Evaluation of AI-based Gleason grading algorithms "in the wild".

The biopsy Gleason score is an important prognostic marker for prostate cancer patients. It is, however, subject to substantial variability among pathologists. Artificial intelligence (AI)-based algorithms employing deep learning have shown their ability to match pathologists' performance in assigning Gleason scores, with the potential to enhance pathologists' grading accuracy. The performance of Gleason AI algorithms in research is mostly reported on common benchmark datasets or within public challenges. In contrast, many commercial algorithms are evaluated in clinical studies, for which data is not publicly released. As commercial AI vendors typically do not publish performance on public benchmarks, comparison between research and commercial AI is difficult. The aim of this study is to evaluate and compare the performance of top-ranked public and commercial algorithms using real-world data. We curated a diverse dataset of whole-slide prostate biopsy images through crowdsourcing, containing images with a range of Gleason scores and from diverse sources. Predictions were obtained from five top-ranked public algorithms from the PANDA challenge and from two commercial Gleason grading algorithms. Additionally, ten pathologists evaluated the data set in a reader study. Overall, the pairwise quadratic weighted kappa among pathologists ranged from 0.777 to 0.916. Both public and commercial algorithms showed high agreement with pathologists, with quadratic kappa ranging from 0.617 to 0.900. Commercial algorithms performed on par or outperformed top public algorithms.

Topographic distribution of inflammation factors in a healing aneurysm.

BACKGROUND: Healing of intracranial aneurysms following endovascular treatment relies on the organization of early thrombus into mature scar tissue and neointima formation. Activation and deactivation of the inflammation cascade plays an important role in this process. In addition to timely evolution, its topographic distribution is hypothesized to be crucial for successful aneurysm healing. METHODS: Decellularized saccular sidewall aneurysms were created in Lewis rats and coiled. At follow-up (after 3 days (n = 16); 7 days (n = 19); 21 days (n = 8)), aneurysms were harvested and assessed for healing status. In situ hybridization was performed for soluble inflammatory markers (IL6, MMP2, MMP9, TNF-α, FGF23, VEGF), and immunohistochemical analysis to visualize inflammatory cells (CD45, CD3, CD20, CD31, CD163, HLA-DR). These markers were specifically documented for five regions of interest: aneurysm neck, dome, neointima, thrombus, and adjacent vessel wall. RESULTS: Coiled aneurysms showed enhanced patterns of thrombus organization and neointima formation, whereas those without treatment demonstrated heterogeneous patterns of thrombosis, thrombus recanalization, and aneurysm growth (p = 0.02). In coiled aneurysms, inflammation markers tended to accumulate inside the thrombus and in the neointima (p < 0.001). Endothelial cells accumulated directly in the neointima (p < 0.0001), and their presence was associated with complete aneurysm healing. CONCLUSION: The presence of proinflammatory cells plays a crucial role in aneurysm remodeling after coiling. Whereas thrombus organization is hallmarked by a pronounced intra-thrombotic inflammatory reaction, neointima maturation is characterized by direct invasion of endothelial cells. Knowledge concerning topographic distribution of regenerative inflammatory processes may pave the way for future treatment modalities which enhance aneurysm healing after endovascular therapy.

Pathologist Computer-Aided Diagnostic Scoring of Tumor Cell Fraction: A Swiss National Study.

