Is there a genomic fingerprint of Radon (Rn)-induced lung cancer (LC)? Comparison of genomic alterations in LC specimens from high and low Rn zones.

Abstract

1572 Background: Rn-222 is a radioactive gas found in rocks and soil. It emits alpha particles that cause dsDNA breaks and increase potential for carcinogenesis. Rn is the 2nd leading cause of LC in the US after smoking. EPA estimates >15,000 deaths/yr (9% of LC deaths) from Rn. We hypothesize that the impact of Rn exposure may be reflected in LC gene mutation (mut) profiles. Methods: Using commercial NGS assays, we retrospectively analyzed genomic DNA alterations in FFPE specimens from 159 LC patients (pts) from the Lifespan Cancer Institute in Rhode Island (2014- 2019), followed by validation in a larger cohort of 5,532 pts using Caris platform. Based on EPA Rn maps, we identified counties with high indoor Rn levels (>4 pci/L; HR), and compared gene mut patterns with those from low Rn zones (<4 pci/L; LR). Based on pt’s zip code of residence, we categorized them to HR and LR. In the validation cohort, p values adjusted for multiple comparison (q) of < .05 were considered significant. Results: In the pilot cohort, 35 pts (22%) were in HR and 124 (78%) in LR zones. Adenocarcinoma histology was most frequent (73%) and smoking prevalence was high (75%) in both groups. Most prevalent alterations were TP53, KRAS and CDKN2A muts. In the HR, we noted more frequent recurrent muts in 2 DNA repair genes (DDR): ATM (11 vs 1%, p= .00086) and CHEK2 (6 vs 0%, p= .047) when compared to LR group. When classified into major pathways implicated in lung carcinogenesis, higher frequency of mutations were seen in DDR in HR zones vs. LR (29 vs 13%, p= .038). In the validation cohort, 1,433 (26%) pts were in HR and 4099 (74%) in LR zones. Among the DDR genes, ATM muts in HR group tended to be more frequent (4.7 vs 3.4% in LR, p= .03) as well as PALB2 (0.9 vs 0.4%, p= .02) while no difference seen in CHEK2. Other genes with significantly higher prevalence in HR were TP53, SMARCA4 and NFE2L2 (q< .05); while KMT2D, KEAP1, CDKN2A, MET, NF2, DNMT3A, CCND1 and FAS show a trend (p< .05). EGFR muts were significantly more frequent in LR zones (8.4 vs 14.6%, q= .001). Similar to the pilot cohort, DDR pathway alterations trend to be higher in HR zones (14 vs 12%, p= .05). Using a high TMB cut-off >10, tumors from HR zones had significantly higher TMB when compared to LR zones (56 vs 48%, q= .0005). Conclusions: To our knowledge, this is the first attempt to elucidate the pathobiology of Rn induced LC using gene mut analyses. Our observations suggest that LC associated with higher Rn exposure may have disabled DNA repair pathways and higher TMB. Assuming uniform tobacco smoke exposure, higher Rn was not associated with EGFR mut.