Emergence of DNA Damage Repair Clonal Hematopoiesis after Peptide Receptor Radionuclide Therapy for Neuroendocrine Tumors

Background. In advanced/metastatic neuroendocrine tumors (NETs), PRRT (Peptide Receptor Radionuclide Therapy), has shown efficacy but with significant hematologic toxicity, including therapy-related myeloid neoplasms (t-MN). Here, we assessed the effect of PRRT on the development of clonal hematopoiesis (CHIP) or clonal cytopenia of undetermined significance (CCUS) known to be a risk factor for t-MN development.

Methods. We performed a targeted 77 genes mutational analysis using Next Generation Sequencing (NGS) in blood samples of 153 patients with metastatic gastroenteropancreatic (GEP) NET, exposed or not to PRRT, without any history of hematological malignancies. Mutations in TP53, PPM1D, CHEK2 and ATM genes were considered as DNA damage repair (DDR) mutations, others as Non-DDR mutations The ethics committee (CPP) Ile de France IV has approved this study (Hematote 2) on January 2022 (ID-RCB: 2022-A00016-37).

Results.

Between January 2022 and April 2024, 153 patients with metastatic GEP NET patients were explored by NGS. Primary NET localization was mainly located in the small intestine (jejuno-ileal in 89/153, 58%) and in the pancreas (45/153, 29%). Seventy-eight patients (51%) were male, median age at NGS was 65.4 years (IQR, 57.3-73.1) and time between NET diagnosis and NGS was 6.5 years (IQR, 3.6-11.3). A total of 102 patients were treated with PRRT, including 44 (43%) who had also received chemotherapy and 16 (16%) chemoembolization; 51 were not treated with PRRT but 21 (41%) received chemotherapy, 4 (8%) chemoembolization and 26 (51%) none of these treatments.

Eighty-two (53.6%) patients had CHIP (n=34, 42%) or CCUS (n=46, 57%), including 33 (40%) patients with DNMT3A mutation, 20 (24%) patients with TET2 mutation, 33 (40%) patients with PPM1D mutation and 6 (7%) with TP53 mutation. Median variant allelic frequency (VAF) of the main clone for DNMT3A, TET2, PPM1D, TP53 was 5% (range, 2-32), 4% (range, 2-40), 6.1% (range, 2-31) and 4.6% (range, 2-27), respectively. Median number of mutations was 2 (range, 0-6).

Patient with CHIP/CCUS were older compared to others with a median age of 69.5 years (IQR, 62.6-75.6) vs 59.6 (IQR, 50.2-67.9) (p<0.0001), received more PRRT treatment (65/102, 64%) vs (17/51, 33%) (p=0.0004) but no differences were seen in term of gender, tumor site localization, history of treatment with chemotherapy, cytopenia at NGS and time between NGS and NET diagnosis. Interestingly, the detection DDR mutations were restricted to patients with history of PRRT treatment (41/102 [40%] vs 2/51 [4%]; p<0.0001), which was not the case for Non-DDR mutations (40/102 [39%] vs 16/102 [31%], p= 0.3).

Median time between first PRRT cycle and NET diagnosis was 4.3 (IQR, 1.6-9.2) years. Patients received a median of 4 (IQR, 4-4) cycles of PRRT for a median total dose of 29.6 Gy (IQR, 29.5-29.8) over a period of 5.7 months (IQR, 5.5-7.3) between the first and last PRRT cycle. Twenty patients (20%) had cytopenia before PRRT treatment, 39 (40%) at the end of PRRT. Median time between PRRT and NGS was 2.4 years (IQR, 1-3.8). Out of the patients treated with PRRT, 41 (40%) had DDR mutations including 33 (32%) patients with PPM1D mutation and 6 (6%) TP53 mutation. Forty (39%) had Non-DDR mutations including 24 (24%) DNMT3A mutation and 13 (13%) TET2 mutation. Except age, the only factor associated with detection of CHIP was the duration of PRRT (5.6 months [IQR, 5.5-6.2] vs 5.9 [IQR, 5.5-7.8], p=0.04).

In patients treated with PRRT with CHIP, DDR mutated patients had lower neutrophil counts (2.7 G/L [1.9-3.6] vs 3.9 G/L [IQR, 2.8-4.8], p= 0.009) and platelets (167.5 G/L [IQR, 134-206] vs 208 G/L [IQR, 151-268], p=0.03) at the end of the PRRT than patient without DDR mutations. Conversely, patients with non-DDR mutations had lower hemoglobin level before initiation of PRRT (12.5 g/dl [IQR, 11.8-13.2] vs 13.8 g/dl [IQR, 12.6-14.2], p=0.0078).

No myeloid transformation occurred during follow up (the median follow-up from NGS in the overall population was 1.24 years [95% CI, 0.97-1.49]).

Conclusion.

Patients with NET have a very high incidence of CH. Patients exposed to PRRT seems to have more CH, especially CHIP in the DDR pathway suggesting a potential clonal selection of this treatment. Association between PRRT, CHIP and early hematological toxicity is unclear and longer follow up is necessary to correlate the presence of CHIP with a higher incidence of t-MN.