Immune checkpoint inhibitors (ICI) provide survival benefits to more than 50% of patients with non-small cell lung cancer (NSCLC)[1]; however, there are still many patients who do not benefit from treatment, and some patients receiving immunotherapy experience severe adverse effects.
Therefore, how to accurately identify which patients can benefit from immunotherapy, maximize efficacy and reduce toxicity, has been one of the directions that scientists strive for.
Tumor mutation load (TMB) is a promising biomarker that is the total number of somatic mutations in each coding region of the tumor genome. High-TMB tumors have been found to produce more neoantigens, which in turn triggers a stronger antitumor immune response [2,3].
Back in 2020, the FDA approved FoundationOne CDx (F1CDx) for assessment based on tissue tumor mutation burden (tTMB) in patients with unresectable or metastatic high-TMB solid tumors, with a cut-off value of 10mut/Mb to determine which patients to use pemberizumab [4].
However, tTMB evaluation requires invasive procedures for patients with an inherent risk of complications. In addition, the biopsy is sparsely obtained, which may result in insufficient AMOUNT of DNA and affect the results of the analysis [5]. The hematologic tumor mutation load (bTMB) can not only compensate for the shortcomings of tTMB detection, but also avoid spatial intratumor heterogeneity between primary and metastatic lesions.
Recently, a research team led by Daniel Klass of Roche published important research results in the Journal for ImmunoTherapy of Cancer[6]. Based on bTMB, they designed lung TMB panels for the detection of patients with stage I to IV NSCLC.
The results showed that when the tTMB critical value of 16mut/Mb was used as the criterion, the 1.1Mb lung TMB panel had a highly accurate TMB detection response, and the positive predictive value (PPV) was 95%, equivalent to 42mut/Mb of bTMB; but the positive compliance rate (PPA) of bTMB was relatively low, at 32%. But in stage IV NSCLC patients with cfDNA input of at least 20 ng, the PPA of bTMB can be increased to 63%, and the detection panel can be minimized to 577 kb.
This result promises to make bTMB a predictive biomarker for immunotherapy in NSCLC patients.
Screenshot of the first page of the paper
First, the researchers analyzed 183 tumor tissue samples using an existing ~2.2M tTMB panel, and the results showed that the median TMB value of the tTMB panel was higher than the total exon sequencing (WES) analysis value (13.9 vs 7.6) of lung, adenocarcinoma and squamous cell carcinoma samples in the TCG database. In addition, an in-silico analog comparison shows that the tTMB panel has a good correlation with the FDA-approved F1CDx panel (R2=0.99).
Next, the Klass team used the CAPP-Seq panel design algorithm to design as small as possible, mutation-maximized plasma panels for lung TMB testing to minimize the cost of testing each sample. The researchers took about 300 Kb covering key genes as the core, adding about 50 Kb each time, and finally forming a lung TMB panel of about 1.1 Mb, covering the exon region of 965 genes.
Immediately afterward, the researchers analyzed 178 plasma samples with a lung TMB panel, and the results showed that the sequencing depth was closely related to the amount of cfDNA input, when the cfDNA input was greater than 30 ng and the sequencing depth was greater than 3000, indicating that the panel was highly accurate in detecting mutations.
To assess the clinical effects of lung TMB panels, the researchers conducted an in-silico simulation assessment using WES data from the NSCLC cohort published by Rizvi et al. [7]. The results showed that the WES critical value (10 mut/Mb) was converted to 19.3 mut/Mb for the lung TMB panel.
In addition, in patients with lasting clinical benefit (DCB) and no lasting clinical benefit (NDB), the separation effect of lung TMB panels on both groups of patients was stronger than THAT (p=0.002), and in the predictive analysis of improvement in progression-free survival (PFS) rates, the separation effect of lung TMB panels was similarly stronger (p=0.0039, HR=4.023).
Lung TMB panels can separate DCB and NDB patients, as well as patients with different PFS rates
To compare the TMB values of the tumor tissue TMB panel with the lung TMB panel, the researchers used the lung TMB panel to recapture the tissue sample and determined the correlation equation for the TMB value between the two panels, and the analysis showed that 16mut/Mb in the tissue panel corresponded to 42mut/Mb in the lung TMB panel.
The researchers used the tTMB critical value of 16 mut/Mb to further evaluate the consistency of plasma and tumor tissue TMB classification (TMB-H and TMB-L), and the results showed that the PPV of lung TMB panel detection of TMB-H was also high (91%-100%), indicating that plasma detection accuracy was good. However, lung TMB panels measured TMB-H with lower PPA (24%-32%), indicating lower sensitivity.
Given the lower plasma panel PPA, the researchers believe it may be associated with a small amount of cfDNA extracted from tumors with early or lower shedding rates. The results of the analysis showed that the correlation between tissues and plasma TMB was significantly improved with an increase in cfDNA input. Through the analysis, the researchers recommend that at least 20 ng cfDNA be injected to accurately and reproducibly assess bTMB [8].
Effect of cfDNA input on bTMB assessment
The researchers then analyzed the effect of tumor burden on bTMB assessment. The results showed that the tumor load was greater than 0.1% of the allele frequency, and the TMB value between the tissues and plasma had a good correlation.
Since tumor burden is related to tumor staging, the researchers further analyzed and found that the PPA of bTMB in patients with stage I, II, III, and IV diseases gradually increased from 0% to 11%, 35%, and 54%, and in the samples of patients with stage IV NSCLC, the TMB value correlation between tissue and plasma was the strongest.
The researchers then combined cfDNA input, tumor burden, and tumor staging to analyze and found that when samples from stage IV, cfDNA > 20ng were limited to samples, the tissue TMB critical value was 16mut/Mb, and the PPV and PPA of patients with high plasma TMB were 100% and 63%, respectively.
Analysis of a combination of cfDNA input, tumor burden, and tumor staging
Finally, to minimize lung TMB sequencing panels, the researchers studied a subset of panels during the design process and found that at least about 577 kb of panels had similar detection performance and clinical relevance compared to 1.1 Mb lung TMB panels.
Overall, the bTMB detection panel developed in this study is cost-effective, has high detection accuracy, and is significantly smaller than the previously set minimum value (1Mb), so bTMB is expected to become a predictive biomarker for judging whether NSCLC patients can receive ICI treatment.
bibliography:
[1] Le DT, Uram JN, Wang H, et al. PD-1 Blockade in Tumors with Mismatch-Repair Deficiency. N Engl J Med. 2015;372(26):2509-2520. doi:10.1056/NEJMoa1500596
[2] Rooney MS, Shukla SA, Wu CJ, Getz G, Hacohen N. Molecular and genetic properties of tumors associated with local immune cytolytic activity. Cell. 2015;160(1-2):48-61. doi:10.1016/j.cell.2014.12.033
[4] Fda approves pembrolizumab for adults and children with TMB H solid tumors. Available: https://www.fda.gov/drugs/drug-approvals-and-databases/fda-approves-pembrolizumab-adults-and-children-tmb-h-solid-tumors2020 [Accessed 13 Dec 2021].
[6] Schuurbiers M, Huang Z, Saelee S, et al. Biological and technical factors in the assessment of blood-based tumor mutational burden (bTMB) in patients with NSCLC. J Immunother Cancer. 2022;10(2):e004064. doi:10.1136/jitc-2021-004064
[7] Rizvi NA, Hellmann MD, Snyder A, et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science. 2015;348(6230):124-128. doi:10.1126/science.aaa1348
Responsible editor 丨IoTalker