1 🧭 Quick Start Guide — Surv Web Application This guide explains step by step how to use the Surv web platform to perform survival analysis, model extrapolation, and export results for health - economic modeling. The platform is currently in open beta and accessible for free using the demo account: Login: admin@admin.com Password: admin111 🧩 Overview Surv is an integrated web - based platform that allows users to conduct the full workflow of survival analysis — from digitizing Kaplan – Meier (KM) curves or importing datasets, to fitting flexible survival models and exporting the results to Excel or CSV files. Unlike traditional workflows that require multiple tools (e.g., WebPlotDigitizer, R, Excel), Surv unifies all analytical steps in one browser - based environment, requiring no coding skills or software installation. 1⃣ Step 1 — Define the Analysis Setup When you first open the app, you will see the Entry Data tab. 1. Select the number of interventions • Choose whether your analysis includes 1 or 2 interventions (for single - arm or two - arm comparisons). • Enter clear, descriptive names for each intervention (e.g., Treatment vs. Control ). 2. Choose the time unit for your graph • Define the time scale corresponding to the x - axis of your Kaplan – Meier data: o Days , Weeks , Months , or Years • This ensures proper extrapolation later (e.g., a model built in “months” will project survival month by month). 3. Select the type of data you plan to upload You can import data into Surv in three different formats: D ata Type Description Notes KM Curves (Graph Image) Upload a published Kaplan – Meier plot (e.g., from an article or report) Requires manual or CSV “At risk” table for better extrapolation CSV with Individual Patient Data (IPD) Import individual time - to - event records with “time”, “event”, and “group” columns Ideal when you have raw data CSV with KM Coordinates Import digitized curve coordinates (time, survival probability, at - risk count) For pre - digitized data . A lso require s “ At risk tabl e ) 2 2⃣ Step 2 — Upload and Prepare the Data Depending on your data type, the upload process differs slightly. If you selected “KM Curves (graph image)” 1. Input number of interventions 2. Input names of interventions 3. Input graph steps dataf rame 4. Upload or manually input the “At risk” table (either via CSV or the in - browser form). 5. Click “ Start” to begin curve extraction. 6. Upload the image file of your Kaplan – Meier curve. o Accepted formats: .png , .jpg , .jpeg Once you enter the digitization interface, follow the on - screen instructions carefully: Axis calibration • Define the X and Y axes by marking two or more known points. • Input their values directly in the pop - up fields. ⚠ Important: Please input Y - axis values in percentage format (0 – 100%), not as decimal fractions. 3 Curve recognition • Choose the recognition algorithm (recommended: X Step , step size = 1). • Pick the color of the survival curve to be recognized. • If the same color appears elsewhere in the image (e.g., in grid lines or labels), use the Pen tool to draw a mask that isolates only the curve. • You can also add or delete point manually • You can preview detected points directly on the image. Run digitization • Click “Run Digitization → Next Step → Check Graph → Proceed Using Results.” • The resulting data (time, survival probability, and number at risk) will be displayed automatically. 4 If you selected “CSV with IPD” or “CSV with KM coordinates” 1. Upload the CSV file with your dataset. 2. Then upload or enter the “At risk” table 3. When prompted, specify: o Column separators — comma , , semicolon ; , or tab \ t o Decimal format — dot or comma , o Encoding — UTF - 8 is recommended to avoid character issues. 4. The system automatically validates column names and content. 5. Once validated, you can move to modeling. 3⃣ Step 3 — Fit the Models After uploading data successfully, you can navigate to the modeling tabs. Available options: • Standard Parametric Models • Spline Models • Mixture Cure Models 🔒 Note: The “Survival Analysis” tab is currently unavailable in the beta version. 5 Standard Parametric Models Use this module for classical parametric distributions and piecewise models. Required parameters: • Model type for extrapolation: Weibull, Gompertz, Log - logistic, Log - normal, Exponential, Generalized Gamma, or Piecewise. • Cycle length — defines the time step used for predicted survival (e.g., 1 month). • Extrapolation horizon — total time length for the projection. The interface displays both observed KM data and fitted curves, along with AIC/BIC and key statistics for model comparison. 6 Spline Models This module allows for flexible, smooth extrapolation using spline - based survival functions. You can define: • Number of knots and/or their coordinates. • Model scale: Normal , Hazard , or Odds • Cycle length and extrapolation horizon. The fitted spline curves provide better fit for datasets with non - monotonic hazard functions or complex shapes. Mixture Cure Models Used when a proportion of patients can be considered “cured” (e.g., long - term survivors in oncology). Required inputs: • Type of model for extrapolation (e.g., Weibull, Log - logistic). • Cycle length. • Mean patient age. • Mortality table — you can use your own or select the built - in US Life Table. The output provides cure fraction estimates and extrapolated survival for both cured and uncured subgroups. 7 4⃣ Step 4 — Run Calculations or Export Results Once all parameters are set: • Click “Calculate” to run the model and display fitted survival curves, parameter tables, and diagnostics directly in the browser. • Click “Download CSV” or “Download XLSX” to export all parameter estimates, survival predictions, and variance - covariance matrices for downstream analysis (e.g., in Excel or R - based cost - effectiveness models). 5⃣ Step 5 — Next Steps You can return to any previous stage (data entry, digitization, or modeling) at any time Future updates will include: • A Survival Analysis module with Kaplan – Meier visualization and Cox regression (currently only one covariate: intervention arm ). • AI - based tools for automatic curve recognition and report generation. 💡 Tips and Notes • The digitization quality depends on the resolution and color contrast of your uploaded KM curve. Prefer clean, high - resolution images with distinct curve colors. • Always verify axis calibration before running digitization — this step critically affects accuracy. • If survival values appear truncated or noisy, reduce the X - step size (e.g., from 2 to 1). • All exports are in plain text or Excel format, compatible with R , Python , and TreeAge economic modeling frameworks. • Demo access ( admin@admin.com / admin111 ) is available for testing but does not store data permanently.