may reflect pool size of CLL-cells in tissues and correlation with lymph node burden is being evaluated. In summary, analysis of CLL-cell birth rates in vivo in these early stage patients provides unique information about the biology of CLL. Correlations with disease progression during the follow-up period will help determine the utility of this test as a prognostic biomarker.