Category: Statistics

Accid Anal Prev. 2021 Dec 8;165:106511. doi: 10.1016/j.aap.2021.106511. Online ahead of print.

ABSTRACT

Real-time crash prediction is a heavily studied area given their potential applications in proactive traffic safety management in which a plethora of statistical and machine learning (ML) models have been developed to predict traffic crashes in real-time. However, one of the fundamental issues relating to the application of these models is spatio-temporal transferability. The present paper attempts to address this gap of knowledge by combining Generative Adversarial Network (GAN) and transfer learning to examine the transferability of real-time crash prediction models under an extremely imbalanced data setting. Initially, a baseline model was developed using Deep Neural Network (DNN) with crash and microscopic traffic data collected from M1 Motorway in the UK in 2017. The dataset utilised in the baseline model is naturally imbalanced with 257 crash cases and 16,359,163 non-crash cases. To overcome data imbalance issue, Wasserstein GAN (WGAN) was utilised to generate synthetic crash data. Non-crash data were randomly undersampled due to computational limitations. The calibrated model was then applied to predict traffic crashes for five other datasets obtained from M1 (2018), M4 (2017 & 2018 separately) and M6 Motorway (2017 & 2018 separately) by using transfer learning. Model transferability was compared with standalone models and direct transfer from the baseline model. The study revealed that direct transfer is not feasible. However, models become transferable temporally, spatially, and spatio-temporally if transfer learning is applied. The predictability of the transferred models outperformed existing studies by achieving high Area Under Curve (AUC) values ranging between 0.69 and 0.95. The best transferred model can predict nearly 95% crashes with only a 5% false alarm rate by tuning thresholds. Furthermore, the performances of transferred models are on par with or better than the standalone model. The findings of this study proves that transfer learning can improve model transferability under extremely imbalanced settings which helps traffic engineers in developing highly transferable models in future.

PMID:34894483 | DOI:10.1016/j.aap.2021.106511

Lung Cancer. 2021 Dec 1;163:27-34. doi: 10.1016/j.lungcan.2021.11.015. Online ahead of print.

ABSTRACT

OBJECTIVE: Two large randomized controlled trials (RCTs) have demonstrated that low dose computed tomography (LDCT) screening reduces lung cancer mortality. Risk-prediction models have been proved to select individuals for lung cancer screening effectively. With the focus on established risk factors for lung cancer routinely available in general cancer screening settings, we aimed to develop and internally validated a risk prediction model for lung cancer.

MATERIALS AND METHODS: Using data from the Cancer Screening Program in Urban China (CanSPUC) in Henan province, China between 2013 and 2019, we conducted a prospective cohort study consisting of 282,254 participants including 126,445 males and 155,809 females. Detailed questionnaire, physical assessment and follow-up were completed for all participants. Using Cox proportional risk regression analysis, we developed the Henan Lung Cancer Risk Models based on simplified questionnaire. Model discrimination was evaluated by concordance statistics (C-statistics), and model calibration was evaluated by the bootstrap sampling, respectively.

RESULTS: By 2020, a total of 589 lung cancer cases occurred in the follow-up yielding an incident density of 64.91/100,000 person-years (pyrs). Age, gender, smoking, history of tuberculosis and history of emphysema were included into the model. The C-index of the model for 1-year lung cancer risk was 0.766 and 0.741 in the training set and validation set, respectively. In stratified analysis, the model showed better predictive power in males, younger participants, and former or current smoking participants. The model calibrated well across the deciles of predicted risk in both the overall population and all subgroups.

CONCLUSIONS: We developed and internally validated a simple risk prediction model for lung cancer, which may be useful to identify high-risk individuals for more intensive screening for cancer prevention.

PMID:34894456 | DOI:10.1016/j.lungcan.2021.11.015

Breast. 2021 Nov 25;61:29-34. doi: 10.1016/j.breast.2021.11.012. Online ahead of print.

ABSTRACT

PURPOSE: To evaluate the readability, understandability, and actionability of online patient education materials (OPEM) related to breast cancer risk assessment.

MATERIAL AND METHODS: We queried seven English-language search terms related to breast cancer risk assessment: breast cancer high-risk, breast cancer risk factors, breast cancer family history, BRCA, breast cancer risk assessment, Tyrer-Cuzick, and Gail model. Websites were categorized as: academic/hospital-based, commercial, government, non-profit or academic based on the organization hosting the site. Grade-level readability of qualifying websites and categories was determined using readability metrics and generalized estimating equations based on written content only. Readability scores were compared to the recommended parameters set by the American Medical Association (AMA). Understandability and actionability of OPEM related to breast cancer high-risk were evaluated using the Patient Education Materials Assessment Tool (PEMAT) and compared to criteria set at ≥70%. Descriptive statistics and inter-rater reliability analysis were utilized.

