Patient-centric Comparability Assessment of Biopharmaceuticals.

The comparability assessment of a biological product after implementing a manufacturing process change should involve a risk-based approach. Process changes may occur at any stage of the product lifecycle: early development, clinical manufacture for pivotal trials, or post-approval. The risk of the change to impact product quality varies. The design of the comparability assessment should be adapted accordingly. A working group reviewed and consolidated industry approaches to assess comparability of traditional protein-based biological products during clinical development and post-approval. The insights compiled in this review article encompass topics such as a risk-evaluation strategy, the design of comparability studies, definition of assessment criteria for comparability, holistic evaluation of data, and the regulatory submission strategy. These practices can be leveraged across the industry to help companies in design and execution of comparability assessments, and to inform discussions with global regulators.

Helping reviewers assess statistical analysis: A case study from analytic methods.

Analytic methods development, like many other disciplines, relies on experimentation and data analysis. Determining the contribution of a paper or report on a study incorporating data analysis is typically left to the reviewer's experience and good sense, without reliance on structured guidelines. This is amplified by the growing role of machine learning driven analysis, where results are based on computer intensive algorithm applications. The evaluation of a predictive model where cross validation was used to fit its parameters adds challenges to the evaluation of regression models, where the estimates can be easily reproduced. This lack of structure to support reviews increases uncertainty and variability in reviews. In this paper, aspects of statistical assessment are considered. We provide checklists for reviewers of applied statistics work with a focus on analytic method development. The checklist covers six aspects relevant to a review of statistical analysis, namely: (1) study design, (2) algorithmic and inferential methods in frequentism analysis, (3) Bayesian methods in Bayesian analysis (if relevant), (4) selective inference aspects, (5) severe testing properties and (6) presentation of findings. We provide a brief overview of these elements providing references for a more elaborate treatment. The robustness analysis of an analytical method is used to illustrate how an improvement can be achieved in response to questions in the checklist. The paper is aimed at both engineers and seasoned researchers.

To tolerate or to agree: A tutorial on tolerance intervals in method comparison studies with BivRegBLS R Package.

The well-known agreement interval by Bland and Altman is extensively applied in method comparison studies. Two clinical measurement methods are considered interchangeable if their differences are not clinically significant. The agreement interval is commonly applied to assess the spread of the differences. However, this interval is approximate (too narrow) and several authors propose calculating a confidence interval around each bound. This article demonstrates that this approach is misleading, awkward, and confusing. On the other hand, tolerance intervals are exact and can include a confidence level if needed. Tolerance intervals are also easier to calculate and to interpret. Real data sets are used to illustrate the tolerance intervals with the R package BivRegBLS under normal or log-normal assumptions. Furthermore, it is also explained how to assess the coverage probabilities of the tolerance intervals with simulations.

Confidence, prediction, and tolerance in linear mixed models.

The literature about Prediction Interval (PI) and Tolerance Interval (TI) in linear mixed models is usually developed for specific designs, which is a main limitation to their use. This paper proposes to reformulate the two-sided PI to be generalizable under a wide variety of designs (one random factor, nested and crossed designs for multiple random factors, and balanced or unbalanced designs). This new methodology is based on the Hessian matrix, namely, the inverse of (observed) Fisher Information matrix, and is built with a cell mean model. The degrees of freedom for the total variance are calculated with the generalized Satterthwaite method and compared to the Kenward-Roger's degrees of freedom for fixed effects. Construction of two-sided TIs are also detailed with one random factor, and two nested and two crossed random variables. An extensive simulation study is carried out to compare the widths and coverage probabilities of Confidence Intervals (CI), PIs, and TIs to their nominal levels. It shows excellent coverage whatever the design and the sample size are. Finally, these CIs, PIs, and TIs are applied to two real data sets: one from orthopedic surgery study (intralesional resection risk) and the other from assay validation study during vaccine development.

Pedicle screw insertion accuracy in terms of breach and reposition using a new intraoperative cone beam computed tomography imaging technique and evaluation of the factors associated with these parameters of accuracy: a series of 695 screws.

