The use of non-steroid anti-inflammatory drugs during radical resection correlated with the outcome in non-small cell lung cancer

Lung cancer is still a great health threat worldwide [1] with non-small cell lung cancer (NSCLC) which accounts as the predominant pattern [2]. In recent years, with the success of neoadjuvant, adjuvant immunotherapies [3, 4], and target therapies [5, 6], the overall survival (OS) of NSCLC patients was greatly improved even in those locally advanced cases. Nonetheless, radical resection was still the most important treatment for the majority of the patients; however, some complications like pain are still a serious problem to harm the quality of life for the patients after surgery.

It was reported the prevalence of clinically relevant postoperative pain for lung cancer patients could be up to 63% for those received thoracotomy [7] and could persist to 36 months (m) in 17.4% patients [8]. Clinically, opioids and non-steroid anti-inflammatory drugs (NSAIDs) are the most used agents to alleviate the postoperative pain in these patients [9], the latter of which include those unselected COX inhibitors like aspirin, diclofenac, and ibuprofen and those selected COX-2 inhibitors like rofecoxib and celecoxib [10]. Interestingly, a great number of previous studies have indicated some of these NSAIDs may have additional role in regulating lung cancer cells besides its function in anti-inflammation and alleviate pain. For example, aspirin could improve cisplatin resistance by inhibiting cancer cell stemness [11] or reduce the metastasis of the cancer cells to regional lymph nodes [12]; ibuprofen could enhance the effect of cisplatin by suppressing the heat shock protein 70 in cancer cells [13], which indicated a positive role of these agents in anti-cancer. On the contrary, celecoxib could induce the epithelial-mesenchymal transition and increase the risks of cancer metastasis [14]. Based on these facts, some clinical observations have conducted to explore the role of the NSAIDs in lung cancer patients as it was found that the combined use of aspirin with osimertinib [15] or immunotherapies [16] contributes to the good survival in patients; however, no such role was detected for the combination with celecoxib [17, 18]. Of note, all these trials were performed in advanced or metastatic settings. Up to date, only one study has explored the use of the NSAIDs in postoperative stages I–III NSCLC patients [19], but it only includes the use of indomethacin and ibuprofen which aimed to alleviate the postoperative fever; other types of NSAIDs were not included. Nonetheless, none of the study has explored the prognostic value of the use of NSAIDs during surgery in these patients.

In this study, we sought to explore the prognostic value of the use of NSAIDs during radical resection in stages I–III NSCLC.

In this study, the use of NSAIDs during radical resection was considered to be correlated with the outcome in NSCLC patients. UI was found to be the most robust prognostic indicator for the use of NSAIDs in these patients, and UI high patients would have the superior outcome in contrast to the low ones. Moreover, UI was recognized as an independent risk factor for DFS.

Previously, a great number of epidemiological investigations have indicated the use of NSAIDs like aspirin and ibuprofen can reduce the incidence of cancers including lung cancer [21‐24]. As in cancer patients, these agents also maintained a positive role in the patients’ outcome. For example, Giampieri et al. in a study with 66 previously heavily treated metastatic colorectal cancer found that aspirin could obviously improve the survival [25]. With regard to lung cancer, Chuang et al. in a large retrospective study with 38,842 inoperable NSCLC patients found the use of aspirin correlated with improved OS [26]; Kanda et al. in a study with 217 stages IIIB and IV NSCLC patients also found the use of loxoprofen sodium that could extend the survival in older NSCLC patients [27]. In recent years, with the popular target and immunotherapies in NSCLC, the value of the use of NSAIDs was also explored with these therapies. For example, Liu et al. in a study with 365 metastatic EGFR-mutant NSCLC patients received osimertinib ± aspirin and found that a combination of aspirin with osimertinib could improve the progression-free survival (PFS) and OS [15]; similarly, Aiad et al. in a study with 500 stages I–IV patients received immunotherapies ± aspirin and also found that the combination of aspirin could extend the outcome [16]. However, it was also notable that some studies suggested that use of celecoxib could have less positive role in combination with gefitinib [17], docetaxel, or other platinum-based chemotherapy [18, 28]. Nonetheless, all these studies were conducted in advance or metastatic staged settings. Up to date, only one study has explored the use of the NSAIDs in postoperative stages I–III NSCLC patients, and the results indicated the use of the NSAIDs correlated with good PFS and OS [19]. In addition, the use or not (corresponding to the US in our study) was found to be an independent risk factor both for PFS and OS [19]. However, this study only includes the use of indomethacin and ibuprofen which aimed to treat the postoperative fever; other types of NSAIDs or other intentions of the use of these agents were not involved. Indeed, some previous studies in colorectal cancer and breast cancer have validated the protective role of aspirin in adjuvant settings [29, 30]. Our study for the first time supported the protective role of the use of NSAIDs in addition to indomethacin and ibuprofen in NSCLC patients in postoperative scenario, which was partly in line with previous studies in adjuvant settings [19, 29, 30]. In addition, none of the patients received celecoxib, which is also in line with previous results with these agents in aforementioned advanced or metastatic cases.