RESULTS: 343 websites were identified, of which 162 met study inclusion criteria. The average grade readability score was 12.1 across all websites (range 10.8-13.4). No website met the AMA recommendation. Commercial websites demonstrated the highest overall average readability of 13.1. Of the 26 websites related to the search term breast cancer high-risk, the average understandability and actionability scores were 62% and 34% respectively, both below criteria.

CONCLUSIONS: OPEM on breast cancer risk assessment available to the general public do not meet criteria for readability, understandability, or actionability. To ensure patient comprehension of medical information online, future information should be published in simpler, more appropriate terms.

PMID:34894464 | DOI:10.1016/j.breast.2021.11.012

Lung Cancer. 2021 Nov 25;163:14-18. doi: 10.1016/j.lungcan.2021.11.010. Online ahead of print.

ABSTRACT

PURPOSE: Response Evaluation Criteria in Solid Tumors (RECIST) has limitations but remains the conventional approach for tumor assessments. We explored whether circulating tumor DNA (ctDNA) can be incorporated into RECIST to provide a more robust measure of tumor response in advanced EGFR-mutant NSCLC.

PATIENTS AND METHODS: In FASTACT-2, patients with advanced NSCLC received platinum/gemcitabine intercalated with erlotinib or placebo. EGFR mutation (tumor and plasma ctDNA) was detected using cobas v2. Patients selected for this hypothesis-generating analysis had EGFR mutations (on either tumor or plasma) at baseline and evaluable week 8 plasma EGFR. Week 8 ctDNA and radiologic response status were correlated with survival using landmark cox regression analyses.

RESULTS: Of the original 451 patients, 86 (19.1%) were eligible for this analysis. 73% (n = 63) had detectable ctDNA at baseline. At week 8, 40% (n = 34) had radiologic partial response (PR), 60% (n = 52) had stable disease (SD); 80% (n = 69) had a ctDNA response (undetectable ctDNA). In patients who had initial PR and undetectable ctDNA, 93% (28/30) had ongoing PR subsequently at week 16. The median duration of response was 14.9 months. In patients with SD and undetectable ctDNA at week 8, 28% had radiological PR at week 16. Amongst those with PR at week 8, survival outcomes for those with undetectable vs detectable ctDNA were not statistically significant (PFS HR 0.49, 95%CI 0.16-1.48, p = 0.21; OS HR 0.39, 95%CI 0.13-1.19, p = 0.10). Amongst those with SD at week 8, there was significantly longer survival for those with undetectable vs detectable ctDNA (PFS HR 0.27, 95% CI 0.13-0.57, p < 0.0001; OS HR 0.40, 95% CI 0.20-0.80, p = 0.009).

CONCLUSION: In patients with SD, undetectable ctDNA at week 8 correlated with survival improvement. Both radiologic and ctDNA responses are prognostic of PFS. Incorporation of ctDNA with RECIST may improve tumor response assessment in EGFR-mutant NSCLC.

PMID:34894454 | DOI:10.1016/j.lungcan.2021.11.010

Ophthalmic Physiol Opt. 2021 Dec 11. doi: 10.1111/opo.12931. Online ahead of print.

ABSTRACT

PURPOSE: To expand the SyntEyes keratoconus (KTC) model to assess the Visual Image Quality (VIQ) of sphero-cylindrical spectacle and rigid contact lens corrections as keratoconus progresses.

METHODS: The previously published SyntEyes KTC eye model to determine best sphero-cylindrical spectacle and rigid contact lens correction in keratoconic eyes was expanded to include the natural progression of keratoconus, thus allowing the assessment of corrected VIQ with disease progression.

RESULTS: As keratoconus progresses, the pattern of visual Strehl ratio (VSX) in correction space for spectacles alters from a typical hourglass into a shell pattern. The former would guide the subjective refraction towards the optimal correction while the latter is relatively insensitive to large dioptric steps. In 15 out of the 20 SyntEyes, the shell pattern eventually produces two foci on different sides of the correction space separated by a clinically significant dioptric difference with a similar, albeit lower VIQ. Wearing the best possible spectacle corrections provided an average gain of up to 3.5 lines of logMAR visual acuity compared to the uncorrected cases, which increased to 5.5 lines for the best rigid contact lens correction. Continuing to wear a spectacle correction as the disease progresses often leads to a VIQ that is almost as bad as the uncorrected case. Continuing to wear a rigid contact lens correction as the disease progresses maintains a relatively high level of VIQ, albeit in the low range for typically well-corrected normal eyes.

