Development and validation of artificial intelligence-based prescreening of large-bowel biopsies taken in the UK and Portugal: a retrospective cohort study

BACKGROUND

Histopathological examination is a crucial step in the diagnosis and treatment of many major diseases. Aiming to facilitate diagnostic decision making and improve the workload of pathologists, we developed an artificial intelligence (AI)-based prescreening tool that analyses whole-slide images (WSIs) of large-bowel biopsies to identify typical, non-neoplastic, and neoplastic biopsies.

METHODS

This retrospective cohort study was conducted with an internal development cohort of slides acquired from a hospital in the UK and three external validation cohorts of WSIs acquired from two hospitals in the UK and one clinical laboratory in Portugal. To learn the differential histological patterns from digitised WSIs of large-bowel biopsy slides, our proposed weakly supervised deep-learning model (Colorectal AI Model for Abnormality Detection [CAIMAN]) used slide-level diagnostic labels and no detailed cell or region-level annotations. The method was developed with an internal development cohort of 5054 biopsy slides from 2080 patients that were labelled with corresponding diagnostic categories assigned by pathologists. The three external validation cohorts, with a total of 1536 slides, were used for independent validation of CAIMAN. Each WSI was classified into one of three classes (ie, typical, atypical non-neoplastic, and atypical neoplastic). Prediction scores of image tiles were aggregated into three prediction scores for the whole slide, one for its likelihood of being typical, one for its likelihood of being non-neoplastic, and one for its likelihood of being neoplastic. The assessment of the external validation cohorts was conducted by the trained and frozen CAIMAN model. To evaluate model performance, we calculated area under the convex hull of the receiver operating characteristic curve (AUROC), area under the precision-recall curve, and specificity compared with our previously published iterative draw and rank sampling (IDaRS) algorithm. We also generated heat maps and saliency maps to analyse and visualise the relationship between the WSI diagnostic labels and spatial features of the tissue microenvironment. The main outcome of this study was the ability of CAIMAN to accurately identify typical and atypical WSIs of colon biopsies, which could potentially facilitate automatic removing of typical biopsies from the diagnostic workload in clinics.

FINDINGS

A randomly selected subset of all large bowel biopsies was obtained between Jan 1, 2012, and Dec 31, 2017. The AI training, validation, and assessments were done between Jan 1, 2021, and Sept 30, 2022. WSIs with diagnostic labels were collected between Jan 1 and Sept 30, 2022. Our analysis showed no statistically significant differences across prediction scores from CAIMAN for typical and atypical classes based on anatomical sites of the biopsy. At 0·99 sensitivity, CAIMAN (specificity 0·5592) was more accurate than an IDaRS-based weakly supervised WSI-classification pipeline (0·4629) in identifying typical and atypical biopsies on cross-validation in the internal development cohort (p<0·0001). At 0·99 sensitivity, CAIMAN was also more accurate than IDaRS for two external validation cohorts (p<0·0001), but not for a third external validation cohort (p=0·10). CAIMAN provided higher specificity than IDaRS at some high-sensitivity thresholds (0·7763 vs 0·6222 for 0·95 sensitivity, 0·7126 vs 0·5407 for 0·97 sensitivity, and 0·5615 vs 0·3970 for 0·99 sensitivity on one of the external validation cohorts) and showed high classification performance in distinguishing between neoplastic biopsies (AUROC 0·9928, 95% CI 0·9927-0·9929), inflammatory biopsies (0·9658, 0·9655-0·9661), and atypical biopsies (0·9789, 0·9786-0·9792). On the three external validation cohorts, CAIMAN had AUROC values of 0·9431 (95% CI 0·9165-0·9697), 0·9576 (0·9568-0·9584), and 0·9636 (0·9615-0·9657) for the detection of atypical biopsies. Saliency maps supported the representation of disease heterogeneity in model predictions and its association with relevant histological features.

INTERPRETATION

CAIMAN, with its high sensitivity in detecting atypical large-bowel biopsies, might be a promising improvement in clinical workflow efficiency and diagnostic decision making in prescreening of typical colorectal biopsies.

