García Alejandro, Recondo Gonzalo, Greco Martín, de la Vega Máximo, Perazzo Florencia, Recondo Gonzalo, Avagnina Alejandra, Denninghoff Valeria
Department of Pathology, Center for Medical Education and Clinical Research (CEMIC), Argentina.
Department of Oncology, Center for Medical Education and Clinical Research (CEMIC), Argentina.
Heliyon. 2020 Jun 1;6(6):e04117. doi: 10.1016/j.heliyon.2020.e04117. eCollection 2020 Jun.
Lung cancer is the leading cause of cancer-related death worldwide. Recent advances in the management of non-small cell carcinoma are focused on the discovery of targeted therapies and novel immunotherapy strategies for patients with advanced disease. Treatment with anti PD-(L)1 immune checkpoint inhibitors requires the development of predictive biomarkers to select those patients that can most benefit from these therapies. Several immunohistochemical biomarkers have been developed in different technological platforms. However, the most useful and accessible for the daily clinical practice need to be selected. The objective of this study was to compare PD-L1 expression by automated immunohistochemistry in lung adenocarcinoma (ADC) FFPE samples with clones 28-8 and SP263 performed with the BenchMark GX automated staining instrument. To further determine interobserver agreement between two pathologists, and to correlate the results with histologic and pathology variables. FFPE tissue from 40 samples obtained from patients with lung ADC were reviewed retrospectively. Among all studied specimens, 53% of samples presented <1% of positive tumor cells with the 28-8 clone and 50% had <1% of PD-L1 expression in tumor cells with the SP263 clone; PD-L1 expression between ≥1 and <5% was observed in 18% and 24%; ≥5 and <50% PD-L1 expression in 18% and 21%; and ≥50% PD-L1 expression in 11% and 5% of samples, respectively. Similar results between antibodies were observed in 84% of cases for each of the four PD-L1 cutoff groups (Pearson's score 0.90, p < 0.00001). The interobserver degree of agreement calculated with Kappa was 0.75 (95%CI: 0.57-0.93), z = 7.08; p < 0.001. Lepidic, acinar and mucinous patterns had predominantly <1% PD-L1 expression, and the solid pattern subtype had high levels of PD-L1 staining using both clones. PD-L1 expression in less than 1% of tumor cells was similar in stages I/II compared to III/IV. No significant differences were observed in PD-L1 staining and quantification pattern between IHC antibodies 28-8 and SP263.
肺癌是全球癌症相关死亡的主要原因。非小细胞癌治疗的最新进展集中在为晚期疾病患者发现靶向治疗和新型免疫治疗策略。使用抗PD-(L)1免疫检查点抑制剂进行治疗需要开发预测性生物标志物,以选择那些能从这些疗法中最大程度获益的患者。已在不同技术平台上开发了几种免疫组化生物标志物。然而,需要选择对日常临床实践最有用且最易获取的标志物。本研究的目的是比较使用BenchMark GX自动染色仪,采用克隆号28-8和SP263对肺腺癌(ADC)FFPE样本进行自动免疫组化检测时的PD-L1表达情况。进一步确定两名病理学家之间的观察者间一致性,并将结果与组织学和病理学变量相关联。对40例肺ADC患者的FFPE组织进行回顾性分析。在所有研究标本中,使用克隆号28-8时,53%的样本肿瘤细胞阳性率<1%,使用克隆号SP263时,50%的样本肿瘤细胞PD-L1表达<1%;18%和24%的样本PD-L1表达在≥1%且<5%之间;18%和21%的样本PD-L1表达在≥5%且<50%之间;11%和5%的样本PD-L1表达≥50%。在四个PD-L1临界值组中,每组84%的病例抗体之间观察到相似结果(Pearson评分0.90,p<0.00001)。用Kappa计算的观察者间一致程度为0.75(95%CI:0.57-0.93),z=7.08;p<0.001。鳞屑状、腺泡状和黏液状模式主要PD-L1表达<1%,实性模式亚型使用两种克隆时PD-L1染色水平较高。与III/IV期相比,I/II期肿瘤细胞中PD-L1表达<1%的情况相似。在IHC抗体28-8和SP263之间,PD-L1染色和定量模式未观察到显著差异。