Although it has been well recognized that EGFR
mutation is strongly associated with the therapeutic effect of TKIs in NSCLC patients, most patients could not provide the tumor tissues that needed for the mutation test [5
]. Prior literatures indicate that it is feasible to use the free DNA in body fluid such as pleural fluid and plasma as alternative clinical specimen for EGFR
mutation analysis [13
], but the procedure still needs to be optimized, standardized and validated.
The major finding of our research was that, when body fluid was used as substitute for EGFR
mutation detection, the positive result was a good indicator for TKIs therapy, no matter it was detected by direct sequencing or ARMS. For patients who provided pleural fluid, the ORR of the 16 mutation positive patients defined by direct sequencing was 81.3%. When ARMS was used, 6 more patients were defined as mutation positive, with the ORR of the 22 patients at 72.7%. For patients who provided plasma, 5 mutation positive patients were detected only by ARMS, with the ORR at 80%. Generally, our result was consistent with that of OPTIMAL and IPASS research, both using tumor tissue for EGFR
mutation analysis [5
]. The ORR for mutation positive patients in OPTIMAL using direct sequencing was 83%, higher than that of IPASS using ARMS strategy (71.2%). Interestingly, such difference also occurred in our study using pleural fluid samples (81.3% Vs 72.7%). The results implied that, more sensitive methods such as ADx-ARMS may find more positive patients, but for them, mutative cells may represent a minority of the whole tumor, which may influence the final clinical outcome of TKIs. The explanation is consistent with the work of Qing Zhou et al. which found that the relative EGFR
mutation abundance could predict benefit from EGFR
-TKIs treatment for advanced NSCLC [19
]. Our data emphasized that, for mutation positive results, the predictive effect of body fluid was no less than that of tumor tissue.
As considered for the two problems mentioned above, our research agreed with former reports that more sensitive method such as ARMS would be one of the feasible solutions [14
]. Compared with direct sequencing, ADx-ARMS assay found 18.8% (6/32) and 27.8% (5/18) more patients to be mutation positive for pleural fluid and plasma, respectively. Direct sequencing is currently the routine method used to detect EGFR
mutations. The merits of this method are readily available and economic, but the procedure is complicated and time-consuming. Meanwhile, the sensitivity of sequencing is about 30%, which tends to cause false negative result [21
]. Given the poor sensitivity of DNA sequencing, many patients and physicians opt to start TKIs treatment even if the sequencing results were negative for EGFR
mutation. If the tumor does not contain activating mutations on EGFR
, treatment with TKIs will most likely be ineffective. In our study, 11 former negative patients (6 pleural fluids, 5 plasmas) defined by sequencing were proved to be positive at last, and the clinical outcome for them was quite satisfactory. If the treatment plan was made according to the result of direct sequencing, those patients may lose the chance of TKIs therapy.
Besides, by using ARMS, we also found 7 samples which harbouring double mutations (2 patients with 19 del and L858R, 1 with L858R and L861Q or S768I, 4 with 19 del and T790M). The clinical evaluations for the former 3 patients were all PR. This result was consistent with the study of Zhang et al. [22
] which showed that patients with double activating mutations involving both exons 19 and 21 tend to respond well to TKIs and the sensitivity to TKIs was enhanced compared with either single mutant. As demonstrated by Qing Zhou et al. that the relative EGFR
mutation abundance could predict benefit from EGFR
-TKIs treatment [19
], we hypothesized that the clinical benefits (3PR, 1SD) of the 4 patients which harbouring both 19 del and T790M may be owing to the dominant composition in 19 del.
Notably, even reclassified by ARMS, no difference was found in PFS among mutation positive and negative patients, the ORR for negative patients was still relatively high, 60% for pleural fluids and 46.2% for plasma, higher than that of IPASS (1.1%) and First-SIGNAL (25.9%) research [5
]. Taking into consideration that all the patients in our research were adenocarcinoma, the well known type of lung cancer that can get maximum benefit from TKIs therapy, and the low abundance of DNA in body fluid, the results indicated that there might still be false negative mutations in these samples. We presumed that the phenomenon can be explained in two aspects.
