Cancer and Sleep Apnea: Cutaneous Melanoma as a Case Study

Melanoma is the most aggressive of all skin cancers and carries a very high mortality rate. Malignant transformation of melanocytes underlies the occurrence of this malignancy, which accounts for approximately 5% of all cutaneous malignancies, but it represents a very large proportion of cutaneous cancer–associated deaths. The incidence of cutaneous melanoma has been on the rise and is currently the 12th most prevalent cancer worldwide, carrying an age-standardized incidence rate of 3.0 per 100,000, reflecting well-established dynamic interdependencies between environmental factors and genetic susceptibility. Known risk factors for cutaneous melanoma include increased lifetime ultraviolet light exposure, family antecedents of melanoma, degree of skin pigmentation, and hair color. The identification of new risk factors for development of melanoma or factors related to melanoma aggressiveness is crucial to the development of new treatment options (1).

Over the last several years, a potential association between cancer and obstructive sleep apnea (OSA) has been advanced and has gained substantial interest in the sleep community (2). Several biologically plausible pathophysiological pathways support this relationship (3–5). However, a critical limitation of the current scientific literature is that the vast majority of these studies were not originally designed to address this association, and they have explored all cancers as the outcome variable rather than focusing on specific cancer types or sites. Indeed, it is definitely reasonable to assume that not all tumor cells originating from different organs and tissues will respond equally to the intermittent hypoxia (IH), sleep fragmentation, and immune dysregulation that characterize OSA, which could account for the conflictive and contradictory results reported by the published studies to date (2).

Among the scarce studies that have specifically examined the links between specific cancer sites and OSA, cutaneous malignant melanoma (hereinafter “melanoma”)—a very lethal and prevalent skin tumor with reliable and extensively validated aggressiveness markers (1)—has probably received the greatest attention (Table 1). This perspective provides a critical review of the pathophysiological mechanisms and epidemiological links between these two disorders, as well as the future challenges that should be assessed to better understand this relationship and its validity.

Table 1.

Most Relevant Animal Models and Pathophysiological and Clinical Studies on Association between Sleep-disordered Breathing and Malignant Cutaneous Melanoma

