Real-world concordance of tumor mutational burden (TMB) between blood (circulating tumor DNA) and tissue, and its association with immunotherapy response.

2580 Background: Tissue-derived tumor mutation burden (tTMB) ≥10 mutations/Mb is a histology-agnostic predictive biomarker for the immune checkpoint inhibitor (ICI) pembrolizumab. It is unknown if circulating tumor DNA–derived TMB (bTMB) is also predictive and current ASCO guidelines recommend against using bTMB alone. Concordance between tTMB and bTMB is only moderate, with reported Pearson correlation coefficients between 0.54-0.70. In prior studies, bTMB was a median 2.4-fold greater than tTMB in matched samples. This study evaluated the correlation and predictive accuracy of immunotherapy response of bTMB vs. tTMB in solid tumors. Methods: 221 patients with solid tumors were included from the Cleveland Clinic genomics repository who had at least one tTMB and one bTMB result between July 1, 2021, and July 1, 2025. High TMB was defined as tTMB ≥10 mutations/Mb. For bTMB, thresholds of ≥10 and ≥16 mutations/Mb were evaluated, as prior studies suggest a bTMB threshold of 16 for Guardant 360 may be comparable to a tTMB of 10. Correlation between bTMB and tTMB was assessed using Pearson’s correlation coefficient. Among patients treated with ICIs, treatment response and time to treatment failure (TTF) were analyzed using both bTMB thresholds and compared with tTMB. Results: Among the 221 patients included in the analysis, 80 received ICIs. Median bTMB was 2.3-fold higher than matched tTMB. A strong positive linear correlation was observed between bTMB and tTMB both among patients who received ICIs (R² = 0.845, p < 0.001) and across the full cohort of 221 patients (R² = 0.818, p < 0.001). High tTMB was associated with numerically longer TTF in those with both low bTMB (median 21.3 months) and high bTMB (median 15.2 months), as shown in Table 1. Those with high tTMB/low bTMB had numerically longer median TTF compared to those with low tTMB/high bTMB (21.3 vs. 7.0 months). Increasing the bTMB threshold from 10 to 16 was associated with improved median TTF from 3.5 to 7.0 months among patients with low tTMB. Conclusions: This study confirms that bTMB values are consistently higher than tTMB and demonstrates a strong positive correlation between the two measures. Raising the bTMB threshold appears to improve prediction of immunotherapy response in patients with low tTMB. However, tTMB continued to show evidence of predictive value for response to ICIs regardless of bTMB, although these subgroup analyses were limited by small sample sizes. Larger, disease-specific studies are needed to better define the independent clinical utility of bTMB. Time to treatment failure in TMB-high vs. TMB-low patient pairs. bTMB -10=H bTMB -16=H Median TTF (months) Median TTF (months) tTMB-L, bTMB-L(n=35) 4.9 tTMB-L, bTMB-L(n=51) 3.9 tTMB-H, bTMB-L(n=3) 21.3 tTMB-H, bTMB-L(n=4) 21.3 tTMB-L, bTMB-H(n=32) 3.5 tTMB-L, bTMB-H(n=16) 7.0 tTMB-H, bTMB-H(n=10) 10.2 tTMB-H, bTMB-H(n=9) 15.2