Figure 1: ctDNA dynamics reflect changes in disease burden across different tissue compartments.

(a) Case CLL020: plasma ctDNA levels by TP53 p.R209fs mutation (left) and IGH (right) decreased over the course of ofatumumab and venetoclax treatment, which paralleled the decline in mutant DNA levels in the MNL, the decrease in PBL count and the reduction in lymphadenopathy as assessed by imaging. (b) Case CLL018: the initial peripheral lymphocytosis observed post ibrutinib treatment coincided with a increase in the abundance of the NOTCH1 p.P2415fs mutation in the MNL. This contrasted with plasma ctDNA levels of both the NOTCH1 p.P2415fs mutation (left) and IGH (right), which instead reflected the reduction in radiological disease burden. (c) Case CLL001: following obinutuzumab therapy, rapid clearance of the circulating leukaemic cells was observed and the NOTCH1 p.P2415fs mutation became undetectable in the MNL. This coincided with a parallel decrease in plasma ctDNA as assessed by NOTCH1 p.P2415fs mutation (left) and IGH (right) levels, although ctDNA did not become undetectable. The patient then had compartmentalized disease progression with increasing lymphadenopathy that was matched by a plasma ctDNA rise. Subsequent ibrutinib treatment resulted in a reduction in lymphadenopathy and plasma ctDNA. These dynamic changes were not reflected in the PBL count or by the NOTCH1 p.P2415fs mutation in MNL DNA. (d) Correlation between the change in ctDNA/MNL MAF versus change in radiological disease burden (cm²) from the maximal value analysed for each individual patient across any time point. Of 38 matched serial time points from a total of 12 patients, the correlation with radiology was significantly better with ctDNA (r²=0.78, P<0.0001) than with MNL DNA (r²=0.24, P=0.0019). (e) Comparison of ctDNA and MNL detection across 111 matched time points. (f) Comparison of ctDNA with detection of CLL by multi-colour flow cytometry on PB or BM.