Author(s), Year (Reference)	Design	Outcomes/Objective	Main Findings	Comments and Limitations
Animal model studies
Almendros and colleagues, 2012 (19)	Mouse melanoma in vivo model	Tumor weight	IH enhanced tumor growth	Descriptive study
Almendros and colleagues, 2012 (20)	Melanoma in vivo model in lean and obese mice	Tumor weight and angiogenesis	Obesity and IH increased tumor growth	The metabolic syndrome mouse model employed causes severe obesity
Almendros and colleagues, 2013 (22)	Mouse melanoma in vivo model	Metastasis and intratumoral oxygen partial pressure	IH induced an increase in lung metastasis	Only lung metastasis was assessed
Eubank and colleagues, 2013 (64)	Mouse melanoma in vivo model	Lung metastasis, tumor necrosis, and angiogenesis	IH induced tumor necrosis, angiogenesis, increased expression of HIF, and lung metastasis	The total exposure to IH was relatively short (10 d)
Almendros and colleagues, 2014 (27)	Melanoma in vitro model in monoculture or in coculture with macrophages	Cancer cell proliferation	Proliferation rates of melanoma cells exposed to IH increased only when macrophages were concurrently present	No tested additional malignant properties in melanoma cells
Perini and colleagues, 2016 (26)	Mouse melanoma in vivo model	Tumor volume and markers of malignancy	IH enhanced expression of several markers of melanoma aggressiveness and promoted a tendency of increase in tumor growth	Low number of mice included in the study, and tumor weight was not measured
Yoon and colleagues, 2017 (21)	Mouse melanoma in vivo model	Tumor growth, angiogenesis, expression of HIF, and apoptosis	IH increased tumor progression and vascularization	IH with longer cycles than experienced by patients with severe OSA
Li and colleagues, 2018 (23)	Mouse melanoma in vivo model	Lung metastases and generation of ROS	IH increased the number and weight of lung metastases and enhanced ROS generation in melanoma cells	Only lung metastases were studied
Pathophysiological studies
Olbryt and colleagues, 2014 (30)	In vitro experimental study in three tumor cell lines (33% melanoma)	Gene expression analysis of tumor cell lines exposed to SH and IH	Expression of IL-8, VEGF, and other genes is more affected by IH than by SH	No long-term effect evaluation
Martínez-García and colleagues, 2017 (34)	Multicentric prospective study in 376 samples	Relationship between HIF-1α or VEGF expression in melanoma	HIF-1α but not VEGF expression was associated with tumor aggressiveness	Descriptive study
Almendros and colleagues, 2018 (33)	Multicentric prospective study in 376 samples	Association between HIF-1α and VEGF expression and SDB	HIF-1α expression in tumoral lesions was associated with DI4%	No discrimination of major cell lineage subsets in the tumor sample
Santamaria-Martos and colleagues, 2018 (37)	Multicentric prospective study in 360 samples	Relationship between OSA and tumor growth–related biomarkers	OSA was associated with elevated circulating concentrations of VCAM-1	No information about the CPAP effect
Cubillos-Zapata and colleagues, 2019 (46)	Multicentric prospective study in 360 samples	Association between circulating concentrations of soluble PD-L1 in patients with or without OSA	Soluble PD-L1 concentrations are increased in patients with melanoma who have severe OSA and related to tumor aggressiveness and invasiveness	No determination of PD-1 expression
Clinical and epidemiological studies
Martinez-Garcia and colleagues, 2014 (54)	Multicentric, prospective, pilot study in 56 patients with melanoma	OSA prevalence and its relationship with aggressiveness markers of melanoma	AHI ≥15 events/h: 30.3%; AHI and Tsat90% were predictors of fast-growing melanoma	Limited number of patients
Cohen and colleagues, 2015 (52)	Epidemiological study including three prospective cohorts (178,633 subjects, including 880 patients with melanoma)	Incident melanoma in patients with OSA	No association between snoring or OSA and risk of incident melanoma	Two of three series exclusively of women; lack of control by potential confounders
Gozal and colleagues, 2016 (50)	Epidemiological case–control study in 5.1 million subjects (with 1.7 million patients with OSA and 19,927 patients with melanoma)	Relationship of OSA diagnosis to incidence of cancer and risk of metastasis	The incidence of melanoma was higher in patients with OSA. No relationship was found between OSA and risk of metastasis in patients with melanoma	Health insurance database with lack of some important confounders
Sillah and colleagues, 2018 (51)	Epidemiological retrospective study of 34,402 patients with sleep apnea (129 patients with melanoma)	Incidence of cancer in patients with OSA compared with population estimate of incidence of cancer	The incidence of melanoma was higher in patients with OSA	Association was particularly prominent in younger patients (<60 yr) and in men
Martinez-Garcia and colleagues, 2018 (53)	Multicentric, prospective study of 443 patients with melanoma	Relationship between OSA and aggressiveness markers	AHI or DI4% values were associated with some markers of melanoma aggressiveness	Association was particularly prominent among patients aged <56 yr with Breslow depth >2 mm

Definition of abbreviations: AHI = apnea–hypopnea index; CPAP = continuous positive air pressure; DI4% = desaturation index at 4%; HIF = hypoxia-inducible factor; IH = intermittent hypoxia; OSA = obstructive sleep apnea; PD-1 = programmed cell death protein 1; PD-L1 = programmed death ligand 1; ROS = reactive oxygen species; SDB = sleep-disordered breathing; SH = sustained hypoxia; Tsat90% = nighttime spent with oxygen saturation below 90%; VCAM-1 = vascular cell adhesion molecule 1; VEGF = vascular endothelial growth factor.

Cancer and OSA: Pathophysiological Pathways

Sleep fragmentation and IH are the two main consequences of OSA that activate intermediate mechanisms which promote oncogenic processes (Figure 1). IH has become one of the fundamental phenomena driving solid tumor phenotypic expression, particularly since the realization that rapidly growing aggressive tumors, such as melanoma, outpace the evolving vascular network and oxygen supply together with the formation of new aberrant vessels (6, 7). The consequences include reduced availability of oxygen and nutrients to the tumor, thereby creating heterogeneous microenvironments. Considering the fact that, similar to all other cells, melanoma tumor cells have an intrinsic ability to sense changes in oxygen tension and adapt to preserve survival, exposures to such dynamic changes in oxygenation will trigger the generation of hypoxia-resistant cells, together with the formation of centric regions of necrotic cells, which are frequently observed upon histological examination of human solid tumors and are generally perceived as poor prognostic indicators (8, 9). Indeed, cancer cells can rapidly develop the ability to adapt to chronic hypoxic stress via several mechanisms, including upregulated expression and activity of angiogenic factors, shifts in metabolic regulation to reduce $\stackrel{.}{\text{V}}$o₂, altered mitophagy pathways, and coordinated changes in signaling pathways aimed at reducing their apoptotic potential. Taken together, the unique temporal trajectories of intratumoral hypoxia and downstream pathways can either foster cancer cell apoptosis or effectively guide phenotypic plasticity of tumor cells, which can give rise to subpopulations of highly invasive cells that are intrinsically less sensitive to novel therapies for melanoma (10, 11).