PURPOSE: The goals of this study were to assess the accuracy of pedicle screw insertion using an intraoperative cone beam computed tomography (CBCT) system, and to analyze the factors potentially influencing this accuracy. METHODS: Six hundred and ninety-five pedicle screws were inserted in 118 patients between October 2013 and March 2016. Screw insertion was performed using 2D-fluoroscopy or CBCT-based navigation. Accuracy was assessed in terms of breach and reposition. All the intraoperative CBCT scans, done after screw insertion, were reviewed to assess the accuracy of screw placement using two established classification systems: Gertzbein and Heary. Generalized linear mixed models were used to model the odds (95% CI) for a screw to lead to a breach according to the independent variables. RESULTS: The breach rate was 11.7% using the Gertzbein classification and 15.4% using the Heary classification. Seventeen screws (2.4%) were repositioned intraoperatively. The only factor affecting statistically the odds to have a breach was the indication of surgery. The patients with non-degenerative disease had a significantly higher risk of breach than those with degenerative disease. CONCLUSION: Use of intraoperative CBCT as 2D-fluoroscopy or coupled with a navigation system for pedicle screw insertion is accurate in terms of breach occurrence and reposition. However, these rates depend on the classification or grading system used. Use of a navigation system does not decrease the risk of breach significantly. And the risk of breach is higher in non-degenerative conditions (trauma, scoliosis, infection, and malignancy disease) than in degenerative diseases.

Accuracy of a new intraoperative cone beam CT imaging technique (Artis zeego II) compared to postoperative CT scan for assessment of pedicle screws placement and breaches detection.

PURPOSE: The goal of this study was to compare the accuracy of a novel intraoperative cone beam computed tomography (CBCT) imaging technique with that of conventional computed tomography (CT) scans for assessment of pedicle screw placement and breach detection. METHODS: Three hundred and forty-eight pedicle screws were inserted in 58 patients between October 2013 and March 2016. All patients had an intraoperative CBCT scan and a conventional CT scan to verify the placement of the screws. The CBCT and CT images were reviewed by two surgeons to assess the accuracy of screw placement and detect pedicle breaches using two established classification systems. Agreement on screw placement between intraoperative CBCT and postoperative CT was assessed using Kappa and Gwet's coefficients. Using CT scanning as the gold standard, the sensitivity, specificity, positive predictive value, and negative predictive value were calculated to determine the ability of CBCT imaging to accurately evaluate screw placement. RESULTS: The Kappa coefficient was 0.78 using the Gertzbein classification and 0.80 using the Heary classification, indicating a substantial agreement between the intraoperative CBCT and postoperative CT images. Gwet's coefficient was 0.94 for both classifications, indicating almost perfect agreement. The sensitivity, specificity, positive predictive value and negative predictive value of the CBCT images were 77, 98, 86, and 96%, respectively, for the Gertzbein classification and 79, 98, 88, and 96%, respectively, for the Heary classification. CONCLUSIONS: Intraoperative CBCT provides accurate assessment of pedicle screw placement and enables intraoperative repositioning of misplaced screws. This technique may make postoperative CT imaging unnecessary.

A Randomized, Single-Blind, Crossover Trial of Recovery Time in High-Flux Hemodialysis and Hemodiafiltration.

BACKGROUND: The choice between hemodiafiltration (HDF) or high-flux hemodialysis (HD) to treat end-stage kidney disease remains a matter of debate. The duration of recovery time after treatment has been associated with mortality, affects quality of life, and may therefore be important in informing patient choice. We aimed to establish whether recovery time is influenced by treatment with HDF or HD. STUDY DESIGN: Randomized patient-blinded crossover trial. SETTINGS & PARTICIPANTS: 100 patients with end-stage kidney disease were enrolled from 2 satellite dialysis units in Glasgow, United Kingdom. INTERVENTION: 8 weeks of HD followed by 8 weeks of online postdilution HDF or vice versa. OUTCOMES: Posttreatment recovery time, symptomatic hypotension events, dialysis circuit clotting events, and biochemical parameters. MEASUREMENTS: Patient-reported recovery time in minutes, incidence of adverse events during treatments, hematology and biochemistry results, quality-of-life questionnaire. RESULTS: There was no overall difference in recovery time between treatments (medians for HDF vs HD of 47.5 [IQR, 0-240] vs 30 [IQR, 0-210] minutes, respectively; P=0.9). During HDF treatment, there were significant increases in rates of symptomatic hypotension (8.0% in HDF vs 5.3% in HD; relative risk [RR], 1.52; 95% CI, 1.2-1.9; P<0.001) and intradialytic tendency to clotting (1.8% in HDF vs 0.7% in HD; RR, 2.7; 95% CI, 1.5-5.0; P=0.002). Serum albumin level was significantly lower during HDF (3.2 vs 3.3g/dL; P<0.001). Health-related quality-of-life scores were equivalent. LIMITATIONS: Single center; mean achieved HDF convection volume, 20.6L. CONCLUSIONS: Patients blinded to whether they were receiving HD or HDF in a randomized controlled crossover study reported similar posttreatment recovery times and health-related quality-of-life scores.

