Advancing Ki67 hotspot detection in breast cancer: a comparative analysis of automated digital image analysis algorithms.

AIM: Manual detection and scoring of Ki67 hotspots is difficult and prone to variability, limiting its clinical utility. Automated hotspot detection and scoring by digital image analysis (DIA) could improve the assessment of the Ki67 hotspot proliferation index (PI). This study compared the clinical performance of Ki67 hotspot detection and scoring DIA algorithms based on virtual dual staining (VDS) and deep learning (DL) with manual Ki67 hotspot PI assessment. METHODS: Tissue sections of 135 consecutive invasive breast carcinomas were immunohistochemically stained for Ki67. Two DIA algorithms, based on VDS and DL, automatically determined the Ki67 hotspot PI. For manual assessment; two independent observers detected hotspots and calculated scores using a validated scoring protocol. RESULTS: Automated hotspot detection and assessment by VDS and DL could be performed in 73% and 100% of the cases, respectively. Automated hotspot detection by VDS and DL led to higher Ki67 hotspot PIs (mean 39.6% and 38.3%, respectively) compared to manual consensus Ki67 PIs (mean 28.8%). Comparing manual consensus Ki67 PIs with VDS Ki67 PIs revealed substantial correlation (r = 0.90), while manual consensus versus DL Ki67 PIs demonstrated high correlation (r = 0.95). CONCLUSION: Automated Ki67 hotspot detection and analysis correlated strongly with manual Ki67 assessment and provided higher PIs compared to manual assessment. The DL-based algorithm outperformed the VDS-based algorithm in clinical applicability, because it did not depend on virtual alignment of slides and correlated stronger with manual scores. Use of a DL-based algorithm may allow clearer Ki67 PI cutoff values, thereby improving the clinical usability of Ki67.

VEGF, EGFR and PSMA as possible imaging targets of lymph node metastases of urothelial carcinoma of the bladder.

BACKGROUND: In this study we investigated the expression of vascular endothelial growth factor (VEGF), epidermal growth factor receptor (EGFR) and prostate-specific membrane antigen (PSMA) to analyze their potency as targets for the detection of lymph node (LN) metastases of urothelial carcinoma of the bladder. METHODS: Antigen expression was determined in 40 samples with urothelial carcinoma and compared to 17 matched samples without metastases by immunohistochemistry. The total immunostaining score (TIS 0-12) was determined as the product of a proportion score (PS 0-4) and intensity score (IS 0-3). RESULTS: VEGF expression was high in primary tumor and LN metastases (median TIS 8 in both) and VEGF expression was also seen in LNs without metastases (median TIS 6). EGFR expression was low in primary tumor and LN metastases (median TIS 3 and 2 respectively) and absent in LNs without metastases. PSMA expression was low in samples with urothelial carcinoma (median TIS 2). CONCLUSION: VEGF shows moderate to high expression levels in both primary tumors and LN metastases and could be a candidate as a target agent for imaging modalities of urothelial carcinoma. EGFR and PSMA do show low staining levels in tumor tissue with urothelial carcinoma and do not seem suitable as target agents. TRIAL REGISTRATION: The Medical Ethics Review Board of the University Medical Center Groningen approved this study on 14 December 2017 (METc UMCG 2017/639). Trial registration number (UMCG Research Register): 201700868.

Systematically higher Ki67 scores on core biopsy samples compared to corresponding resection specimen in breast cancer: a multi-operator and multi-institutional study.

Ki67 has potential clinical importance in breast cancer but has yet to see broad acceptance due to inter-laboratory variability. Here we tested an open source and calibrated automated digital image analysis (DIA) platform to: (i) investigate the comparability of Ki67 measurement across corresponding core biopsy and resection specimen cases, and (ii) assess section to section differences in Ki67 scoring. Two sets of 60 previously stained slides containing 30 core-cut biopsy and 30 corresponding resection specimens from 30 estrogen receptor-positive breast cancer patients were sent to 17 participating labs for automated assessment of average Ki67 expression. The blocks were centrally cut and immunohistochemically (IHC) stained for Ki67 (MIB-1 antibody). The QuPath platform was used to evaluate tumoral Ki67 expression. Calibration of the DIA method was performed as in published studies. A guideline for building an automated Ki67 scoring algorithm was sent to participating labs. Very high correlation and no systematic error (p = 0.08) was found between consecutive Ki67 IHC sections. Ki67 scores were higher for core biopsy slides compared to paired whole sections from resections (p ≤ 0.001; median difference: 5.31%). The systematic discrepancy between core biopsy and corresponding whole sections was likely due to pre-analytical factors (tissue handling, fixation). Therefore, Ki67 IHC should be tested on core biopsy samples to best reflect the biological status of the tumor.

