When you start using DeepLR, you will be asked to upload your demographics data file and CT screening data file. Both files must be in CSV format with file extensions “.csv”. If your data files are not prepared according to instructions from the Tools page of the main menu, you will get error message and DeepLR will not give you correct estimates. If there is no error message, DeepLR will return 3 panels: “Input demo”, “Input CT”, and “DeepLR Output”.

Panels “Input demo” and “Input CT” display the demographics data and CT screening data you have uploaded. Please note that variable names in these two panels may not be the same as the variable names you used. The purpose to display these two panels is for you to check if you have used exact the same variable order as the instruction in Tools page of the main menu. If not, then DeepLR output will not be interpretable.

“DeepLR Output” panel displays a table with DeepLR predictions. Due to web server computational capacity, you may need to wait for a few minutes or even longer before they appear. DeepLR gives the following 4 estimates in this panel:

  1. DeepLR score = risk of lung cancer development within 3 years after (current) S2 screening visit date. We define high-, median-, and low-risk scores by DeepLR>0.3, 0.1 to 0.3, and <0.1 respectively. However, you can make an adjustment to these threshold values.
  2. interval.fu = recommended lung screening follow up interval in months for the next lung CT screening. If this value is zero, it means that DeepLR recommends immediate cancer treatment or diagnostic referral.
  3. aggressive = risk of aggressive tumor development. A tumor is considered to be aggressive if it is associated with lung cancer mortality. Low, median, and high risk of aggressive tumor development is defined by the aggressive risk score of <0.00565, between 0.00565 and 0.19500, and >0.19500 respectively.
  4. aggr.group = aggressive tumor risk category: 1=low risk, 2=moderate risk, 3=high risk.

You can also download DeepLR output table to your local computer by clicking “Download” button.

Note:

  • DeepLR scores and aggressive scores are not perfectly correlated. This is because that DeepLR scores estimate the risk of lung tumor development while aggressive scores estimate the risk of aggressive tumor development. A person with a high DeepLR score but a low aggressive score implies that the person is predicted to have a high risk of developing a non-aggressive tumor. However, a person with a high aggressive score must have a high DeepLR score.
  • DeepLR can be used recursively after the second screening visits. For example, applying DeepLR to screening visits 1 and 2 to determine the timing of visit 3 (if interval.fu >0). Next, applying DeepLR to screening visits 2 and 3 to determine the timing of visit 4 (if interval.fu >0), and so on.

Please note: this website provides DeepLR predictions only for your research purpose. We are not responsible for any consequence resulted from your use of our prediction.

To start using DeepLR, click “CONTINUE” button below.

Continue