Analysis and Validation of Image Search Engines in Histopathology.

Searching for similar images in archives of histology and histopathology images is a crucial task that may aid in patient tissue comparison for various purposes, ranging from triaging and diagnosis to prognosis and prediction. Whole slide images (WSIs) are highly detailed digital representations of tissue specimens mounted on glass slides. Matching WSI to WSI can serve as the critical method for patient tissue comparison. In this paper, we report extensive analysis and validation of four search methods bag of visual words (BoVW), Yottixel, SISH, RetCCL, and some of their potential variants. We analyze their algorithms and structures and assess their performance. For this evaluation, we utilized four internal datasets (1269 patients) and three public datasets (1207 patients), totaling more than 200, 000 patches from 38 different classes/subtypes across five primary sites. Certain search engines, for example, BoVW, exhibit notable efficiency and speed but suffer from low accuracy. Conversely, search engines like Yottixel demonstrate efficiency and speed, providing moderately accurate results. Recent proposals, including SISH, display inefficiency and yield inconsistent outcomes, while alternatives like RetCCL prove inadequate in both accuracy and efficiency. Further research is imperative to address the dual aspects of accuracy and minimal storage requirements in histopathological image search.

Creating an atlas of normal tissue for pruning WSI patching through anomaly detection.

Patching whole slide images (WSIs) is an important task in computational pathology. While most of them are designed to classify or detect the presence of pathological lesions in a WSI, the confounding role and redundant nature of normal histology are generally overlooked. In this paper, we propose and validate the concept of an "atlas of normal tissue" solely using samples of WSIs obtained from normal biopsies. Such atlases can be employed to eliminate normal fragments of tissue samples and hence increase the representativeness of the remaining patches. We tested our proposed method by establishing a normal atlas using 107 normal skin WSIs and demonstrated how established search engines like Yottixel can be improved. We used 553 WSIs of cutaneous squamous cell carcinoma to demonstrate the advantage. We also validated our method applied to an external dataset of 451 breast WSIs. The number of selected WSI patches was reduced by 30% to 50% after utilizing the proposed normal atlas while maintaining the same indexing and search performance in leave-one-patient-out validation for both datasets. We show that the proposed concept of establishing and using a normal atlas shows promise for unsupervised selection of the most representative patches of the abnormal WSI patches.

Seeing the random forest through the decision trees. Supporting learning health systems from histopathology with machine learning models: Challenges and opportunities.

This paper discusses some overlooked challenges faced when working with machine learning models for histopathology and presents a novel opportunity to support "Learning Health Systems" with them. Initially, the authors elaborate on these challenges after separating them according to their mitigation strategies: those that need innovative approaches, time, or future technological capabilities and those that require a conceptual reappraisal from a critical perspective. Then, a novel opportunity to support "Learning Health Systems" by integrating hidden information extracted by ML models from digitalized histopathology slides with other healthcare big data is presented.

Performance of externally validated machine learning models based on histopathology images for the diagnosis, classification, prognosis, or treatment outcome prediction in female breast cancer: A systematic review.

Numerous machine learning (ML) models have been developed for breast cancer using various types of data. Successful external validation (EV) of ML models is important evidence of their generalizability. The aim of this systematic review was to assess the performance of externally validated ML models based on histopathology images for diagnosis, classification, prognosis, or treatment outcome prediction in female breast cancer. A systematic search of MEDLINE, EMBASE, CINAHL, IEEE, MICCAI, and SPIE conferences was performed for studies published between January 2010 and February 2022. The Prediction Model Risk of Bias Assessment Tool (PROBAST) was employed, and the results were narratively described. Of the 2011 non-duplicated citations, 8 journal articles and 2 conference proceedings met inclusion criteria. Three studies externally validated ML models for diagnosis, 4 for classification, 2 for prognosis, and 1 for both classification and prognosis. Most studies used Convolutional Neural Networks and one used logistic regression algorithms. For diagnostic/classification models, the most common performance metrics reported in the EV were accuracy and area under the curve, which were greater than 87% and 90%, respectively, using pathologists' annotations/diagnoses as ground truth. The hazard ratios in the EV of prognostic ML models were between 1.7 (95% CI, 1.2-2.6) and 1.8 (95% CI, 1.3-2.7) to predict distant disease-free survival; 1.91 (95% CI, 1.11-3.29) for recurrence, and between 0.09 (95% CI, 0.01-0.70) and 0.65 (95% CI, 0.43-0.98) for overall survival, using clinical data as ground truth. Despite EV being an important step before the clinical application of a ML model, it hasn't been performed routinely. The large variability in the training/validation datasets, methods, performance metrics, and reported information limited the comparison of the models and the analysis of their results. Increasing the availability of validation datasets and implementing standardized methods and reporting protocols may facilitate future analyses.

Obesity as an independent risk factor for COVID-19 severity and mortality.