Immune checkpoint inhibitor treatment induces colitis with heavy infiltration of CD8 + T cells and an infiltration pattern that resembles ulcerative colitis.

Colitis is a common, but poorly understood, adverse event of immune checkpoint inhibitors that are standard-of-care for an expanding range of cancer types. This explorative study aimed to describe the immune infiltrates in the colon from individuals developing checkpoint inhibitor colitis and compare them to well-known immunophenotypes of acute graft-versus-host disease, ulcerative colitis, and Crohn's disease. Colon biopsies (n = 20 per group) of patients with checkpoint inhibitor colitis, acute graft-versus-host disease, ulcerative colitis and Crohn's disease, all colitis treatment-naïve, and of individuals with a normal colon were analyzed using immunohistochemistry: CD8 for cytotoxic T cells, CD4 for T helper cells, and CD68 to identify cells of macrophage lineage. CD8 + T cell, CD4 + T cell, and CD68 + cell counts were performed. Cell infiltration was scored as scattered/patchy or band-like in the superficial and deep gut mucosa. Checkpoint inhibitor colitis was found to be heavily infiltrated by CD8 + T cells. Comparative analysis between groups showed that both CD8 + T cell counts (P < 0.01) and immune cell infiltration patterns in checkpoint inhibitor colitis were most similar to those observed in ulcerative colitis, with a deep band-like CD4 + T cell infiltration pattern and a superficial band-like CD68 + cell infiltration pattern in both. In conclusion, this is the first immunohistopathological study comparing infiltrate characteristics of checkpoint inhibitor colitis, acute graft-versus-host disease, ulcerative colitis, and Crohn's disease. Checkpoint inhibitor colitis samples are heterogeneous, heavily infiltrated by CD8 + T cells, and show an immune cell infiltration pattern that is more similar to ulcerative colitis than to colonic acute graft-versus-host disease or colonic Crohn's disease.

Potential Receptors for Targeted Imaging of Lymph Node Metastases in Penile Cancer.

Imaging modalities using tumor-directed monoclonal antibodies may be of value to improve the pre- and intraoperative detection and resection of lymph node (LN) metastatic disease in penile squamous cell carcinoma (PSCC). We investigated the expression of prostate-specific membrane antigen (PSMA), vascular endothelial growth factor (VEGF), epidermal growth factor receptor (EGFR) and epithelial cell adhesion molecule (EpCAM) to analyze their potency for diagnostic applications. Antigen expression was determined in primary tumors and LNs with and without metastases of 22 patients with PSCC. The total immunostaining score (TIS, 0-12) was determined as the product of a proportion score (PS, 0-4) and an intensity score (IS, 0-3). EGFR and VEGF expression were high in primary tumor (median TIS 8) and LN metastases (median TIS 6 and 8, respectively). No EGFR expression was seen in LNs without metastases. However, LNs without metastases did show VEGF expression (median TIS 6). No EpCAM or PSMA expression was seen in PSCC. This study shows that VEGF and EGFR expression is moderate to high in LN metastases of PSCC. Both VEGF and EGFR warrant further clinical evaluation to determine their value as a target for pre- and intraoperative imaging modalities in the detection of LN metastases in PSCC.

What is the added value of digital image analysis of HER2 immunohistochemistry in breast cancer in clinical practice? A study with multiple platforms.