Figure 1.

Potential mechanisms linking obstructive sleep apnea and oncogenesis. Sleep fragmentation and intermittent hypoxia are two main consequences of obstructive sleep apnea that activate intermediate mechanisms, which promote oncogenic processes. HIF = hypoxia-inducible factor; ROS = reactive oxygen species; VEGF = vascular endothelial growth factor.

Alongside the effects of intratumoral hypoxia on cancer cells, as reflected by the emergence of the Warburg effect and glutamine turnover, there is also evidence of enhanced metabolism of glucose from glycolysis to oxidative phosphorylation that will result in the activation of signaling pathways, upregulation of oncogenes and transcriptional factors, and inactivation of tumor suppressor genes. In parallel, immune cells (natural killer cells, macrophages, and T-cell lymphocytes) are also susceptible to the hypoxic conditions within the tumor and will undergo metabolic reprogramming when switching to an activated state. Such reprogramming markedly affects the ability of these cells to effectively induce cytotoxicity, thereby illustrating how hypoxia-induced deregulated metabolism in the tumor microenvironment exerts a global immunosuppressive effect (12–15).

On the basis of the aforementioned considerations, the superimposition of cycling patterns of hypoxia that characterize OSA, albeit at much higher frequencies than those occurring within the tumor, could markedly influence the ongoing processes within solid tumors such as melanoma. Such a hypothetical framework is reviewed below. However, in addition to the aforementioned considerations involving intratumoral IH, other hallmark perturbations of OSA could also play a role in altering tumor biological properties and the defense mechanisms against them. For example, sleep fragmentation has been implicated in the emergence of reduced anticancer immunosurveillance, and neoplastic growth control may also be less effective, despite increased systemic inflammatory activity (16–18).

Animal Model Studies

The experimental findings substantiating that melanoma progression is enhanced by OSA were obtained by employing a conventional animal model. Specifically, melanoma cells were subcutaneously injected into the flank of the animal, and the in vivo evolution of these malignant cells (growth of primary tumor and distant organ metastasis) in animals challenged with IH-mimicking OSA was compared with that in normoxic mice. A variant of this model consisted of injecting melanoma cells into a peripheral vein to directly establish tumors in the lung.

There is evidence that IH with magnitude and frequency similar to those experienced by patients with severe OSA increases the growth rate of primary melanoma tumors. Almendros and colleagues (19, 20) reported that 2 weeks after injecting melanoma cells, the weight of the tumor was approximately twofold greater in severe OSA model mice (60 events/h; 20 s at 5% O₂ and 40 s at 21% O₂) than in the normoxic control animals (Figure 2). IH-increased melanoma tumor growth was also recently reported by Yoon and colleagues (21), using the same model but with less severe nadir hypoxemic events (12% O₂ instead 5% O₂ nadir) and with reduced frequency of hypoxic cycles (20 events/h vs. 60 events/h in Almendros and colleagues [19, 20]). Of note, no increases in melanoma tumor weight were detected when the rate of IH was 10 events per hour, approaching that of mild OSA. In addition to increased tumor growth rates, IH also appears to enhance melanoma metastasis to the lung. Almendros and colleagues (22) reported that both spontaneous metastasis to the lung from the primary tumor and induced lung metastasis were significantly increased in mice subjected to IH compared with corresponding normoxic control animals (Figure 2). Similarly, Li and colleagues (23) have recently shown that IH significantly increased both the number and the weight of mouse melanoma metastatic colonies in the lung. In terms of spontaneous tumorigenesis, a recent study has shown that chronic IH in old mice was able to induce spontaneously occurring tumors in skin, liver, and lung (24).

Figure 2.

Melanoma metastasis in mice. (A) Spontaneous (left) and induced (right) metastasis models in animals subjected to intermittent hypoxia mimicking obstructive sleep apnea and in normoxic control animals. (B) Total number of melanoma metastases per lung area in the spontaneous metastasis model. (C) Total number of lung metastases in the induced metastasis model. Reprinted by permission from Reference 22.