Tumor cell fraction (TCF) estimation is a common clinical task with well-established large interobserver variability. It thus provides an ideal test bed to evaluate potential impacts of employing a tumor cell fraction computer-aided diagnostic (TCFCAD) tool to support pathologists' evaluation. During a National Slide Seminar event, pathologists (n = 69) were asked to visually estimate TCF in 10 regions of interest (ROIs) from hematoxylin and eosin colorectal cancer images intentionally curated for diverse tissue compositions, cellularity, and stain intensities. Next, they re-evaluated the same ROIs while being provided a TCFCAD-created overlay highlighting predicted tumor vs nontumor cells, together with the corresponding TCF percentage. Participants also reported confidence levels in their assessments using a 5-tier scale, indicating no confidence to high confidence, respectively. The TCF ground truth (GT) was defined by manual cell-counting by experts. When assisted, interobserver variability significantly decreased, showing estimates converging to the GT. This improvement remained even when TCFCAD predictions deviated slightly from the GT. The standard deviation (SD) of the estimated TCF to the GT across ROIs was 9.9% vs 5.8% with TCFCAD (P < .0001). The intraclass correlation coefficient increased from 0.8 to 0.93 (95% CI, 0.65-0.93 vs 0.86-0.98), and pathologists stated feeling more confident when aided (3.67 ± 0.81 vs 4.17 ± 0.82 with the computer-aided diagnostic [CAD] tool). TCFCAD estimation support demonstrated improved scoring accuracy, interpathologist agreement, and scoring confidence. Interestingly, pathologists also expressed more willingness to use such a CAD tool at the end of the survey, highlighting the importance of training/education to increase adoption of CAD systems.

Effects of lenalidomide on the bone marrow microenvironment in acute myeloid leukemia: Translational analysis of the HOVON103 AML/SAKK30/10 Swiss trial cohort.

This translational study aimed at gaining insight into the effects of lenalidomide in acute myeloid leukemia (AML). Forty-one AML patients aged 66 or older of the Swiss cohort of the HOVON-103 AML/SAKK30/10 study were included. After randomization, they received standard induction chemotherapy with or without lenalidomide. Bone marrow biopsies at diagnosis and before the 2nd induction cycle were obtained to assess the therapeutic impact on leukemic blasts and microenvironment. Increased bone marrow angiogenesis, as assessed by microvessel density (MVD), was found at AML diagnosis and differed significantly between the WHO categories. Morphological analysis revealed a higher initial MVD in AML with myelodysplasia-related changes (AML-MRC) and a more substantial decrease of microvascularization after lenalidomide exposure. A slight increase of T-bet-positive TH1-equivalents was identifiable under lenalidomide. In the subgroup of patients with AML-MRC, the progression-free survival differed between the two treatment regimens, showing a potential but not significant benefit of lenalidomide. We found no correlation between the cereblon genotype (the target of lenalidomide) and treatment response or prognosis. In conclusion, addition of lenalidomide may be beneficial to elderly patients suffering from AML-MRC, where it leads to a reduction of microvascularization and, probably, to an intensified specific T cell-driven anti-leukemic response.

Prognostic implications of the microenvironment for follicular lymphoma under immunomodulation therapy.

Follicular lymphoma (FL) constitutes a significant proportion of lymphomas and shows frequent relapses. Beyond conventional chemotherapy, new therapeutic approaches have emerged, focussing on the interplay between lymphoma cells and the microenvironment. Here we report the immunophenotypic investigation of the microenvironment of a clinically well-characterized prospective cohort (study SAKK35/10, NCT01307605) of 154 treatment-naïve FL patients in need of therapy, who have been treated with rituximab only or a combination of rituximab and the immunomodulatory drug lenalidomide/Revlimid® A high ratio of CD4- to CD8-positive T cells (P = 0·009) and increased amounts of PD1+ tumour-infiltrating T cells (P = 0·007) were associated with inferior progression-free survival in the whole cohort. Interestingly, the prognostic impact of PD1+ T cells and the CD4/CD8 ratio lost its significance in the subgroup treated with R2 . In the latter group, high amounts of GATA3+ T helper (Th2) equivalents were associated with better progression-free survival (P < 0·001). We identified tumour microenvironmental features that allow prognostic stratification with respect to immuno- and combined immuno- and immunomodulatory therapy. Our analysis indicates that lenalidomide may compensate the adverse prognostic implication of higher amounts of CD4+ and, particularly, PD1+ T cells and that it has favourable effects mainly in cases with higher amounts of Th2 equivalents. [Correction added on 11 February 2020, after online publication: The NCT-trial number was previously incorrect and has been updated in this version].

Influence of regular aspirin intake on PSA values, prostate cancer incidence and overall survival in a prospective screening trial (ERSPC Aarau).

