A Deep Learning-Based Assay for Programmed Death Ligand 1 Immunohistochemistry Scoring in Non-Small Cell Lung Carcinoma: Does it Help Pathologists Score?

Several studies have developed various artificial intelligence (AI) models for immunohistochemical analysis of programmed death ligand 1 (PD-L1) in patients with non-small cell lung carcinoma; however, none have focused on specific ways by which AI-assisted systems could help pathologists determine the tumor proportion score (TPS). In this study, we developed an AI model to calculate the TPS of the PD-L1 22C3 assay and evaluated whether and how this AI-assisted system could help pathologists determine the TPS and analyze how AI-assisted systems could affect pathologists' assessment accuracy. We assessed the 4 methods of the AI-assisted system: (1 and 2) pathologists first assessed and then referred to automated AI scoring results (1, positive tumor cell percentage; 2, positive tumor cell percentage and visualized overlay image) for final confirmation, and (3 and 4) pathologists referred to the automated AI scoring results (3, positive tumor cell percentage; 4, positive tumor cell percentage and visualized overlay image) while determining TPS. Mixed-model analysis was used to calculate the odds ratios (ORs) with 95% CI for AI-assisted TPS methods 1 to 4 compared with pathologists' scoring. For all 584 samples of the tissue microarray, the OR for AI-assisted TPS methods 1 to 4 was 0.94 to 1.07 and not statistically significant. Of them, we found 332 discordant cases, on which the pathologists' judgments were inconsistent; the ORs for AI-assisted TPS methods 1, 2, 3, and 4 were 1.28 (1.06-1.54; P = .012), 1.29 (1.06-1.55; P = .010), 1.28 (1.06-1.54; P = .012), and 1.29 (1.06-1.55; P = .010), respectively, which were statistically significant. For discordant cases, the OR for each AI-assisted TPS method compared with the others was 0.99 to 1.01 and not statistically significant. This study emphasized the usefulness of the AI-assisted system for cases in which pathologists had difficulty determining the PD-L1 TPS.

Patch-Based Cervical Cancer Segmentation using Distance from Boundary of Tissue.

Pathological diagnosis is used for examining cancer in detail, and its automation is in demand. To automatically segment each cancer area, a patch-based approach is usually used since a Whole Slide Image (WSI) is huge. However, this approach loses the global information needed to distinguish between classes. In this paper, we utilized the Distance from the Boundary of tissue (DfB), which is global information that can be extracted from the original image. We experimentally applied our method to the three-class classification of cervical cancer, and found that it improved the total performance compared with the conventional method.

Validation Study of the International Association for the Study of Lung Cancer Histologic Grading System of Invasive Lung Adenocarcinoma.

INTRODUCTION: A histologic grading system for invasive lung adenocarcinoma (ADC) has been proposed by the International Association for the Study of Lung Cancer (IASLC) Pathology Committee in June 2020. This study evaluated the prognostic value of the IASLC histologic grading system (the IASLC system) in a large Japanese cohort. METHODS: We performed comprehensive histologic subtyping using the semiquantitative estimation of five major patterns and complex glandular patterns in patients with a completely resected lung ADC and determined the histologic grade using the IASLC system. Concordance index and receiver-operating characteristic curves were used to evaluate the clinical utility of the IASLC system for recurrence and death; the comparison was performed with the architectural-pattern system (the Arch system) and the grading system on the basis of the two most predominant patterns (the Sica's system). RESULTS: Of 1002 patients with invasive ADC, 235 had recurrent disease and 166 died of lung cancer. The concordance index and area under the curve of the IASLC system were 0.777 and 0.807 for recurrence and 0.767 and 0.776 for death, respectively. These were similar to those of the Arch system (0.763 and 0.796 for recurrence, 0.743 and 0.755 for death) and the Sica's system (0.786 and 0.814 for recurrence, 0.762 and 0.773 for death). CONCLUSIONS: We reported that the IASLC system for invasive lung ADC has prognostic significance by evaluating a large Japanese cohort. We believe that the IASLC grading system will provide physicians with better information for postsurgery treatment.

Prognostic significance of cribriform adenocarcinoma of the lung: validation analysis of 1,057 Japanese patients with resected lung adenocarcinoma and a review of the literature.

