We have carefully reviewed the letter from Li et al and believe that there were misunderstandings of the results and interpretation of the findings, which we would like to take the opportunity to clarify.
We agree with Li et al that a prognostic biomarker provides information about the patient’s overall cancer outcome, regardless of therapy, whilst a predictive biomarker gives information about the effect of a therapeutic intervention. As we discuss in the paper, mutations in STK11/KEAP1 have been associated with inferior outcomes in patients treated with pembrolizumab-based chemotherapy, including among patients with high tissue TMB and PD-L1–positivity, suggesting a predictive role1. Additionally, we acknowledge that the biomarkers that have been studied are both predictive and prognostic. The results from the MYSTIC trial confirmed the negative prognostic role of KEAP1 using plasma NGS in patients with metastatic NSCLC receiving combination immunotherapy; however, it did not clearly confirm the predictive role for STK11, but rather showed that this may be a prognostic biomarker, with overall worse outcomes seen in patients with STK11 mutation2. The authors provide an example of a study where STK11 has been shown to be a prognostic biomarker, however, two factors must be considered when taking this study into consideration. Firstly, patients on this study were not treated with pembrolizumab-based therapy, and secondly, role of co-existing mutations, PD-L1 and TMB has not been reported. Therefore, no direct conclusions can be made. The divergent data on the role of STK11/KEAP mutations and their interplay with outcomes following immunotherapy-based combinations can be potentially explained by the complex molecular interactions that exist within the tumor microenvironment.
Regarding their second critique that in our study overall survival (OS) was insignificantly linked to plasma TMB, we would like to point out that previous studies cited by Li et al have also failed to show this correlation with OS and in fact, the results from our study are consistent. For example, on the B-F1RST study, median OS for blood TMB high vs. low was NE vs. 13.1 months; HR = 0.77 (90% CI, 0.41–1.43; P = 0.48)3.
Finally, we would like to clarify that pTMB is currently a research only assay, and is not ready for primetime use to guide clinical practice. We recognize that our observations are based on small numbers and, as stated in our paper, these analyses should be considered exploratory. We raise an idea for evaluation of these findings in a larger, prospective trial, and if validated, could suggest that including these genomic biomarkers may improve identification of pTMB-high patients unlikely to respond from immunotherapy.
Acknowledgments
We appreciate the interest of Dr. Li and colleagues in the research findings, and acknowledge that many of their critiques have been addressed in our manuscript.
Footnotes
Disclosures
C. Aggarwal is an advisory board member/unpaid consultant for Bristol-Myers Squibb, Celgene, Eli Lilly, Merck, Roche, and AstraZeneca. J.C. Thompson is an advisory board member for Guardant Health. E. L. Carpenter disclosed no potential conflicts of interest.
References:
- 1.Skoulidis F, et al. , STK11/LKB1 Mutations and PD-1 Inhibitor Resistance in KRAS-Mutant Lung Adenocarcinoma. Cancer Discov, 2018. 8(7): p. 822–835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Rizvi NA, et al. LBA6Durvalumab with or without tremelimumab vs platinum-based chemotherapy as first-line treatment for metastatic non-small cell lung cancer: MYSTIC. in ESMO Immuno-Oncology Congress. 2018. Geneva, Switzerland: EMSO. [Google Scholar]
- 3.Kim ES, et al. , LBA55Primary efficacy results from B-F1RST, a prospective phase II trial evaluating blood-based tumour mutational burden (bTMB) as a predictive biomarker for atezolizumab (atezo) in 1L non-small cell lung cancer (NSCLC). Annals of Oncology, 2018. 29(suppl_8). [Google Scholar]