Validation of prostate and breast cancer detection artificial intelligence algorithms for accurate histopathological diagnosis and grading: a retrospective study with a Japanese cohort.

Prostate and breast cancer incidence rates have been on the rise in Japan, emphasising the need for precise histopathological diagnosis to determine patient prognosis and guide treatment decisions. However, existing diagnostic methods face numerous challenges and are susceptible to inconsistencies between observers. To tackle these issues, artificial intelligence (AI) algorithms have been developed to aid in the diagnosis of prostate and breast cancer. This study focuses on validating the performance of two such algorithms, Galen Prostate and Galen Breast, in a Japanese cohort, with a particular focus on the grading accuracy and the ability to differentiate between invasive and non-invasive tumours. The research entailed a retrospective examination of 100 consecutive prostate and 100 consecutive breast biopsy cases obtained from a Japanese institution. Our findings demonstrated that the AI algorithms showed accurate cancer detection, with AUCs of 0.969 and 0.997 for the Galen Prostate and Galen Breast, respectively. The Galen Prostate was able to detect a higher Gleason score in four adenocarcinoma cases and detect a previously unreported cancer. The two algorithms successfully identified relevant pathological features, such as perineural invasions and lymphovascular invasions. Although further improvements are required to accurately differentiate rare cancer subtypes, these findings highlight the potential of these algorithms to enhance the precision and efficiency of prostate and breast cancer diagnosis in Japan. Furthermore, this validation paves the way for broader adoption of these algorithms as decision support tools within the Asian population.

Validation and real-world clinical application of an artificial intelligence algorithm for breast cancer detection in biopsies.

Breast cancer is the most common malignant disease worldwide, with over 2.26 million new cases in 2020. Its diagnosis is determined by a histological review of breast biopsy specimens, which can be labor-intensive, subjective, and error-prone. Artificial Intelligence (AI)-based tools can support cancer detection and classification in breast biopsies ensuring rapid, accurate, and objective diagnosis. We present here the development, external clinical validation, and deployment in routine use of an AI-based quality control solution for breast biopsy review. The underlying AI algorithm is trained to identify 51 different types of clinical and morphological features, and it achieves very high accuracy in a large, multi-site validation study. Specifically, the area under the receiver operating characteristic curves (AUC) for the detection of invasive carcinoma and of ductal carcinoma in situ (DCIS) are 0.99 (specificity and sensitivity of 93.57 and 95.51%, respectively) and 0.98 (specificity and sensitivity of 93.79 and 93.20% respectively), respectively. The AI algorithm differentiates well between subtypes of invasive and different grades of in situ carcinomas with an AUC of 0.97 for invasive ductal carcinoma (IDC) vs. invasive lobular carcinoma (ILC) and AUC of 0.92 for DCIS high grade vs. low grade/atypical ductal hyperplasia, respectively, as well as accurately identifies stromal tumor-infiltrating lymphocytes (TILs) with an AUC of 0.965. Deployment of this AI solution as a real-time quality control solution in clinical routine leads to the identification of cancers initially missed by the reviewing pathologist, demonstrating both clinical utility and accuracy in real-world clinical application.

An artificial intelligence algorithm for prostate cancer diagnosis in whole slide images of core needle biopsies: a blinded clinical validation and deployment study.

