TP53 Mutation Variant Allele Frequency in Cell-Free DNA Is Increased in Patients with Transformed Follicular Lymphoma

Introduction

Although the prognosis of follicular lymphoma (FL) is improved, that of transformed FL (trFL) is still poor. Distinguishing between FL and trFL is crucial for determining appropriate treatment strategies. However, due to tumor heterogeneity, a single-site biopsy may be insufficient to differentiate between FL and trFL. Cell-free DNA (cfDNA) contains tumor-derived DNA and has the potential to capture heterogeneous tumor profiles. We compared mutation profiles in cfDNA and tissue-derived DNA from FL patients and evaluated the utility of cfDNA in assessing transformation.

Materials and methods

This is a prospective, non-interventional, observational study. Patients with newly diagnosed or relapsed FL were enrolled. Clinical trFL was defined based on previously reported criteria. cfDNA was extracted from the 33 patients. Mutations in 33 genes associated with FL transformation were analyzed using targeted-capture sequencing.

Results

The study included 33 patients with FL, of whom 13 were trFL (5 clinically diagnosed and 8 histologically diagnosed). The median age at cfDNA collection was 71 (36-93) years, and 15 were males. Twenty-two patients were classified as Ann Arbor stage III/IV. Twenty-three patients underwent PET scans within three months before and after cfDNA sample collection, with a median SUV max of 15. There were no significant differences in age, gender, PS, FLIPI, or stage between FL and trFL groups.

The amount of cfDNA in trFL was higher than that of FL (median:23.2 ng/mL in trFL vs 4.2 ng/mL in FL, p<0.001). Even after adjusting for the metabolic tumor volume of PET scans, the amount of cfDNA was significantly higher in trFL than in FL (p=0.017). The optimal cut-off value for distinguishing between FL and trFL based on cfDNA amount was determined to be 16.6 ng/mL using ROC curve analysis.

Targeted sequencing was performed on cfDNA from 30 patients, and for 16 of these patients, DNA extracted from FFPE tissue biopsies was also analyzed using targeted sequencing. At least one somatic mutation was detected in all cfDNA samples, with a mean of 11.7 mutations, whereas a mean of 2.1 mutations were detected per FFPE sample. One hundred mutations were detected in cfDNA, 13 of which were also identified in FFPE-derived DNA. Among mutations in cfDNA, TP53 was most frequent (76%), followed by CREBBP in 60%, KMT2D in 60%, CARD11 in 43%, CIITA in 26%. In FFPE-derived DNA, CREBBP (37%) and KMT2D (18%) were also frequent. However, TP53 was detected in only one case (6%) from FFPE-derived DNA, and the same TP53 mutation was also observed in the cfDNA of that case. Although there was no significant difference in the detection rate of TP53 mutations between FL and trFL, the TP53 variant allele frequency (VAF) in cfDNA was significantly higher in histologically trFL compared to other FL (p=0.0472). Interestingly, a TP53 G245D substitution was observed in 14 samples (63%).

Discussion

In cfDNA, TP53 mutations were detected even in cases lacking TP53 mutations in FFPE samples. The higher VAF of TP53 in trFL suggests that cfDNA analysis can detect subclonal TP53 mutations and may enable early identification of transformation.

No relevant conflicts of interest to declare.