Distinct Genomic Profile of Lung Cancer in Never Smokers Comes Into Focus

During the past 2 decades, the treatment of patients with lung cancer has undergone a paradigm shift, a transformation fueled by genome sequencing efforts that have revealed oncogenic driver mutations and fostered targeted drug development.

Tobacco smoking is the most significant risk factor for lung cancer development¹; therefore, findings from studies of smoking-associated disease largely have informed our current understanding of the genomic background of this cancer type.²

Now a growing recognition of the distinct clinical, pathological, and biological features of lung cancers that arise in nonsmokers³ is fostering greater interest in examining the molecular underpinnings of lung cancer in this patient subset, which by itself ranks among the top 10 causes of cancer-related mortality.⁴

Findings from recent large-scale genome sequencing studies of tumors from never smokers have demonstrated that differences between tumors from smokers and nonsmokers extend to the molecular level.^2,5 Data from these studies have important implications for the way that patients with lung cancers are treated and highlight the need for earlier identification of disease in never smokers.

A Changing Disease

The era of precision medicine for non–small cell lung cancer (NSCLC) has resulted in the development of therapies for numerous targetable molecular alterations, including EGFR mutations, ALK and ROS1 translocations, and even the once-elusive KRAS mutations.^6,7 The development of immune checkpoint inhibitors targeting the PD-1 protein and its ligand, PD-L1, has been another significant breakthrough in the treatment of NSCLC.⁸

Despite these therapeutic achievements, accompanied by advances in screening, lung cancer remains the leading cause of cancer- related mortality in the United States,⁴ and the prognosis for patients with advanced disease continues to be poor.

The majority of cases (56%) are diagnosed in the distant stage, after the cancer has metastasized, according to the Surveillance, Epidemiology, and End Results Program data and staging criteria.⁹ The 5-year relative survival rate (2010-2016) for patients with distant-stage disease is 7% for NSCLC and 3% for small cell lung cancer.¹⁰ The 60-month overall survival rate ranges from 68% in patients with stage IB disease to 0% to 10% in patients with stage IVA and IVB disease.¹¹

Tobacco smoking is the leading cause of lung cancer; there are more than 60 known carcinogens in tobacco smoke, which are thought to contribute to the development of lung cancer principally by generating DNA adducts that foster mutagenesis.¹²

However, declining rates of smoking mean that a growing proportion of lung cancers are found in patients who have smoked fewer than 100 cigarettes in their lifetime (never smokers). Never smokers made up 12.5% of patients diagnosed with lung malignancies (N = 129,309), according to results of an analysis of data from cancer registries of 7 states from 2011 to 2016. The proportion of never smokers was highest among patients with lung cancer who received a diagnosis at aged 20 to 49 years, with 27.7% of cases in women and 18.6% in men (Figure¹³).^4,13

The drivers of lung cancer in these patients are less clear than they are in smokers, although a number of potential risk factors have been identified, including secondhand smoke, environmental exposures, and genetic factors.^3,14

What has become increasingly clear is that lung cancer in never smokers is distinct from that which arises in smokers in a number of important ways (Table^1,3,5,14-18). In addition to histological and epidemiological differences in lung cancers from never smokers, distinctions in underlying molecular alterations have been noted in many studies.^3,19

Comprehensive molecular profiling of lung adenocarcinoma has provided valuable insights into the genomic landscape of this cancer type in the past decade^20,21; however, the majority of samples used in sequencing studies have been from smokers. Until recently, instances of whole-genome sequencing of lung cancer from never smokers had been much more limited.²

The publication of several comprehensive sequencing studies focusing on lung cancer in never smokers during the past 2 years reflects a growing interest in illuminating the molecular pathogenesis underlying this disease and how it differs from lung cancer in smokers.