AIMS: We aimed to compare digital image analysis (DIA) of human epidermal growth factor receptor 2 (HER2) immunohistochemistry (IHC) in breast cancer by two platforms: (i) to validate DIA against standard diagnostics; and (ii) to evaluate the added value of DIA in clinical practice. METHODS AND RESULTS: HER2 IHC and in-situ hybridisation (ISH) were performed on 152 consecutive invasive breast carcinomas. IHC scores were determined with DIA using two independent platforms. Manual scoring was performed by two independent observers. HER2 status was considered positive in 3+ and ISH-positive 2+ cases. HER2 status using DIA was compared to HER2 status with standard diagnostics (manual scoring with ISH in 2+ cases). Interplatform agreement of IHC scores was 'moderate' (linear weighted κ = 0.58), agreement between manual scoring and platform A was 'moderate' (κ = 0.60) and between manual scoring and platform B 'almost perfect' (κ = 0.85). Compared to manual scoring, DIA resulted in a reduction of 2+ cases from 17.1 to 1.3% with platform A and from 17.1 to 15.8% with platform B. However, compared to standard diagnostics, there were three false-negative cases with DIA using platform A [81.3% sensitivity, 100% specificity, 100% positive predictive value (PPV), 97.8% negative predictive value (NPV)]. Sensitivity, specificity, PPV and NPV were 100% with DIA using platform B. CONCLUSIONS: DIA of HER2 IHC is a valid tool in determining HER2 status in breast carcinoma. Algorithms in different platforms can behave differently, and optimal calibration is essential. In clinical practice, DIA offers an objective alternative to manual scoring, but a reduction in 2+ cases could result in loss of sensitivity.

Digital image analysis of Ki67 proliferation index in breast cancer using virtual dual staining on whole tissue sections: clinical validation and inter-platform agreement.

PURPOSE: The Ki67 proliferation index is a prognostic and predictive marker in breast cancer. Manual scoring is prone to inter- and intra-observer variability. The aims of this study were to clinically validate digital image analysis (DIA) of Ki67 using virtual dual staining (VDS) on whole tissue sections and to assess inter-platform agreement between two independent DIA platforms. METHODS: Serial whole tissue sections of 154 consecutive invasive breast carcinomas were stained for Ki67 and cytokeratin 8/18 with immunohistochemistry in a clinical setting. Ki67 proliferation index was determined using two independent DIA platforms, implementing VDS to identify tumor tissue. Manual Ki67 score was determined using a standardized manual counting protocol. Inter-observer agreement between manual and DIA scores and inter-platform agreement between both DIA platforms were determined and calculated using Spearman's correlation coefficients. Correlations and agreement were assessed with scatterplots and Bland-Altman plots. RESULTS: Spearman's correlation coefficients were 0.94 (p < 0.001) for inter-observer agreement between manual counting and platform A, 0.93 (p < 0.001) between manual counting and platform B, and 0.96 (p < 0.001) for inter-platform agreement. Scatterplots and Bland-Altman plots revealed no skewness within specific data ranges. In the few cases with ≥ 10% difference between manual counting and DIA, results by both platforms were similar. CONCLUSIONS: DIA using VDS is an accurate method to determine the Ki67 proliferation index in breast cancer, as an alternative to manual scoring of whole sections in clinical practice. Inter-platform agreement between two different DIA platforms was excellent, suggesting vendor-independent clinical implementability.

Digital image analysis of HER2 immunohistochemistry in gastric- and oesophageal adenocarcinoma: a validation study on biopsies and surgical specimens.

AIMS: To test the validity of diagnostics incorporating digital image analysis (DIA) for human epidermal growth factor 2 (HER2) immunohistochemistry (IHC) in gastro-oesophageal adenocarcinomas, as an alternative to current standard diagnostics using manual scoring. METHODS AND RESULTS: We included 319 consecutive gastro-oesophageal adenocarcinomas (232 biopsies and 87 surgical specimens). DIA was applied to determine HER2 IHC classification, using both standard breast cancer (BC) and modified gastro-oesophageal cancer (GEC) cut-offs. Consensus manual scores were established by four independent observers. Chromogenic in-situ hybridization (CISH) was performed on all 2+ cases by manual scoring, DIA or both. HER2 status was considered positive in 3+ and CISH-positive 2+ cases. Overall agreement between DIA and consensus manual scores was 76.5% (weighted κ = 0.66, BC cut-offs) and 85.6% (weighted κ = 0.80, GEC cut-offs). Agreement was similar for biopsies and surgical specimens. All disagreement occurred in the manual IHC equivocal cases. DIA resulted in a reduction of 2+ cases: 75.8% with BC cut-offs and 46.5% with GEC cut-offs. HER2 status was positive in 48 cases (15%) with standard diagnostics and DIA using GEC cut-offs, and 46 cases (14.4%) using BC cut-offs (all with CISH in 2+ cases). Considering standard diagnostics as a reference, DIA showed 93.8% sensitivity and 99.6% specificity (BC cut-offs) or 97.9% sensitivity and 99.6% specificity (GEC cut-offs). CONCLUSIONS: DIA is a reliable and feasible alternative to manual HER2 IHC scoring in gastro-oesophageal adenocarcinoma, both in biopsies and surgical specimens, leading to a reduction of 2+ cases for which subsequent ISH testing is required.

