Faculty Bibliography

Objectives/UNASSIGNED:To evaluate whether non-small cell lung carcinoma (NSCLC) cytology specimens are reliable for programmed death-ligand 1 (PD-L1) immunohistochemical (IHC) testing. Methods/UNASSIGNED:Fifty-two cell blocks (CBs) with corresponding surgical pathology PD-L1 IHC testing were stained with a Dako PD-L1 pharmDX antibody (clone-22C3). Tumor cellularity was recorded as <100 or ≥100 cells. PD-L1 IHC was scored by percentage of tumor cells staining (<1%, ≥1%-49%, ≥50%) and compared between matched cases. Results/UNASSIGNED:Substantial agreement (κ = 0.63; 95% CI, 0.53-0.73) was reached between matched CB and surgical cases in CBs with ≥100 tumor cells compared to CBs with <100 tumor cells (slight agreement, κ = 0.19; 95% CI, 0.04-0.35). Overall, there was 67% agreement among paired cases (35/52 cases, κ = 0.51; 95% CI, 0.42-0.60). Conclusions/UNASSIGNED:CBs can be utilized for PD-L1 IHC testing, as illustrated by the 67% agreement between CB and surgical cases in our study. Disagreement is attributable to intratumoral heterogeneity and CB cellularity.

Since the 2015 WHO classification was introduced into clinical practice, the importance of immunohistochemistry (IHC) has figured prominently in lung cancer diagnosis. In addition to distinction of small versus non-small cell carcinoma (NSCC), patients' treatment of choice is directly linked to histological subtypes of NSCC, which pertains to IHC results, particularly for poorly-differentiated tumors. The use of IHC has improved diagnostic accuracy in the lung carcinoma classification, but the interpretation remains challenging in some instances. Also, pathologists must be aware of many interpretation pitfalls, and the use of IHC should be efficient to spare the tissue for molecular testing. The IASLC Pathology Committee received questions on practical application and interpretation of IHC in lung cancer diagnosis. After discussions in several IASLC Pathology Committee meetings, the issues and caveats were summarized as eleven key questions, which cover common and important diagnostic situations in a daily clinical practice with some relevant challenging queries. The questions included best IHC markers for distinguishing NSCLC subtypes, differences in TTF1 clones, utility of IHC in diagnosing uncommon subtypes of lung cancer and distinguishing primary from metastatic tumors." This article provides answers and explanations for the key questions about the use of IHC in lung carcinoma diagnosis representing viewpoints of experts in thoracic pathology that should assist the community in the appropriate use of IHC in diagnostic pathology.

Visual inspection of histopathology slides is one of the main methods used by pathologists to assess the stage, type and subtype of lung tumors. Adenocarcinoma (LUAD) and squamous cell carcinoma (LUSC) are the most prevalent subtypes of lung cancer, and their distinction requires visual inspection by an experienced pathologist. In this study, we trained a deep convolutional neural network (inception v3) on whole-slide images obtained from The Cancer Genome Atlas to accurately and automatically classify them into LUAD, LUSC or normal lung tissue. The performance of our method is comparable to that of pathologists, with an average area under the curve (AUC) of 0.97. Our model was validated on independent datasets of frozen tissues, formalin-fixed paraffin-embedded tissues and biopsies. Furthermore, we trained the network to predict the ten most commonly mutated genes in LUAD. We found that six of them-STK11, EGFR, FAT1, SETBP1, KRAS and TP53-can be predicted from pathology images, with AUCs from 0.733 to 0.856 as measured on a held-out population. These findings suggest that deep-learning models can assist pathologists in the detection of cancer subtype or gene mutations. Our approach can be applied to any cancer type, and the code is available at https://github.com/ncoudray/DeepPATH .