OBJECTIVES: To analyze the influence of aspirin (ASA) intake on PSA values and prostate cancer (PCa) development in a prospective screening study cohort. METHODS: 4314 men from the Swiss section of the European Randomized Study of Screening for Prostate Cancer (ERSPC) were included. A transrectal prostate biopsy was performed in men with a PSA level ≥ 3 ng/ml. Mortality data were obtained through registry linkages. PCa incidence and grade, total PSA, free-to-total PSA and overall survival were compared between ASA users and non-users. RESULTS: Median follow-up time was 9.6 years. In 789 men (18.3%) using aspirin [ASA +], the overall PCa incidence was significantly lower (6.8% vs. 9.6%, p = 0.015), but the multivariate Cox regression analysis showed no significant decrease in risk of PCa diagnosis (HR 0.84, p = 0.297). Total PSA values were significantly lower in ASA users for both baseline (1.6 vs. 1.8 ng/ml, p = 0.007) and follow-up visits (1.75 vs. 2.1 ng/ml, p < 0.001). Multivariate Cox regression analysis predicted significantly higher overall mortality risk among ASA users (HR 1.46, p = 0.009). CONCLUSIONS: In our study population, PCa incidence was significantly reduced among patients on aspirin. While we did not observe a statistically significant PCa risk reduction during the follow-up period, we found lower PSA values among ASA users compared to non-users, with a more distinct difference after 4 years of ASA intake, suggesting a cumulative effect and a potential protective association between regular ASA intake and PCa development. As for clinical practice, lowering PSA cutoff values by 0.4 ng/ml could be considered in long-term ASA users to avoid a potential bias towards delayed PCa detection.

Donor-derived, metastatic urothelial cancer after kidney transplantation associated with a potentially oncogenic BK polyomavirus.

BK polyomavirus has been linked to urothelial carcinoma in immunosuppressed patients. Here, we performed comprehensive genomic analysis of a BK polyomavirus-associated, metachronous, multifocal and metastatic micropapillary urothelial cancer in a kidney transplant recipient. Dissecting cancer heterogeneity by sorting technologies prior to array-comparative genomic hybridization followed by short tandem repeat analysis revealed that the metastatic urothelial cancer was of donor origin (4-year-old male). The top 50 cancer-associated genes showed no key driver mutations as assessed by next-generation sequencing. Whole genome sequencing and BK polyomavirus-specific amplification provided evidence for episomal and subgenomic chromosomally integrated BK polyomavirus genomes, which carried the same unique 17-bp deletion signature in the viral non-coding control region (NCCR). Whereas no role in oncogenesis could be attributed to the host gene integration in chromosome 1, the 17-bp deletion in the NCCR increased early viral gene expression, but decreased viral replication capacity. Consequently, urothelial cells were exposed to high levels of the transforming BK polyomavirus early proteins large tumour antigen and small tumour antigen from episomal and integrated gene expression. Surgery combined with discontinuation of immunosuppression resulted in complete remission, but sacrificed the renal transplant. Thus, this report links, for the first time, BK polyomavirus NCCR rearrangements with oncogenic transformation in urothelial cancer in immunosuppressed patients. Copyright © 2017 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Prognostic value of MIB-1 proliferation index in solitary fibrous tumors of the pleura implemented in a new score - a multicenter study.

BACKGROUND: Although the majority of solitary fibrous tumors of the pleura (SFTP) follow a benign course, 10-25% of patients suffer from recurrence or metastatic disease. Several scoring models have been proposed to predict the outcome. However, none of these included immunohistochemical (IHC) markers as possible prognosticators. METHODS: In this multicenter study, we collected clinical data and formalin-fixed and paraffin-embedded (FFPE) tissue blocks of patients with histologically proven SFTP which had been surgically resected between 2000 und 2015. After systematic and extensive IHC staining on tissue microarrays, the results were analyzed and compared to histomorphological and clinical data for their possible prognostic value. RESULTS: In total, 78 patients (mean age 61 ± 11 years) were included. Of these, 9 patients (11%) had an adverse outcome including SFTP recurrence (n = 6) or SFTP-related death (n = 3). Mean overall survival was 172 ± 13 months. 1 and 10-year event-free survival rates were 99% and 93%. In the multivariable analysis only MIB-1 proliferation index (Ki-67) ≥10% (HR 12.3, CI 1.1-139.5, p = 0.043), ≥4 mitoses per 10 high power fields (HR 36.5, CI 1.2-1103.7, p = 0.039) and tumor size larger than 10 cm (HR 81.8, CI 1.7-4016.8, p = 0.027) were independently associated with adverse outcome. CONCLUSION: A high proliferation rate by MIB-1 IHC was associated with impaired outcome. Upon this, we established a new score using mitosis, necrosis, size of the tumor and MIB-1, which performed better than the traditional scores in our data set. This prognostic score could help to better evaluate outcome of SFTP, but requires external validation.