FUNDING

The Pathology Image Data Lake for Analytics, Knowledge and Education Centre of Excellence; the UK Government's Industrial Strategy Challenge Fund; and Innovate UK on behalf of UK Research and Innovation.

Screening of normal endoscopic large bowel biopsies with interpretable graph learning: a retrospective study

OBJECTIVE

To develop an interpretable artificial intelligence algorithm to rule out normal large bowel endoscopic biopsies, saving pathologist resources and helping with early diagnosis.

DESIGN

A graph neural network was developed incorporating pathologist domain knowledge to classify 6591 whole-slides images (WSIs) of endoscopic large bowel biopsies from 3291 patients (approximately 54% female, 46% male) as normal or abnormal (non-neoplastic and neoplastic) using clinically driven interpretable features. One UK National Health Service (NHS) site was used for model training and internal validation. External validation was conducted on data from two other NHS sites and one Portuguese site.

RESULTS

Model training and internal validation were performed on 5054 WSIs of 2080 patients resulting in an area under the curve-receiver operating characteristic (AUC-ROC) of 0.98 (SD=0.004) and AUC-precision-recall (PR) of 0.98 (SD=0.003). The performance of the model, named Interpretable Gland-Graphs using a Neural Aggregator (IGUANA), was consistent in testing over 1537 WSIs of 1211 patients from three independent external datasets with mean AUC-ROC=0.97 (SD=0.007) and AUC-PR=0.97 (SD=0.005). At a high sensitivity threshold of 99%, the proposed model can reduce the number of normal slides to be reviewed by a pathologist by approximately 55%. IGUANA also provides an explainable output highlighting potential abnormalities in a WSI in the form of a heatmap as well as numerical values associating the model prediction with various histological features.

CONCLUSION

The model achieved consistently high accuracy showing its potential in optimising increasingly scarce pathologist resources. Explainable predictions can guide pathologists in their diagnostic decision-making and help boost their confidence in the algorithm, paving the way for its future clinical adoption.

Deep learning based digital cell profiles for risk stratification of urine cytology images

Urine cytology is a test for the detection of high-grade bladder cancer. In clinical practice, the pathologist would manually scan the sample under the microscope to locate atypical and malignant cells. They would assess the morphology of these cells to make a diagnosis. Accurate identification of atypical and malignant cells in urine cytology is a challenging task and is an essential part of identifying different diagnosis with low-risk and high-risk malignancy. Computer-assisted identification of malignancy in urine cytology can be complementary to the clinicians for treatment management and in providing advice for carrying out further tests. In this study, we presented a method for identifying atypical and malignant cells followed by their profiling to predict the risk of diagnosis automatically. For cell detection and classification, we employed two different deep learning-based approaches. Based on the best performing network predictions at the cell level, we identified low-risk and high-risk cases using the count of atypical cells and the total count of atypical and malignant cells. The area under the receiver operating characteristic (ROC) curve shows that a total count of atypical and malignant cells is comparably better at diagnosis as compared to the count of malignant cells only. We obtained area under the ROC curve with the count of malignant cells and the total count of atypical and malignant cells as 0.81 and 0.83, respectively. Our experiments also demonstrate that the digital risk could be a better predictor of the final histopathology-based diagnosis. We also analyzed the variability in annotations at both cell and whole slide image level and also explored the possible inherent rationales behind this variability.

Multiplex Cellular Communities in Multi-Gigapixel Colorectal Cancer Histology Images for Tissue Phenotyping

In computational pathology, automated tissue phenotyping in cancer histology images is a fundamental tool for profiling tumor microenvironments. Current tissue phenotyping methods use features derived from image patches which may not carry biological significance. In this work, we propose a novel multiplex cellular community-based algorithm for tissue phenotyping integrating cell-level features within a graph-based hierarchical framework. We demonstrate that such integration offers better performance compared to prior deep learning and texture-based methods as well as to cellular community based methods using uniplex networks. To this end, we construct celllevel graphs using texture, alpha diversity and multi-resolution deep features. Using these graphs, we compute cellular connectivity features which are then employed for the construction of a patch-level multiplex network. Over this network, we compute multiplex cellular communities using a novel objective function. The proposed objective function computes a low-dimensional subspace from each cellular network and subsequently seeks a common low-dimensional subspace using the Grassmann manifold. We evaluate our proposed algorithm on three publicly available datasets for tissue phenotyping, demonstrating a significant improvement over existing state-of-the-art methods.

