Patients and samples
A total of ninety-three patients received nivolumab between December 2015 and March 2018. Only adenocarcinoma or squamous cell carcinoma 46 cases with extra cytological slides were selected, considering the exhaustion of formalin-fixed paraffin-embedded blocks for previous companion diagnostics and specimen storage for future clinical trials. During the molecular analysis, four cases with low quantities of mRNA or abnormal signals were excluded, and the final cohort consisted of 42 cases (20 adenocarcinomas and 22 squamous cell carcinomas).
The clinical characteristics of the 42 patients are presented in Table 1. The median age was 69 years (range, 43–85 years). Most patients were male (34/42, 81.0%) and had smoked (37/42, 88.1%) at some point. Squamous cell carcinoma (22/42, 52.48%) was predominant. EGFR mutations were detected in two cases (2/42, 4.8%). As there is essentially no need to confirm EGFR mutation in squamous cell carcinoma, the EGFR mutation status of many of the cases (18/42, 42.8%) remains unknown. Further, during the study period (between 2015 and 2018), measurement of ROS1 and BRAF mutations was not covered and required by the insurance system. Moreover, there were few residual specimens. As a result, additional confirmation of these mutations was not possible.Most patients received nivolumab as second-line chemotherapy (22/42, 52.4%), and the mean number of chemotherapy cycles was 9 (range, 1–64 cycles).However, measurement of PD-L1 was not mandatory in second-line chemotherapy with nivolumab and beyond during this study period, and there were few residual specimens, making additional confirmation impossible. As a result, the PD-L1 status was also predominantly unknown (32/42, 76.2%).
Most of the cytological materials analyzed consisted of brush from bronchoscopy (29; 69.1%) followed by imprinting from transbronchial biopsy (8; 19%), fine needle aspiration from lymph node deposit (4; 9.5%), and imprinting from CT-guided biopsy (1; 2.4%).
Treatment efficacy
The antitumor activity results are summarized in Table 2. The overall response rate (ORR) was 42.8%, and the overall disease control rate was 69.0%. In adenocarcinoma, 7 cases with PR were assigned to responders, and the remaining 13 cases were assigned to non-responders. In squamous cell carcinoma, 11 cases with PR were assigned to responders, and the other 11 cases were assigned to non-responders.
Table 3 shows the differences in ORR due to each clinicopathological factor. There was no significant difference in response rate according to the histological type. There were two cases withEGFR mutations, both of which were PD. There were two PD-L1 negative cases, both of which were evaluated for PR.
In all patients, the median PFS was 4.32 months (95% confidence interval [CI]: 2.80‒10.5) (Fig. 1A). The median OS was not reached (95% CI: 21.5‒not reached) (Fig. 1B). There was no significant difference in median PFS between histological subtypes (adenocarcinoma; 5.70 months, 95% CI: 2.33–15.4 and squamous cell carcinoma; 3.85 months, 95% CI: 1.87–11.6, p = 0.386; Fig. 1C). There was no significant difference in median OS between histological subtypes (adenocarcinoma; 23.4 months, 95% CI: 7.97‒not reached and squamous cell carcinoma; not reached, 95% CI: 23.2‒not reached, p = 0.458; Fig. 1D).
Expression profiles of adenocarcinoma and squamous cell carcinoma
Volcano plot of p‑value versus log2 fold change of the differential expression between responders and non‑responders in adenocarcinoma revealed that the numbers of genes with high expression in responders and non-responders werefour and five, respectively (Fig. 2A).In squamous cell carcinoma cases, the corresponding numbers were 12 and 5 (Fig. 2B).
T cell–inflamed 18 gene expression profile
Unsupervised hierarchical clustering analysis data between responders and non-responders are shown for all cases (Fig. 3A), adenocarcinoma (Fig. 3B), and squamous cell carcinoma (Fig. 3C). In all cases, no significant cluster formation was observed. No significant cluster formation was observed even when adenocarcinoma and squamous cell carcinoma data were separately analyzed.