Distinct Molecular Drivers

In one such study, investigators performed tumor and matched germline whole-exome sequencing (WES) and RNA sequencing of lung adenocarcinoma samples from never smokers received from 3 US institutions (dubbed the “institutional cohort”). They also analyzed WES and RNA sequencing data from never-smoker samples included in prior sequencing efforts by The Cancer Genome Atlas (TCGA) and Clinical Proteomic Tumor Analysis Consortium (CPTAC) (external cohort).⁵

The analysis included 160 never-smoker samples (84 from the institutional cohort and 76 from the external cohort), in addition to 299 smoker samples from the external cohort.⁵ The median age at diagnosis was similar in the smoking and nonsmoking patients, but there was a higher prevalence of female patients among the never smokers (69.4% vs 48.5%; P < .001).⁵

Among tumor samples in the institutional cohort (all never smokers), the median tumor mutation burden (TMB) was 2.93 mutations per megabase (Mb).⁵ Notably, TMB in lung adenocarcinoma samples from never smokers was significantly lower than in those from smokers in the sequencing study; median TMB was 1.25 and 1.95 mutations per Mb in the CPTAC and TCGA samples of the external cohort for never smokers compared with 9.1 and 10.83 mutations per Mb, respectively, for smokers (P < .0001).⁵ Overall, lung cancers tend to have among the highest TMBs among malignancies, according to a 2017 analysis of diverse cases. The median TMB ranged from 5.4 for lung adenosquamous carcinoma to 12.2 for lung large cell carcinoma, including 6.3 for lung adenocarcinoma.²²

Importantly, a driver alteration was identified in 81% of never-smoker samples in the institutional cohort, the vast majority of which were therapeutically actionable with FDA-approved targeted therapies.⁵

A comparison of the mutation frequency of 20 genes known to be recurrently altered in lung adenocarcinomas revealed that EGFR was the most commonly mutated gene in never smokers (51% of samples), whereas KRAS mutations were the most common among smokers (35% of samples).⁵

This confirms findings from numerous other studies establishing EGFR mutations as more common in lung adenocarcinomas from never smokers. Other key therapeutically actionable driver alterations, including rearrangements involving the ALK and ROS1 genes, have also consistently been found more frequently in never smokers.

Meanwhile, the likelihood of harboring KRAS mutations was lower for never smokers than for ever smokers among patients with lung cancer both for those of Caucasian/mixed ethnicity (odds ratio [OR], 0.33) and Asian ethnicity (OR, 0.39), according to findings from a 2016 meta-analysis.²³ In another study, investigators found that KRAS mutations were less frequent in samples from never smokers (15%) than in former (22%) or current smokers (25%), although the difference was not statistically significant.²⁴ The types of KRAS mutations observed do appear to differ, however, with transition mutations more common in never smokers and transversion mutations in smokers.¹⁴

Study data also have shown that lung adenocarcinomas from never smokers are less likely to express high levels of PD-L1 compared with those from smokers; consistent with this finding, never smokers are less responsive to immunotherapy.^1,14 The authors of the sequencing study discussed above found that lung adenocarcinomas from never smokers could be categorized into 3 different immune subtypes (IM-1, IM-2, and IM-3), which were immunologically “cold” (IM-3 and select IM-2 samples) or “hot” (IM-1). The use of appropriate biomarkers could help identify never smokers within the IM-1 subtype, who would be more likely to respond to immune checkpoint inhibitor therapy.⁵

Immunotherapy is less effective in patients with EGFR mutations; therefore, patients with these alterations have been excluded from clinical trials of immune checkpoint inhibitors in lung cancer, which thereby largely rules out never smokers.²⁵ Some investigators are seeking to change that equation. A phase 2 trial (NCT03786692) evaluating the combination of carboplatin, pemetrexed, and bevacizumab (Avastin) with or without the PD-L1 inhibitor atezolizumab (Tecentriq) includes treatment-naïve never smokers with wild-type tumors along with participants with EGFR exon 19 or 21 mutations irrespective of smoking history. Eligibility requirements for all participants include stage IV nonsquamous NSCLC, no ALK or ROS1 rearrangements, and no prior treatment with chemotherapy, a VEGF inhibitor, or immunotherapy, except tyrosine kinase inhibitors in patients with EGFR-mutant tumors.

Divide and Conquer?