Delta method and bootstrap in linear mixed models to estimate a proportion when no event is observed: application to intralesional resection in bone tumor surgery.

Resecting bone tumors requires good cutting accuracy to reduce the occurrence of local recurrence. This issue is considerably reduced with a navigated technology. The estimation of extreme proportions is challenging especially with small or moderate sample sizes. When no success is observed, the commonly used binomial proportion confidence interval is not suitable while the rule of three provides a simple solution. Unfortunately, these approaches are unable to differentiate between different unobserved events. Different delta methods and bootstrap procedures are compared in univariate and linear mixed models with simulations and real data by assuming the normality. The delta method on the z-score and parametric bootstrap provide similar results but the delta method requires the estimation of the covariance matrix of the estimates. In mixed models, the observed Fisher information matrix with unbounded variance components should be preferred. The parametric bootstrap, easier to apply, outperforms the delta method for larger sample sizes but it may be time costly. Copyright © 2016 John Wiley & Sons, Ltd.

How to regress and predict in a Bland-Altman plot? Review and contribution based on tolerance intervals and correlated-errors-in-variables models.

Two main methodologies for assessing equivalence in method-comparison studies are presented separately in the literature. The first one is the well-known and widely applied Bland-Altman approach with its agreement intervals, where two methods are considered interchangeable if their differences are not clinically significant. The second approach is based on errors-in-variables regression in a classical (X,Y) plot and focuses on confidence intervals, whereby two methods are considered equivalent when providing similar measures notwithstanding the random measurement errors. This paper reconciles these two methodologies and shows their similarities and differences using both real data and simulations. A new consistent correlated-errors-in-variables regression is introduced as the errors are shown to be correlated in the Bland-Altman plot. Indeed, the coverage probabilities collapse and the biases soar when this correlation is ignored. Novel tolerance intervals are compared with agreement intervals with or without replicated data, and novel predictive intervals are introduced to predict a single measure in an (X,Y) plot or in a Bland-Atman plot with excellent coverage probabilities. We conclude that the (correlated)-errors-in-variables regressions should not be avoided in method comparison studies, although the Bland-Altman approach is usually applied to avert their complexity. We argue that tolerance or predictive intervals are better alternatives than agreement intervals, and we provide guidelines for practitioners regarding method comparison studies. Copyright © 2016 John Wiley & Sons, Ltd.

Efficacy and safety of ivabradine in patients with chronic systolic heart failure and diabetes: an analysis from the SHIFT trial.

AIMS: To evaluate clinical profiles and outcomes in patients with systolic heart failure (HF) with or without diabetes, and the efficacy and safety of ivabradine (heart rate-lowering agent) with respect to diabetic status. METHODS AND RESULTS: This is a post hoc analysis on patients in SHIFT, a randomized controlled trial in adults in sinus rhythm with systolic HF, left ventricular ejection fraction ≤35%, and resting heart rate ≥70 b.p.m. Patients were randomized to ivabradine (titrated to 7.5 mg bid) or placebo. Diabetic status was established by medical history at baseline. The primary composite endpoint (PCE) was cardiovascular death or hospitalisation for worsening HF. Of 6505 patients, 30% had diabetes, 32% of whom used insulin. The PCE was more frequent in patients with diabetes [adjusted hazard ratio (HR) 1.18, 95% confidence interval (CI) 1.07-1.31; p = 0.001], as was hospitalization for worsening HF (adjusted HR 1.28, 95% CI 1.13-1.44; P < 0.001), and was increased in patients treated with insulin (adjusted HR 1.43, 95% CI 1.23-1.66; P < 0.01 vs. non-diabetics). Ivabradine significantly reduced the PCE in patients with and without diabetes (adjusted HR 0.80, 95% CI 0.68-0.94 and HR 0.84, 95% CI, 0.75-0.95, respectively; interaction P was non-significant) vs. placebo. Adverse events were significantly more frequent in patients with diabetes (78%) than without (74%) (P < 0.001). Regardless of diabetic status, the incidence of serious adverse events was not significantly different between ivabradine and placebo. CONCLUSIONS: Comorbid diabetes in chronic HF worsens the prognosis of systolic HF patients. Irrespective of diabetic status, ivabradine is effective and safe in these patients.