Comparing the Cervista HPV HR test and Hybrid Capture 2 assay in a Dutch screening population: improved specificity of the Cervista HPV HR test by changing the cut-off.

The diagnostic performance of the widely-used Cervista HPV HR test was compared to the Hybrid Capture 2 (HC2) test in a Dutch population-based cervical cancer screening program. In 900 scrapings of women with normal cytomorphology, specificity was 90% (95%CI: 87.84-91.87) for the Cervista HPV HR test and 96% (95%CI: 94.76-97.37) for the HC2 test with 93% agreement between both tests (κ = 0.5, p<0.001). The sensitivity for CIN2+ using 65 scrapings of women with histological-confirmed CIN2+ was 91% (95%CI: 80.97-96.51) for the Cervista HPV HR test and 92% (95%CI: 82.94-97.43) for the HC2 test with 95% agreement between both tests (κ = 0.7, p<0.001). Fifty-seven of 60 HC2 negative/Cervista positive cases tested HPV-negative with PCR-based HPV assays; of these cases 56% were defined as Cervista triple-positive with FOZ values in all 3 mixes higher than the second cut-off of 1.93 (as set by manufacturer). By setting this cut-off at 5.0, specificity improved significantly without affecting sensitivity. External validation of this new cut-off at 5.0 in triple-positive scrapings of women selected from the SHENCCASTII database revealed that 22/24 histological normal cases now tested HPV-negative in the Cervista HPV HR test, while CIN2+ lesions remained HPV-positive. The intra-laboratory reproducibility of the Cervista HPV HR test (n = 510) showed a concordance of 92% and 93% for cut-off 1.93 and 5.0 (κ = 0.83 and κ = 0.84, p<0.001) and inter-laboratory agreement of the Cervista HPV HR test was 90% and 93% for cut-off 1.93 and 5.0 (κ = 0.80 and κ = 0.85, p<0.001). In conclusion, the specificity of the Cervista HPV HR test could be improved significantly by increasing the second cut-off from 1.93 to 5.0, without affecting the sensitivity of the test in a population-based screening setting.

Increased protein aggregation in Zucker diabetic fatty rat brain: identification of key mechanistic targets and the therapeutic application of hydrogen sulfide.

BACKGROUND: Diabetes and particularly high blood glucose levels are implicated in neurodegeneration. One of the hallmarks of neurodegeneration is protein aggregation. We investigated the presence of protein aggregation in the frontal brain of Zucker diabetic fatty (ZDF) rats, an animal model for diabetes. Further, the effect of NaHS in suppressing protein aggregation in cultured brain slices from ZDF was assessed. RESULTS: The levels of protein synthesis, protein/gene expression, autophagy and anti-oxidant defense were evaluated in ZDF and control (Lean) brains.Compared to Lean, ZDF brains displayed a significant increase in protein aggregates, p-tau, fibronectin expression and protein glycosylation. Increased phosphorylation of mTOR and S6 ribosomal protein in ZDF indicated higher protein synthesis, while the increase in ubiquitinated proteins and LC3-I in ZDF brains accompanied by lower LC3-II expression and LC3-II/LC3-I levels indicated the blockage of proteolytic pathways. CBS (cystathionine beta synthase) protein and mRNA expression and thiol group levels in ZDF brains were lower compared to Lean. ZDF brains show a higher level of reactive oxygen species. In vitro NaHS treatment normalized proteostasis while counteracting oxidative stress. CONCLUSION: Our data demonstrate increased protein synthesis and aggregation in the diabetic ZDF rat brain, which was reversible by NaHS treatment.This is the first report on the potential use of NaHS as a novel strategy against protein aggregation in diabetic brain.

Methylation markers for CCNA1 and C13ORF18 are strongly associated with high-grade cervical intraepithelial neoplasia and cervical cancer in cervical scrapings.