A positive family history as a risk factor for prostate cancer in a population-based study with organised prostate-specific antigen screening: results of the Swiss European Randomised Study of Screening for Prostate Cancer (ERSPC, Aarau).

OBJECTIVE: To assess the value of a positive family history (FH) as a risk factor for prostate cancer incidence and grade among men undergoing organised prostate-specific antigen (PSA) screening in a population-based study. SUBJECTS AND METHODS: The study cohort comprised all attendees of the Swiss arm of the European Randomised Study of Screening for Prostate Cancer (ERSPC) with systematic PSA level tests every 4 years. Men reporting first-degree relative(s) diagnosed with prostate cancer were considered to have a positive FH. Biopsy was exclusively PSA triggered at a PSA level threshold of 3 ng/mL. The primary endpoint was prostate cancer diagnosis. Kaplan-Meier and Cox regression analyses were used. RESULTS: Of 4 932 attendees with a median (interquartile range, IQR) age of 60.9 (57.6-65.1) years, 334 (6.8%) reported a positive FH. The median (IQR) follow-up duration was 11.6 (10.3-13.3) years. Cumulative prostate cancer incidence was 60/334 (18%, positive FH) and 550/4 598 (12%, negative FH) [odds ratio 1.6, 95% confidence interval (CI) 1.2-2.2, P = 0.001). In both groups, most prostate cancer diagnosed was low grade. There were no significant differences in PSA level at diagnosis, biopsy Gleason score or Gleason score on pathological specimen among men who underwent radical prostatectomy between both groups. On multivariable analysis, age (hazard ratio [HR] 1.04, 95% CI 1.02-1.06), baseline PSA level (HR 1.13, 95% CI 1.12-1.14), and FH (HR 1.6, 95% CI 1.24-2.14) were independent predictors for overall prostate cancer incidence (all P < 0.001). Only baseline PSA level (HR 1.14, 95% CI 1.12-1.16, P < 0.001) was an independent predictor of Gleason score ≥7 prostate cancer on prostate biopsy. The proportion of interval prostate cancer diagnosed in-between the screening rounds was not significantly different. CONCLUSION: Irrespective of the FH status, the current PSA-based screening setting detects the majority of aggressive prostate cancers and missed only a minority of interval cancers with a 4-year screening algorithm. Our results suggest that men with a positive FH are at increased risk of low-grade but not aggressive prostate cancer.

Is further screening of men with baseline PSA < 1 ng ml(-1) worthwhile? The discussion continues-Results of the Swiss ERSPC (Aarau).

