肿瘤组织及血浆KRAS突变与结直肠癌患者预后的相关性研究

【作者】 刘晓静；

【导师】 宋三泰； 徐建明；

【作者基本信息】 中国人民解放军军事医学科学院 ， 药理学， 2012， 博士

【摘要】 随着新型抗肿瘤药物的不断出现，晚期结直肠癌治疗的有效率及患者的总生存时间不断提高，特别是靶向药物EGFR（Epidermal growth factor receptor，表皮生长因子受体）单抗的应用，联合化疗可以使患者的生存期延长到两年，而且研究已经发现了EGFR单抗的疗效预测因子，即KRAS基因突变。在用药前，通过检测患者肿瘤组织的KRAS基因是否发生突变，来筛选敏感人群，初步达到了个体化应用。但是在个体化应用过程中，也存在一些很大的缺陷，由于肿瘤组织取材比较困难，而且不具备实时性，因此，找到能够满足实时性、方便可取，含有肿瘤组织遗传信息的替代物来代替肿瘤组织进行基因检测，就成了目前亟待解决的一个问题。随着检测技术的发展，KRAS基因突变的检测方法也面临着改革，有研究提出，目前公认的测序方法，灵敏度较低，可能会导致假阴性结果的出现，应该用灵敏度高的方法来检测。我们考虑灵敏度不同的两种方法检测，肯定会存在一类患者，即灵敏度低的方法检测为野生型，但灵敏度高的方法可能检测为突变型，也就是说这部分患者突变的量可能比较低。对于这部分患者，是否适合应用EGFR单抗，除了KRAS是否发生突变之外，KRAS突变的量（突变丰度）是否会对EGFR单抗的疗效及患者的预后产生影响？这在目前还没有任何相关的研究。KRAS突变虽然是EGFR单抗的疗效预测因子，但与单纯化疗患者预后的相关性目前仍然存在分歧，存在分歧的原因复杂，我们认为其中较为重要的一点，可能是因为各种研究采用的检测方法没有统一，检测方法的灵敏度高低不同。因此，为了探讨检测方法的不同对KRAS突变与单纯化疗患者预后相关性的影响；KRAS突变丰度对单纯化疗的患者疗效和预后的关系；以及血浆替代肿瘤组织进行KRAS基因突变检测的可行性，我们开展了此项研究。从2008年1月到2011年6月筛选接受过至少两个方案的单纯化疗（奥沙利铂为主或者伊立替康为主的方案）的晚期结直肠癌患者117例，根据RECIST（Response Evaluation Criteria in Solid Tumors，实体瘤疗效评价标准）评价标准，一线化疗有37例患者获得部分缓解（Partial response，PR）（31.6%），全体患者的中位PFS（Progression free survival，无进展生存时间）是6.3个月，中位OS（Overall survival，总生存时间）是19.5个月。用直接测序法和PNA-PCR（Peptide NucleicAcid-mediated PCR Clamping，肽核酸钳制PCR）方法分别检测了患者的肿瘤组织和配对血清中的KRAS突变状态，直接测序法检测肿瘤组织KRAS突变的患者46例，检出率为39.3%，血浆中KRAS突变的患者18例，检出率15.4%，与肿瘤组织的一致率为32.6%，野生与突变整体的一致率为70.9%，一致性较差（Kappa0.318，p<0.001）。用PNA-PCR方法检测肿瘤组织和血浆的KRAS基因突变，肿瘤组织KRAS突变的患者56例，检出率为47.9%，血浆中KRAS突变的患者35例，检出率为29.9%，与肿瘤组织的一致率为55.4%，野生与突变整体的一致性率为75.2%，一致性一般（Kappa0.496，p<0.001）。肿瘤组织的KRAS突变用直接测序法检测，而血浆用PNA-PCR方法检测，发现血浆与肿瘤组织的KRAS突变的一致率为63.0%，整体一致率为80.3%，一致性较好（Kappa0.570，p<0.001）。从以上结果我们可以看出，因为血浆游离DNA中含有的肿瘤组织特异性DNA量很低，因此更适合用灵敏度高的检测方法进行KRAS基因突变的检测，提高与肿瘤组织KRAS突变的一致率。将肿瘤组织的KRAS突变状态与单纯化疗的疗效进行统计分析发现，无论用直接测序法或者PNA-PCR方法，野生型与突变型的患者在ORR（Over responserate，总有效率）及PFS均没有统计学差异。在患者的预后方面，直接测序法检测的野生型与突变型患者在OS上的差异接近统计学意义（p=0.052），PNA-PCR方法检测的野生型与突变型患者的OS具有明显的统计学差异(p=0.010)。根据直接测序法和PNA-PCR方法的KRAS检测结果（KRAS突变丰度）将患者分为野生型组（两种方法检测的KRAS均为野生型）、高突变组（两种方法检测的KRAS均为突变型）和低突变组（直接测序法检测的为野生型，而PNA-PCR方法为突变型），三组患者在ORR及PFS上仍然没有差异，但野生型组患者的中位OS可长达21.3个月，低突变组的17.1个月，而高突变组的15.6个月，差异具有明显的统计学意义（p=0.033）。将这三组进行两两比较，结果显示，野生型组与高突变组OS的差异具有显著的统计学意义（p=0.029），与低突变组OS的差异接近统计学意义（p=0.058）。而高突变组和低突变组间的OS没有统计学差异。由此我们可以看出，低突变组的OS接近高突变组，野生型组的OS还是最长的。直接测序法或PNA-PCR方法检测的血浆中KRAS突变状态对单纯化疗的疗效没有影响，但与患者的OS具有明显的相关性。直接测序法检测的野生型患者比突变型患者中位OS长5个月，有统计学差异（p=0.009）。同样的，用PNA-PCR方法检测的野生型和突变型患者的OS分别是21.3个月和15.1个月，差异也具有明显的统计学意义（p=0.001），以上结果说明血浆中的KRAS基因突变状态与患者的OS存在相关性，KRAS突变的患者预后很差。血浆中KRAS突变可以作为单纯化疗患者预后的预测因子。以上结果表明，用PNA-PCR方法检测血浆中的KRAS突变与肿瘤组织的一致性较高，在肿瘤组织获取困难时，可以考虑通过检测血浆中KRAS基因的突变来指导个体化用药。不论何种方法检测KRAS基因突变，KRAS突变与单纯化疗的疗效没有相关性，不能作为疗效的预测因子。但是对于KRAS突变与单纯化疗患者预后的相关性，不同检测方法导致了不同的结论，从而证实目前研究中KRAS突变与单纯化疗患者的预后存在分歧的原因之一，可能就是检测方法的不同导致的。在本研究中，用直接测序法检测的野生型和突变型患者的OS没有统计学差异，而PNA-PCR方法检测的野生型和突变型患者的OS有明显的差异，因此，灵敏度高的PNA-PCR方法可能比直接测序法更适合肿瘤组织KRAS突变的检测。通过分析野生型组、高突变组和低突变组的OS发现，野生型患者的OS最长，低突变组的次之，高突变组的最短。因此如果按照KRAS基因突变的丰度进行更详细的分组，可能对预测患者的预后具有更重要的意义。这提示我们应该联合两种方法进行KRAS基因突变的检测，野生型患者最好用PNA-PCR方法来确定，从而能够降低假阴性结果的出现，而突变型患者则用传统的方法--直接测序法来确定，避免灵敏度高的方法，如PNA-PCR方法对KRAS微量突变的夸大，更好的预测患者的预后。而对于血浆中的KRAS突变，不论用哪种方法，KRAS突变都与患者的疗效没有相关性，但与患者的预后却存在明显的相关性，说明血浆中的KRAS突变是患者预后的预测因子。结论：不同的检测方法对KRAS突变预测单纯化疗患者的预后带来不同的结论，可能是导致目前KRAS突变与患者预后存在分歧的重要原因之一。灵敏度高的PNA-PCR方法联合传统的直接测序法进行KRAS突变的检测，将患者按照KRAS突变丰度进行分组，可能会对患者的预后具有更大的帮助。血浆中KRAS突变的检测推荐使用灵敏性高的方法，突变的患者预后较差。当患者的肿瘤组织获取困难时，血浆也可以考虑作为肿瘤组织的替代物进行KRAS基因突变的检测。更多还原