Normal limits of the electrocardiogram in Indians.

AIMS: The objective of the study was to develop normal limits of the ECG in an apparently healthy population of South Asians living in India. METHODS: Three centres contributed to recording 12 lead ECGs on identical digital electrocardiographs. Apparently healthy volunteers were recruited and ECGs were first transferred to a local database and then to Glasgow where all ECGs were analysed by the same University of Glasgow ECG Interpretation Program. RESULTS: A total of 963 individuals were recruited into the study (30.4% female) with an age range of 18-83 years. QRS duration was longer in males than females, QT interval was longer in females than males, and QRS voltages in general were higher in males than females and in younger compared to older individuals. CONCLUSION: Findings in general paralleled those in other populations and suggested that criteria for a white Caucasian population could be applied to a South Asian Indian population.

Local complications of massive bone allografts: an appraisal of their prevalence in 128 patients.

Bone allografts were used in our department since twenty-five years to reconstruct segmental bone loss and our data were retrospectively reviewed to assess the complications related to the use of a bone allograft. A consecutive series of 128 patients who received a bone allograft was analyzed. The minimal follow-up was 18 months. Fracture, nonunion, infection and explantation were investigated using a multivariate analysis and logistical regression. Kaplan-Meier survival of the allograft was performed, using allograft removal as the end point. Tumour disease was excluded from this study. Patients were followed up for an average 103 months. Bone tumour occurred in 78% of the patients whereas revision arthroplasty was the cause of implantation in 15% of them. Nonunion was the most prevalent complication, occurring in 35% of the grafts. For nonunion occurrence, the type of reconstruction was found to be a significant variable, the intercalary allograft being the most exposed. Primary bone autografting at the anastomotic site was not significant to prevent nonunion. Fracture of the allograft was the second most frequent complication with a prevalence of 16.4%. The length of the allograft and an osteoarticular allograft were two significant variables in that occurrence. Infection of the allograft was present with a rate of 5.4% of patients. Explantation of failed allografts occurred in 30% of them. The duration of the frozen storage of the allograft and the donor age of the allograft were not significant on any local complication occurrence. Bone allografts are a reliable material but a high rate of local complications must be anticipated.

Improved accuracy with 3D planning and patient-specific instruments during simulated pelvic bone tumor surgery.

In orthopaedic surgery, resection of pelvic bone tumors can be inaccurate due to complex geometry, limited visibility and restricted working space of the pelvis. The present study investigated accuracy of patient-specific instrumentation (PSI) for bone-cutting during simulated tumor surgery within the pelvis. A synthetic pelvic bone model was imaged using a CT-scanner. The set of images was reconstructed in 3D and resection of a simulated periacetabular tumor was defined with four target planes (ischium, pubis, anterior ilium, and posterior ilium) with a 10-mm desired safe margin. Patient-specific instruments for bone-cutting were designed and manufactured using rapid-prototyping technology. Twenty-four surgeons (10 senior and 14 junior) were asked to perform tumor resection. After cutting, ISO1101 location and flatness parameters, achieved surgical margins and the time were measured. With PSI, the location accuracy of the cut planes with respect to the target planes averaged 1 and 1.2 mm in the anterior and posterior ilium, 2 mm in the pubis and 3.7 mm in the ischium (p < 0.0001). Results in terms of the location of the cut planes and the achieved surgical margins did not reveal any significant difference between senior and junior surgeons (p = 0.2214 and 0.8449, respectively). The maximum differences between the achieved margins and the 10-mm desired safe margin were found in the pubis (3.1 and 5.1 mm for senior and junior surgeons respectively). Of the 24 simulated resection, there was no intralesional tumor cutting. This study demonstrates that using PSI technology during simulated bone cuts of the pelvis can provide good cutting accuracy. Compared to a previous report on computer assistance for pelvic bone cutting, PSI technology clearly demonstrates an equivalent value-added for bone cutting accuracy than navigation technology. When in vivo validated, PSI technology may improve pelvic bone tumor surgery by providing clinically acceptable margins.