Recent studies indicate frequent early PSA retesting unrelated of men's baseline PSA, which increases the harms of early detection especially among men with low PSA. The current study investigates the PCa incidence among men with baseline PSA <1.0 ng ml(-1) in order to adjust retest intervals for more targeted early detection. Between 1998 and 2012, 2,416 men with baseline PSA <1.0 ng ml(-1) were prospectively observed. Primary endpoint was PCa diagnosis. Negative predictive value (NPV) and number needed to screen (NNS) to detect one PCa were calculated. During a median follow-up time of 12.1 years, 54 (2.2%) PCa were diagnosed with n = 26 (48.1%) among men with baseline PSA of 0.75 ≤ 1.0 ng ml(-1) (upper baseline PSA quartile). The 10-year probability of being diagnosed with PCa increased significantly from 0.19% (baseline PSA < 0.40 ng ml(-1) ) to 2.0% (baseline PSA 0.40 ≤ 0.56 ng ml(-1) ), 2.5% (baseline PSA 0.56 ≤ 0.75 ng ml(-1) ) over 4.4% (baseline PSA 0.75 ≤ 1.0 ng ml(-1) ) (all p values <0.0001), respectively. The frequency of Gleason ≥7 PCa increased from 1 (0.17%) to 8 (1.4%), 5 (0.8) over 11 (1.8%) in these groups. The 8-year NPV for Gleason ≥ 7 PCa were 99.8 (baseline PSA < 0.40 ng ml(-1) ), 99.8 (baseline PSA 0.40 ≤ 0.56 ng ml(-1) ), 100 (baseline PSA 0.56 ≤ 0.75 ng ml(-1) ) and 99.5 (baseline PSA 0.75 ≤ 1.0 ng ml(-1) ), respectively. During 12 years, the numbers were 99.8, 98.6, 99.2, and 98.2, respectively. Therefore, due to the very low rate of Gleason ≥ 7 PCa, further screening might be omitted in men with baseline PSA < 0.4 ng ml(-1) . Between 0.4 and 1.0 ng ml(-1) , an 8-year interval can be discussed.

Influence of metformin use on PSA values, free-to-total PSA, prostate cancer incidence and grade and overall survival in a prospective screening trial (ERSPC Aarau).

PURPOSE: To analyze the effect of the oral antidiabetic drug metformin on PSA level, free-to-total PSA ratio (f/t-ratio), PCa incidence and grade as well as mortality in men participating in a population-based screening trial. METHODS: Data from 4,314 men aged 55-70 years from a population-based PSA-screening trial (ERSPC Aarau) were analyzed. Information on metformin exposure was obtained by a self-administered questionnaire. Serum PSA threshold at ≥3 ng/ml triggered prostate biopsy. Data on PCa incidence and mortality were obtained through registry linkages. RESULTS: Median follow-up time was 7.6 years. Mean age at baseline was 65.5 years (±SD 4.4). In all, n = 150 (3.5 %) men used metformin [metf+]. Mean baseline PSA levels were comparable between both groups ([metf+] 1.6 ng/ml ± 2.4 vs. [metf-] 1.8ug/l ± 2.2, p = 0.4) while f/t-ratio was slightly higher in metformin users ([metf+] 30.7 % ± 10.9 vs. [metf-] 27.3 % ± 10.9, p = 0.01). Overall, n = 372 (8.6 %) PCa cases were detected. Neither cumulative PCa incidence (n = 11; 7.3 % [metf+] vs. n = 361 8.7 % [metf-]; p = 0.5) nor d`Amico risk groups were significantly different between both groups. One man in each group (metf+ 0.7 % and metf- 0.02 %) died from PCa (p < 0.0001), respectively. All-cause mortality was significantly higher among met + compared to met- (adjusted OR 2.50, 95 %CI 1.59-3.82; p = 0.0001). CONCLUSION: No significant differences in PSA levels or PCa incidence and grade were observed. The slightly higher f/t-ratio did not result in lower PCa detection rate. Metformin users were at significantly higher risk of all-cause mortality. The relatively small number of men on metformin is a main limitation of the study.

Pulmonary tularaemia: all that looks like cancer is not necessarily cancer - case report of four consecutive cases.