Firstly, the sensitivity of ARMS is 1%, nevertheless, if the abundance of the mutation DNA was below this limitation, false negative results were inevitable. Prior literature indicated that, using ARMS for plasma samples, the false negative rate was still relatively high, which was about 30% as compared with tumor tissue [13
]. Recently, Yung TK et al. reported a method named Microfluidics Digital PCR, which could detect a single-mutant DNA molecule and precisely determine the quantities of mutant and wild-type sequences. By using this method, the sensitivity and specificity of plasma EGFR
mutation analysis reached 92% and 100% respectively, as compared with the sequencing results of tumor samples [18
]. This method may be more suitable than ARMS for EGFR
mutation analysis using body fluid samples, but it is not readily available now and more stringent clinical evidence is still needed in the future.
Secondly, regardless of the sensitivity of detection method, if tumor-derived DNA was not contained in the body fluid sample, the mutation analysis was obviously in vain.
For pleural fluid samples, it is well recognized that cell pellets could be used to ensure tumor cells was contained in the sample. Nevertheless, in a significant proportion of patients (30-40%), the yield of malignant cells from thoracentesis is inadequate for cytological and molecular diagnostic testing. We used cell-free pleural fluids in this study because it is abundant. Meanwhile, prior literature demonstrated that when sensitive genotyping assays was used, cell-free pleural fluid could provide the same mutational information as pleural effusion cells [15
]. The problem is that, when cell-free pleural fluid was used, it was impossible to precisely evaluate whether the tumor-derived DNA was adequately contained, since the extracted free DNA arises not only from tumor cells, but also from the necrotic or apoptotic nontumor cells. Recently, free RNA in pleural fluid as a favouring material for EGFR
mutation analysis was attracting more and more attention. The high EGFR
mutation rate of free RNA in pleural effusion has been reported in the article by Wu et al. [24
]. Later on, the same research group found out that the mutation-detection yield of sequencing from RNA was coupled with the superior prediction of clinical efficacy to first-line TKIs [25
]. The explanation was that, contaminated nontumor cells within pleural fluid may have no or lower EGFR
expression, using RNA instead of genomic DNA as the source for EGFR
mutation analysis could minimize the influence of nontumor cells.
For blood samples, most reports used plasma rather than cell pellets for mutation analysis, because tumor cells in the blood are rare as compared with the cells of hematopoietic lineages. The documented sensitivity of plasma varied from 33% to 100%, which may be resulted from various detection methods or from different patients enrolled [17
]. But using plasma encounter the same problem as using cell-free pleural fluid, namely, it is impossible to precisely evaluate whether the tumor-derived DNA was adequately contained. The characterization of circulating tumor cell might resolve the problem ultimately, since it is ascertain that the test was done on tumor cells. In the study by Maheswaran et al, there were 12 patients for whom specimens of the primary tumor, CTCs, and plasma were all available for EGFR
mutation analysis. The genotyping of CTCs appeared to be more sensitive than plasma (92% Vs 33%, P
= 0.009) [27
]. The main problem now is that the technology of CTC enrichment still needs to be standardized and generalized. In recent years, tremendous efforts have been made on CTC detection and characterization [28
]. In the near future, EGFR
mutation analysis on CTC may become a reality in the routine clinical practice.
Our study had two limitations, which hindered us from verifying the hypothesis mentioned above. First, although we and others have demonstrated that body fluid is feasible [13
], analysis for EGFR
mutations with DNA extracted from tumor tissue remains the gold standard. Nevertheless, since all the patients enrolled in this study couldn't provide sufficient tumor tissue after routine pathological examination was done, the mutation status of the tumor tissue were not available and we could not testify whether there were still false negative results left after the extracted DNA were re-examined by ARMS. Second, although it is necessary to re-extract the nucleic acid with an optimized procedure by RNA or CTC, and then, to compare the mutation analysis with current study, the original body fluid samples of the patients were not preserved after the mutation analysis was done, the comparison could not be carried out. In order to address the two issues above, we had set a new research plan and the patients were now under enrolling.