Lung carcinoma is a leading cause of cancer related death worldwide. Despite the growing body of evidence in research, the molecular mechanisms of lung cancer chemoresistance remains elusive [4]. Thus, accurate biomarker assessment is critical for individually tailored disease management. In the current study, we assessed the clinical prognostic and predictive values of ROS1, ALK, PD-1, and PD-L1 protein expressions in NSCLC using IHC along with ROS1 gene rearrangement using FISH. Our major finding showed that positive ALK expression was significantly associated with poor treatment response and shorter PFS in NSCLC patients. Moreover, both the EGFR mutation and ALK expression were independent predictors for NSCLC.
Since IHC staining for protein expression has the advantages of being less costly and higher in accessibility [18], we examined the of ROS1, ALK, PD1 and PD-L1 proteins expression using IHC in relation to clinic-pathological features of the patients, EGFR mutations, survival rates, and response to treatment in a cohort of Egyptian NSCLC patients who were given standard platinum-doublet chemotherapy as a front-line treatment.
Our study revealed that ROS1, ALK, PD-1, and PD-L1 protein expression levels were 18.57%, 54.29%, 84.29%, and 87.14%, respectively, in NSCLC patients. In 15.71% of patients, a ROS1 gene rearrangement was found. The current study showed a high level of PD-1/PD-L1 expression along with a low level of ROS1 expression. This is consistent with previously published data. In Mahoney and Atkins’ study, the prevalence of PD-L1 protein expression in NSCLC ranged from 24 to 60% [19].
The increased expression of PD-1 and PD-L1 in our current cohort of NSCLC patients compared to previous studies may be explained by the fact that 100% of our study cohort had AC, 61.4% had advanced disease stage (III–IV), and 72.9% were male. Since a previous study by Sorensen, Zhou [20], demonstrated that PD-L1 expression was present in 75% of patients with advanced NSCLC and that higher levels of PD-1 expression were observed in male smokers with adenocarcinoma [21], this could confirm the higher levels of PD-1/PD-L1 expression observed in our study.
In the present study, no significant association was found between PD-1/PD-L1 expression and any of the studied clinicopathological characteristics which was consistent with the previous study Furthermore, they have discovered non-significant relationship between PD-L1 expression and relevant clinicopathological features of the patients as described in our study. Other studies have found that PD-L1 is highly expressed in male smokers with high histologic grade, positive lymph node metastasis, and advanced stage [18, 22].
Till now, the prognostic significance of PD-L1 is still obscure. PD-L1 expression has been associated with a favorable prognosis, a poor prognosis, or no prognostic significance in a variety of studies [23]. The current use of non-standardized IHC techniques to determine the levels of PD-L1 protein in tissue may explain some of these discrepancies. Additionally, PD-L1 expression may vary between lung carcinoma cohorts due to the presence of histological subtypes and different patient selection criteria [24].
We found no significant association between PD-1/PD-L1 expression and patient PFS or response to chemotherapy. This could be explained by the fact that all the patients in our study had been diagnosed with adenocarcinoma on histological examination. The adjusted HRs for PD-1 and PD-L1 positive groups were 1.85 (95% CI: 0.81–4.26; median PFS, 15 months, p = 0.24) and 1.11 (95% CI: 0.40–3.08; median PFS, 17 months, p = 0.31), respectively, when compared to PD-1/PD-L1 negative groups. Additionally, a previous study found a stronger correlation between PD-L1 expression and survival outcome in patients with squamous cell carcinoma compared to adenocarcinoma [20]. Consistent with the current findings, a previous meta-analysis study involving 64 patients found no statistically significant relationship between PD-L1 expression and survival outcome [25]. On the contrary, other meta-analyses have discovered a strong association between PD-L1 overexpression and poor survival outcomes [26].
A complex relationship may exist between oncogenic driver mutations and PD-1/PD-L1 expression. It has been reported that positive PD-L1 expression was not associated with major oncogenic driver mutations like EGFR, ALK, KRAS, and BRAF in NSCLC patients [27]. However, a study by D’Incecco, Andreozzi [21], has linked PD-1/PD-L1 expression to EGFR and KRAS mutations. In the current study, there was no relationship between PD-1 or PD-L1 expression and EGFR mutation, ROS1, or ALK positive expression. Differences in the used staining antibodies, scoring criteria, oncogene analyzed, and mutation rates among patients of different ethnicities may explain these contradictory results. Due to the small number of mutation positive cases, more research is required to confirm our preliminary findings.