【Abstract】 With the constant emergence of new anti-tumor drugs, the efficiency of thetreatment for advanced colorectal cancer and the whole survival time of the patientsare increasing constantly, especially with the use of targeted drug EGFR (Epidermalgrowth factor receptor) McAb, combined chemotherapy can extend the patients’survival time as long as two years. Besides, the study has discovered the outcomepredictor of the EGFR McAb, which is KRAS genetic mutation. Before takingmedication, susceptive group can be screened through detecting whether the KRASgenes of the patients’ tumor tissues have mutated, so EGFR McAb has realizedpreliminary individual application.However, in the process of individual application, there exist some big defects.For the reason that it is relatively difficult to obtain tumor tissues and that it does nothave the real-time, it is an urgent problem to find out the substitute containing thegenetic information of the tumor tissues, which satisfies the need of real-time andeasy access, to replace the tumor tissues for genetic detection.With the development of detecting techniques, the detecting methods of KRASgenetic mutation are also faced with reform. A study shows that the commonlyaccepted sequencing methods for the moment has low sensitivity, which might lead tothe existence of false-negative results, so it is necessary to detect with methods ofhigh sensitivity. We considerd that there is bound to be a kind of patients whendetected with two methods of different sensitivity. When applied to the method of lowsensitivity, the patients belong to the wild type. As for the method of high sensitivity,the patients may belong to the mutant type. That is to say, this kind of patients mayhave relatively low quantity of mutation. Whether it is suitable to apply EGFR McAbto this kind of patients and whether the quantity of the KRAS mutation (mutationabundance), except for whether the KRAS mutation happens, will have influence onthe curative effect of the EGFR McAb and the patients’ prognosis? No relative studyhas been done by now. Though KRAS mutation is the outcome predictor for EGFRMcAb, it still has difference with the relativity of the prognosis of the simple chemotherapy patients. The reasons for differences, we believed that one of theimportant point may be the detection methods used by the various studies was notunified, he detection sensitivity of the method was different.So we carry out the study in order to investigate the different detection methodsto KRAS mutation and patients’ prognosis; and the relations between the KRASmutation abundance and the curative effects and prognosis of the simplechemotherapy patients, and the feasibility of using plasma to replace tumor tissues inthe process of detecting KRAS genetic mutation.From January,2008to June,2011, we have screened117advanced colorectalcancer patients, who have received at least two schemes of simple chemotherapy(schemes centered on oxaliplatin or on irinotecan). Based on RECIST(ResponseEvaluation Criteria in Solid Tumors),37patients in the first-line chemotherapy havegained partial response (PR)(31.6%). The median PFS(Progression free survival) forall the patients is6.3months and the time of median OS(Overall survival) is19.5months. We have applied methods of direct sequencing and PNA-PCR(PeptideNucleic Acid-mediated PCR Clamping) to detect the KRAS mutation status in thepatients’ tumor tissues and in the paired serum. There are46patients of KRASmutation detected in the tumor tissues with the method of direct sequencing and thedetection rate is39.3%. As for the KRAS mutation in the plasma,18patients havebeen detected and the detection rate is15.4%. Its concordance rate