The experimental research performed so far linking melanoma and OSA has provided some clues on the potential mechanisms involved (25). In particular, activation of the immune system seems to also boost melanoma progression (26), because Almendros and colleagues (27) showed that IH increased the proliferation of these cells in coculture with macrophages. Furthermore, IH has been widely associated with increased reactive oxygen species production in several tissues. Considering that the IH pattern applied to mice produced tissue oxygen tension swings from approximately 45 mm Hg to approximately 5 mm Hg in melanoma tumors (22), it is plausible that reactive oxygen species will play a relevant role in IH-induced malignancy. Interestingly, and according to this hypothesis, the treatment of mice with an antioxidant drug was able to ameliorate IH-induced melanoma lung metastasis (23). Although the evidence linking OSA and melanoma as derived from animal models appears to be consistent and reproducible, several methodological questions remain to be explored. Indeed, only one mouse strain (C57BL6) and one melanoma cell type (B16F10) have been employed. This problem is probably relatively minor, because the basic response mechanisms to IH are not anticipated to markedly differ from one mouse strain to another or from one melanoma cell line to the next. This assumption is further supported by data in BALB/c nu/nu mice bearing A-07 human melanoma xenografts that were subjected to IH exposures which also differed from the aforementioned profiles (i.e., cycles of 10 min at 8% O₂ followed by 10 min of normoxia) and also were far removed from the typical oxygenation profiles seen in patients with OSA (28).

Interestingly, the hypoxic profile is an important issue to be considered in IH-induced cancer malignancy according to available studies performed to date (29). Olbryt and colleagues characterized the global gene expression profiling in melanoma cells exposed to either sustained hypoxia or IH; they reported specific genes expressed only in IH exposures; and they further uncovered that HIF-1α (hypoxia-inducible factor 1α) is activated strongly by sustained hypoxia compared with IH (30). The differential effects of IH and sustained hypoxia on cancer malignancy have also been reported in the context of breast cancer (31) and lung cancer (32) studies. However, the gene expression profile triggered by IH and sustained hypoxia in melanoma cells compared with ovarian and prostate cancer was also strongly cell dependent (29), indicating that the increased malignancy observed in melanoma cannot be directly translated to other types of cancers. Notwithstanding this, in the context of the multitude of melanoma-associated mutations and their roles in the biological properties of the malignancy, it would be necessary to systematically explore at least the epidemiologically dominant melanoma cell lines to either corroborate or dispel this notion. In addition, another relevant experimental limitation is that the melanoma–OSA relationship has been studied by challenging mice with IH exposures only, with no data collected on the effects of sleep fragmentation or episodic hypercapnia, two other major hallmark characteristic perturbations in OSA. In fact, it is plausible that melanoma could also be affected by sleep fragmentation, as previously observed in a lung cancer model (16). Finally, in-depth exploration of whether IH promotes the occurrence of melanoma in genetically susceptible animal models and whether the unique biological properties of such naturally occurring tumors are altered by IH, including their responses to therapy, has yet to be undertaken, thereby seriously restricting the extrapolation of currently available findings to the clinical setting.

Biomarkers and Histologic Studies

Potential biomarkers of tumor aggressiveness have been identified, consisting of both histological and systemic markers. The contribution of HIF-1α to the aggressiveness of melanoma in patients with OSA was confirmed by the significant association of HIF-1α expression in tumor cells with the desaturation index (33) and with different indicators of tumor aggressiveness, such as ulcer rate, mitotic index, Breslow depth, sentinel lymph node involvement, or locoregional or metastatic extension (34). In turn, because many HIF-1α target genes are involved in the regulation of cellular bioenergetics and angiogenesis, their contribution to the aggressiveness of melanoma in patients with OSA has also been explored. Among these, VEGF (vascular endothelial growth factor) plays a recognized role in the promotion of tumor growth and metastasis by inducing angiogenesis (35, 36). The analysis of postexcisional melanoma tumor tissues has shown elevated VEGF expression in most tumor cells (33, 34); surprisingly, however, no significant relationship was identified with the apnea–hypopnea index (37), hypoxemia variables (37), or indicators of melanoma aggressiveness (37). These findings are overall concordant with previous studies in which the relationship between VEGF expression and malignancy characteristics was inconsistent (38, 39), and they suggest an apparent discrepancy between the consequences of IH on the expression of HIF-1α and VEGF at the tumor level. Therefore, more detailed studies are required.

Among the systemic biomarkers of IH-related aggressiveness, higher VCAM-1 (vascular cell adhesion molecule 1) plasma concentrations are found in patients with melanoma who have OSA than in those without OSA (40). Although the VCAM-1 relationship with melanoma growth rate is not statistically significant, VCAM-1 could be of interest because it plays various roles in tumor growth, formation of metastatic niches, and support of the angiogenic process (40).