BACKGROUND: Since December 2019, the world has struggled with the COVID-19 pandemic. Even after the introduction of various vaccines, this disease still takes a considerable toll. In order to improve the optimal allocation of resources and communication of prognosis, healthcare providers and patients need an accurate understanding of factors (such as obesity) that are associated with a higher risk of adverse outcomes from the COVID-19 infection. OBJECTIVES: To evaluate obesity as an independent prognostic factor for COVID-19 severity and mortality among adult patients in whom infection with the COVID-19 virus is confirmed. SEARCH METHODS: MEDLINE, Embase, two COVID-19 reference collections, and four Chinese biomedical databases were searched up to April 2021. SELECTION CRITERIA: We included case-control, case-series, prospective and retrospective cohort studies, and secondary analyses of randomised controlled trials if they evaluated associations between obesity and COVID-19 adverse outcomes including mortality, mechanical ventilation, intensive care unit (ICU) admission, hospitalisation, severe COVID, and COVID pneumonia. Given our interest in ascertaining the independent association between obesity and these outcomes, we selected studies that adjusted for at least one factor other than obesity. Studies were evaluated for inclusion by two independent reviewers working in duplicate.  DATA COLLECTION AND ANALYSIS: Using standardised data extraction forms, we extracted relevant information from the included studies. When appropriate, we pooled the estimates of association across studies with the use of random-effects meta-analyses. The Quality in Prognostic Studies (QUIPS) tool provided the platform for assessing the risk of bias across each included study. In our main comparison, we conducted meta-analyses for each obesity class separately. We also meta-analysed unclassified obesity and obesity as a continuous variable (5 kg/m2 increase in BMI (body mass index)). We used the GRADE framework to rate our certainty in the importance of the association observed between obesity and each outcome. As obesity is closely associated with other comorbidities, we decided to prespecify the minimum adjustment set of variables including age, sex, diabetes, hypertension, and cardiovascular disease for subgroup analysis.  MAIN RESULTS: We identified 171 studies, 149 of which were included in meta-analyses.  As compared to 'normal' BMI (18.5 to 24.9 kg/m2) or patients without obesity, those with obesity classes I (BMI 30 to 35 kg/m2), and II (BMI 35 to 40 kg/m2) were not at increased odds for mortality (Class I: odds ratio [OR] 1.04, 95% confidence interval [CI] 0.94 to 1.16, high certainty (15 studies, 335,209 participants); Class II: OR 1.16, 95% CI 0.99 to 1.36, high certainty (11 studies, 317,925 participants)). However, those with class III obesity (BMI 40 kg/m2 and above) may be at increased odds for mortality (Class III: OR 1.67, 95% CI 1.39 to 2.00, low certainty, (19 studies, 354,967 participants)) compared to normal BMI or patients without obesity. For mechanical ventilation, we observed increasing odds with higher classes of obesity in comparison to normal BMI or patients without obesity (class I: OR 1.38, 95% CI 1.20 to 1.59, 10 studies, 187,895 participants, moderate certainty; class II: OR 1.67, 95% CI 1.42 to 1.96, 6 studies, 171,149 participants, high certainty; class III: OR 2.17, 95% CI 1.59 to 2.97, 12 studies, 174,520 participants, high certainty). However, we did not observe a dose-response relationship across increasing obesity classifications for ICU admission and hospitalisation. AUTHORS' CONCLUSIONS: Our findings suggest that obesity is an important independent prognostic factor in the setting of COVID-19. Consideration of obesity may inform the optimal management and allocation of limited resources in the care of COVID-19 patients.

Pancreatic Cancer in Iran - Result of the Iranian National Cancer Registry Program.

OBJECTIVE: In this article, we aimed to report the incidence rate of PC at the national and regional levels of Iran from 2014 to 2017 for the first time based on the IARC protocols. METHODS: The data was recruited from the Iranian national program of cancer registry, a national cancer registry program reformed in 2014 after including cancer diagnosis based on clinical judgment and death certificates. This registry includes data from the pathology laboratories and clinical sectors included with death certificates from 60 medical universities in 31 provinces of Iran. Age-standardized incidence rates were calculated at the national and regional levels. RESULTS: From 2014 to 2017, 8851 new cases (males=60.46%) were diagnosed, with a mean age of 66.2 ± 19.6. Forty-one percent of the patients were diagnosed by microscopic verification, and 51% were diagnosed based on clinical judgment without microscopic verification and death certificates. The age-standardized incidence rate was measured as 3.45 per 100,000 in 2017, with the highest rates in individuals older than 85 (30.91 per 100,000), and the provinces of Qom, Tehran, and Isfahan recorded the highest incidence rates with 3.87, 3.85, and 3.66 per 100,000 respectively. CONCLUSIONS: PC incidence in Iran is still lower than in western countries. However, the incidence from 2014 to 2017 is higher than previous national and regional reports and should not be overlooked. Improvement in the national cancer registry program and documentation may be reasons for this difference.