BACKGROUND: There is high demand to develop computer-assisted diagnostic tools to evaluate prostate core needle biopsies (CNBs), but little clinical validation and a lack of clinical deployment of such tools. We report here on a blinded clinical validation study and deployment of an artificial intelligence (AI)-based algorithm in a pathology laboratory for routine clinical use to aid prostate diagnosis. METHODS: An AI-based algorithm was developed using haematoxylin and eosin (H&E)-stained slides of prostate CNBs digitised with a Philips scanner, which were divided into training (1 357 480 image patches from 549 H&E-stained slides) and internal test (2501 H&E-stained slides) datasets. The algorithm provided slide-level scores for probability of cancer, Gleason score 7-10 (vs Gleason score 6 or atypical small acinar proliferation [ASAP]), Gleason pattern 5, and perineural invasion and calculation of cancer percentage present in CNB material. The algorithm was subsequently validated on an external dataset of 100 consecutive cases (1627 H&E-stained slides) digitised on an Aperio AT2 scanner. In addition, the AI tool was implemented in a pathology laboratory within routine clinical workflow as a second read system to review all prostate CNBs. Algorithm performance was assessed with area under the receiver operating characteristic curve (AUC), specificity, and sensitivity, as well as Pearson's correlation coefficient (Pearson's r) for cancer percentage. FINDINGS: The algorithm achieved an AUC of 0·997 (95% CI 0·995 to 0·998) for cancer detection in the internal test set and 0·991 (0·979 to 1·00) in the external validation set. The AUC for distinguishing between a low-grade (Gleason score 6 or ASAP) and high-grade (Gleason score 7-10) cancer diagnosis was 0·941 (0·905 to 0·977) and the AUC for detecting Gleason pattern 5 was 0·971 (0·943 to 0·998) in the external validation set. Cancer percentage calculated by pathologists and the algorithm showed good agreement (r=0·882, 95% CI 0·834 to 0·915; p<0·0001) with a mean bias of -4·14% (-6·36 to -1·91). The algorithm achieved an AUC of 0·957 (0·930 to 0·985) for perineural invasion. In routine practice, the algorithm was used to assess 11 429 H&E-stained slides pertaining to 941 cases leading to 90 Gleason score 7-10 alerts and 560 cancer alerts. 51 (9%) cancer alerts led to additional cuts or stains being ordered, two (4%) of which led to a third opinion request. We report on the first case of missed cancer that was detected by the algorithm. INTERPRETATION: This study reports the successful development, external clinical validation, and deployment in clinical practice of an AI-based algorithm to accurately detect, grade, and evaluate clinically relevant findings in digitised slides of prostate CNBs. FUNDING: Ibex Medical Analytics.

Unbiased transcriptome signature of in vivo cell proliferation reveals pro- and antiproliferative gene networks.

Different types of mature B-cell lymphocytes are overall highly similar. Nevertheless, some B cells proliferate intensively, while others rarely do. Here, we demonstrate that a simple binary classification of gene expression in proliferating vs. resting B cells can identify, with remarkable selectivity, global in vivo regulators of the mammalian cell cycle, many of which are also post-translationally regulated by the APC/C E3 ligase. Consequently, we discover a novel regulatory network between the APC/C and the E2F transcription factors and discuss its potential impact on the G1-S transition of the cell cycle. In addition, by focusing on genes whose expression inversely correlates with proliferation, we demonstrate the inherent ability of our approach to also identify in vivo regulators of cell differentiation, cell survival, and other antiproliferative processes. Relying on data sets of wt, non-transgenic animals, our approach can be applied to other cell lineages and human data sets.

Gas2l3, a novel constriction site-associated protein whose regulation is mediated by the APC/C Cdh1 complex.

Growth arrest-specific 2-like protein 3 (Gas2l3) was recently identified as an Actin/Tubulin cross-linker protein that regulates cytokinesis. Using cell-free systems from both frog eggs and human cells, we show that the Gas2l3 protein is targeted for ubiquitin-mediated proteolysis by the APC/C(Cdh1) complex, but not by the APC/C(Cdc20) complex, and is phosphorylated by Cdk1 in mitosis. Moreover, late in cytokinesis, Gas2l3 is exclusively localized to the constriction sites, which are the narrowest parts of the intercellular bridge connecting the two daughter cells. Overexpression of Gas2l3 specifically interferes with cell abscission, which is the final stage of cell division, when the cutting of the intercellular bridge at the constriction sites occurs. We therefore suggest that Gas2l3 is part of the cellular mechanism that terminates cell division.

