Our study found a significant association between the degree
Our study found a significant association between the degree of primary tumor lobulation and EGFR mutations status (P < 0.001). Shallow lobulation was present in 62.62% of EGFR+ primary tumors but only in 21.28% of primary tumors with wild-type EGFR. Deep lobulation was present in 63.83% of tumors with wild-type EGFR but only in 29.91% of EGFR+ tumors. Several studies attempted to describe the relationship between CT-features and EGFR salvinorin a status. However, only a small portion of these previous studies included the evaluation of lobulation, and almost all of them reported no association between EGFR mutations and lobulation, which was simply described as presence or absence of lobulation [, , ]. In our study, only one patient lacked lobulation based on a presence/absence interpretation; thus, based on this qualitative assessment, there was no association between EGFR mutation status and lobulation, which is concordant with previous studies. However, when we graded lobulation (by defining the edge types of the “lobulated” class), EGFR + patients had significantly higher percentages of shallow lobulation and lower percentages of deep lobulation. To the best of our knowledge, no previous study used graded lobulation to evaluate the association between lobulation and EGFR mutations. Hasegawa et al. found a significant association between EGFR mutations and notch on multivariate analysis (p = 0.0428), although there was no statistically significant difference based on univariate analysis (p = 0.9722) (the “notch sign” is a particular form of lobulation defined as an abrupt bulging of the lesion contour) . Some studies equated the notch with lobulation [, , ]; thus, our results are not comparable with these studies. Hsu et al. reported that an irregular shape of primary tumor was more common in adenocarcinomas with wild-type EGFR than in those with EGFR mutations . We postulate that the irregular shape may be similar to or reflect the sign of deep lobulation. It has been observed that cancer patients with EGFR gene mutations have longer survival times than those without EGFR gene mutations [28,29], which agrees with our results. The lobulation sign has the appearance of uneven arcs in the edge of tumors, which is attributable to different or uneven growth rates and is strongly associated with malignancy and invasiveness [, , ]. Relative to shallow lobulation, deep lobulation indicates greater malignancy and invasiveness and therefore reflects a worse prognosis. Our study demonstrated that smaller lymph node size and lower percentages of partial fusion of lymph nodes were significantly associated with EGFR-mutation positive adenocarcinoma, based on both univariate and multivariate analyses. However, when lymph nodes were described as presence or absence of swollen lymph nodes, there was no association between EGFR mutations and lymph nodes. To the best of our knowledge, this finding has not been previously reported. However, Guan et al. reported that EGFR+ patients had an earlier N stage than the group with wild-type EGFR based on PET/CT scans , which may be similar to our findings. This observation also reflects the lower malignancy and invasiveness of tumors in the EGFR+ group as compared with the group with wild-type EGFR. The mechanism of this finding was uncertain and its value on clinical application needs to be further investigated. Previous studies demonstrated that cancer patients with EGFR mutations, including NSCLC and thyroid carcinoma, often developed multiple pulmonary metastases, particularly miliary metastases [, , ]. Our study showed that EGFR mutations were associated with the presence of pulmonary and/or pleural metastatic nodules (P = 0.006) and with small metastatic nodule sizes (P = 0.021) based on univariate analysis, but subsequent multivariate analysis completely negated these associations. No significant associations between EGFR mutations and the distribution of metastatic nodules in lungs and pleura were observed in this study. Guan et al. reported that EGFR mutations in patients with NSCLC are associated with a higher incidence of distant brain and bone metastases but are not associated with distant liver and adrenals metastases . In contrast to the study by Guan et al., our study showed no significant association between EGFR mutations and organs that are common sites of distant metastasis, such as bone, brain, adrenals, and liver. However, it is worth noting that the subjects in our study all had advanced pulmonary adenocarcinoma (stage IIIB, IIIC, or IV) whereas the Guan et al. study included patients with all stages of NSCLC (stage I-IV), which may account for the observed differences between our study and the Guan et al. study.