BACKGROUND: Cribriform-predominant adenocarcinoma of the lung (Cribri-ADC) is a recently described tumor growth pattern. However, its prognostic impact has not been clearly determined. We analyzed the data of a series of 1,057 Japanese patients with resected lung adenocarcinoma to identify the clinical significance of Cribri-ADC. METHODS: Cribriform pattern (Cribri-p) is defined as invasive back-to-back fused tumor glands with poorly formed glandular spaces or invasive tumor nests comprising tumors cells that produced glandular lumina. We investigated the correlations of Cribri-p and Cribri-ADC with clinicopathological factors as well as disease-free survival (DFS) and overall survival (OS). RESULTS: Cribri-p was present in 217 patients (20.5%) and Cribri-ADC was determined in 25 patients (2.4%). Cribri-p was associated with larger tumor size, pleural invasion, vascular invasion, lymphatic invasion, and spreading through air spaces (STAS) (all, P<0.0001). Cribri-ADC was associated with younger age (P=0.019), vascular invasion (P=0.0025), STAS (P<0.0001), and ALK rearrangement (P=0.012). The DFS curve of patients with Cribri-ADC was identical to that of patients with solid adenocarcinoma; however, the OS curve was located between that of patients with papillary and acinar adenocarcinoma. Of the 10 patients who had tumor recurrences, eight had EGFR mutations or ALK rearrangement, six of whom achieved relatively long survival (median, 64.6, range, 37.4-113 months) following treatment with tyrosine kinase inhibitors (TKIs). In multivariate analysis, Cribri-ADC was not an independent prognostic factor of either recurrence or death. CONCLUSIONS: Cribri-ADC is associated with a higher risk of recurrence; however, most patients can be successfully treated with TKIs.

Large nest micropapillary pattern of lung adenocarcinoma has poorer prognosis than typical floret pattern: analysis of 1,062 resected tumors.

BACKGROUND: A micropapillary pattern (MP-p) is related to poor prognosis in patients with lung adenocarcinoma (L-ADC). In 2015, the WHO defined the MP-p as "papillary tufts forming florets that lack fibrovascular cores and may appear detached from alveolar walls"; however, the sizes of tumor clusters in air space were not mentioned in this classification. METHODS: We evaluated the MP-p dividing the cluster sizes in the air space by reviewing 1,062 cases of resected L-ADCs. We classified MP-p into two types according to cluster size as follows: typical floret MP-p, tumors with small-to-medium-sized clusters (1-20 tumor cells); and large nest MP-p, tumors with large-sized clusters (>20 tumor cells, large nest). We then recorded the frequency of each type and investigated the association between the MP-p type and clinicopathological factors. RESULTS: Twenty-nine percent of L-ADCs (n=308) were MP-p-positive. Typical floret MP-p and large nest MP-p were observed in 244 tumors (22.9%) and 64 tumors (6.0%), respectively. Only 7 additional micropapillary ADCs were detected when we reclassified ADCs in addition to large nest MP-p. Tumors with large nest MP-p showed the highest frequency of node metastasis and worse prognosis compared to those with typical floret MP-p and absent (P<0.001). In multivariate analysis, patients with L-ADC with typical floret MP-p and large nest MP-p showed a higher recurrence rate [hazard ratio (HR): 1.762 (type 1 vs. absent), HR: 2.450 (type 2 vs. absent)]. CONCLUSIONS: Large nest MP-p should be included in the original MP-p and recorded separately.

Uterine cervical squamous cell carcinoma without p16 (CDKN2A) expression: Heterogeneous causes of an unusual immunophenotype.

Immunohistochemically p16 (CDKN2A)-negative uterine cervical squamous cell carcinoma (SCC) is uncommon, and there are few reports about its pathological features. This study explored the causes of p16 negativity in such cases. We analyzed diagnostic tissue samples of five cases of p16-negative cervical SCC among 107 patients who underwent hysterectomy at Kyoto University Hospital between January 2010 and December 2015. The samples were subjected to immunohistochemical staining, in situ hybridization and a genetic analysis. Two of five cases were positive for human papilloma virus (HPV) by genotyping. One was positive for HPV56 with promoter hypermethylation of CDKN2A and co-existing Epstein-Barr virus infection. Another was positive for HPV6 categorized as low-risk HPV with condylomatous morphology. Among the remaining three cases, one had amplification of the L1 gene of HPV with promoter hypermethylation of CDKN2A and TP53 mutation, and one of the other two HPV-negative cases had a homozygous CDKN2A deletion, while the other was positive for p53 and CK7. p16-negativity of cervical SCC is often associated with an unusual virus infection status and CDKN2A gene abnormality.