CONCLUSIONS: The results reflect the clinical experience that subjective refraction is difficult in highly-aberrated keratoconic eyes, the benefit of spectacle correction is short lived and that rigid contact lenses provide better and more stable VIQ with disease progression. Other aspects, such as the presence and behaviour of the second focus in some cases, remain to be confirmed clinically.

PMID:34894167 | DOI:10.1111/opo.12931

Int J Cancer. 2021 Dec 11. doi: 10.1002/ijc.33893. Online ahead of print.

ABSTRACT

The increases observed in incidence and survival of oropharyngeal squamous cell carcinoma (OPSCC) have been attributed to human papillomavirus (HPV) infection, but the survival-impact of specific genotypes is poorly understood. We investigated the potential influence of HPV genotypes on survival in HPV-positive (HPV+) OPSCC. All patients with HPV+/p16+ OPSCC and available genotype data within the period 2011-2017 in Eastern Denmark were included. Descriptive statistics on clinical and tumor data, as well as overall survival (OS) and recurrence-free survival (RFS) with Cox hazard models and Kaplan-Meier plots were performed. Overall, 769 HPV+/p16+ OPSCC patients were included of which genotype HPV16 accounted for 86% (n=662). Compared to high-risk non-HPV16 genotypes (HR non-HPV16), HPV16 patients were younger at diagnosis (median years, 60 vs. 64), had a higher male to female ratio (3.7:1 vs. 2.1:1), and lower performance scores of <1 (90%, n=559, vs. 81%, n=74). Regarding 5-year OS and RFS, no difference was observed between HPV16 and HR non-HPV16 patients. Subgrouping the HR non-HPV16 group into HPV33 (n=57), HPV35 (n=26) and “other genotypes” (n=24) a significantly worse OS in the “other genotypes” group (hazard rate: 2.33, p=0.027) was shown. With similar survival results between HPV16 and non-HPV16 genotypes, genotyping in OPSCC is interesting from an epidemiological point of view as well as in vaccination programs, but not a necessary addition in prognostication of HPV+/p16+ OPSCC.

PMID:34894151 | DOI:10.1002/ijc.33893

Laryngoscope. 2021 Dec 11. doi: 10.1002/lary.29965. Online ahead of print.

ABSTRACT

INTRODUCTION: Lipoinjection is one of the available treatments for unilateral vocal fold paralysis.

OBJECTIVE: To evaluate lipoinjection predictability, and analyze the differences in safety and efficacy of the different techniques.

STUDY DESIGN: Systematic review and meta-analysis.

METHODS AND RESULTS: A systematic review on Medline, Cochrane, and Scopus databases included 49 articles analyzing the data of 1,166 patients, concerning technical details and voice parameters changes. Lipoinjection used a mean volume of 1.3 mL, 95% confidence interval (CI) (0.92, 1.69)-average overcorrection of 30%. Meta-analysis of pre- and postoperative voice parameters’ means showed a significant improvement at 6 months of mean phonation time (preoperative: 5.12, 95% CI [4.48, 5.76]-6 months: 10.46, 95% CI [9.18, 11.75]), Jitter (preoperative: 2.71, 95% CI [2.08, 3.33])-6 months: 1.37, 95% CI [1.05, 1.70]), Shimmer (preoperative: 4.55, 95% CI [3.04, 6.07]-6 months: 2.57, 95% CI [1.69, 3.45]), grade (preoperative: 2.15, 95% CI [1.73, 2.57]-6 months: 0.12, 95% CI [0.97, 1.43]), breathiness (preoperative: 2.012, 95% CI [1.48, 2.55]-6 months: 0.99, 95% CI [0.58, 1.40]), and asthenia (preoperative: 1.90, 95% CI [1.33, 2.47]-6 months: 0.75, 95% CI [0.17, 1.33]) of GRBAS (Grade, Roughness, Breathiness, Asthenia and Strain), and Voice Handicap Index-30 (preoperative: 72.06, 95% CI [54.35, 89.76]-6 months: 26.24, 95% CI [19.58, 32.90]). Subgroup analysis by harvesting technique concluded in no statistically significant difference between them. Few complications were reported. Reintervention was only required for 86 patients.

CONCLUSION: Lipoinjection seems a safe therapeutic option for unilateral vocal fold paralysis, with available data showing an efficacy lasting 6 months to 1 year. Laryngoscope, 2021.

PMID:34894158 | DOI:10.1002/lary.29965

J Burn Care Res. 2021 Nov 19:irab227. doi: 10.1093/jbcr/irab227. Online ahead of print.

