Cannabis Use Is Common in Young Lung Cancer Patients and Linked to More Aggressive Disease

A French multicenter prospective cohort finds that dual cannabis-tobacco smokers develop lung cancer earlier and with more emphysema, worse pulmonary function, and more rare aggressive tumor types than tobacco-only smokers, though the inability to separate cannabis effects from tobacco exposure remains a fundamental limitation.

Why This Matters

Lung cancer in younger adults represents a distinct and increasingly recognized clinical challenge. Patients diagnosed before age 60 often present with different molecular profiles, later-stage disease, and exposures that do not fit the classic epidemiological picture of a decades-long, tobacco-only smoker. At the same time, cannabis use has risen substantially across much of the world, driven by expanding legalization, cultural normalization, and a widespread public perception that cannabis is fundamentally safer than tobacco. The intersection of these two trends raises an urgent clinical question: does chronic cannabis smoking contribute to lung cancer development, alter its phenotype, or worsen its severity? The endocannabinoid system is expressed throughout the pulmonary epithelium, and combusted cannabis smoke contains many of the same carcinogens found in tobacco smoke, along with compounds unique to the cannabis plant. Preclinical evidence has suggested both pro-inflammatory and immunomodulatory effects of chronic cannabinoid exposure in pulmonary tissue, but prospective clinical data in human lung cancer populations remain remarkably sparse. This study arrives at a moment when clinicians urgently need data to inform honest conversations with patients about dual smoking and lung cancer risk, and when the absence of such data has left a vacuum filled largely by assumption.

Clinical Summary

Lung cancer remains the leading cause of cancer death worldwide, and while overall incidence has declined modestly in many high-income countries due to tobacco control measures, the disease continues to exact a devastating toll. Among younger patients, the clinical picture is evolving. Increasing numbers of never-smokers are being diagnosed, molecular driver mutations are found at higher rates, and exposure histories are more complex. Cannabis smoking, which often co-occurs with tobacco use, has been hypothesized to contribute to pulmonary carcinogenesis through mechanisms that overlap with but are not identical to those of tobacco. Yet rigorous prospective data on the prevalence and clinical significance of cannabis co-use among lung cancer patients have been lacking, leaving clinicians to counsel patients in an evidence vacuum.

This prospective, multicenter, observational cohort study, conducted across three French hospitals (Gustave Roussy, Paris-Saint-Joseph, and Marie Lannelongue) between 2021 and 2023, enrolled 150 consecutive lung cancer patients aged 18 to 60 years. The investigators classified patients into three groups: dual cannabis-tobacco smokers (CTS), tobacco-only smokers (TS), and never-smokers (NS). A key methodological strength was the use of hair analysis by liquid chromatography-tandem mass spectrometry (LC-MS/MS) to objectively confirm self-reported exposure to THC, CBD, nicotine, and cotinine, thereby reducing misclassification bias. The mechanistic rationale for investigating cannabis as a modifier of lung cancer phenotype rests on the known presence of cannabinoid receptors (CB1 and CB2) in bronchial and alveolar epithelium, the documented carcinogenic content of combusted cannabis smoke, and the immunomodulatory effects of chronic cannabinoid exposure on pulmonary immune surveillance. Cannabis smoke delivers particulate matter and polycyclic aromatic hydrocarbons to the airways in a manner that differs from tobacco partly because of inhalation technique: deeper inhalation, longer breath-hold, and lack of filtration may increase per-joint carcinogen deposition.

Among the 148 analyzable patients, 39% were classified as dual cannabis-tobacco smokers, a prevalence far exceeding general-population estimates of cannabis use and underscoring the potential clinical relevance of this exposure in young lung cancer patients. These dual smokers reported heavy, chronic use, averaging approximately 4 joints per day over a median of 26 years. Compared with tobacco-only smokers, the CTS group was diagnosed approximately 3 years younger (median age 53 versus 56; P=.006), had nearly double the rate of emphysema on CT imaging (64% versus 38%; P=.003), significantly lower diffusing capacity for carbon monoxide (DLCO 63% versus 70%; P=.004), more chest pain at diagnosis (22% versus 8%; P=.03), and a markedly higher prevalence of rare and aggressive tumor histologies (17% versus 4%; P=.007). Crucially, tobacco burden was statistically equivalent between the CTS and TS groups as measured by both pack-years and hair cotinine concentrations, meaning that differential tobacco exposure does not readily explain the observed clinical differences.

Despite these striking associations, the study’s limitations are substantial and must be weighed carefully. Because this is a case-only cohort in which all participants already have lung cancer, the study cannot estimate whether cannabis increases the risk of developing lung cancer in the general population. It can only describe associations between cannabis exposure and disease characteristics among those already diagnosed. Furthermore, because every cannabis smoker in the cohort also smoked tobacco, it is impossible to isolate an independent cannabis effect. Unmeasured confounders, including differences in socioeconomic status, occupational exposures, genetic susceptibility, diet, and other substance use, could contribute to the observed clinical differences. The sample size of 148 patients, while reasonable for a prospective multicenter effort, limits statistical power for subgroup analyses and raises the possibility that some significant findings reflect type I error. The restriction to patients aged 60 and younger means findings may not generalize to the broader lung cancer population. The authors appropriately conclude that larger, controlled studies with better confounder adjustment are needed before any causal inferences or clinical recommendations can be drawn.