Using standard optical flow cytometry for synchronizing proliferating cells in the G1 phase.

Cell cycle research greatly relies on synchronization of proliferating cells. However, effective synchronization of mammalian cells is commonly achieved by long exposure to one or more cell cycle blocking agents. These chemicals are, by definition, hazardous (some more than others), pose uneven cell cycle arrest, thus introducing unwanted variables. The challenge of synchronizing proliferating cells in G1 is even greater; this process typically involves the release of drug-arrested cells into the cycle that follows, a heterogeneous process that can truly limit synchronization. Moreover, drug-based synchronization decouples the cell cycle from cell growth in ways that are understudied and intolerable for those who investigate the relationship between these two processes. In this study we showed that cell size, as approximated by a single light-scatter parameter available in all standard sorters, can be used for synchronizing proliferating mammalian cells in G1 with minimal or no risk to either the cell cycle or cell growth. The power and selectivity of our method are demonstrated for human HEK293 cells that, despite their many advantages, are suboptimal for synchronization, let alone in G1. Our approach is readily available, simple, fast, and inexpensive; it is independent of any drugs or dyes, and nonhazardous. These properties are relevant for the study of the mammalian cell cycle, specifically in the context of G1 and cell growth.

Ex vivo treatment with a novel synthetic aminoglycoside NB54 in primary fibroblasts from Rett syndrome patients suppresses MECP2 nonsense mutations.

BACKGROUND: Nonsense mutations in the X-linked methyl CpG-binding protein 2 (MECP2) comprise a significant proportion of causative MECP2 mutations in Rett syndrome (RTT). Naturally occurring aminoglycosides, such as gentamicin, have been shown to enable partial suppression of nonsense mutations related to several human genetic disorders, however, their clinical applicability has been compromised by parallel findings of severe toxic effects. Recently developed synthetic NB aminoglycosides have demonstrated significantly improved effects compared to gentamicin evident in substantially higher suppression and reduced acute toxicity in vitro. RESULTS: We performed comparative study of suppression effects of the novel NB54 and gentamicin on three MECP2 nonsense mutations (R294X, R270X and R168X) common in RTT, using ex vivo treatment of primary fibroblasts from RTT patients harboring these mutations and testing for the C-terminal containing full-length MeCP2. We observed that NB54 induces dose-dependent suppression of MECP2 nonsense mutations more efficiently than gentamicin, which was evident at concentrations as low as 50 µg/ml. NB54 read-through activity was mutation specific, with maximal full-length MeCP2 recovery in R168X (38%), R270X (27%) and R294X (18%). In addition, the recovered MeCP2 was translocated to the cell nucleus and moreover led to parallel increase in one of the most important MeCP2 downstream effectors, the brain derived neurotrophic factor (BDNF). CONCLUSION: Our findings suggest that NB54 may induce restoration of the potentially functional MeCP2 in primary RTT fibroblasts and encourage further studies of NB54 and other rationally designed aminoglycoside derivatives as potential therapeutic agents for nonsense MECP2 mutations in RTT.

Epilepsy in Rett syndrome---the experience of a National Rett Center.

