You must login before you can run this tool.
PET2TET: A tool that uses SUVmax to predict clinical parameters for TETThe PET2TET tool will let you
- Submit a spreadsheet of patient data which includes SUVmax, WHO class, clinical Masaoka-Koga stage, pathologic Masaoka-Koga stage, and resection status
- Analyze the predictive power of SUVmax for each of the variables: WHO class, clinical Masaoka-Koga stage, pathologic Masaoka-Koga stage, and resection status
- Review and graph the results of the analysis using histograms, "box and whisker" plots, AUC prediction graphs & SUVmax prediction power (percent)
How to use PET2TETYou can load spreadsheet data from any source into the PET2TET tool!
- data from the ITMIG Prospective Database
- data from ANY database, such as the ESTS Database
- your own hospital patient data
- data from ANY research study
To run the PET2TET tool:
STEP 1. Click on the Launch Tool button.
STEP 2. Provide two inputs: source of your data spreadsheet and then the data spreadsheet itself.
First: you will identify the source of your data spreadsheet. There are two choices:
- ITMIG Prospective Database
If you select this source, you will input a spreadsheet (in CSV format) from the ITMIG Prospective Database "dataview". This option is only available for authorized members of the ITMIG Prospective Database. See the following section for instructions.
- External 6 Column Data
If you select this source, you will input a 6 column spreadsheet (in CSV format) containing your own data, which can be taken from any source. See the following section for instructions.
Second: you will upload the data spreadsheet. The format of the spreadsheet you upload MUST MATCH the source you selected, or the tool will fail to run.
- If you selected source = ITMIG Prospective Database, your spreadsheet must be from the ITMIG dataview.
- If you selected source=External 6 Column Data, your spreadsheet must follow the external data 6 column format.
STEP 3: Click the Simulate button to run the analysis
STEP 4: Look at the reports and graphs output by the analysis
To understand the results, read the Case Study below. It identifies all the graphs, reports and information output by the tool.
How to get a spreadsheet of data to input to PET2TET
Data from the ITMIG Propspective Database!This option is only available for authorized members of the ITMIG Prospective Database.
You can download a spreadsheet of the required data variables from patients in the ITMIG Prospective Database. Just click the link:
PET SUVmax relationship to TET workup .
and then click the Download button to the left of the dataview title. Upload this CSV format file directly into the PET2TET tool.
Data in this dataview shows all current ITMIG prospective patient data that have data entered for SUVmax , along with the patient clinical parameters for WHO class, clinical and pathologic staging, and R0/R1/R2 resection.
Users can download data from the dataview into a spreadsheet (in CSV format) by clicking the Download button. The CSV format spreadsheet can be loaded directly into the PET2TET tool for analysis.
Your own data, from any sourcelYou can use a CSV format spreadsheet with patient data collected from any source. The spreadsheet must be saved in CSV format with 6 columns of data. The columns and valid entries are listed below:
Column 1: ID (unique identifier for each row)
Column 2: SUVmax (a real number, e.g., 10.4)
Column 3: WHO class (A, AB, B1, B2, B3, Thymic Carcinoma, Thymic Malignancy NOS, Other)
Column 4: clinical Masaoka-Koga stage (I, II, IIA, IIB, III, IV, IVA, IVB)
Column 5: pathologic Masaoka-Koga stage (I, II, IIA, IIB, III, IV, IVA, IVB)
Column 6: final resection status (R0, R1, R2)
Click Example 6 Column Spreadsheet
for an example spreadsheet with the required headers. Replace the example data with your own data.
A Case Study for PET2TET: Input, Analysis and OutputThis section describes the analysis process used in the PET2TET tool, with examples of the output. Studying this use case will help you better understand the results of the PET2TET tool.
Histogram of SUVmax (0 - 20)
Area Under the Curve (AUC) is a very good way to evaluate the prediction performance of a binary classifier. if AUC=0.7 we can say we are 70% sure that the prediction of our model is correct.
Patient Data for Study Variables
(real-time patient data in this view)
Counts of data used in this research study:
WHO class (109)
Clinical MK Stage (100)
Pathologic MK Stage (97)
Final Resection Status (95)
Click for the Final Report on the analysis of the case study data (completed October 11, 2015).
Cite this work
Researchers should cite this work as follows: