Article Text

Download PDFPDF

135 Integration of T-cell activation and tumor mutation burden predicts anti-PD-1 response
  1. Fiona Ehrich1,
  2. Matthew Adamow1,2,
  3. Colleen Maher1,2,
  4. Jasme Lee1,
  5. Xiyu Peng1,
  6. James W Smithy1,
  7. Samuel A Funt1,3,
  8. Michael Postow1,3,
  9. Niamh Keegan1,
  10. Katherine S Panageas1,
  11. Ronglai Shen1 and
  12. Margaret K Callahan1,2,3
  1. 1Memorial Sloan Kettering Cancer Center, New York, NY, USA
  2. 2Parker Institute for Cancer Immunotherapy, San Francisco, CA, USA
  3. 3Weill Cornell Medical College, New York, NY, USA
  • Journal for ImmunoTherapy of Cancer (JITC) preprint. The copyright holder for this preprint are the authors/funders, who have granted JITC permission to display the preprint. All rights reserved. No reuse allowed without permission.


Background While immune checkpoint blockade (ICB) therapies have revolutionized cancer care, most patients do not benefit from these agents. Existing biomarkers are limited. An early indicator of treatment response could inform treatment modifications, such as adaptive dosing or the use of combination therapy.1 To that end, there is opportunity to enhance the prediction of response by integrating multiple modes of data providing orthogonal information. This retrospective study aimed to identify an early signal of response to anti-PD-1 monotherapy by leveraging two complementary modalities: peripheral blood T-cell dynamics and tumor mutation burden (TMB).

Methods We identified patients with unresectable or metastatic urothelial carcinoma or melanoma who received anti-PD-1 monotherapy in clinical trials open at Memorial Sloan Kettering Cancer Center. Pretreatment and on-treatment peripheral blood samples were analyzed by flow cytometry. T-cell activation was defined as the absolute change from baseline in the proportion of Ki67+ cells of PD-1+ CD8+ T cells prior to administration of the second dose. Additionally, we performed targeted next-generation sequencing (MSK-IMPACT) on tumor tissue DNA and matched normal DNA.2 TMB was calculated as the number of nonsynonymous coding mutations per megabase and was log transformed for analysis. Treatment response was defined as complete or partial response using Response Evaluation Criteria in Solid Tumors (RECIST) best overall response.3 To assess the predictive performance of T-cell activation alone, TMB alone, and the two in combination, logistic regression was employed with response as the endpoint. For each configuration, median test area under the receiver operating characteristic curve (AUC) across 100 train-test splits was calculated.

Results Our cohort consisted of 67 patients with urothelial carcinoma (n=53, 79%) or melanoma (n=14, 21%) treated with anti-PD-1 monotherapy (table 1). The response rate was 39% (95% confidence interval, 27–52%). Patients were accrued from 2012 to 2018 across five trials. Using T-cell activation alone as a predictor of response achieved a median test AUC of 0.72. Similarly, using TMB alone achieved a median test AUC of 0.68. When these two modalities were combined, median test AUC increased to 0.78, indicating improved performance (figure 1).

Conclusions A blood-based signal of T-cell activation after one dose of treatment and TMB provide valuable and complementary information in predicting anti-PD-1 response. Integrating these modalities enhances predictive performance beyond that of either modality individually. An early indicator of anti-PD-1 response holds potential to guide clinical decision-making in an area where biomarkers are needed.


  1. Postow MA, Goldman DA, Shoushtari AN, et al. Adaptive dosing of nivolumab + ipilimumab immunotherapy based upon early, interim radiographic assessment in advanced melanoma (the ADAPT-IT study). J Clin Oncol. 2022;40:1059–1067.

  2. Zehir A, Benayed R, Shah RH, et al. Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat Med. 2017;23:703–713.

  3. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45:228–247.

Ethics Approval This study was approved by the Institutional Review Board of Memorial Sloan Kettering Cancer Center; approval number 19–114.

Abstract 135 Table 1

Patient characteristics.

Abstract 135 Figure 1

Test AUC of each model configuration across 100 train-test splits. The horizontal axis indicates the predictor(s) included in the logistic regression model predicting response. The value in the center of each boxplot indicates median test AUC.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.