PURPOSE: Rett syndrome (RTT), an X-linked, dominant neurodevelopmental disorder caused by mutations in the methyl-CpG-binding protein 2 (MECP2) gene, presents with acquired microcephaly, autistic regression, hand usage loss, and stereotypies. Epilepsy is frequent and has been reported to correlate with mutation type, general disease severity, and BDNF polymorphism. Our purpose was a comprehensive description of epilepsy features and course in RTT. METHODS: Retrospective review of charts and electroencephalography (EEG) studies in 97 patients with RTT. RESULTS: Seventy-two percent of patients had epilepsy, appearing at a median age of 3 years. According to age of onset, we divided patients into three groups: 6 with early epileptic variant (0-1 year), 42 with early epilepsy (1-5 years), and 20 with late epilepsy (after 5 years). Early epileptic variant had severe seizure types in the first year of life, followed by a typical RTT picture; all were MECP2 negative. Early epilepsy and late epilepsy groups were similar with respect to Rett-related symptoms, but seizures were better controlled in the second group (p < 0.05). Epileptiform activity appeared earlier and was more confluent in the early epilepsy group, including nine patients with electrical status epilepticus during sleep (ESES) versus one in the late epilepsy group (p < 0.05). No correlation was found between epilepsy onset or severity and genotype. BDNF val/met polymorphism correlated with earlier onset of seizures (p < 0.05). DISCUSSION: Epilepsy appears earlier than described previously, frequently during the regression stage. Early age of onset predicts a more severe course of seizures. ESES is common among those with early onset epilepsy. BDNF polymorphism was the only genetic correlate with seizure onset, whereas MECP2 mutation type and location did not influence epilepsy.

MeCP2 deficiency downregulates specific nuclear proteins that could be partially recovered by valproic acid in vitro.

MeCP2, the major causative factor of Rett syndrome and related phenotypes including autism, is a two-face nuclear modulator acting via transcriptional and chromatin remodeling mechanisms. This study investigated the expression of several nuclear proteins and their dependence on MeCP2 dose and presence of the Rett causative R306C mutation. To this end, we developed in vitro models representing MeCP2 deficiency induced by siRNAs, and cells expressing the R306C mutation. Using an extended antibody microarray validated by specific assays, revealed that MeCP2 dose was correlated with specific nuclear proteins profiles including the BRM/SNF2 component of SWI/SNF complex, PRMT1 methyl transferase and HDAC2. Furthermore, while exposing the MeCP2 knock-down system to therapeutic concentrations of valproic acid (VPA), a known HDACs inhibitor, we observed a partial restoration of MeCP2 expression levels. Exposure to VPA also increased the levels of BRM, as well as of BDNF, an important co-factor in MeCP2-mediated pathway. Our findings provide additional evidence of diverse mechanisms of MeCP2 function as transcriptional repressor and activator of specific genes. As it has been recently demonstrated that post-natal restoration of MeCP2 deficiency may reverse neurological defects in a mouse model of Rett syndrome, we suggest to study the restorative effect of HDAC inhibitors in MeCP2-deficient mouse model.

A coding VKORC1 Asp36Tyr polymorphism predisposes to warfarin resistance.

CYP2C9 and VKORC1 genetic variants are associated with low and intermediate warfarin dose requirements, but markers of high doses are less well characterized. We analyzed the VKORC1 coding sequence and known CYP2C9 and VKORC1 polymorphisms in 15 selected warfarin-resistant (dose, 80 to 185 mg/wk) and 8 warfarin-sensitive patients (7 to 13 mg/wk) and 99 unselected controls (8 to 105 mg/wk). We identified a coding VKORC1 Asp36Tyr polymorphism in 7 of 15 resistant compared with 0 of 8 sensitive patients (P = .026) Carriers of Asp36Tyr in the control group (8 of 99) required significantly higher warfarin doses of 80.9 +/- 10.1 mg/wk compared with 42.7 +/- 7.5 mg/wk in noncarriers (F = 9.79, P = .002). Asp36Tyr was significantly associated with doses of more than 70 mg/wk (odds ratio, 13.0; 95% confidence limit, 1.3 to 124.2), while doses of 20 to 70 mg/wk were associated with Asp36Tyr (partial r(2) = .11; P = .004), CYP2C9*2 and *3 (r(2) = .08; P = .01), and VKORC1*2 and *3 markers (r(2) = .05; P = .05). All Asp36Tyr carriers also had VKORC1*1 tag-single nucleotide polymorphisms (tag-SNPs) indicating a new haplotype. Asp36Tyr was common in Jewish ethnic groups of Ethiopian (15%) and Ashkenazi (4%) origin. We suggest that Asp36Tyr is a new marker of the high end of the warfarin dosing range.