Cellular community detection for tissue phenotyping in colorectal cancer histology images

Classification of various types of tissue in cancer histology images based on the cellular compositions is an important step towards the development of computational pathology tools for systematic digital profiling of the spatial tumor microenvironment. Most existing methods for tissue phenotyping are limited to the classification of tumor and stroma and require large amount of annotated histology images which are often not available. In the current work, we pose the problem of identifying distinct tissue phenotypes as finding communities in cellular graphs or networks. First, we train a deep neural network for cell detection and classification into five distinct cellular components. Considering the detected nuclei as nodes, potential cell-cell connections are assigned using Delaunay triangulation resulting in a cell-level graph. Based on this cell graph, a feature vector capturing potential cell-cell connection of different types of cells is computed. These feature vectors are used to construct a patch-level graph based on chi-square distance. We map patch-level nodes to the geometric space by representing each node as a vector of geodesic distances from other nodes in the network and iteratively drifting the patch nodes in the direction of positive density gradients towards maximum density regions. The proposed algorithm is evaluated on a publicly available dataset and another new large-scale dataset consisting of 280K patches of seven tissue phenotypes. The estimated communities have significant biological meanings as verified by the expert pathologists. A comparison with current state-of-the-art methods reveals significant performance improvement in tissue phenotyping.

The Polycystic Ovary Syndrome Health-Related Quality of Life Questionnaire (PCOSQ): a validation

BACKGROUND

We wished to evaluate the psychometric properties of the Polycystic Ovary Syndrome Questionnaire (PCOSQ), a questionnaire developed to measure the health-related quality of life (HRQoL) of women with polycystic ovary syndrome.

METHOD

To assess reliability and validity, women recruited from an outpatient gynaecology clinic at the Jessop Wing, Royal Hallamshire Hospital, Sheffield completed two copies of the PCOSQ and the Short Form-36 (SF-36). Secondary factor analysis was carried out to verify the composition of the dimensions. Semi-structured interviews were conducted to assess face validity.

RESULTS

Of the 92 women who consented, 82 women (89%) returned questionnaires at time 1, and 69 women (75%) returned questionnaires at time 2. All five PCOSQ dimensions were internally reliable with Cronbach's alpha scores ranging from 0.70 to 0.97. Intra-class correlation coefficients to evaluate test-retest reliability were high (range 0.89-0.95, P < 0.001). Construct validity was demonstrated by high correlations for all comparisons of similar scales of the SF-36 and PCOSQ (0.49 and 0.54). Acne was identified as an important area of HRQoL missing from the questionnaire.

CONCLUSIONS

The PCOSQ is a reliable instrument for measuring the HRQoL in women with PCOS. However, the validity of the questionnaire needs to be improved by incorporating a dimension on acne into the instrument.

Cardiovascular load of competitive golf in cardiac patients and healthy controls

PURPOSE

Sports in cardiovascular patients (CVP) should serve for risk factor management, increase of exercise capacity, and reintegration into daily life. Competition of cardiac patients with healthy sportsmen is often discouraged and thus reintegration hampered. Golf, with its endurance component and exceptional rules (e.g., the handicap) should be an alternative.

METHODS

In 20 male golfers (65.2 +/- 6.1 yr, 1.4 +/- 0.3 W x kg(-1) body weight (approximately 4.8 METs)) with cardiovascular diseases and eight controls (C) (62 +/- 5 yr, 2 +/- 0.4 W x kg(-1) body weight (approximately 6.9 METs)), the performance assessed in the laboratory (ergospirometry, serum lactate) allowed for comparison of the cardiovascular load on the golf course (lactate, Holter monitoring, blood pressure, urine catecholamines).