with the tumortissues is32.6%. The general concordance rate of the wild type and the mutant type is70.9%and the consistency is relatively bad（Kappa0.318，p<0.001）. By using themethod of PNA-PCR to detect KRAS genetic mutation of the tumor tissues and theplasma,56patients of KRAS mutation of tumor tissues have been detected, whosedetection rate is47.9%, and35patients of KRAS mutation in the plasma have beendetected, whose detection rate is29.9%and whose concordance rate with tumortissues is55.4%. The general concordance of the wild type and the mutant type is75.2%and the consistency is ordinary（Kappa0.496，p<0.001）. By applying themethod of direct sequencing to detect KRAS mutation in tumor tissues and themethod of PNA-PCR to detect the plasma, we find that the concordance rate of KRAS mutation in the plasma and tumor tissues is63.0%. The general concordance rate is80.3%and the consistency is good（Kappa0.570，p<0.001）. We can conclude fromthe above results that because the plasma DNA contains very low amount of tumortissue-specific DNA, it is more suitable to use the detecting methods of highsensitivity to detect KRAS genetic mutation and so as to increase the concordancerate with the KRAS mutation of tumor tissues.Through measurement and analysis of the KRAS mutation conditions of thetumor tissues and the curative effects of the simple chemotherapy, we find that neitherpatients of the wild type nor patients of the mutant type have statistical difference inORR(over response rate) and PFS when referred to both methods of direct sequencingand PNA-PCR. In the aspect of the patients’ prognosis, the wild type patients and themutant type patients detected with the method of direct sequencing approachstatistical meaning in the difference of OS (p=0.052), while patients of wild type andmutant type detected with the method of PNA-PCR have obvious statistical differencein OS(p=0.010). According to results of the KRAS detected with the methods ofdirect sequencing and PNA-PCR(abundance of KRAS mutation), we divide thepatients into wild type group(wide type for the two methods), high mutationgroup(mutation for the two methods) and low mutation group(wild type for the directsequencing, while mutation for the PNA-PCR method). The patients of the threegroups have no difference in efficiency and PFS, but the median OS for the wild typegroup is as long as21.3months,17.1months for the low mutation group,and15.6months for the high mutation group, the difference has obvious statistical meaning(p=0.033). Comparing each two of the three groups, we find that the OS differencebetween the wild type group and the high mutation group has obvious statisticalmeaning (p=0.029) and the OS difference between the wild type group and the lowmutation group approaches statistical meaning(p=0.058). Besides, there is nostatistical difference of OS between high mutation group and low mutation group. Aswe can see, the OS of the low mutation group approaches that of the high mutationgroup and the OS of the wild type group is still the longest. The KRAS mutation conditions in the plasma detected with the method of eitherdirect sequencing or PNA-PCR have no influence on the curative effects of the simplechemotherapy, but it has obvious relativity with the patients’ OS. The median OSdetected by the method of direct sequencing is five months longer than that of thepatients of mutant type and there exists statistical difference (p=0.009). Similarly, theOS for the patients of the wild type and the mutant