BACKGROUND: Pulmonary tularaemia is a very rare disease with only a small number of cases described in the literature. So far, to our knowledge, there exists no case report of pulmonary tularaemia where PET-CT scans and follow up CT scans are available. CASE PRESENTATION: We present four consecutive cases of pulmonary tularaemia. All patients suffered from non-specific symptoms. All patients were referred to our institution with strong suspicions of malignancy, particularly lung cancer. Diagnosis of tularaemia was made by typical findings in the aspirate of EBUS guided fine needle aspiration (necrosis, epithelioid cell aggregation) and surgical biopsy respectively, and a positive serology. In three of the four cases, the diagnosis was confirmed by positive PCR results of the tissue. PET-CT scans obtained in all four cases were indistinguishable from lesions typically seen in patients suffering from lung cancer. One of the four patients suffered from recurrence of the disease after antibiotic treatment; also this patient finally recovered after initiation of a second antibiotic regimen. One case became asymptomatic spontaneously, but this patient still received an antibiotic treatment. In one case, a follow up CT scan was unchanged compared to the initial PET-CT scan; in all other cases, the lesions disappeared almost completely. CONCLUSIONS: Symptoms of patients suffering from pulmonary tularaemia are non-specific and can be of prolonged character. PET-CT scans in these cases are indistinguishable from lung cancer. The diagnosis can be established when typical findings in EBUS guided fine needle aspirates or surgical biopsies are found in combination with a positive serology. In most cases the lesions disappear in follow up CT scans after clinically successful treatment.

Differences among men on active surveillance for very low-risk prostate cancer detected through population-based versus opportunistic prostate-specific antigen-screening.

INTRODUCTION: Very low-risk prostate cancer (PCa) is being increasingly managed by active surveillance (AS). Our aim was to assess the influence of the origin of diagnosis on PCa characteristics and treatment rates among men with very low-risk PCa in our prospective AS cohort. METHODS: Overall, 191 men with very low-risk PCa fulfilling Epstein-criteria underwent protocol-based AS. These men originated either from the prospective population-based screening program (P-AS) or were diagnosed by opportunistic screening (O-AS). RESULTS: Overall, n = 86 (45.0%) originated from the P-AS group, whereas n = 105 (55.0%) from the O-AS group. On univariate Cox regression analysis, age (HR 0.96, 95% CI 0.92-1.00; p = 0.05), origin of diagnosis (HR 0.72, 95% CI 0.41-1.28; p = 0.001), number of positive cores (HR 2.15, 95% CI 1.18-3.90; p = 0.01) and maximum core involvement (HR 1.03, 95% CI 0.99-1.05; p = 0.05) were predictors for treatment necessity. On multivariate analysis, age (HR 0.95, 95% CI 0.89-0.99; p = 0.05), number of positive cores (HR 2.07, 95% CI 1.10-3.88; p = 0.02), maximum core involvement (HR 1.03, 95% CI 1.00-1.06; p = 0.04) but not origin of diagnosis were independent predictors for treatment necessity. Four men developed biochemical recurrence (all from O-AS group [p = 0.05]). CONCLUSION: The origin of PCa diagnosis in men undergoing AS had no influence on disease progression and treatment necessity.

Margin status of the vas deferens in radical prostatectomy specimens: relevant or waste of time?

AIMS: Positive surgical margins (PSM) after radical prostatectomy are of great interest, but investigation of the vas deferens (VD) is not recommended. This study examined the VD margins in radical prostatectomy patients to report the incidence of PSM and their clinical staging. METHODS AND RESULTS: A total of 2701 consecutive specimens (1995-2009) were reviewed for tumour infiltration of the VD margin and correlated with clinicopathological data. Forty-one of 2701 cases (1.5%) had a positive VD margin. Thirteen cases had bilateral infiltration. All tumours were locally advanced [pT3a (n = 1), pT3b (n = 34), pT4 (n = 6)]; 15 (37%) had lymph node metastases. While Gleason scores ranged from 7 to 9, mean PSA was 22.3 ng/ml (1.68-127 ng/ml). In all cases with seminal vesicle infiltration (40 of 41) the PSM of the VD was seen ipsilaterally. In 11 of 15 patients (73%) with pN1 status, seminal vesicle infiltration and PSM of the VD were seen on the same side. In 16 cases (39%) the VD was the only PSM. CONCLUSIONS: A PSM of the VD is an infrequent finding, but might appear as the only PSM. Histological evaluation of the VD therefore seems reasonable, especially as biochemical recurrence has been reported with positive VD margins, and awareness of them might assist in making clinical decisions for adjuvant therapy.

Differentiation between calcification and hemorrhage in brain tumors using susceptibility-weighted imaging: a pilot study.