Discohesive growth pattern (Disco-p) as an unfavorable prognostic factor in lung adenocarcinoma: an analysis of 1062 Japanese patients with resected lung adenocarcinoma.

Discohesive growth pattern (Disco-p) is often observed in lung adenocarcinoma (ADC) and mimics tumor budding (TB), stromal invasive-type micropapillary pattern (SMPP), and complex glandular pattern. However, the clinical impact of Disco-p in lung ADC has not been well studied. To investigate the prognostic significance of Disco-p, we analyzed 1062 Japanese patients with resected lung ADC. Disco-p was defined as an invasive growth pattern composed of single tumor cells, or trabeculae or small nests of tumor cells associated with desmoplastic fibrous stroma. We recorded the percentage of Disco-p in 5% increments independent of the major histologic pattern and investigated its correlation with different clinicopathological factors. We also analyzed the overall survival (OS) and disease-free survival (DFS). Disco-p was observed in 203 tumors (19.1%). Disco-p was significantly associated with male sex, smoking, lymph node metastasis, large tumor size, high TNM stage, lymphovascular and pleural invasion, spread through air spaces, and TB (all, p < 0.001). Of the total cases, only eight cases exhibited a dubious pattern between SMPP and Disco-p. Disco-p was also associated with wild-type EGFR (p < 0.001) and ALK fusion (p = 0.008). Patients harboring tumors with Disco-p had significantly worse prognoses (OS and DFS (both, p < 0.001)) compared with those without Disco-p. On multivariate analysis, Disco-p was an independent prognostic factor of worse OS (hazard ratio (HR), 2.572; 95% confidence interval (CI), 1.789-3.680; p < 0.001), and DFS (HR, 3.413; 95% CI, 2.482-4.683; p < 0.001), whereas TB was not an independent unfavorable prognostic factor. Disco-p was an independent unfavorable prognostic factor in patients with resected lung ADC, although a careful evaluation is necessary to distinguish it from similar patterns. We proposed that Disco-p should be recognized as a new invasive pattern and accurately recorded for the better management of patients with lung ADCs.

GATA6-positive lung adenocarcinomas are associated with invasive mucinous adenocarcinoma morphology, hepatocyte nuclear factor 4α expression, and KRAS mutations.

AIMS: GATA6 is known to play a role in lung development. However, its role in the carcinogenesis of lung cancer is not well studied. The aim of this study was to analyse GATA6 expression in lung adenocarcinomas (LAs) by immunohistochemistry (IHC) in order to define its association with clinicopathological characteristics. METHODS AND RESULTS: IHC analysis of GATA6 was performed with tissue microarray slides containing 348 LAs. The association between GATA6 expression and clinicopathological parameters was evaluated. GATA6 expression in epithelial tumours other than lung cancer was also evaluated. GATA6 expression was found in 47 LAs (13.5%). This occurred more frequently in younger patients (P = 0.005), and was associated with the absence of lymph node metastasis (P =0.024), well-differentiated to moderately differentiated tumours (P < 0.001), the absence of lymphatic invasion (P = 0.020), and the absence of vascular invasion (P = 0.011). GATA6 expression was associated with mucin production (P < 0.001), the invasive mucinous adenocarcinoma subtype (P < 0.001), KRAS mutations (P = 0.026), expression of MUC2 (P < 0.001), CDX2 (P = 0.049), and MUC5AC (P < 0.001), and absence of expression of TTF-1 (P = 0.002). GATA6 expression was also associated with hepatocyte nuclear factor 4α (HNF4α) expression (P < 0.001). GATA6 expression tended to indicate better prognoses, whereas patients with HNF4α expression had significantly worse prognoses (P = 0.033). Of 270 tumours other than lung cancer, 110 expressed GATA6. CONCLUSIONS: These findings suggest that GATA6 might interact with HNF4α and contribute to the development of mucinous-type LAs.

Killer cell immunoglobulin-like receptor 2DL4 is expressed in and suppresses the cell growth of Langerhans cell histiocytosis.