type are21.3months and15.1months respectively, and the difference also has obvious statistical meaning (p=0.001).The above results show that the conditions of the KRAS genetic mutation in theplasma have relativity with the OS of patients and that the prognosis for the KRASmutation patients is bad. The KRAS mutation in the plasma is the outcome predictorfor the patients of simple chemotherapy.The results show that the KRAS mutation in the plasma and tumor tissueshas relatively high consistency when detected by PNA-PCR, so detecting the KRASgenetic mutation in the plasma can be considered for guiding individual medicationwhen it is difficult to access the tumor tissues. In whichever way to detect KRASgenetic mutation, there is no relativity between KRAS mutation and the curativeeffects of the simple chemotherapy, so KRAS mutation cannot be regarded as thepredictor of the curative effects. However, as for the relativity between KRASmutation and the OS of patients of simple chemotherapy, different detecting methodsmay lead to different conclusions, which confirmed that one of the reasons fordifferent conclusions of the relationship between KRAS mutation and patients’prognosis. In this study, the overall survival for the patients of wild type and mutanttype detected by direct sequencing has no statistical difference, while the overallsurvival for the patients of the wild type and mutant type detected by PNA-PCR hasobvious difference. Therefore, the method of PNA-PCR with high sensitivity may bemore suitable than the method of direct sequencing for the detection of KRASmutation in the tumor tissues. But through analyzing the OS of the three groups, wefind that the OS of the wild type patients is the longest, and that of the low mutationgroup and high mutation group come as the second and third longest respectively.Therefore, it may have more important predicting meaning if we carry out detailed categorization according to the abundance of the KRAS genetic mutation. Thissuggests that we should combine the two methods to detect KRAS genetic mutation,which means that the patients of the wild type should be applied to the method ofPNA-PCR, which can reduce the false negative results. and that the patients of themutant type should be applied to the traditional method-----direct sequencing,to avoid the exaggeration of KRAS mutation detected by high sensitivity methods,such as the PNA-PCR method. As for the KRAS mutation in the plasma, bothmethods have showed that KRAS mutation has no relativity with the patients’curative effects, but it has obvious relativity with the patients’ prognosis, suggestingthat the KRAS mutation in the plasma is the predictor for the prognosis of thepatients.In conclusion, different detecting methods may lead to different conclusions,which confirmed that one of the reasons for different conclusions of therelationship between KRAS mutation and patients’ prognosis. it may have betterpredicting effects for the patients’ prognosis by detecting the KRAS mutation with thecombination of PNA-PCR of high sensitivity and the traditional direct sequencing,and by categorizing the patients according to the abundance of the KRAS mutation. Itwould be better to use methods of high sensitivity to detect KRAS mutation in theplasma and the prognosis for the mutant patients is relatively bad. When it is difficultto gain the patients’ tumor tissues, plasma can be taken as the replacement of thetumor tissues for the detection of the KRAS genetic mutation.更多还原