2110. Novel Measure of Therapy Impact (TXI): Does Early Restaging Imaging Predict LENGTH OF STAY on Nivolumab in Patients with Advanced Cancer?
Authors * Denotes Presenting Author
  1. Milos Janicek *; Boston Medical Center
Cross-sectional Imaging (CI) plays an essential role in early assessment of patients on immune checkpoint inhibitors (ICIs), yet often remains inconclusive in the setting of advanced disease, prior cancer treatment, toxicity, and cancer unrelated findings. We propose a robust, reproducible score based on measurable tumor response on the first two re-staging scans that may predict TXI. We are testing prognostic value of early imaging, time lag of ICIs effect, against Duration of clinically effective treatment as a Measure of Treatment Impact.

Materials and Methods:
Retrospective review of CI studies in 92 patients with advanced stage solid tumors treated with Nivolumab (Nivo) during 6/2015-8/2018 at BMC on standard clinical protocols. Aero-digestive tract tumors represented majority of our group.Time on therapy (TOTx) was used as a surrogate marker of therapy impact. Modified imRECIST score utilizing ? of 4 discrete cancer lesions at baseline, 1st, and 2nd CI since initiation of Nivo were calculated. Score changes from baseline were compared with length of effective therapy. Finally, we computed the best fit function of tumor change/day [%] at each time point vs. TOTx for entire group.

Median TOTx was 70 days (range 1-1029). Differences in imRECIST scores derived from two early restaging CI studies were able to identify with high statistical significance two groups of "SHORT" vs "LONG" stay on Nivo around Median of 70 days. Early Progression curve was associated with significantly shorter benefit of Nivo; Pseudo-progression curve showed significantly shorter benefit from Nivo compared to initial Regression or Stable score curves (P < .002).

Highly potent yet toxic treatments with ICIs need tools that allow early prediction of ICIs treatment response and clinical outcome. Our novel scoring tool validates early restaging imaging after initiation of ICI identifying those, who may not benefit. Further correlation with clinical biomarkers may lead to development of predictive nomograms of "short" versus "long" therapy impact. The imRECIST based score may differentiate with high likelihood patients with short versus long stay on Nivo as a novel measure of TXI. Our study validates early imaging with a response curve and standardized analysis of the 1st and 2nd post-treatment scans in patients on immunotherapy.