Computer-assisted planning and navigation improves cutting accuracy during simulated bone tumor surgery of the pelvis.

BACKGROUND: Resection of bone tumors within the pelvis requires good cutting accuracy to achieve satisfactory safe margins. Manually controlled bone cutting can result in serious errors, especially due to the complex three-dimensional geometry, limited visibility, and restricted working space of the pelvic bone. This experimental study investigated cutting accuracy during navigated and non-navigated simulated bone tumor cutting in the pelvis. METHODS: A periacetabular tumor resection was simulated using a pelvic bone model. Twenty-three operators (10 senior and 13 junior surgeons) were asked to perform the tumor cutting, initially according to a freehand procedure and later with the aid of a navigation system. Before cutting, each operator used preoperative planning software to define four target planes around the tumor with a 10-mm desired safe margin. After cutting, the location and flatness of the cut planes were measured, as well as the achieved surgical margins and the time required for each cutting procedure. RESULTS: The location of the cut planes with respect to the target planes was significantly improved by using the navigated cutting procedure, averaging 2.8 mm as compared to 11.2 mm for the freehand cutting procedure (p < 0.001). There was no intralesional tumor cutting when using the navigation system. The maximum difference between the achieved margins and the 10-mm desired safe margin was 6.5 mm with the navigated cutting process (compared to 13 mm with the freehand cutting process). CONCLUSIONS: Cutting accuracy during simulated bone cuts of the pelvis can be significantly improved by using a freehand process assisted by a navigation system. When fully validated with complementary in vivo studies, the planning and navigation-guided technologies that have been developed for the present study may improve bone cutting accuracy during pelvic tumor resection by providing clinically acceptable margins.

Accuracy in planar cutting of bones: an ISO-based evaluation.

BACKGROUND: Computer- and robot-assisted technologies are capable of improving the accuracy of planar cutting in orthopaedic surgery. This study is a first step toward formulating and validating a new evaluation methodology for planar bone cutting, based on the standards from the International Organization for Standardization. METHODS: Our experimental test bed consisted of a purely geometrical model of the cutting process around a simulated bone. Cuts were performed at three levels of surgical assistance: unassisted, computer-assisted and robot-assisted. We measured three parameters of the standard ISO1101:2004: flatness, parallelism and location of the cut plane. RESULTS: The location was the most relevant parameter for assessing cutting errors. The three levels of assistance were easily distinguished using the location parameter. CONCLUSIONS: Our ISO methodology employs the location to obtain all information about translational and rotational cutting errors. Location may be used on any osseous structure to compare the performance of existing assistance technologies.

Surgical inaccuracy of tumor resection and reconstruction within the pelvis: an experimental study.

BACKGROUND AND PURPOSE: Osseous pelvic tumors can be resected and reconstructed using massive bone allografts. Geometric accuracy of the conventional surgical procedure has not yet been documented. The aim of this experimental study was mainly to assess accuracy of tumoral resection with a 10-mm surgical margin, and also to evaluate the geometry of the host-graft reconstruction. METHODS: An experimental model on plastic pelvises was designed to simulate tumor resection and reconstruction. 4 experienced surgeons were asked to resect 3 different tumors and to reconstruct pelvises. 24 resections and host-graft junctions were available for evaluation. Resection margins were measured. Several methods were created to evaluate geometric properties of the host-graft junction. RESULTS: The probability of a surgeon obtaining a 10-mm surgical margin with a 5-mm tolerance above or below, was 52% (95% CI: 37-67). Maximal gap, gap volume, and mean gap between host and graft was 3.3 (SD 1.9) mm, 2.7 (SD 2.1) cm3 and 3.2 (SD 2.1) mm, respectively. Correlation between these 3 reconstruction measures and the degree of contact at the host-graft junction was poor. INTERPRETATION: 4 experienced surgeons did not manage to consistently respect a fixed surgical margin under ideal working conditions. The complex 3-dimensional architecture of the pelvis would mainly explain this inaccuracy. Solutions to this might be to increase the surgical margin or to use computer- and robotic-assisted technologies in pelvic tumor resection. Furthermore, our attempt to evaluate geometry of the pelvic reconstruction using simple parameters was not satisfactory. We believe that there is a need to define new standards of evaluation.