The oncogenic ROS1 protein is overexpressed on the tumor cells in most malignant tumors tougher with ROS1 fusions. ROS1 rearrangement has been identified as a druggable target in the NSCLC patients, among others [28]. However, the prognostic value of ROS1 protein in NSCLC and its potential therapeutic significance have received relatively little attention until now. The frequency of ROS1 protein expression and ROS1 rearrangement in the current study were 18.57% and 15.71%, respectively, which is higher in comparison to other previous studies. ROS1 rearrangements were found in 0.9–1.7% of NSCLC patients in the study by Chen, Hsieh [29]. This could be partly explained by the fact that ROS1 fusions are common in adenocarcinoma [30], since all of the patients in this study had NSCLC adenocarcinoma.
ROS1 kinase overexpression has also been discovered in primary and recurrent NSCLC, and it has a potential role as an independent prognostic factor for survival in adenocarcinoma NSCLC cases [31]. In contrast, we found no significant association between ROS1 protein expression and survival outcomes of NSCLC patients. Most studies in literatures have shown that ROS1 fusions are mutually exclusive in relation to EGFR, KRAS mutations, or ALK fusions [32, 33]. In concordance with these studies, using spearman correlation, we found no significant correlation between ROS1 protein expression and EGFR mutations or ALK protein expression suggesting that ROS1-positive NSCLC patients do not benefit from EGFR tyrosine kinase inhibitor therapy. Furthermore, we found no significant relationship between ROS1 expression and gender, smoking status, or tumor stage. The data in this context contrasts to previous findings which show that ROS1 rearrangement was common in female patients, nonsmokers, and patients with advanced disease stage and triple wild-type EGFR/KRAS/ALK genotype [32]. Based on this data, the clinicopathological features of NSCLC patients with positive ROS1 expression differ across ethnicities [34].
Thus, patient choice, which strongly depends on the clinicopathological features, would be ineffective for detecting ROS1 genetic alterations because the incidence of ROS1 alterations in NSCLCs is extremely low. Therefore, the molecular tests cannot be used on all patients and an effective screening method is highly required. Since ROS1 rearrangement occurs at a rate of 0.5 to 2% in NSCLC, IHC appears to be a cost-effective screening assay, allowing for rapid results at a lower cost. Since FISH using dual color “break-apart” probes is the “gold standard” for detecting ROS1 gene rearrangement [35], we used spearman correlation to examine the correlation between FISH and IHC approaches for detecting ROS1 rearrangement. We discovered a strong positive correlation between ROS1 protein expression and ROS1 gene rearrangement (rho = 0.702, p < 0.001). In consistent with our findings, a study, conducted by Shan, Lian [35], compared IHC to FISH and real-time RT-PCR in 60 lung adenocarcinomas, including 16 with ROS1 protein expression. Their results indicated that 75–100% of patients with tumors having positive IHC scores were also FISH positive. Consequently, IHC could be considered as an effective and convenient technique for detection of ROS1 rearrangement in NSCLC rather than the expensive, more difficult, time-consuming, and high expertise demanding molecular tests. However, further studies are needed to compare between the sensitivity and specificity of the two techniques.
Despite the fact that patients with EGFR mutations and ALK rearrangements shared several clinicopathological features, such as never or light smoking status and adenocarcinoma histology, a previous meta-analysis revealed that ALK gene rearrangement was mutually exclusive for EGFR mutations, implying a distinct genetic subtype of lung adenocarcinoma [36]. In line with this finding, we found no significant correlation between EGFR mutations and positive ALK expression, implying that ALK expression is mutually exclusive of EGFR. Furthermore, we did not find any relation between ALK expression and gender or smoking status. Since all the NSCLC patients assessed here were adenocarcinoma, we were unable to detect the relationship between ALK expression and histological subtype. However, we found that positive ALK expression was significantly associated with advanced T stage and left-sided tumors.
We also found that EGFR mutations and positive ALK expression were significantly associated with shorter PFS in NSCLC patients. Furthermore, positive LN metastasis, EGFR mutations, and ALK expression were independent predictors for PFS in NSCLC patients in a multivariate survival analysis. A multi-variate analysis study by Ito, Miyata [37], demonstrated poor survival outcome and increased rate of metastatic recurrence in EGFR-mutated NSCLC patients which is consistent with our findings.
The findings of our study should open the path for future prospective studies of the predictive and prognostic role of PD1/PD-L1 expression along with ROS1 and ALK as measured by immunohistochemistry in patients with NSCLC treated with chemotherapy. Such studies could yield interesting results, particularly regarding the efficacy of immunotherapy and/or receptor tyrosine kinase (RTK) inhibitors in combination with chemotherapy. Because the immune system is so dynamic, future studies should include repeat biopsies during treatment and at progression, as well as sequential analysis of circulating biomarkers.
This study was limited in that it included only adenocarcinoma NSCLC patients, the most common histopathological subtype among Egyptian population. Our study will also open the way for future studies using molecular technique (RNA and/or DNA) to validate the IHC results.