ABSTRACT

Burn injuries are among the most traumatic incidences which could be associated with substantial morbidity and mortality. The incidence of these injuries differs based on population socioeconomic characteristics. In this study, we aim to evaluate the distribution of burn injuries based on socioeconomic and causative factors. Relative features from the targeted population were obtained and evaluated. Data were extracted in August 2019 and statistical analysis of the data was conducted from February 2020 to April 2020. This cross-sectional record-based study was conducted from 2008 to 2016 in a main referral center for burn injuries in the southern Iran. Participants included burn survivors and burn patients who were registered as outpatients or inpatients. A total of 4919 burn cases with a mean age of 31.91 ± 17.33, including 2926 (59.5%) males and 1993 (40.5%) females was reported. The majority of our cases had an educational level of under diploma (40.72%). A significant correlation between age, gender, and percentage of burn with the level of education was recorded (P < .001). The most frequent cause of burn injuries was flame with 2537 (51.9%) cases. The most susceptible population to burn injuries were reported to be poorly educated individuals, which emphasizes the preventive role of education.

PMID:34894143 | DOI:10.1093/jbcr/irab227

J Chromatogr Sci. 2021 Dec 11:bmab131. doi: 10.1093/chromsci/bmab131. Online ahead of print.

ABSTRACT

An isocratic reversed-phase high performance liquid chromatographic method has been developed and validated to simultaneously determine nicotinic acid, pravastatin sodium, rosuvastatin calcium, atorvastatin calcium, pitavastatin calcium, lovastatin sodium and simvastatin sodium in focus on counterfeit drug detection. Thin-layer chromatography, nuclear magnetic resonance and mass spectrometry have been additionally performed to verify the identification of adulterants of counterfeit herbal medicines. Chromatographic separation was carried out on Inertsil® ODS-3 C18 (4.6 × 150 mm, 5 μm) with isocratic mobile phase elution containing a mixture of acetonitrile: methanol: 25 mM potassium dihydrogen phosphate buffer, pH 2.86 adjusted with 0.1 M o-phosphoric acid (48: 30: 22, v/v/v), at a flow rate of 1 mL/min and with UV detection at 238 nm. The design of experiment methodology, Plackett-Burman and Box-Behnken designs, was used to screen and optimize the mobile phase composition. The validation of the method was also carried out under the International Conference on Harmonization guidelines. The developed method was sensitive, accurate, simple, economical and highly robust, in addition to the comprehensiveness and novelty of this method for separating the seven drugs. The results were statistically compared with the reference methods used Student’s t-test and variance ratio F-test at P < 0.05.

PMID:34894146 | DOI:10.1093/chromsci/bmab131

Physiol Rep. 2021 Dec;9(23):e15141. doi: 10.14814/phy2.15141.

ABSTRACT

We conducted a meta-analysis to synthesize the best available evidence comparing cardiac biventricular structure and function using cardiac magnetic resonance imaging (CMR) and transthoracic echocardiography (TTE) in elite female athletes and healthy controls (HC). Chronic exposure to exercise may induce cardiac chamber enlargement as a means to augment stroke volume, a condition known as the “athlete’s heart.” These changes have not been clearly characterized in female athletes. Multiple databases were searched from inception to June 18, 2019. Outcomes of interest included left ventricular (LV) and right ventricular (RV) dimensional, volumetric, mass, and functional assessments in female athletes. Most values were indexed to body surface area. The final search yielded 22 studies, including 1000 female athletes from endurance, strength, and mixed athletic disciplines. CMR-derived LV end-diastolic volume (LVEDV) and RV end-diastolic volume (RVEDV) were greater in endurance athletes (EA) versus HC (17.0% and 18.5%, respectively; both p < 0.001). Similarly, TTE-derived LVEDV and RVEDV were greater in EA versus HC (16.8% and 28.0%, respectively; both p < 0.001). Both LVEF and RVEF were lower in EA versus HC, with the most pronounced difference observed in RVEF via TTE (9%) (p < 0.001). LV stroke volume was greater in EA versus HC via both CMR (18.5%) and TTE (13.2%) (both p < 0.05). Few studies reported data for the mixed athlete (MA) population and even fewer studies reported data for strength athletes (SA), therefore a limited analysis was performed on MA and no analysis was performed on SA. This evidence-synthesis review demonstrates the RV may be more susceptible to ventricular enlargement. General changes in LV and RV structure and function in female EA mirrored changes observed in male counterparts. Further studies are needed to determine if potential adverse outcomes occur secondary to these changes.

PMID:34894105 | DOI:10.14814/phy2.15141