Killer cell immunoglobulin-like receptor (KIR) 2DL4 (CD158d) is a receptor for human leukocyte antigen-G. The function of KIR2DL4 has been reported in human natural killer cell lymphoma and mastocytosis, but not in Langerhans cell histiocytosis (LCH). Herein, we examined the expression and function of KIR2DL4 in LCHs. In pathological specimens, 27 of 36 LCH cases (75.0%) were immunohistochemically positive for KIR2DL4. Its expression was independent of age, gender, location, multi- or single-system, and the status of BRAFV600E immunostaining. We also confirmed the expression of KIR2DL4 mRNA and protein in the human LCH-like cell lines ELD-1 and PRU-1. KIR2DL4 protein was distributed in the membrane and cytoplasm of ELD-1 cells, but only in the cytoplasm of PRU-1 cells. An agonistic antibody against KIR2DL4 reduced phosphorylation of extracellular signal-regulated kinases (ERKs) and suppressed the cell growth of ELD-1 cells in a Src homology region 2 domain-containing phosphatase-2 dependent manner, but it had no effect in PRU-1 cells. These results suggest that KIR2DL4-mediated ERK suppression is a possible therapeutic target for LCH cells.

Evaluating the effectiveness of RNA in-situ hybridization for detecting lung adenocarcinoma with anaplastic lymphoma kinase rearrangement.

AIMS: An easy and rapid assay for detecting mRNA in formalin-fixed paraffin-embedded samples [RNA in-situ hybridization (ISH)] has been reported recently. The aim of this study was to investigate the diagnostic accuracy of RNA ISH in detecting lung adenocarcinoma (LA) with anaplastic lymphoma kinase (ALK) gene rearrangement. METHODS AND RESULTS: We tested ALK RNA ISH on 11 resected LAs for which ALK fusion was confirmed by immunohistochemistry (IHC) and/or fluorescence in-situ hybridization (FISH). ALK mRNA expression was detected by RNA ISH in all 11 ALK-positive LAs, with a mean positive cell proportion of 68.4% (median, 75.3%; range, 3-98.8%), by counting 100 tumour cells at 10 different loci; RNA ISH did not detect ALK mRNA expression in the normal surrounding lung cells. Next, we explored the concordance between ALK RNA ISH and IHC/FISH tests by using tissue microarrays (TMAs) containing 294 LAs. In the TMA slides, we found five ALK-positive cases with IHC and/or FISH. The mean proportion of ALK RNA ISH-positive cells in these five cases was 75.6% (median, 82%; range, 40-94%), whereas the proportion of ALK RNA ISH-positive cells in the remaining 289 cases was 0.3% (median 0%; range, 0-15%). When the cutoff value was set at 15%, ALK RNA ISH-positive and ALK RNA ISH-negative cases were distinguishable with 100% sensitivity and specificity relative to the IHC/FISH tests. CONCLUSIONS: Our findings show that RNA ISH is useful for detecting ALK rearrangement with high sensitivity and specificity relative to conventional IHC/FISH tests. Thus, RNA ISH, which is an easy and rapid assay, could be an alternative method to IHC and FISH.

Lung Adenocarcinoma With MUC4 Expression Is Associated With Smoking Status, HER2 Protein Expression, and Poor Prognosis: Clinicopathologic Analysis of 338 Cases.

BACKGROUND: MUC4 is a transmembrane glycoprotein that plays a role in the cell growth signaling pathway and has been studied in various organ malignancies. This study aimed to analyze MUC4 expression in resected lung adenocarcinomas (ADCs) to define the clinicopathologic characteristics of MUC4-positive cancers. PATIENTS AND METHODS: Immunohistochemical MUC4 analysis was performed using tissue microarray slides containing 338 lung ADCs. Associations between MUC4 expression and the following clinicopathologic parameters were evaluated: sex; age; smoking status; tumor stage; tumor grade; lymphovascular invasion; pleural invasion; TTF-1 and HNF4α expression; EGFR, KRAS, BRAF, and HER2 mutation status; and ALK and ROS1 fusion status. RESULTS: Ninety-four tumors (27.8%) were MUC4 positive. Most patients with MUC4-positive tumors were male (P < .001) and smokers (P = .006). Moreover, MUC4 expression was significantly associated with solid ADCs (P < .001) and vascular invasion (P = .001). MUC4 expression inversely correlated with TTF-1 expression (P = .020) and EGFR mutations (P = .004). Interestingly, MUC4 expression correlated with HER2 protein expression (P = .042), although MUC4 expression did not correlate with HER2 DNA amplification or HER2 gene mutations. Patients with MUC4-positive tumors had significantly worse prognoses compared to patients with MUC4-negative tumors (P = .025). CONCLUSION: The present study showed that MUC4-positive lung ADCs correlated with male smokers, solid ADCs, negative TTF-1 expression, the EGFR wild-type gene, HER2 protein expression, and poorer prognoses. These results suggest that MUC4-positive lung ADC may be a distinct subtype found in patients with smoking-related poor outcomes, mediated by HER2 signaling pathway.