PURPOSE: Recently, we reported 13 possible cervical cancer-specific methylated biomarkers identified by pharmacologic unmasking microarray in combination with large-genome computational screening. The aim of the present study was to perform an in-depth analysis of the methylation patterns of these 13 candidate genes in cervical neoplasia and to determine their diagnostic relevance. EXPERIMENTAL DESIGN AND RESULTS: Five of the 13 gene promoters (C13ORF18, CCNA1, TFPI2, C1ORF166, and NPTX1) were found to be more frequently methylated in frozen cervical cancer compared with normal cervix specimens. Quantitative methylation analysis for these five markers revealed that both CCNA1 and C13ORF18 were methylated in 68 of 97 cervical scrapings from cervical cancer patients and in only 5 and 3 scrapings, respectively, from 103 healthy controls (P < 0.0005). In cervical scrapings from patients referred with an abnormal Pap smear, CCNA1 and C13ORF18 were methylated in 2 of 43 and 0 of 43 CIN 0 (no cervical intraepithelial neoplasia) and in 1 of 41 and 0 of 41 CIN I, respectively. Furthermore, 8 of 43 CIN II, 22 of 43 CIN III, and 3 of 3 microinvasive cancer patients were positive for both markers. Although sensitivity for CIN II or higher (for both markers 37%) was low, specificity (96% and 100%, respectively) and positive predictive value (92% and 100%, respectively) were high. CONCLUSION: Methylation of CCNA1 and C13ORF18 in cervical scrapings is strongly associated with CIN II or higher-grade lesions. Therefore, these markers might be used for direct referral to gynecologists for patients with a methylation-positive scraping.

Assessment of gene promoter hypermethylation for detection of cervical neoplasia.

Current cervical cancer screening is based on morphological assessment of Pap smears and associated with significant false negative and false positive results. Previously, we have shown that detection of hypermethylated genes in cervical scrapings using quantitative methylation-specific PCR (QMSP) is a promising tool for identification of squamous cell cervical cancer. Aim of the present pilot-study was to evaluate presence of hypermethylated genes in cervical carcinogenesis, both in squamous cell as well as adenocarcinomas. Cervical scrapings were obtained from 30 patients diagnosed with cervical cancer (20 squamous cell carcinomas and 10 adenocarcinomas) and 19 women with histologically normal cervices. The scraped cells were used for determination of promoter hypermethylation by QMSP for 12 genes and for morphological assessment. Overall, CALCA, DAPK, ESR1, TIMP3, APC and RAR-beta2 promoters were significantly more often hypermethylated in cancers than in controls, while adenocarcinomas were more often hypermethylated above the highest control ratio for APC, TIMP3 and RASSF1A promoters. Combining 4 genes (CALCA, DAPK, ESR1 and APC) yielded a sensitivity of 89% (with all adenocarcinomas identified), equal to cytomorphology (89%) and high-risk human papilloma virus (Hr-HPV; 90%). The 4-gene QMSP proved theoretically superior to cytomorphology as well as Hr-HPV in specificity (100% vs. 83 and 68%, respectively), because cytology identified 3 controls as moderate or severe dyskaryosis and 6 controls were positive for Hr-HPV. In conclusions, QMSP of 4 gene promoters combined appears to have comparable sensitivity and potentially better specificity in comparison to "classic" cytomorphological assessment and Hr-HPV detection. QMSP holds promise as a new diagnostic tool for both squamous cell carcinoma and adenocarcinoma of the cervix.

Detecting cervical cancer by quantitative promoter hypermethylation assay on cervical scrapings: a feasibility study.

Current morphology-based cervical cancer screening is associated with significant false-positive and false-negative results. Tumor suppressor gene hypermethylation is frequently present in cervical cancer. It is unknown whether a cervical scraping reflects the methylation status of the underlying epithelium, and it is therefore unclear whether quantitative hypermethylation specific PCR (QMSP) on cervical scrapings could be used as a future screening method augmenting the current approach. Cervical scrapings and paired fresh frozen cervical tissue samples were obtained from 53 cervical cancer patients and 45 controls. All scrapings were morphologically scored and analyzed with QMSP for the genes APC, DAPK, MGMT, and GSTP1. To adjust for DNA input, hypermethylation ratios were calculated against DNA levels of a reference gene. Hypermethylation ratios of paired fresh frozen tissue samples and scrapings of cervical cancer patients and controls were strongly related (Spearman correlation coefficient, 0.80 for APC, 0.98 for DAPK, and 0.83 for MGMT; P < 0.001). More cervical cancer patients than controls were DAPK positive (P < 0.001). When cutoff levels for ratios were defined to be above the highest ratio observed in controls, QMSP in cervical scrapings identified 32 (67%) of 48 cervical cancer patients. This feasibility study demonstrates that QMSP on cervical scrapings holds promise as a new diagnostic tool for cervical cancer. The addition of more genes specifically methylated in cervical cancer will further improve the assay.