OBJECTIVE: Our aim was to develop a robust method to differentiate calcification from hemorrhage in gliomas. Histopathologic examination was performed to validate hemorrhage and calcification. CONCLUSION: Phase images from eleven patients with glioma yielded statistically significant phase-shift values for calcification and hemorrhage compared with normal brain, whereas CT showed substantial overlap of Hounsfield units. Phase image analysis correctly differentiated between intratumoral calcification and hemorrhage in 86% of cases.

Swiss digital pathology recommendations: results from a Delphi process conducted by the Swiss Digital Pathology Consortium of the Swiss Society of Pathology.

Integration of digital pathology (DP) into clinical diagnostic workflows is increasingly receiving attention as new hardware and software become available. To facilitate the adoption of DP, the Swiss Digital Pathology Consortium (SDiPath) organized a Delphi process to produce a series of recommendations for DP integration within Swiss clinical environments. This process saw the creation of 4 working groups, focusing on the various components of a DP system (1) scanners, quality assurance and validation of scans, (2) integration of Whole Slide Image (WSI)-scanners and DP systems into the Pathology Laboratory Information System, (3) digital workflow-compliance with general quality guidelines, and (4) image analysis (IA)/artificial intelligence (AI), with topic experts for each recruited for discussion and statement generation. The work product of the Delphi process is 83 consensus statements presented here, forming the basis for "SDiPath Recommendations for Digital Pathology". They represent an up-to-date resource for national and international hospitals, researchers, device manufacturers, algorithm developers, and all supporting fields, with the intent of providing expectations and best practices to help ensure safe and efficient DP usage.

Standardization of Gleason grading among 337 European pathologists.

AIMS: The 2005 International Society of Urological Pathology (ISUP) modification of Gleason grading recommended that the highest grade should always be included in the Gleason score (GS) in prostate biopsies. We analysed the impact of this recommendation on reporting of GS 6 versus 7. METHODS AND RESULTS: Fifteen expert uropathologists reached two-thirds consensus on 15 prostate biopsies with GS 6-7 cancer. Eighty-five microphotographs were graded by 337 of 617 members of the European Network of Uropathology (ENUP), representing 19 countries. There was agreement between expert and majority member GS in 12 of 15 cases, while members upgraded in three cases. Among members and the expert consensus, a GS >6 was assigned by 64.5% and 60%, respectively. Mean member GS was higher than consensus GS in nine of 15 cases. A Gleason pattern (GP) 5 was reported by 0.3-5.6% in 10 cases. Agreement between consensus and member GS was 58.2-89.3% (mean 71.4%) in GS 6 cases and 46.3-63.8% (mean 56.4%) in GS 7 cases (P = 0.009). CONCLUSIONS: While undergrading of prostate cancer used to be prevalent, some now tend to overgrade. Minimum diagnostic criteria for GP 4 and 5 in biopsies need to be better defined. Image libraries reviewed by experts may be useful for standardization.

Hepatocellular carcinoma as extracolonic manifestation of Lynch syndrome indicates SEC63 as potential target gene in hepatocarcinogenesis.

OBJECTIVE: Lynch syndrome is a cancer predisposition syndrome caused by germline mutations in DNA mismatch repair (MMR) genes with microsatellite instability (MSI) as its molecular hallmark. Hepatocellular carcinoma (HCC) has not been considered part of the tumor spectrum. The aim was to provide a detailed molecular characterization of an HCC associated with Lynch Syndrome (Muir-Torre variant). MATERIALS AND METHODS: HCC samples were analyzed for MSI, MMR protein expression and coding microsatellite instability (cMSI). Since cMSI also affected SEC63 coding for an endoplasmic reticulum membrane protein with implications for intracellular protein translocation, its impact on hepatocyte growth control was assessed in an established short-term model. Recombinant inbred mouse lines (BXD) showing different basal SEC63 expression levels were treated with the chemocarcinogen diethylnitrosamine (DEN) intraperitoneally. Proliferation and apoptosis of hepatocytes were determined after 48 h using Ki67 and TUNEL assays. RESULTS: The HCC was high-grade microsatellite unstable with loss of MSH2 expression. cMSI was detected in four genes (ASTE1, SEC63, TAF1B, TGFBR2). However, only TGFBR2 is known to be involved in hepatocarcinogenesis. When investigating the impact of SEC63 expression on hepatocyte growth control in the murine model, low hepatic expression correlated significantly (p < 0.05) with a decrease in apoptosis and increased proliferative activity. CONCLUSIONS: For the first time, an HCC with characteristic molecular features of association with Lynch syndrome is described. The pro-carcinogenic growth behavior of hepatocytes with low SEC63 expression in the murine model indicates a potential role for SEC63 in hepatocarcinogenesis in general, but this needs further functional validation.