In the Sherlock-Lung study, investigators at the National Cancer Institute conducted whole- genome sequencing on tumors and matched healthy tissues from 232 never smokers with NSCLC, predominantly adenocarcinoma.²

Again, the investigators found that more than half of the tumor samples had known driver alterations in components of receptor tyrosine kinase signaling pathways; EGFR was the most frequently altered gene (30.6% of samples), followed by KRAS (7.3%), ALK (6.0%), MET (4.3%), HER2 (3.9%), ROS1 (2.6%), and RET (1.3%).²

Most notably, however, the investigators performed a cluster analysis of somatic copy number alterations and identified 3 molecular subtypes of NSCLC in never smokers. They named these subtypes using terms from musical dynamics: piano, mezzo-forte, and forte.²

The piano subtype was the most common, representing almost half of all tumors. Piano tumors were characterized by the lowest median TMB (0.7 mutations/Mb), with few known drivers. One exception to this was KRAS mutations, which were overrepresented in this subtype (more than 75% of KRAS-mutant tumors were in the piano subtype).²

The mezzo-forte subtype represented approximately one-third of tumors and had a higher median TMB (1.6 mutations/Mb). More than half (51.4%) of the tumors in this subtype harbored EGFR mutations, whereas few had KRAS mutations.²

Of the 3 subtypes, forte bore the most genomic similarity to lung cancers from smokers. Nonetheless, this subtype’s median TMB (1.4 mutations/Mb) was still significantly lower than that in smokers. Tumors in this subtype were the most likely to have TP53 mutations (31.9% of cases) and were characterized by frequent whole-genome doubling (in 95.7% of cases). EGFR mutations were observed in 38.3% of forte tumors and KRAS mutations in 6.4%.²

The investigators applied their clustering analysis to lung adenocarcinoma data from a TCGA study,²⁶ and in addition to identifying the same 3 subtypes with similar features among the never smokers, they found that the majority of lung adenocarcinomas from smokers fell into the forte subtype. Interestingly, among the smoker samples characterized as the mezzo-forte subtype, the frequencies of EGFR and KRAS mutations were different from never smokers—9% and 34%, respectively, compared with 51.4% and 1.4% for the never smokers.²

These findings suggest that subtypes of lung adenocarcinoma in never smokers may benefit from different therapeutic strategies. Piano tumors have a dearth of driver mutations and are therefore likely to be less amenable to targeted therapies. However, newly approved KRAS G12C inhibitors may prove to be a valuable therapeutic strategy in these tumors, given the prevalence of KRAS mutations. The study authors also highlighted the possible role of stem cells in the development of piano tumors and suggested that targeting stem cellassociated signaling pathways might represent another future therapeutic avenue.²

The mezzo-forte and forte subtypes appeared to be more readily amenable to existing targeted therapies, particularly EGFR inhibitors, and driver mutations in these tumors were more clonal in nature, suggesting that a therapeutic strategy could be identified with a single biopsy.²

Currently no therapies are approved specifically for never smokers with NSCLC, but given the growing burden of this type of lung cancer, there is increasing interest in studying drugs in this patient population. Lantern Pharma is developing LP-300, a multitargeted modulator of EGFR, MET, and ROS1.²⁷

Originally developed by BioNumerik Pharmaceuticals, Inc as BNP7787 (Tavocept),²⁸ the agent did not meet the prespecified end point in a previous phase 3 study (NCT00966914) in patients with advanced lung adenocarcinoma. However, retrospective analyses demonstrated that median overall survival was significantly improved in a subset of never-smoker patients.²⁷ Lantern Pharma is repositioning this drug as a treatment for never smokers, with plans for launching a phase 2 clinical trial.²⁹

Screening Implications

Another notable finding from the Sherlock-Lung study was that the 3 never-smoker subtypes varied significantly in their latency period. By using clock-like mutations to determine the time of appearance of the most recent common ancestor (MRCA) of all tumor cells and then grouping tumors according to common driver events, the study authors determined that tumors in the piano subtype had a much longer (P = .0008) median latency period (9.10 years) than mezzoforte (0.28 years) and forte tumors (0.08 years).²

This indicates that tumors in the piano subtype are much slower growing, with almost a decade between emergence of the MRCA and diagnosis. These findings have potentially important implications for the detection of these tumors; there is a window for early detection when they may be more amenable to curative surgical resection.²

Although current screening recommendations apply only to smokers, there have been calls to consider expanding screening criteria due to the growing proportion of cases in never smokers. However, with their rapid growth, tumors in the mezzo-forte and forte subtypes may not be identified by annual screening.²

Jane de Lartigue, PhD, is a freelance medical writer based in Gainesville, Florida.