Ciliated muconodular papillary tumors of the lung with KRAS/BRAF/AKT1 mutation

Cases of surgically resected CMPT from 2012 and 2016 were searched in the pathology archive of Nagasaki University Hospital. Hematoxylin and eosin-stained slides of each case were reviewed by two pathologists specializing in thoracic medicine. We identified four CMPT cases; two of these had been originally diagnosed as CMPT, whereas the other cases had been diagnosed as glandular papilloma and mucinous adenocarcinoma. Clinical data were extracted from the hospital’s electronic medical records. Sections with a thickness of 4 μm from formalin-fixed, paraffin-embedded (FFPE) tissue samples were examined by immunohistochemistry, and 15 sections with a thickness of 5 μm were subjected to NGS of 50 cancer-related genes using the Ion Torrent PGM system (Life Technologies, Carlsbad, CA), and10 μm thick section on the conventional glass slide were submitted for 29 genes analysis using MiSeq (Illumina, San Diego, CA, USA).

Genomic DNA was extracted from FFPE tumor samples using the QIAamp DNA FFPE kit (Qiagen, Valencia, CA, USA), and the concentration was determined using the Quant-iT PicoGreen dsDNA Assay kit (Life Technologies).

Genomic DNA was subjected to whole-exome sequencing using the Ion AmpliSeq Cancer Hotspot Panel v.2 (Life Technologies) according to the manufacturer’s instructions. The purified library was sequenced on an Ion PGM instrument using Ion PGM Hi-Q Sequencing kit and Ion 318 Chip Kit v.2 (all from Life Technologies). DNA sequencing data were accessed with the Torrent Suite v.5.0 program (Life Technologies). The coverage analysis was performed using the coverage analysis plug-in v5.0. Reads were aligned with the hg19 human reference genome, and potential mutations were identified using Variant Call Format v.5.0. Raw variant calls were annotated with CLC Genomics Workbench software (CLC bio, Aarhus, Denmark). Variants were manually verified using the integrative genomics viewer (Broad Institute, Cambridge, MA, USA). Known single nucleotide polymorphisms were identified using the Human Genetic Variation Database (http://​www.​hgvd.​genome.​med.​kyoto-u.​ac.​jp/​) [7] and were excluded.

We also performed sequencing on Illumina MiSeq instrument. As recommended by the manufacturer, 50 ng of dsDNA was used for generating sequencing libraries using Find-It cancer hotspot panel (Contextual Genomics, Vancouver, BC, Canada). The panel targets 90+ cancer hotspot mutations including eight coding exons in 29 cancer-related genes. DNA libraries were denatured and diluted as per Illumina’s recommendations. Samples were run on MiSeq machine using 300 cycle MiSeq Reagent Kit V2 (Cat.​No.​MS-102-2002, Illumina).

Immunohistochemical analysis of 4-μm tissue sections was carried out using the Ventana Bench Mark XT Automated stainer (Ventana Medical Systems, Tucson, AZ, USA) and BOND III fully automated stainer (Leica Biosystems, Melbourne, Australia) with antibodies against thyroid transcription factor (TTF)-1 (clone SPT24; Leica Novocastra, Newcastle Upon Tyne, UK), hepatocyte nuclear factor (HNF)-4α (clone H1415; Perseus Proteomics, Tokyo, Japan), mucin (MUC)5B (Santa Cruz Biotechnology, Santa Cruz, CA, USA), MUC5A (clone CLH2; Leica Novocastra), cytokeratin (CK)7 (clone OV-TL12/30; Dako Japan, Tokyo, Japan), CK20 (clone Ks20.8; Dako Japan), p53 (clone DO-7; Leica Novocastra), Ki-67 (clone MIB-1; Dako Japan), Caudal-type homeobox (CDX)2 (clone EPR2764Y; Roche Diagnostics, Indianapolis, IN, USA), ALK (clone D5F3; Roche Diagnostics), BRAF V600E (clone VE1; Roche Diagnostics) and programmed death ligand (PD-L)1 (clone SP142; Spring Bioscience, Pleasanton, CA, USA). Details of optimized staining protocols are listed in Additional file 1. HNF4α, MUC5B, MUC5A, CK7, and CK20 staining was scored as follows: −, negative; 1+, focal; and 2+, diffuse. The intensity of p53 expression was assessed as follows: 1+, faint and sporadic; 2+, frequently positive with no or equivocal overexpression; and 3+, unequivocal and diffuse overexpression. BRAF V600E and ALK expression was judged as positive when strong and diffuse staining was observed in cancer cells. The percentage of Ki-67-immunopositive cells was analyzed using a nuclear algorithm (Nuclear_v.9_v.10.0.0.1798; Imagescope, Leica Biosystems, Nussloch, Germany). PD-L1 expression was scored as the percentage of PD-L1-positive cells among tumor cells. CDX2 expression was scored as follows: −, negative; 1+, ≤ 25% positive; 2+, > 25 and <50% positive; and 3+, ≥ 50% positive. Two observers (E.U. and J.F.) independently scored the staining and a consensus was obtained through discussion when there was a discrepancy.

Clinical data for the four CMPT patients are summarized in Table 1. The patients were all female with a median age of 67 years old. None had a history of smoking. Tumors ranged in size from 8 to 25 mm (median: 11 mm). Lobectomy was performed in three cases, and segmentectomy was carried out in one case.

The histological features of the four CMPT cases were consistent with those previously reported [1, 8, 9]. That is, the tumors typically exhibited a mixture of acinar and papillary growth patterns without clear evidence of invasion, and were surrounded by abundant mucus pools in the alveolar spaces (Fig. 1a). The tumors were composed of two basal cell layers and surface epithelia. The latter consisted of an uneven mosaic of ciliated columnar, goblet, and mucin-producing epithelial cells of the gastric type (Fig. 1b). Few tumor cells showed nuclear atypia, and no mitosis or necrosis was observed.

All four cases were positive for nuclear HNF4α and TTF-1. Two cases showed diffuse and strong staining for HNF4α, as seen in mucinous adenocarcinoma or colorectal carcinoma (Fig. 1c). The Ki-67 index was low (2.5%–10%), consistent with a less aggressive nature (Fig. 1d). In contrast to mucinous adenocarcinoma, all cases examined in this study showed diffuse TTF-1 positivity and one case showed possible overexpression. There was sparse cytoplasmic expression of MUC5AC (Fig. 1e) while diffuse MUC5B—which is very rare in the normal lung parenchyma—was detected in all cases (Fig. 1f). Nuclear expression of p53 was sporadic in one case and more frequent in two cases, and in one case p53 was clearly overexpressed. There were no or only a few cells that were positive for CDX2 and CK20. p63 expression highlighted the basal layers in all cases, although the area of coverage was focal in two cases and broad in the two others (Table 2).

One case with BRAF V600E mutation identified by NGS showed diffuse and strong BRAF V600E staining, while the other three cases showed none (Fig. 1g). There was no ALK and EGFR mutations identified in our cases. ALK status was also confirmed by immunohistochemistry.

All cases exhibited <1% PD-L1 positivity in tumor cell membranes (Fig. 1h). These findings suggest that CMPTs are tumors that originated from a terminal respiratory unit and are showed features of gastric-type glands.

Gene mutations were detected in two of the four cases (50%) (Table 1). One case harbored BRAF V600E and AKT1 E17K mutations—which was interestingly identical as what have been recently reported [4, 5]—and another had a KRAS G12D mutation. The detected mutation status was identical on both Ion PGM instrument system and Illumina MiSeq system. A KRAS G12C mutation was found in one of the two other cases; however, the significance of this mutation is unclear due to its low frequency.