Unusual Masquerading Infraorbital Mass-A Case Report of Human Ocular Dirofilariosis Found in A Ukrainian Patient in Switzerland.

A 78-year-old Ukrainian woman who had immigrated to Switzerland presented with a rapid growing subcutaneous infraorbital mass. Surgical excision of the mass revealed a well-circumscribed, encapsulated tumor, adherent to the skin. The excision showed a soft tissue inflammation with parts of Dirofilaria spp. The number of cases of human dirofilariosis reported in the last 50 years has gradually increased. Dirofilaria repens is now endemic in many countries and is currently considered to be one of the fast spreading zoonoses in Central, Eastern and Northern Europe. The first empirical evidence of Swiss spreading of D. repens infections was in a dog from southern Switzerland in 1998. Ours is the first case of human orbital dirofilariosis found in a Ukranian patient reported in Switzerland. Our purpose is to inform the ophthalmologist to consider orbital dirofilariosis in the differential diagnosis of inflammatory masses of the orbit and to warn about the spread of this infection in Switzerland.

Multi-site cross-organ calibrated deep learning (MuSClD): Automated diagnosis of non-melanoma skin cancer.

Although deep learning (DL) has demonstrated impressive diagnostic performance for a variety of computational pathology tasks, this performance often markedly deteriorates on whole slide images (WSI) generated at external test sites. This phenomenon is due in part to domain shift, wherein differences in test-site pre-analytical variables (e.g., slide scanner, staining procedure) result in WSI with notably different visual presentations compared to training data. To ameliorate pre-analytic variances, approaches such as CycleGAN can be used to calibrate visual properties of images between sites, with the intent of improving DL classifier generalizability. In this work, we present a new approach termed Multi-Site Cross-Organ Calibration based Deep Learning (MuSClD) that employs WSIs of an off-target organ for calibration created at the same site as the on-target organ, based off the assumption that cross-organ slides are subjected to a common set of pre-analytical sources of variance. We demonstrate that by using an off-target organ from the test site to calibrate training data, the domain shift between training and testing data can be mitigated. Importantly, this strategy uniquely guards against potential data leakage introduced during calibration, wherein information only available in the testing data is imparted on the training data. We evaluate MuSClD in the context of the automated diagnosis of non-melanoma skin cancer (NMSC). Specifically, we evaluated MuSClD for identifying and distinguishing (a) basal cell carcinoma (BCC), (b) in-situ squamous cell carcinomas (SCC-In Situ), and (c) invasive squamous cell carcinomas (SCC-Invasive), using an Australian (training, n = 85) and a Swiss (held-out testing, n = 352) cohort. Our experiments reveal that MuSCID reduces the Wasserstein distances between sites in terms of color, contrast, and brightness metrics, without imparting noticeable artifacts to training data. The NMSC-subtyping performance is statistically improved as a result of MuSCID in terms of one-vs. rest AUC: BCC (0.92 vs 0.87, p = 0.01), SCC-In Situ (0.87 vs 0.73, p = 0.15) and SCC-Invasive (0.92 vs 0.82, p = 1e-5). Compared to baseline NMSC-subtyping with no calibration, the internal validation results of MuSClD (BCC (0.98), SCC-In Situ (0.92), and SCC-Invasive (0.97)) suggest that while domain shift indeed degrades classification performance, our on-target calibration using off-target tissue can safely compensate for pre-analytical variabilities, while improving the robustness of the model.