Combined genetic profiles of components and regulators of the vitamin K-dependent gamma-carboxylation system affect individual sensitivity to warfarin.

We examined the influence of combined genotypes on interindividual variability in warfarin dose-response. In 100 anticoagulated patients we quantified the effects of polymorphisms in: CYP2C9, VKORC1, calumenin (CALU), gamma-glutamyl carboxylase (GGCX) and microsomal epoxide hydrolase (EPHX1) on warfarin dose requirements. The G(1542)C VKORC1 polymorphism was associated with decreased warfarin doses in the hetero- and homozygous mutant patients (21% and 50% lower, respectively; p < 0.0001). Warfarin daily dose was predominantly determined by VKORC1 and CYP2C9 genotypes (partial r(2) = 0.21; 0.20, respectively). Together with age and body weight, these two genotypes explained 63% of the dose variance. A single patient, homozygous for G(11)A CALU mutant allele, required an exceptionally high warfarin dose (20 mg/day) and the prevalence of heterozygous (11)A allele carriers in the upper 10(th) dose percentile was significantly higher (0.27 vs. 0.18, p < 0.02). Combined genotype analysis revealed that CYP2C9 andVKORC1 wild type and CALU mutant patients required the highest warfarin doses (7.8 +/- 1.5mg/day; n = 9) as compared to the CYP2C9 and VKORC1 mutant and CALU wild type genotypes (2.8 +/- 0.3 mg/day; n = 18; p < 0.01). The odds ratio for doses <3mg/day was 5.9 (1.9-18.4) for this genotype. Compound genetic profiles comprising VKORC1, CALU and CYP2C9 improve categorization of individual warfarin dose requirements in more than 25% of patients at steady-state anticoagulation.

Common genetic variants of microsomal epoxide hydrolase affect warfarin dose requirements beyond the effect of cytochrome P450 2C9.

BACKGROUND: Warfarin dose response is partially explained by the polymorphisms in the cytochrome P450 (CYP) 2C9 gene, affecting S -warfarin clearance, as well as by age and body weight. We examined the influence on warfarin dose requirements of candidate genes encoding microsomal epoxide hydrolase (mEH), as well as glutathione S -transferase A1 (GSTA1) components of vitamin K epoxide reductase and the gamma-glutamylcarboxylase (GGCX) gene. METHODS: We studied the effects of CYP2C9, mEH, GSTA1, and GGCX genotypes on warfarin maintenance doses, accounting for age, weight, vitamin K plasma concentrations and concurrent medications, in 100 patients undergoing therapeutic anticoagulation. RESULTS: Allele frequencies were 76.5%, 12.5%, and 11% for CYP2C9*1 , *2 , and *3 , respectively; 75% and 25% for mEH T 612 C; 75.8% and 24.2% for mEH A 691 G; 73.5% and 26.5% for GSTA1 T 631 G; and 70.5% and 29.5% for GGCX G 8762 A. Warfarin doses differed among the CYP2C9 ( 2C9*1 , 2C9*2 , and 2C9*3 ) genotype groups: 6.3 +/- 1.9 mg/d, 5.3 +/- 1.8 mg/d, and 3.8 +/- 1.7 mg/d, respectively (F = 4.83, P < .01). There were no differences in any of the other genotype groups. Among the 62 wild-type CYP2C9 patients, variant mEH T 612 C homozygotes required higher doses than heterozygotes and wild-type patients (7.5 +/- 2.9 mg/d, 6.5 +/- 4.2 mg/d, and 6.0 +/- 2.6 mg/d, respectively [F = 3.57, P = .03]). The odds ratio for requiring greater than 7 mg/d in variant mEH T 612 C patients versus wild-type patients was 3.14 (95% confidence interval, 1.47-6.67), accounting for CYP2C9. CONCLUSIONS: Variant mEH T 612 C genotypes are associated with warfarin doses of greater than 50 mg/wk beyond the effect of CYP2C9.