In our current study, we found that the prognosis of patients with IMA who received palliative chemotherapy was variable in accordance with the type of chemotherapeutic intervention. Patients with IMA treated by immunotherapy appeared to have a better outcome than those who received chemotherapy. Of note in particular, immunotherapy was found to be associated with improved survival in multivariable analysis. On the other hand, targeted therapy did not improve the survival outcomes of the advanced IMA cases in our present series, although the PFS was better in these patients compared with those without targeted therapy. In addition, no OS differences were evident based on conventional agents. We thus contend from our current data that immunotherapy should be considered, if possible, as a principal treatment option in patients with advanced IMA.
In our present analyses, the median survival of patients with IMA was 20.1 months. When compared with the findings of previous reports, our measured OS is similar to that reported by Cha et al of 17.9–20.9 months from patients with metastatic IMA, with no OS differences found between IMA and non-IMA adenocarcinoma cases in that prior study . Several previous clinical trials that compared treatment efficacies reported a median survival of 14.2–21.5 months from the use of conventional chemotherapy [22–24]. On assessing the conventional chemotherapy regimens in our present IMA population, we noted that the combination of pemetrexed or gemcitabine with platinum agents was the most common first line protocol for these treatments. Real-world practices reported in other studies have also shown a similar pattern. In prior reports on non-squamous NSCLC, the most common regimen was carboplatin plus pemetrexed (25.7%) in the US, and pemetrexed-based therapy (68%) in China [25, 26]. Notably, the proportion (about 42%) of patients who received second-line therapies in the American study  and the Chinese study  was lower than that (75.9%) in our present study. In our current cohort, only 13 patients (16.5%) were found to have oncogenic mutations whereas 11 KRAS mutations were detected in a setting involving 12 patients with IMA in previous reports [27, 28]. Considering that the patients in our current cohort had similar clinicopathologic features and treatment environments to those previously reported for IMA, the clinical outcomes of IMA seems to be comparable to those for non-squamous NSCLC.
It has been well established that KRAS mutations are the most frequent genetic alternations seen in IMA . On the other hand, IMAs usually lack other oncogenic mutations such as EGFR variants, or translocations of ALK or ROS1 . KRAS mutations are mutually exclusive from EGFR mutations or ALK/ROS1 rearrangements based on oncogene addiction theory . The molecular mechanisms driving IMA remain unclear, but some reports have suggested possible mechanisms for identifying novel therapeutic targets. A previous animal study has suggested that the induction of the signature genes FOXA3 or SPDEF, which are enriched in mucin-producing cancers, along with a KRAS mutation in the lung epithelium, is sufficient to develop mucinous lung tumors in transgenic mice . Other studies have revealed that a loss of TTF-1 expression owing to an NKX2-1 mutation, which occurs in approximately 19% of IMA cases, would de-suppress the expression of the mucin-related genes MUC5Ac, MUC5B, and MUC3 . In addition, novel driver mutations, NRG1 fusions, have been recurrently identified in IMA, i.e. as a regulator of goblet-cell formation in human bronchial epithelial cells . The NRG1 protein mediates juxtacrine signals through the HER2:HER3 receptors that may play a role in the transformation and acquisition of goblet-cell morphology in IMA . Recently, several clinical trials have been performed on drugs that target KRAS-mutated IMAs. A phase 1 trial of sotorasib, a small molecule that selectively targets the KRAS G12C subtype, revealed an 88.1% disease control response in NSCLC . Driver mutations such as KRAS thus represent promising future therapeutic targets for the treatment of IMA.
The OS outcomes were found in our present series to be significantly improved in patients treated with immunotherapy. Numerous prior studies have reported that immunotherapy produces favorable outcomes in NSCLCs and that the clinical response predictions in these patients depend on the expression of biomarkers such as PD-L1 or on the tumor mutation burden [35–37]. Previous analysis of the expression of PD-L1 in IMA has indicated that PD-L1 positive tumors are infrequent . However, other findings suggest the possibility of a good response to immunotherapy by IMA as KRAS mutations show an association with this response . Although most prior randomized trials have not been designed to examine treatment differences between molecular subgroups, immunotherapy showed a significantly greater benefit for KRAS mutant tumors in one previous meta-analysis . However, another study using real-world data demonstrated that a KRAS mutation did not confer any significant OS difference . An attractive biological explanation for these discrepant findings is the molecular and environmental diversity of KRAS mutation subgroups . Three of these subgroups have been defined in accordance with the presence of co-mutations, and have differences in terms of the immune environment and the responses to immunotherapy. In addition, IMAs have distinct expression profiles of immune checkpoint regulators such as VTCN1, which represents a potential immunotherapy target . Although the biology to support the benefits of immunotherapy in IMA remains unclear, further studies are warranted to address whether key mutations such as KRAS can offer predictive insights into the immunotherapy responses of IMA tumors.
Our present study had several limitations of note. First, it was a retrospective study with a relatively small sample size from a single institution. However, we tried to include patients with IMA having similar characteristics to minimize possible bias and investigated all of the regimens which patients received as palliative chemotherapy. This is also, to our knowledge, the first study to identify the clinical outcomes of patients with IMA who had received immunotherapy. Prospective, randomized studies will be needed to clarify the clinical outcomes of patients with IMA in accordance with their palliative treatments. A second limitation was that we included patients who had been diagnosed prior to the WHO reclassification with mucinous adenocarcinoma, not IMA. Patients with non-IMA could therefore have been included in our cohort. We did however exclude patients with other variants of invasive adenocarcinoma such as colloids who has been diagnosed before 2015, which would have helped to reduce the heterogeneity of our cohort. A third limitation was that targeted therapies did not show survival benefits in our present series, which contrasts with the findings of previous studies . Although we did observe an improved PFS from 1st line targeted therapies among our current patients, these treatments were not found to be associated with the IMA prognosis in our multivariate analysis. The distinct characteristics of EGFR mutations that have been found to have no female predominance and operate in a different tumorigenic pathway in IMA could have impacted on these targeted treatment responses . Finally, PD L-1 expression was not examined in all the patients in our cohort. This marker is used as a predictor of the response to immunotherapy and a future examination of its status in IMA could provide insights into the causality behind the good response of these cancers to immunotherapy.