Updates to PK Model Fitting

Revised sequencing of PK model fitting to run faster. Added ability to enable/disable PK model fitting from the PK analysis configuration page. Default setting is PK model fitting is disabled.  

New Data Cleaning Rule: Count records

New rule in Data Cleaning to count number of records based on a set of grouping variables. This can be used to identify profiles that do not have adequate samples for PK analysis in a large dataset. 

Updates to AUCtau calculation

AUCtau was only be calculated when a terminal elimination rate constant could be estimated. The calculation was updated to be calculated if data is available for the full tau interval regardless of ability to estimate the terminal slope, or extrapolated to the full interval using the terminal slope.

Partial AUC name detection

If the user provides a paramcd value that matches an existing paramcd name, the text will be updated so that it has a unique name. For example, if the user provides "Cmax" for the paramcd name it will be modified to "custom_Cmax" to distinguish it from the "Cmax" variable.

Data Cleaning History page added

Users can now view previous data cleaning actions from the Data Cleaning History tab on the Input Data page. Future updates are coming in this area to support other exploratory analyses.

Updates for SSO login support

Updates to support SSO login were made. These will only be visible to organizations that have implemented SSO. 

Bug fixes

Minor bugs fixes to the user interface to disable buttons when executions are in progress, enable validation package download button. Bug in final results file preparation that excluded some individual parameters was fixed and better error handling for datasets with duplicate samples.

Performance updates to Analysis History Page

Speeds up loading of prior analyses and related steps in the user interface. Allows faster launching of re-analysis steps. 

Update user roles

User roles (user vs administrator) now handled with a single source of truth. 

Added Grouping Variables to Ratio Calculations

Grouping variables can now be added to the Ratio calculation in Custom Calculations. This allows you to calculate the same parameter across multiple groups. For example, you can now calculate a dose-normalized partial AUC across both fed and fasted treatments for the same subject with a single ratio calculation call.

Custom Report Templates

Users can now upload customized report templates for use with Aplos NCA Reporting feature. All templates are versioned. Users can upload templates for their own use. Administrators can upload templates for use across the organization. A sample template file is provided for use.

Added Best Fit Terminal Slope Replacement

Users can now replace the automated "best-fit" terminal slope with a different terminal slope of their selection in the Custom Calculation engine. 

Bug Fix: All concentrations are missing

Fixed bug that caused failure when a subject had all missing concentration values. Now profile will have all results reported as "Not Determined" (ND). 

Added Duplicate button

Users can now duplicate any object in Data Cleaning, Custom Calculations, Plots, and Tables. This creates an identical object with a new description. All selections of the duplicated object are carried to the new object. 

Added Series Selection to Plots

Users can now select different series to group and color plots. For example, if you have multiple study days and multiple dose levels, you can group data by Study Day and Dose, but only color by Dose. 

Added Non-Linear Model Fitting

Non-linear model fitting results will be provided for all single-dose profiles. IV bolus and IV infusion administration will be fit to 1-compartment, 2-compartment and 3-compartment PK models. For the extravascular route, the same compartment models will be tested with and without a lag time (6 models in total). The preferred model will be identified by the lowest AIC and lowest BIC values for all models. Parameter estimates will be provided for all models as will observed and predicted concentrations. 

Updated Error Handling in the Post Processing Step

When generating tables and plots during post-processing, a failed table or plot will not result in a complete failure of the entire step. Instead, a diagnostic file will be provided that gives an error message on why the specific plot or table failed. Other plots and tables will be produced. 

Added Folders to Analysis History

Users can now move analyses to folders to organize their results. Users can add, rename, and delete folders. Analyses can be moved from one folder to another. If a folder is deleted, analyses within that folder are returned to the "No folder" area. Analyses are never deleted.

Added Data Manipulation Functions to Custom Calculations

Users can now manipulate results file to add columns and perform other actions to support production of tables and plots. Many of these data manipulation functions are derived from the Data Cleaning tool.

Data Cleaning user interface

The user interface was added to enable users to create the data cleaning JSON from the user interface. Users can select the data cleaning rules, enter the required information and then process the data for cleaning. This includes creation of complex data cleaning rules like BLQ handling with just a few questions.

Updates to Input Data page

Users have many new options from the input data page. These include the ability to select data for cleaning, view the data in tabular format, download the input data file, and archive the data file. 

Improved error messages

Error messages have been improved to provide more information to the user when errors are encountered.

Additional filtering options for Analysis History page

Users can now limit the analysis history page to selected time windows to make it easier to find the analysis of interest. Users can select last 7 days, 30 days, 90 days, 365 days, or set custom start and stop dates to filter the data.

Coming soon: Parallel processing

We are making big changes in how we process analyses in the coming months. Currently, each unique PK profile is processed in sequential fashion. We are working to create parallel processing for profiles so groups of 5-10 profiles are processed in parallel with each other. This should provide significant performance advantages for large datasets including simulations. Watch for updates in the coming weeks and months.