The analysis of circulating immune system activity is another interesting approach to identifying subjects at risk, because the evasion of immune surveillance is an essential pathway for the development and progression of cancer. In animal models, IH has been shown to favor the polarization of tumor-associated macrophages toward a protumoral phenotype (M2), with a lower phagocytic capability of tumor cells, which facilitates their proliferation (27). Patients with OSA also exhibit a higher percentage of M2 polarization, together with a defect in the maturation of natural killer cells induced by HIF-1α, which limits their ability to lyse tumor cells (41). Interestingly, these alterations revert after 6 months of continuous positive airway pressure (CPAP) treatment (41). Accordingly, further studies are needed to determine whether tumor-associated macrophage characteristics can provide more precise indicators of tumorigenesis or tumor aggressiveness.

Over the last decade, immune checkpoint blockade therapy has revolutionized melanoma treatment (42). Its main target is the system formed by PD-1 (programmed cell death protein 1) and its ligand PD-L1 (PD-1/PD-L1), which play major roles in melanoma development (43), aggressiveness (44), and survival (45). Given that HIF-1α upregulation induces expression of the PD-1/PD-L1 pathway in patients with severe OSA, compromising the proliferation capacity and cytotoxicity of their T cells (46, 47), it has been proposed that soluble PD-L1 plasma concentrations could be a potential biomarker of melanoma aggressiveness and metastasis. A recent study showed that soluble PD-L1 plasma concentrations are higher in patients with OSA with melanoma; moreover, in these patients, these concentrations were related to several aggressiveness markers and sentinel lymph node involvement (48).

Undoubtedly, improved understanding of the pathogenic pathways involved in the OSA–melanoma relationship should enable delineation of additional biomarkers, providing not only clues to the consequences of IH but also inferences on the impact of sleep fragmentation and the combined effects on subdermal adipose tissue and malignancy. Finally, the impact of CPAP treatment on potential biomarkers of tumor aggressiveness has not been fully elucidated. An effect of CPAP treatment on gene pathways related to tumorigenesis has been reported (41, 49). Nevertheless, although there are data suggesting an association between OSA and cancer risk, there is a lack of evidence on the possible positive effect of CPAP treatment as far as decreasing tumor aggressiveness in patients with OSA.

Clinical and Epidemiological Studies in Humans

Despite the interest in the association between OSA and cancer, human studies are scarce, particularly those specifically designed to address the relationship between OSA and specific cancer sites. Two retrospective cohorts using large administrative databases reported that OSA was independently associated with a higher risk of incident melanoma but not with poorer melanoma outcomes (50, 51). Gozal and colleagues (50) analyzed a cohort of 5.6 million individuals and found that those with a diagnosis of OSA were 1.14 (95% confidence interval [CI], 1.10–1.18; P < 0.0001) times more likely to develop melanoma than the non-OSA group after follow-up of 4 years. In patients with melanoma, however, the presence of OSA was not associated with a higher mortality risk; rather, it was associated with a lower risk of metastatic spreading. Sillah and colleagues (51) analyzed a cohort of 34,402 patients with OSA and linked it to a population-based cancer registry. After a follow-up of 5.3 ± 2.0 years, a significantly elevated incidence was observed for melanoma (incidence ratio, 1.71; 95% CI, 1.42–2.03), with similar risks in men and women. Finally, a study that evaluated three prospective U.S. cohorts during 2,301,445 person-years of follow-up did not observe any relationship between sleep duration and the risk of melanoma, although OSA was not properly investigated (52). Although these epidemiological studies suggest an association of OSA with greater melanoma incidence, but not with poorer outcomes, all of these studies are either retrospective, rely on administrative databases, were not originally designed to address this association, or did not adequately control for confounders such as obesity, thereby precluding any definitive conclusions on this relationship.

To clarify whether OSA has any influence on the aggressiveness and prognosis of melanoma, the Spanish Sleep Network conducted one large multicenter study specifically designed to address these issues (53). This study prospectively enrolled 443 newly diagnosed patients with melanoma who underwent respiratory polygraphy. Compared with patients in the lower tertiles of apnea–hypopnea index and 4% oxygen desaturation index, those in the upper tertiles were 1.94 (95% CI, 1.14–3.32) and 1.93 (95% CI, 1.14–3.26) times more likely, respectively, to present with aggressive melanoma as defined by a Breslow depth >1 mm, after multiple adjustments. This association was replicated for other markers of aggressiveness, such as ulceration, high mitotic index, sentinel lymph node affected, locoregional extension, and metastatic spreading, supporting the concept that the severity of OSA was independently associated with the aggressiveness of the melanoma. Most of these findings were previously suggested by a small pilot study that preceded the larger multicenter study (54). Currently, this large cohort of patients with melanoma is being followed to evaluate whether this association between OSA severity and greater melanoma aggressiveness translates into a poorer prognosis of the tumor over time.

