Non-small-cell lung carcinomas are typically identified late, have unfavourable prognosis and rapidly metastasize into other tissues. The paucity of curable approaches has recently advanced cancer immunotherapy as a novel therapeutic option. As in other cancers, therapeutic outcomes are often assessed by changes in standardized uptake values (SUVs). For tumors, SUV measurements taken prior to immunotherapy application are compared to SUVs after application. Their differential is used to assess the therapy’s effectiveness.

However, a meaningful estimate of the effectiveness of the therapy would require to know the SUVs of tumors at the time of immunotherapy application –the reference SUVs– and to monitor the change in SUV over the time period following immunotherapy. Substituting these reference values by values measured substantially before immunotherapy , can severely bias effectiveness estimates towards underestimates. 

In this study, we present a method to correct for this bias and to estimate immunotherapeutic effectiveness within a patient from SUVs. These method relies on two basic assumptions.

We applied the methods to a set of data from 97 patients, with a total of ca. 1300 lesion mea- surements. Fits were performed on pairs of measurements taken before and after treatment to assess immunotherapeutic efficiency. The tumor killing efficacies by the immunotherapy derived from these fits correlate with treatment outcome, and improve upon comparable RECIST-based approaches.

The here presented methods presents a simple approach to make use of the information stored in the times elapsed prior to and after immunotherapy start. Thes parameters of our model satisfyingly captures immunotherapeutic efficiency and offers a mathematically sound way to remove biases contained in contemporary effectiveness estimation approaches.