RESULTS

In comparison with in the hospital, resting heart rates were significantly (P < 0.001) elevated in both groups immediately before the tournament (CVP: 76.1 +/- 10.8 vs 90.1 +/- 8.6 bpm; C: 74.8 +/- 6.3 vs 92.3 +/- 9.7 bpm). On the course, the mean heart rates of the patients were closer (P < 0.01) to the anaerobic threshold (105.4 +/- 11.0 vs 115.3 +/- 10.8 bpm) in comparison with controls (100.5 +/- 7.3 vs 125.6 +/- 16.6 bpm) corresponding to 0.9 +/- 0.3 W x kg(-1) (approximately 3.1 METs) or 76.0 +/- 13.1%VO2max (CVP) and to 0.9 +/- 0.2 W x kg(-1) (approximately 3.1 METs) or 55.3 +/- 9.1%VO2max (C). Serum lactate levels were 1.36 +/- 0.7 mmol x L(-1) (approximately 12.4 +/- 6.4 mg x dL(-1)) (CVP) and 1.1 +/- 0.4 mmol x L(-1) (approximately 9.1 +/- 3.6 mg x dL(-1)) (C). In patients, arrhythmias were lower in quantity and quality (LOWN) in comparison with other activities as registered by means of the 24-Holter-ECG.

CONCLUSION

In cardiovascular patients, competitive golf reaches an intensity that may positively influence cardiovascular risk factors, depending on the type of the course and may provide patients the desired integration with healthy sportsmen.

Analysis of genetically modified plant gene expression using GUS fluorimetry

A fluorimetric assay method for the analysis of beta-glucuronidase (GUS) reporter gene expression in genetically modified plants is described. Optimization of this method for woody plants and a statistical approach suitable for comparisons of gene expression in different transformants or tissues of the same plant is described. Example data from elm (Ulmus procera) SR4 regenerant plants, shown to be genetically modified by PCR and DNA-DNA hybridizations, in which higher GUS expression levels are found in stems than in leaves demonstrates the utility of this approach.

[Initial experiences with ambulatory irrigation hysteroscopy]

During one year, 120 patients were referred for outpatient diagnostic hysteroscopy. The most common indications were abnormal uterine bleeding (67%), followed by sterility and infertility, respectively (18%). The uterine cavity was visualised in 103 patients, and about half of the patients (50.5%) had demonstrable uterine pathology. 75 patients (72.8%) did not experience any pain during or after the examination and 28 women (27.2%) reported varying levels of pain. In 17 cases (14%) passage through the internal cervical os was either not possible for anatomical reasons or was not tolerated by the patients. However, diagnostic hysteroscopy, combined with directed biopsy where appropriate, is now considered the method of choice for identifying intrauterine pathology. Performed in an outpatient setting, diagnostic hysteroscopy using a fluid distention medium is largely pain-free and hardly accompanied by side effects, provided patients are thoroughly counseled beforehand. Essential preliminaries for a successful examination are careful patient selection, a clear clinical indication and, above all, the skilled coordination between eye and hand. More widespread use of diagnostic hysteroscopy may spare numerous women the stress of curettage while at the same time facilitating optimal selection of those cases really requiring further medical examination or therapy.

The effect of Ca-formaldehyde fixation on the activity of the reporter enzyme beta-glucuronidase (GUS) in transgenic tobacco

The inhibition of beta-glucuronidase (GUS) activity by Ca-formaldehyde fixation was investigated in leaves of transgenic tobacco (Nicotiana tabacum L.) expressing the beta-glucuronidase uidA reporter gene from Escherichia coli. One-hour fixation of a free-floating leaf disc in a cold fixative decreased the activity of GUS by 60% as compared with the untreated control disc kept in a moist chamber. The inhibition was identical in the periphery and centre of the disc. Fixation including initial vacuum infiltration of the fixative led to almost complete inactivation of the enzyme. The incubation of discs in H2O, either with or without vacuum infiltration, did not lower GUS activity.