Whether the positive relationship observed between OSA and melanoma in epidemiological and clinical studies also occurs for other cancer sites or tumor types is unknown, because well-designed studies are lacking. For example, a very recent study designed to investigate breast cancer did not find any significant association between the presence or severity of OSA and several markers of breast cancer aggressiveness (55).

Role of Main Confounders: Obesity and Age

More than 80% of patients with OSA are overweight, and more than half are obese (56). There is now sufficient evidence supporting an association between overweight and 11 types of cancer (for both incidence and aggressiveness) (57), including melanoma. Moreover, there is an increase in the global prevalence of OSA and cancer with age (2). Therefore, overweight/obesity and age are probably the most important confounders in any analysis aiming to establish the validity of a relationship between OSA and cancer, or more specifically melanoma, and therefore adjustment for these confounders is essential.

Of note, Almendros and colleagues (20) showed that both IH and obesity increased melanoma growth patterns in mice, but they did not identify a synergistic effect between IH and obesity. More specifically, IH-mimicking OSA increased the growth rate of melanoma in lean mice, but not in obese mice.

Regarding age as a confounder, the relationship between OSA and melanoma aggressiveness was stronger in patients younger than 56 years old than in older subjects in a large multicenter Spanish study (53). This association between OSA and cancer in younger rather than in older subjects has previously been reported in other clinical studies (58, 59). In agreement with these data (60), IH-mimicking OSA enhanced tumor progression in young but not in aged female mice. This finding was attributable to changes in the recruitment and function of tumor-associated macrophages and impaired immune responses in aged female mice. Furthermore, Cubillos-Zapata and colleagues reported that PD-L1 upregulation in patients with OSA as a consequence of HIF-1α activation occurs mainly in young patients and is lacking in older patients, probably owing to impaired oxygen sensing (48).

Future Challenges

Notwithstanding the aforementioned evidence supporting a link between OSA and melanoma characteristics, it is clear that such association has only been very preliminarily explored and that there are more questions arising than answers. Further cellular and animal studies are needed to investigate the potential roles of other components of OSA, such as intermittent hypercapnia and sleep fragmentation. Although animal models have consistently shown that IH enhances melanoma progression, there is no evidence of its effect on the risk of developing melanoma.

In addition, a large number of studies have identified the role of microRNAs (miRNAs) as specific biomarkers in oncogenic processes (61). For example, previous studies have identified miR-17-92 cluster as a novel target for p53-mediated transcriptional repression under hypoxia (62). Nevertheless, as far as we know, the role of miRNAs in oncogenic processes in patients with sleep apnea has not been specifically explored. Therefore, investigation of the role of circulating miRNAs as both biomarkers (63) and causal effectors in the OSA–melanoma relationship may unravel potential novel therapeutic targets and a more personalized approach.

Regarding clinical studies, longitudinal assessments are needed to determine whether this association between OSA and greater aggressiveness of the tumor eventually has any influence on melanoma outcomes, such as relapse, metastatic dissemination, and death, as well as whether OSA treatment may have any beneficial effects on these outcomes, in light of the CPAP-induced reductions in cancer-related gene expression in circulating leukocytes (49). Finally, we are fully aware that the vast majority of the available evidence linking OSA with melanoma has originated from the collaborative work of the authors. Undoubtedly, additional well-designed epidemiological studies will be necessary to investigate whether the presence or severity of OSA may be a risk factor for development of melanoma.

Conclusions

The relationship between OSA and cancer has been a subject of study in recent years, as well as within the various types of cancer. The existence of pathophysiological pathways related to IH, sleep fragmentation, and immune dysregulation that may underlie the relationship between OSA and melanoma agrees with most of the animal and epidemiological findings, suggesting that the presence and severity of OSA could be associated with faster tumor growth and greater invasiveness. However, although the current data are insufficient to conclusively establish the validity and causal relationship between OSA and melanoma, given the epidemiological importance of both diseases and the current level of evidence, further and intense exploration of the role of OSA in cancer appears justified.