**4. Discussion**

In cases of multiple lung cancers, clinical differentiation between primary and metastatic tumors can be difficult, rendering treatment selection challenging. Furthermore, in patients with multiple lung cancers metastasized to the lymph nodes or distal sites, the focus of treatment varies depending on the cancer that has metastasized. Thus, determining the origin of the metastasizing cancer is clinically important. Therefore, we performed lung cancer mutation analysis through targeted deep sequencing and demonstrated that mutations of individual lung cancers are able to provide clonal markers, enabling discrimination of the clonal origin of multiple lung cancers and their metastases.

The consistency of mutations across multiple sites, with complete concordance in the position and patterns of base-pair substitutions or indels, cannot be a coincidental phenomenon. Although discordance between two tumors was noted in mutations with an allele fraction <20%, this can be interpreted as tumor heterogeneity [14]. In general, cancers comprise populations of cells with various molecular and phenotypic features, a phenomenon termed intratumor heterogeneity [14,15]. This may bolster tumor adaptation, cancer progression and metastasis, and/or therapeutic failure through negative selection [14,16]. Conversely, a driver mutation triggers clonal expansion and is retained ubiquitously within the tumors of the same clone [16,17]. These theories can be interpreted as the "trunk and branch" mutation models; early somatic events that drive tumor progress in early clonal founders are represented by the "trunk" of the tumor [18,19]. Such trunk somatic mutations to be found at the early stages of tumor development are ubiquitous events occurring at all sites of disease. Meanwhile, later somatic events that occur in the wake of branched separation of subclones represent heterogeneity. Such subclonal heterogeneity may be spatially divided among regions of the same tumor or its metastatic sites [18–20]. In this context, clonally dominant mutations are important clonal markers. Primary and metastatic tumors can be differentiated by determining whether such ubiquitous driver mutations are consistent.

It is relatively straightforward to diagnose multicentric primary lung cancers of different histological type. However, it is often difficult to differentiate between multiple primary lung cancers and intrapulmonary metastases having the same histological type. In particular, in cases of multiple tumors classified as squamous cell carcinoma (such as cases C–F), differentiation based on pathological features alone is extremely difficult. Even when the morphological and immunohistological features are non-homogeneous among different parts of the tumors (e.g., cases A and B), the driver mutation is ubiquitously retained within the tumors of the same clone [16,17]. Therefore, distinction of clonality on the basis of mutation analysis is more specific and definitive than histological examination.

Detterbeck et al. reviewed the clinical and pathological criteria to distinguish second primary tumors from metastatic tumors [21]. They reported that it is impracticable to define criteria that conclusively establish the identical nature of tumors; merely finding observable similarities between tumors is insufficient. Using the method described, comprehensive mutation analysis is initially performed to identify the driver mutations in each cancer, which are subsequently compared to define their clonal origin. These criteria are definitive and reliable. Moreover, the decision criteria are generally clear and intuitive. In fact, this method yielded clear genetic diagnosis in all patients. In other words, no equivocal or ambiguous diagnosis was obtained in any of the cases. In our previous study, we had demonstrated that this method allows bronchoscopic biopsy samples and other small samples to be used for discrimination between primary and metastatic tumors [1]. Thus, our method may enable both flexible and rational decision-making based on accurate diagnosis. For example, a preoperative diagnosis of metastatic tumors may make it possible to avoid surgery, whereas a preoperative diagnosis of primary tumors would lead to surgical treatment. This novel approach may help resolve the dilemma of misdiagnosis in the clinical setting. Thus, we anticipate that it will come to be utilized as a standard diagnostic approach in daily clinical practice in the near future.

When selecting treatment methods for multiple lung cancers, it is necessary to consider the cancer type will markedly affect the prognosis. In cases D–F that have lymph node metastasis, a factor responsible for progression to an advanced stage was identified in the two tumors. Furthermore, the tumors exhibit different mutation profiles and PD-L1 staining properties. Therefore, the lesions targeted for treatment and the options selected for subsequent treatment (e.g., molecular-targeted drugs and immune checkpoint inhibitors) vary depending on the type of tumor that has metastasized to the lymph nodes. This suggests that accurate understanding of the pathology gained by performing a genetic diagnosis can exert a powerful effect on the clinical outcome. Although the use of immunotherapy has revolutionized the treatment of non-small-cell lung cancer, patterns of immunostaining with PD-L1, a biomarker for treatment response, may vary in tumor cells across individual primary tumors (e.g., (,) cases E and F) [22]. At present, molecularly targeted therapies are also rapidly evolving. The development of novel molecularly targeted therapies would enable the treatment to be specifically tailored to the features of mutations detected in individual cancers. Thus, in patients with multiple lung cancers, performing a mutation analysis helps select the medical treatment most likely to be effective.