Dr. Caplan’s Take

This study makes a genuinely important contribution by doing something that is harder than it sounds: prospectively enrolling consecutive lung cancer patients, systematically documenting cannabis exposure, and then confirming those self-reports with validated hair biomarker testing. The use of LC-MS/MS to verify THC, CBD, nicotine, and cotinine levels is a methodological step forward that meaningfully reduces the exposure misclassification that has plagued this area of research. The finding that 39% of young lung cancer patients were heavy, long-duration cannabis smokers is not a number that can be dismissed as background noise.

I encounter patients regularly who ask whether their cannabis use could have contributed to their lung disease, or conversely, who are confident that cannabis is protective or at least harmless because they have read headlines suggesting anticancer properties of cannabinoids. These are patients who deserve a precise answer, and what I have to tell them right now is that we genuinely do not know. The mechanistic plausibility is real: combusted plant material delivers carcinogens to the airways regardless of the plant species, and the deep-inhalation technique characteristic of cannabis smoking likely increases particulate deposition. But plausibility is not proof, and the clinical data remain insufficient to quantify an independent risk.

The central problem with this study, and with much of the existing literature, is that cannabis and tobacco exposure are almost always entangled in the same patients. Even though pack-year exposure was equivalent between the two groups in this cohort, equivalence on a single metric does not guarantee equivalence on all relevant confounders. We do not know whether the dual smokers differed in ways that independently predispose to earlier diagnosis or more aggressive disease: occupational chemical exposures, air pollution burden, family history density, or genetic polymorphisms in carcinogen metabolism. The higher rate of rare tumor histologies in the CTS group is intriguing and hypothesis-generating, but with only a handful of cases driving that comparison, it requires replication before it can be interpreted as a signal rather than noise.

In practice, what I do with this information is straightforward. I document cannabis exposure history carefully in every patient, including duration, frequency, route of administration, and co-use with tobacco. I counsel patients that combusted cannabis smoke carries known carcinogens and that the absence of proven independent risk is not the same as evidence of safety. For patients who choose to continue using cannabis for symptom management or other reasons, I recommend non-combustion routes. And I watch for emerging data from larger cohorts that may eventually allow us to disentangle these exposures. We are not at a point where we can make causal claims, but we are well past the point where we can ignore the question.

Clinical Perspective

This study sits at the leading edge of a research area that has been characterized more by speculation than by data. For decades, the question of whether cannabis smoking independently contributes to lung cancer has lingered without resolution, largely because the epidemiological studies that could answer it have been methodologically difficult to execute. Most large lung cancer cohorts do not collect detailed cannabis exposure data, and those that do rarely confirm self-report with biomarkers. This French cohort, while modest in size, demonstrates that rigorous prospective characterization of dual-smoker lung cancer patients is feasible and can yield clinically meaningful phenotypic distinctions. It confirms what smaller retrospective series have suggested: that dual cannabis-tobacco smokers may present with a more severe clinical profile than tobacco-only smokers, including earlier age at diagnosis, more emphysema, and potentially different tumor biology.

Clinicians should understand clearly what this evidence does and does not support. It supports the recommendation to systematically assess cannabis exposure in all lung cancer patients, particularly younger ones. It supports the clinical observation that dual smokers may have worse baseline pulmonary function and more emphysema, which has direct implications for surgical candidacy, radiation tolerance, and perioperative risk stratification. It does not support telling patients that cannabis causes lung cancer, nor does it support quantifying an independent cannabis-attributable risk. When an informed patient asks whether their cannabis use contributed to their diagnosis, an honest response acknowledges the biological plausibility and the clinical associations while being transparent that the evidence cannot yet distinguish cannabis-specific effects from the consequences of chronic dual combustion exposure.

From a pharmacological and safety standpoint, the finding that dual smokers have significantly lower DLCO deserves particular attention. Reduced gas transfer capacity affects not only respiratory reserve but also drug metabolism and tolerance to cytotoxic chemotherapy, immunotherapy-related pneumonitis risk, and recovery from thoracic surgery. Clinicians managing these patients should ensure that baseline pulmonary function testing includes DLCO, not just spirometry, and should have a low threshold for pulmonary rehabilitation referral. The higher rate of rare and aggressive histologies in the CTS group, if confirmed in larger studies, could have implications for molecular testing strategies and treatment selection, as some rare histologies have distinct mutational landscapes and different responses to targeted therapies.

The actionable recommendation for practicing clinicians is to integrate structured cannabis exposure assessment into the standard intake for all lung cancer patients, using validated questionnaires that capture route, frequency, duration, and co-use patterns. Document these exposures in the medical record in a way that facilitates future research extraction. When available, consider biomarker confirmation of exposure status, as this study demonstrates the feasibility and value of hair toxicology. Communicate findings from studies like this one to patients in calibrated language: the association between dual smoking and worse clinical profiles is real and concerning, the inability to prove causation is also real, and the safest course is to avoid inhaling any combustion products. Finally, encourage eligible patients to participate in prospective registries or cohort studies that are working to resolve the cannabis-lung cancer question with adequate statistical power and confounder control.

Study at a Glance

Study Type

Prospective multicenter observational cohort

Sites

3 French hospitals (Gustave Roussy, Paris-Saint-Joseph, Marie Lannelongue)

Enrollment Period

2021 to 2023

Population

148 analyzable lung cancer patients aged 18 to 60 years

Exposure Groups

Cannabis-tobacco smokers (CTS, 39%), tobacco-only smokers (TS, 52%), never-smokers (NS, 9%)

Exposure Confirmation

Hair LC-MS/MS for THC, CBD, nicotine, and cotinine

Cannabis Exposure in CTS Group

Median 26 years duration, approximately 4 joints per day

Primary Comparisons

Clinical, radiologic,