Persona IQ Cohort Study

The Canary Tibial Extension (CTE)/canturio Smart Extension (CSE) with Canary Health Implanted Reporting Processor (CHIRP) System are intended to provide objective kinematic data from the implanted medical device during a patient's total knee arthroplasty (TKA) post-surgical care. The kinematic data are an adjunct to other physiological parameter measurement tools applied or utilized by the physician during the course of patient monitoring and treatment post-surgery.

The device is indicated for use in patients undergoing a cemented TKA procedure that are normally indicated for at least a 30 mm sized tibial stem extension and/or normally indicated for at least a 58mm sized tibial stem extension. The objective kinematic data generated by the CTE/CSE with CHIRP System are not intended to support clinical decision-making and have not been shown to provide any clinical benefit. The CTE/CSE with CHIRP System is compatible with Zimmer Persona® Personalized Knee System

The CTE/CSE with CHIRP System uses external OR and Home Base Station units to query the CTE/CSE implant (which has an internal radio and antenna) and upload the data collected by the CTE/CSE implant to the Canary Cloud data management platform (CDMP). Information from the implant is processed by the system's Canary Medical Gait Parameter (CMGP) software, located in the Cloud, into clinically relevant metrics.

Follow-up Procedures:

Follow-up evaluations will be conducted at 1 month (± 14 days), and 3 months (± 30 days) after surgery. At every follow-up visit, pain and function of the study knee will be evaluated, and subject's quality of life will be assessed.

RTM assessments will be collected pre-operatively and post-operatively through the mymobility® care plan pathway as defined within the application for each specified assessment. The interval window will open and close automatically within the software for each specified target and captured clinically up-to 6-months post-operatively and indefinitely if the patient continues to utilize the mymobility App. Data will also be captured via the CTE/CSE sensors at variable intervals post-operatively and transmit data indefinitely if the home base station is connected and has power.

Data collected through mymobility® will include, but is not limited to the following:

• KOOS, JR
• EQ-5D-5L
• FJS-12
• Pain \& Opioid Tracking

Data collected passively through CTE/CSE Sensors will include, but is not limited to the following:

• Walking Speed
• Cadence
• Stride Length
• Functional Knee ROM
• Tibial ROM
• Step Count
• Distance Traveled In-Person clinical assessments will be collected at 1-month and 3-months post-operatively.

Clinical Data Collection: The Sponsor will collect all data within multiple databases (e.g. including but not limited to: EDC, mymobility®, HealthKit, Google Fit, PIQ, ROSA). The management of all study data received by the Sponsor will be the responsibility of the Sponsor or its designee. The use or disclosure of all protected health information will comply with all relevant data privacy and data security laws and requirements. All information will be treated with strict adherence to professional standards of confidentiality and will be held by the Sponsor under adequate security and restricted accessibility.

Sample Size Calculation:

The study objective is to assess the seven kinematic metrics (step count, functional knee ROM, tibial ROM, distance walked, walking speed, cadence and stride length) to be captured via the CTE/CSE sensor in combination with the mymobility® platform and develop correlative insights to assist surgeons in understanding and managing recovery in their patient populations through post-operative gait metrics. There will be no primary hypothesis test in this study. All the clinical outcomes and device safety data to be collected will be summarized by the applicable univariate statistics and analyzed by the appropriate statistical models. In general, the logistic regression models for the categorical data points, especially binary data for rare events such as revision, hospital readmission etc., tend to require relatively large sample sizes to have enough power to perform conclusive analyses. According to Bujang et al. [5], 100 + 50i patients are needed to effectively use logistics regression in predicting outcomes, where i refers to the number of independent variables in the final model. It is expected between 15 and 20 independent variables will be available for the model. A sample size of 200 subjects will be sufficient for a logistic regression model with 2 independent variables.

General Statistical Methods:

Data collected in this study will be summarized descriptively. Categorical data (such as biological sex, complications, satisfaction status, etc.) will be summarized using counts and percentages over the time periods of interest. Continuous data (such as age, ROM, etc.) will be summarized by using means, medians, standard deviation, minimum, maximum, and 95% CI over the time periods of interest.

Appropriate linear regression and generalized linear regression models including logistic regression, cox regression, mixed model for repeated measurements will be applied for the primary, secondary and exploratory endpoints analysis per the data type and analysis goals. Subgroup analysis on a variety of subsets of the collected data will also be performed. The applied statistical models will be used to evaluate and quantify the relationship among the differences in the episode of care data, patient pre-op physical condition and medical history, gait parameters, range of motion, patient satisfaction, nine most common post TKR complication incident rates, PROs, etc. Implant survival will be evaluated using Kaplan-Meier product limit estimate and raw survival rates. In addition to inference models, the predictive models will also be built to predict the patient satisfaction and PROs using patient baseline status and early PIQ data.

Data Analysis:

Primary Objectives: The Persona IQ gait metrics from the Canary Smart Tibial Stem when used in conjunction with the Persona IQ Personalized Knee System and mymobility® Platform for primary total knee arthroplasty will be analyzed by the appropriate linear and generalized linear regression models per the data type. The descriptive statistics will be provided as needed to individually describe the data points. The incidence of the most common post TKR medical complications will be analyzed by binary logistic regression model and time to event model such as cox regression. The relationship between pre-op, episode of care and post-operative gait recoveries will be evaluated by both binary logistic regression model and mixed model. The correlation between continuous 14 to 30-days Persona IQ data and 90-days post-operative sagittal ROM data will be evaluated by Pearson correlation, Intraclass Correlation Coefficient (ICC) with and without patient key baseline features adjusted. The relationship between categorical 14 to 30-days Persona IQ data and 90-days post-operative sagittal ROM data will be evaluated by either ANOVA or ANCOVA model per the analysis goal. The ordinary logistic regression will be applied to build predicting model for patient satisfaction at 42 and 90-days.

Secondary Objectives: The evaluation of value related to the use of Remote Therapeutic Monitoring and healthcare utilization in the form of non-standard of care surgeon office visits, post-operative physical therapy, emergency department visits, and hospital readmissions associated with the index procedure will be evaluated using descriptive statistics as described in Section 16.2 along with appropriate statistical models.

Exploratory Objectives: Implant survival and safety will be analyzed by Kaplan-Meier product limit estimate and raw survival rates with 95% confidence interval. Patient functional performance, clinical benefit measurements and healthcare utilization with index procedure will be analyzed for the whole cohort and for varied subgroups stratified by important surgical or clinical or patient physical features via both appropriate statistical models and descriptive statistics. The continuous clinical outcome differences between the subjects utilizing ROSA and conventional implantation will be evaluated by ANCOVA model with baseline function score, patient key baseline features and visit time adjusted. Other statistical methods will be utilized as appropriate to address the exploratory hypotheses.