It continued to be lacking of well-acknowledged definition concerning the use and use intensity of NSAIDs in cancer patients particularly in postoperative background. Previously, some studies have defined the aspirin or other NSAIDs exposure in cancer patients. For example, Liao et al. defined the use of aspirin as one or more prescriptions recorded before and after the diagnosis with any dosage [31], whereas in Giampieri et al.’s study, the definition of exposure was the ones taken for at least 12 months at a dose of at least 100 mg/day [25]; other studies defined various criteria of exposure of these agents [30, 32]. As in lung cancer, Chuang et al. defined the aspirin users as those who used it for > 28 defined daily doses after diagnosis [26]; Jiang et al. defined the use or not of NSAIDs (only indomethacin and ibuprofen) as the standardized dosage in clinic [19], whereas in reports about the combination of it with osimertinib or immunotherapies, the definition of exposure was obscure [15, 16]. Interestingly, all these studies are conducted retrospectively. However, in prospective clinical trials about celecoxib, its exposure was rigorously defined with fixed dosage concurrent with gefitinib [17] or chemotherapy [18]. In our study, all the patients received daily dose of the NSAIDs irrespective of the categories, and we thus refer to Jiang et al.’s and Liao et al.’s study [19, 31] to define the exposure as US; in addition, we also explored the prognostic value of AD and UI in these patients. The results indicated UI displayed the highest AUC among these indexes and was found to be an independent risk factor for survival. Our results for the first time indicated that the UI may a reasonable index for the use of NSAIDs in these patients; however, more studies are still required to validate our speculation in future.

Mechanically, the role of NSAIDs in regulating lung cancer cells has been under extensive study. For example, aspirin could manipulate the miR-98/WNT1 axis to inhibit cancer progression [33]; it could also suppress the growth of cancer cells via targeting the TAZ/PD-L1 axis [34]; in addition, other NSAIDs like ibuprofen could enhance the effect of cisplatin by suppressing the heat shock protein 70 [13], and acetaminophen could promote ferroptosis by regulating Nrf2/heme oxygenase-1 signaling pathway [35], and loxoprofen sodium could inhibit tumor growth by suppressing vascular endothelial growth factor in a mouse model [36]. In recent years, the key role of circulating tumor cells (CTCs) has been identified in tumor recurrence and metastasis [37, 38]; in particular, some of these cells presented features that are similar to the so-called cancer stem cells (CSCs) in many cancers [39, 40] including lung cancer [41, 42], the latter of which are characterized by high potential of self-renew and multiple treatment resistance, and complete removal of these cells was regarded as the ultimate approach to cure the disease [43‐45]. In lung cancer, it was notable that the CTCs could be found in up to 51.8% stages I–III radically resected patients before surgery [38], and these cells could be even found in 96.5% patients during surgery in pulmonary vein [46]. Notably, the use of some NSAIDs like aspirin could decrease the CTCs numbers in colorectal and breast cancer patients [47]; moreover, aspirin could have a broad spectrum of inhibition for many CSCs (including lung cancer) by complex mechanisms [48, 49]. Based on these facts, it was notable that the use of NSAIDs after surgery could potentially reduce the quantity of the CTCs and in particular eliminate a cluster of these cells featured like CSCs, which could then improve the patients’ outcome. Except these, it was well established that inflammation plays an important role in cancer initiation, recurrence, and metastasis [50]. Some cytokines, like IL-6, could be remarkably elevated in lung cancer patients [51] in particular for those who underwent surgery [52]. It is noteworthy that IL-6 could not only promote the cell proliferation [53], induce treatment resistance [54, 55], and promote metastasis [56] in lung cancer but also promote the expansion of the CSCs [57]. Interestingly, some NSAIDs like aspirin could not only cancel the pro-tumorigenic effects of IL-6 [58] but also downregulate the IL-6-STAT3 signaling pathway to induce cancer cell apoptosis, which have been validated in colorectal cancer [59] and glioblastoma A172 cells [60]. Based on these facts, it was also plausible that the use of NSAIDs may also have a role in interrupting the correlation of inflammation and cancer in NSCLC, which may result in a sound outcome for these patients.

Our study could have some clinic implications. First, taking into consideration the evolution of lung cancer cells during its development [61] and the role of NSAIDs in NSCLC, it would be reasonable to take these agents routinely as a tertiary chemoprevention of the disease immediately after surgery; second, since low-dose aspirin (< 75 mg/day) rarely contribute to improve cancer-specific mortality in lung cancer [62] and could even promote the cell growth [63], it was recommend that these patients should take the NSAIDs with a relatively high dose (like UI ≥ 74.55 mg/day in our study); however, adverse effects like gastric erosions and ulcers should also be balanced in clinic [64]. There are also some limitations for present study. First, it was a retrospective study with limited sample, the cases in US no groups were only 77, and potential bias cannot be satisfactorily ruled out; second, the pharmacokinetics and pharmacodynamics of different NSAIDs were not identical, we still lack strong evidence to support the convert ratio as 1 for each other in our study, and there are still lack of relevant studies to full support our explanations for the role of these agents in NSCLC except aspirin; and third, although UI was found to be a robust prognostic index compared to US and AD, the question for the duration of these agents cannot be answered at present. We advocated more studies; in particular, those prospective clinical trials should be carried out to validate our results in future.

As a summary, we found that the use of NSAIDs during radical resection in NSCLC patients correlated with the outcome, and patients with a relative high UI have better outcome.